Cascade control system, cascade control method, and non-temporary computer-readable medium

Abstract

A cascade control system for coordinating and controlling carbon emissions associated with operating building equipment distributed across multiple subsystems includes a first controller configured to generate carbon emissions targets for each of the multiple subsystems using a predictive control process that accounts for aggregate carbon emissions of the multiple subsystems predicted to result from the carbon emissions targets. The cascade control system also includes a plurality of second controllers, each corresponding to one of the multiple subsystems, configured to generate control decisions for the building equipment of the corresponding subsystem that are predicted to cause the building equipment to achieve the carbon emissions target of the corresponding subsystem, and to operate the building equipment of the corresponding subsystem using the control decisions.

Description

[Technical Field]

[0001] Cross-reference of related applications This application claims the benefits and priority of U.S. Provisional Patent Application No. 63 / 194,771 filed on 28 May 2021, U.S. Patent Application No. 17 / 686,320 filed on 3 March 2022, U.S. Patent Application No. 17 / 668,791 filed on 10 February 2022, and U.S. Provisional Patent Application No. 63 / 220,878 filed on 12 July 2021, all of which are incorporated herein by reference.

[0002] This disclosure relates, in general, to modular energy units and building equipment having sustainable energy functions, such as functions related to reducing carbon emissions for building operations and / or achieving carbon neutrality. [Background technology]

[0003] Energy consumption related to buildings, including heating and cooling, accounts for a large portion of global energy consumption. In addition, due to the link between energy consumption and production and carbon dioxide emissions (and other pollutant emissions), energy consumption and generation related to building operations now add a significant amount of carbon dioxide to the atmosphere, which contributes to climate change.

[0004] The environmental and ecological impacts of carbon dioxide emissions present technical challenges in reducing or eliminating carbon dioxide emissions associated with building operations, or in achieving carbon neutrality for building operations. For example, building owners may desire to reduce carbon emissions from their buildings or premises, or achieve carbon neutrality, due to consumer demand, regulatory requirements, personal beliefs, etc. Due to their connection to and dependence on utility grids, which are beyond the control of most building owners, they typically lack the technical capability to significantly reduce their carbon footprint using existing technologies. While solar panels, wind turbines, batteries, etc., can be installed by building owners, such products are typically offered as separate components that are difficult for building owners to install and integrate into existing building systems. Therefore, systems and methods for integrated, modular, and easily installable solutions that optimally address building carbon emissions would be desirable. Widespread deployment of such solutions could have a positive impact on the environment while also reducing the operating costs for building owners. [Prior art documents] [Patent Documents]

[0005] [Patent Document 1] U.S. Patent Application Publication No. 2016 / 195866 [Patent Document 2] U.S. Patent Application Publication No. 2004 / 011066 [Patent Document 3] U.S. Patent Application Publication No. 2013 / 013123 [Overview of the project]

[0006] One implementation of the present disclosure is a cascaded control system for coordinating and controlling carbon emissions associated with operating building equipment distributed across multiple subsystems. The cascaded control system includes a first controller configured to generate carbon emission targets for each of the multiple subsystems, and a plurality of second controllers, each corresponding to one of the multiple subsystems, and is configured to generate a control decision for the building equipment of the corresponding subsystem that is expected to cause the building equipment to achieve the carbon emission target for the corresponding subsystem, and to use the control decision to operate the building equipment of the corresponding subsystem.

[0007] In some embodiments, the first controller generates a carbon emissions target based on a time-varying value of the emission rate associated with electricity from the utility grid. One or more devices of the building equipment consume electricity. In some embodiments, the first controller generates a carbon emissions target using a predictive control process that considers the aggregate carbon emissions of several subsystems that are expected to result from the carbon emissions target, and further considers the comfort of one or more building occupants provided by the building equipment based on several reductions that are expected to occur to meet the carbon emissions target. In some embodiments, the first controller generates a carbon emissions target using a multi-objective optimization process with multiple objectives, including a carbon objective based on the aggregate carbon emissions of several low-level subsystems, and a comfort objective. The multiple objectives may include the cost of purchasing resources consumed by the building equipment.

[0008] In some embodiments, the first controller examines and generates carbon emission targets based on both carbon emissions associated with multiple subsystems and other carbon emissions that are not controllable by the cascaded control system. These other carbon emissions may be attributable to the transport of goods or people. In some embodiments, the first controller is configured to generate carbon emission targets based on a budget or goal for total emissions over a certain period.

[0009] Another implementation of the present disclosure is a method for controlling building equipment to reduce or eliminate carbon emissions. The method includes generating time-varying setpoints for building equipment by processing an objective function that takes into account the total carbon emissions or emission savings that are expected to result from resource consumption over a future time range, based on a time-varying value of carbon emissions per unit of resource consumption, and operating the building equipment according to the time-varying setpoints.

[0010] In some embodiments, the time-varying setpoints for building equipment are carbon emission targets for each of the multiple subsystems of the building equipment. Operating the building equipment according to the time-varying setpoints involves determining control decisions for the building equipment that are expected to achieve the carbon emission targets for the multiple subsystems. The time-varying value of carbon emissions per unit of resource consumption includes the value of the critical operating emission rate.

[0011] In some embodiments, the method includes predicting time-varying values ​​of carbon emissions based on historical emission rate data and weather forecasts. In some embodiments, the building equipment includes batteries, and operating the building equipment includes charging the batteries. In some embodiments, the building equipment includes air conditioning equipment. Operating the building equipment according to a time-varying setpoint includes pre-cooling the building during the first period to reduce the operation of the air conditioning equipment during a subsequent second period when it is predicted that the time-varying values ​​of carbon emissions per unit of resource consumption will be higher during the first period than during a subsequent period.

[0012] Another implementation of the present disclosure is a method comprising: obtaining a time-varying critical operating emission rate that indicates the carbon emissions associated with the electricity consumed from a utility; operating the device in energy storage mode when the time-varying critical operating emission rate is less than a first value; and operating the device in energy release mode when the time-varying critical operating emission rate is greater than a second value.

[0013] In some embodiments, the method also includes implementing an objective function-based control process that uses a time-varying limit operating emission rate to determine a time-varying setpoint for the equipment in energy storage mode and energy release mode. The control process may be a cascaded control process. In some embodiments, the equipment includes heating and / or cooling equipment, and the energy storage mode includes pre-cooling or pre-heating a building. In some embodiments, the equipment includes a battery, and the energy storage mode includes charging the battery, and the energy release mode includes discharging the battery. The second value may be greater than the first value and may be equal to the first value. In some embodiments, the method includes determining the first and second values ​​by performing optimization.

[0014] Another implementation of the present disclosure is a method for controlling building equipment. The method includes providing a user interface that includes a graphical representation of the relationship between a carbon emissions control objective and a second control objective that competes with the carbon emissions control objective over a range of control strategies for the building equipment, and assigning weights to the carbon emissions control objective or the second control objective in an objective function. The weights are associated with control strategies corresponding to user selections based on the graphical representation. The method also includes generating control decisions for building equipment using an objective function having weights assigned to the carbon emissions control objective or the second control objective. The method also includes operating building equipment in accordance with the control decisions.

[0015] In some embodiments, the method also includes automatically adjusting weights over time based on the difference between actual performance and the objective associated with user selection. In some embodiments, a second control objective considers at least one of occupant comfort, operating costs, and energy consumption. In some embodiments, the range of the control strategy corresponds to a range of weight values.

[0016] In some embodiments, generating a control decision involves performing an optimization of an objective function with weights assigned to a carbon emissions control objective or a second control objective. In some embodiments, the method also involves generating different points in a graphical representation by performing a simulation of a range of control strategies for building equipment. Performing a simulation of a range of control strategies for building equipment may include performing an optimization of an objective function with different weight values ​​to generate a simulated control decision for building equipment.

[0017] Another implementation of the present disclosure is a method for controlling building equipment, comprising providing an objective function that takes into account at least two of the following: carbon emissions over a time range, operating costs over a time range, and occupant comfort over a time range. The objective function includes one or more tunable parameters indicating the relative importance of at least two of the following: carbon emissions, operating costs, and occupant comfort. The method also includes automatically tuning one or more tunable parameters based on target operating costs, target emissions, target net energy, or target occupant comfort metrics; generating building setpoints by implementing a control process using the objective function; and operating building equipment according to the building setpoints.

[0018] In some embodiments, the target occupant comfort metric is a target number of reduction actions. In some embodiments, the control process includes generating emission targets associated with multiple subsets of building equipment and determining building setpoints based on the emission targets. Automatic tuning of one or more adjustable parameters is based on a target net energy, which is zero. In some embodiments, the control process includes predicting future time-varying values ​​of the critical operating emission rates of energy that will be consumed by building equipment over a time range and performing predictive optimization of the objective function using the future time-varying values.

[0019] In some embodiments, automatically tuning one or more adjustable parameters includes moving the value of a first parameter in a first direction if the limit operating emission rate is greater than expected, and moving the value of the first parameter in a second direction if the limit operating emission rate is less than expected. In some embodiments, the building equipment includes heating, ventilation, or air conditioning equipment, and the building setpoint is a temperature setpoint.

[0020] Another implementation of the present disclosure is one or more non-temporary computer-readable media for storing program instructions, which cause one or more processors to perform an action when the program instructions are executed by one or more processors. The action includes providing a user interface that includes a graphical representation of the relationship between a carbon emissions control objective and a second control objective that competes with the carbon emissions control objective over a range of control strategies for building equipment, and assigning weights to the carbon emissions control objective or the second control objective in an objective function. The weights relate to a control strategy corresponding to a user selection based on the graphical representation. The method includes generating a control decision for building equipment using an objective function having weights assigned to the carbon emissions control objective or the second control objective, and controlling the building equipment in accordance with the control decision.

[0021] In some embodiments, the operation further includes automatically adjusting weights over time based on the difference between actual performance and objectives related to user selection. In some embodiments, the second control objective takes into account at least one of occupant comfort, operating costs, and energy consumption. In some embodiments, generating a control decision involves performing an optimization of an objective function with weights assigned to the carbon emissions control objective or the second control objective.

[0022] In some embodiments, the operation also includes generating different points in the graph representation by performing simulations of a range of control strategies for building equipment. In some embodiments, performing simulations of a range of control strategies for building equipment includes performing optimization of an objective function with different weight values ​​to generate simulated control decisions for building equipment. [Brief explanation of the drawing]

[0023] [Figure 1] These are drawings of buildings equipped with HVAC systems, according to several embodiments. [Figure 2]This is a diagram of a central energy facility (CEF) that may be used to provide heating or cooling to the building shown in Figure 1, according to several embodiments. [Figure 3] This is a diagram of a CEF having a battery unit and a predictive CEF controller according to several embodiments. [Figure 4] Figure 3 is a block diagram of a predictive CEF control system including a battery unit and a predictive CEF controller, according to several embodiments. [Figure 5] This block diagram illustrates the predictive CEF controller shown in Figure 3 in more detail, based on several embodiments. [Figure 6] Figure 3 shows a graph of user interfaces that can be generated by the predictive CEF controller in several embodiments. [Figure 7] This is a diagram of an air-cooled chiller unit having a battery unit and a predictive chiller controller, according to several embodiments. [Figure 8] Figure 7 is a block diagram of an air-cooled chiller unit according to several embodiments. [Figure 9] Figure 7 is a block diagram of a predictive chiller control system, including a battery unit and a predictive chiller controller, according to several embodiments. [Figure 10] This block diagram illustrates the predictive chiller controller of Figure 7 in more detail, according to several embodiments. [Figure 11] This is a diagram of a pump unit having a battery unit and a predictive pump controller according to several embodiments. [Figure 12] This is a block diagram of the pump unit shown in Figure 11, according to several embodiments. [Figure 13] Figure 11 is a block diagram of a predictive pump control system including a battery unit and a predictive pump controller, according to several embodiments. [Figure 14] This block diagram illustrates the predictive pump controller of Figure 11 in more detail, according to several embodiments. [Figure 15]This is a diagram of a cooling tower unit having a battery unit and a predictive cooling tower controller according to several embodiments. [Figure 16] Figure 15 is a block diagram of a predictive cooling tower control system including a battery unit and a predictive cooling tower controller, according to several embodiments. [Figure 17] This block diagram illustrates the predictive cooling tower controller of Figure 15 in more detail, according to several embodiments. [Figure 18] This is a diagram of a valve unit having a battery unit and a predictive valve controller according to several embodiments. [Figure 19] This is a block diagram of the valve unit shown in Figure 18, according to several embodiments. [Figure 20] Figure 18 is a block diagram of a predictive valve control system including a battery unit and a predictive valve controller, according to several embodiments. [Figure 21] This block diagram illustrates the predictive valve controller of Figure 18 in more detail, according to several embodiments. [Figure 22] This is a flowchart of the process for optimizing fuel cells according to several embodiments. [Figure 23] This is an illustrative diagram of a modular energy unit according to several embodiments. [Figure 24] This is a block diagram of a modular energy unit according to several embodiments. [Figure 25] This is a schematic diagram of another embodiment of a modular energy unit according to several embodiments. [Figure 26] This is a flowchart of a process for controlling a modular energy unit according to several embodiments. [Figure 27] This is a flowchart of the process for achieving net-zero carbon emissions using modular energy units, according to several embodiments. [Figure 28]This is a block diagram of a modular energy unit adapted to optimize building settings, according to several embodiments. [Figure 29] This is a block diagram of a system architecture for multiple modular energy units connected to cloud-based optimization resources, according to several embodiments. [Figure 30] This is a block diagram of a building management system connected to cloud-based optimization resources and a system architecture for multiple modular energy units, according to several embodiments. [Figure 31A] This is a flowchart of the process for operating building equipment to achieve an optimal level of carbon emissions or carbon-to-electricity ratio, according to several embodiments. [Figure 31B] This is a description of exemplary time-varying limiting operational discharge rates through several exemplary scenarios. [Figure 32] This is a flowchart of the cost optimization process for building equipment, taking into account the cost of carbon offsetting to achieve net-zero emissions, according to several embodiments. [Figure 33] This is a flowchart of a process for optimizing the operation of building equipment subject to carbon emission constraints, according to several embodiments. [Figure 34] This is a flowchart of a process for customized optimization based on user input indicating weighted preferences for reducing emissions, saving costs, and / or improving occupant comfort, according to several embodiments. [Figure 35A] This is a flowchart of the process for selecting and controlling equipment to achieve a target point on the cost-to-carbon curve, according to several embodiments. [Figure 35B] This is an illustrative diagram of an exemplary cost-to-carbon curve in several embodiments. [Figure 36] This is a flowchart of the process for selecting and controlling equipment to achieve a target point on the comfort-to-carbon curve, according to several embodiments. [Figure 37] This is a flowchart of a process for automatically generating recommendations for new energy assets to be added to a building in order to achieve technical benefits such as optimal carbon emission reduction, through several embodiments. [Figure 38] This is a flowchart of the process for generating an enterprise-wide carbon emissions dashboard and initiating carbon footprint reduction, based on several embodiments. [Figure 39] This is a block diagram of a system architecture for multiple building edge devices in a supervisory control scheme including a carbon management system, according to several embodiments. [Figure 40] This is a flowchart of a process for providing normalized carbon emissions metrics based on production or usage data, according to several embodiments. [Figure 41] An illustrative diagram of an exemplary dashboard that may be provided based on the exemplary processes of Figures 38 and 40 according to several embodiments. [Figure 42] This is a flowchart of a process for controlling a battery based on the limiting operating discharge rate, according to several embodiments. [Modes for carrying out the invention]

[0024] Buildings and HVAC systems Referring here to Figure 1, a perspective view of building 10 is shown. Building 10 is serviced by a BMS. A BMS is generally a system of devices configured to control, monitor, and manage facilities within or around a building or building area. A BMS may include, for example, an HVAC system, a security system, a lighting system, a fire alarm system, any other system capable of managing building functions or devices, or any combination thereof.

[0025] The BMS for building 10 includes an HVAC system 100. The HVAC system 100 may include a number of HVAC devices (e.g., heaters, chillers, air treatment units, pumps, fans, thermal energy storage, etc.) configured to provide heating, cooling, ventilation, or other services to building 10. For example, the HVAC system 100 is shown to include a waterside system 120 and an airside system 130. The waterside system 120 may supply heated or cold fluid to the air treatment units of the airside system 130. The airside system 130 may use the heated or cold fluid to heat or cool the airflow supplied to building 10.

[0026] The HVAC system 100 is shown to include a chiller 102, a boiler 104, and a rooftop air treatment unit (AHU) 106. The waterside system 120 may use the boiler 104 and chiller 102 to heat or cool a working fluid (e.g., water, glycol, etc.) and circulate the working fluid to the AHU 106. In various embodiments, the HVAC devices of the waterside system 120 may be located inside or around the building 10 (as shown in Figure 1), or in an off-site location such as a central plant (e.g., a chiller plant, steam plant, heat plant, etc.). The working fluid may be heated in the boiler 104 or cooled in the chiller 102, depending on whether heating or cooling is required in the building 10. The boiler 104 may heat the circulating fluid, for example, by burning a flammable material (e.g., natural gas) or by using an electric heating element. The chiller 102 may absorb heat from the circulating fluid by placing it in a heat exchange relationship with another fluid (e.g., a refrigerant) in a heat exchanger (e.g., an evaporator). The working fluid from the chiller 102 and / or boiler 104 can be transported to the AHU 106 via piping 108.

[0027] The AHU 106 can place the working fluid in a heat exchange relationship with an airflow passing through the AHU 106 (for example, via one or more stages of a cooling coil and / or heating coil). The airflow can be, for example, outside air, return air from inside the building 10, or a combination of both. The AHU 106 can transfer heat between the airflow and the working fluid to provide heating or cooling to the airflow. For example, the AHU 106 may include one or more fans or blowers configured to pass the airflow over or through a heat exchanger containing the working fluid. The working fluid can then return to the chiller 102 or boiler 104 via piping 110.

[0028] The airside system 130 can deliver the airflow supplied by the AHU 106 (i.e., the supply airflow) to the building 10 via the air supply duct 112 and provide return air from the building 10 to the AHU 106 via the air return duct 114. In some embodiments, the airside system 130 includes a plurality of variable airflow (VAV) units 116. For example, the airside system 130 is shown to include a separate VAV unit 116 for each floor or zone of the building 10. The VAV unit 116 may include dampers or other flow control elements that can operate to control the amount of supply airflow supplied to individual zones of the building 10. In other embodiments, the airside system 130 delivers the supply airflow to one or more zones of the building 10 (e.g., via the supply duct 112) without using intermediate VAV units 116 or other flow control elements. The AHU 106 may include various sensors (e.g., temperature sensors, pressure sensors, etc.) configured to measure attributes of the supply airflow. AHU106 can receive input from sensors located within AHU106 and / or within the building zone, and can adjust the flow rate, temperature, or other attributes of the supply airflow through AHU106 to achieve setpoint conditions for the building zone.

[0029] Central Energy Facilities Referring here to Figure 2, block diagrams of a central energy facility (CEF) 200 according to several embodiments are shown. In various embodiments, the CEF 200 may complement or replace a waterside system 120 within an HVAC system 100, or it may be implemented separately from the HVAC system 100. When implemented in an HVAC system 100, the CEF 200 may include a subset of the HVAC devices within the HVAC system 100 (e.g., a boiler 104, a chiller 102, pumps, valves, etc.) and may operate to supply heated or cooled fluids to the AHU 106. The HVAC devices of the CEF 200 may be located within the building 10 (e.g., as components of the waterside system 120) or at an off-site location.

[0030] CEF200 is shown to include several subplants 202-212, including a heater subplant 202, a heat recovery chiller subplant 204, a chiller subplant 206, a cooling tower subplant 208, a high-temperature thermal energy storage (TES) subplant 210, and a low-temperature thermal energy storage (TES) subplant 212. Subplants 202-212 consume resources from utilities (e.g., water, natural gas, electricity, hydrogen, etc.) to supply thermal energy loads (e.g., hot water, chilled water, heating, cooling, etc.) within a building or premises. For example, the heater subplant 202 may be configured to heat water in a hot water loop 214 that circulates hot water between the heater subplant 202 and the building 10. The chiller subplant 206 may be configured to cool water in a chilled water loop 216 that circulates chilled water between the chiller subplant 206 and the building 10. The heat recovery chiller subplant 204 may be configured to transfer heat from the chilled water loop 216 to the hot water loop 214, providing additional heating to the hot water and additional cooling to the chilled water. The condenser water loop 218 may absorb heat from the chilled water in the chiller subplant 206 and discharge the absorbed heat in the cooling tower subplant 208, or transfer the absorbed heat to the hot water loop 214. The high-temperature TES subplant 210 and the low-temperature TES subplant 212 may store high-temperature and low-temperature thermal energy, respectively, for subsequent use.

[0031] The hot water loop 214 and the cold water loop 216 can deliver heated and / or cooled water to an air handler located on the roof of the building 10 (e.g., AHU 106) or to individual floors or zones of the building 10 (e.g., VAV units 116). The air handler pushes air through a heat exchanger (e.g., a heating coil or cooling coil) through which water flows to provide heating or cooling to the air. The heated or cooled air can be delivered to individual zones of the building 10 to supply the building's thermal energy load. The water then returns to subplants 202-212 for further heating or cooling.

[0032] Subplants 202–212 are shown and described as heating and cooling water for circulation to the building, but it will be understood that any other type of working fluid (e.g., glycol, CO2, etc.) may be used instead of or in addition to water to provide the thermal energy load. In other embodiments, subplants 202–212 may provide heating and / or cooling directly to the building or premises without requiring an intermediate heat transfer fluid. These and other variations of CEF200 are within the scope of the teachings of this disclosure.

[0033] Each of the subplants 202-212 may include various devices configured to facilitate the function of the subplant. For example, the heater subplant 202 is shown to include several heating elements 220 (e.g., boilers, electric heaters, etc.) configured to heat the hot water in the hot water loop 214. The heater subplant 202 is also shown to include several pumps 222 and 224 configured to circulate the hot water in the hot water loop 214 and to control the flow rate of the hot water through the individual heating elements 220. The chiller subplant 206 is shown to include several chillers 232 configured to remove heat from the cold water in the cold water loop 216. The chiller subplant 206 is also shown to include several pumps 234 and 236 configured to circulate the cold water in the cold water loop 216 and to control the flow rate of the cold water through the individual chillers 232.

[0034] The heat recovery chiller subplant 204 is shown to include a plurality of heat recovery heat exchangers 226 (e.g., a refrigerant circuit) configured to transfer heat from the chilled water loop 216 to the hot water loop 214. The heat recovery chiller subplant 204 is also shown to include several pumps 228 and 230 configured to circulate hot and / or chilled water through the heat recovery heat exchangers 226 and to control the flow rate of water through the individual heat recovery heat exchangers 226. The cooling tower subplant 208 is shown to include a plurality of cooling towers 238 configured to remove heat from condenser water in the condenser water loop 218. The cooling tower subplant 208 is also shown to include several pumps 240 configured to circulate condenser water in the condenser water loop 218 and to control the flow rate of condenser water through the individual cooling towers 238.

[0035] The high-temperature TES subplant 210 is shown to include a high-temperature TES tank 242 configured to store hot water for later use. The high-temperature TES subplant 210 may also include one or more pumps or one or more valves configured to control the flow rate of hot water to or from the high-temperature TES tank 242. The low-temperature TES subplant 212 is shown to include a low-temperature TES tank 244 configured to store cold water for later use. The low-temperature TES subplant 212 may also include one or more pumps or one or more valves configured to control the flow rate of cold water to or from the low-temperature TES tank 244.

[0036] In some embodiments, one or more pumps within the CEF200 (e.g., pumps 222, 224, 228, 230, 234, 236, and / or 240) or piping within the CEF200 include associated insulating valves. The insulating valves may be integrated with the pumps or positioned upstream or downstream of the pumps to control the fluid flow within the CEF200. In various embodiments, the CEF200 may include more, fewer, or different types of devices and / or subplants, depending on the specific configuration of the CEF200 and the type of load supplied by the CEF200.

[0037] Central energy facility with battery unit and predictive control unit Referring here to Figure 3, a central energy facility (CEF) 300 having a battery unit 302 and a predictive CEF controller 304 is shown according to several embodiments. The CEF 300 may be configured to provide cooling to a cooling load 322. The cooling load 322 may include, for example, a building zone, a supply air stream flowing through an air duct, airflow in an air processing unit or rooftop unit, fluid flowing through a heat exchanger, a refrigerator or freezer, a condenser or evaporator, a cooling coil, or any other type of system, device, or space requiring cooling. In some embodiments, a pump 318 circulates a cold fluid to the cooling load 322 via a cold fluid circuit 336. The cold fluid can absorb heat from the cooling load 322, thereby providing cooling to the cooling load 322 and warming the cold fluid.

[0038] CEF300 is shown to include a cooling tower 312 and a chiller 320. The cooling tower 312 may be configured to cool the water in the cooling tower circuit 332 by transferring heat from the water to the outside air. In some embodiments, a pump 316 circulates water through the cooling tower circuit 332 and through the cooling tower 312. The cooling tower 312 may include a fan 314 that causes cooling air to flow through the cooling tower 312. The cooling tower 312 places the cold air in a heat exchange relationship with the warmer water, thereby transferring heat from the warmer water to the colder air. The cooling tower 312 can provide cooling to the condenser 326 of the chiller 320. The condenser 326 can transfer heat from the refrigerant in the refrigerant circuit 334 to the water in the cooling tower circuit 332. Although the cooling tower circuit 332 is shown and described as using circulating water, it should be understood that any type of coolant or working fluid (e.g., water, glycol, CO2, etc.) may be used in the cooling tower circuit 332.

[0039] Chiller 320 is shown to include a condenser 326, a compressor 328, an evaporator 330, and an expansion device 324. The compressor 328 may be configured to circulate the refrigerant between the condenser 326 and the evaporator 330 via a refrigerant circuit 334. The compressor 328 operates to compress the refrigerant to a high-pressure, high-temperature state. The compressed refrigerant flows through the condenser 326, which transfers heat from the refrigerant in the refrigerant circuit 334 to the water in the cooling tower circuit 332. The cooled refrigerant then flows through the expansion device 324, which expands the refrigerant to a low-temperature, low-pressure state. The expanded refrigerant flows through the evaporator 330, which transfers heat from the cold fluid in the cold fluid circuit 336 to the refrigerant in the refrigerant circuit 334.

[0040] In some embodiments, the CEF300 includes a plurality of chillers 320. Each chiller 320 may be arranged in parallel and configured to provide cooling to the fluid in the chilled fluid circuit 336. In some embodiments, the set of chillers 320 may have a cooling capacity of approximately 1 to 3 MW or 1,000 to 6,000 tons. Similarly, the CEF300 may include a plurality of cooling towers 312. Each cooling tower 312 may be arranged in parallel and configured to provide cooling to the water in the cooling tower circuit 332. Although only cooling components are shown in Figure 3, it is intended that the CEF300 may include heating components in some embodiments. For example, the CEF300 may include one or more boilers, heat recovery chillers, steam generators, or other devices configured to provide heating. In some embodiments, the CEF300 includes some or all of the components of the CEF200 described with reference to Figure 2.

[0041] Referring further to Figure 3, the CEF300 is shown to include a battery unit 302. In some embodiments, the battery unit 302 includes one or more photovoltaic (PV) panels 308. A PV panel 308 may include an assembly of photovoltaic cells. Photovoltaic cells are configured to convert solar energy (i.e., sunlight) into electricity using photovoltaic materials such as monocrystalline silicon, polycrystalline silicon, amorphous silicon, cadmium telluride, copper indium gallenium / sulfide selenide, or other materials that exhibit a photovoltaic effect. In some embodiments, the photovoltaic cells are contained within a packaged assembly that forms the PV panel 308. Each PV panel 308 may include a plurality of linked photovoltaic cells. The PV panels 308 can be combined to form a photovoltaic array.

[0042] In some embodiments, the PV panel 308 is configured to maximize solar energy collection. For example, the battery unit 302 may include a solar tracker (e.g., a GPS tracker, a sunshine sensor, etc.) that adjusts the angle of the PV panel 308 so that it is directly facing the sun throughout the day. The solar tracker may allow the PV panel 308 to receive direct sunlight for a longer period of the day, which may increase the total amount of energy produced by the PV panel 308. In some embodiments, the battery unit 302 includes a collection of mirrors, lenses, or sun concentrators configured to guide and / or concentrate sunlight onto the PV panel 308. The energy generated by the PV panel 308 may be stored in the battery cell 306 and / or used to power various components of the CEF 300.

[0043] In some embodiments, the battery unit 302 includes one or more battery cells 306. The battery cells 306 are configured to store and release electrical energy (i.e., electricity). In some embodiments, the battery unit 302 is charged using electricity from an external energy grid (e.g., provided by an electric utility). The electricity stored in the battery unit 302 can be released to power one or more powered components of the CEF 300 (e.g., cooling towers 312, fans 314, chillers 320, pumps 316-318, etc.). Advantageously, the battery unit 302 allows the CEF 300 to shift its electrical load over time by drawing electricity from the energy grid and charging the battery unit 302 when energy prices are low, and releasing the stored electricity when energy prices are high. In some embodiments, the battery unit 302 has sufficient energy capacity (e.g., 6–12 MW hours) to power the CEF 300 for approximately 4–6 hours when operating at maximum capacity, so that it can be utilized during high-energy-cost periods and charged during low-energy-cost periods.

[0044] In some embodiments, the predictive CEF controller 304 performs an optimization process to determine whether to charge or discharge the battery unit 302 during each of several time steps that occur during the optimization period. The predictive CEF controller 304 may use weather and price data 310 to predict the amount of heating / cooling and the cost of electricity required during each of the several time steps. The predictive CEF controller 304 can optimize an objective function that takes into account the cost of electricity purchased from the energy grid over the optimization period. In some embodiments, the objective function also takes into account the cost of operating the various components of the CEF 300 (e.g., the cost of natural gas used to fuel the boiler). During each time step, the predictive CEF controller 304 can determine the amount of electricity to purchase from the energy grid and the amount of electricity to store or release from the battery unit 302. The objective function and optimization performed by the predictive CEF controller 304 are described in more detail with reference to Figures 4-5.

[0045] Predictive CEF control system Referring here to Figure 4, block diagrams of predictive CEF control systems 400 according to several embodiments are shown. Some of the components shown in the control system 400 may be part of the CEF 300. For example, the CEF 300 may include a powered CEF component 402, a battery unit 302, a predictive CEF controller 304, a power inverter 410, and a power junction 412. The powered CEF component 402 may include any component of the CEF 300 that consumes power (e.g., electricity) during operation. For example, the powered CEF component 402 is shown to include a cooling tower 404, a chiller 406, and a pump 408. These components may be analogous to the cooling tower 312, chiller 320, and pumps 316-318 as described with reference to Figure 3.

[0046] The power inverter 410 may be configured to convert power between direct current (DC) and alternating current (AC). For example, the battery unit 302 may be configured to store and output DC power, while the energy grid 414 and the powered CEF component 402 may be configured to consume and supply AC power. The power inverter 410 may be used to convert DC power from the battery unit 302 to a sinusoidal AC output synchronized to the grid frequency of the energy grid 414 and / or the powered CEF component 402. The power inverter 410 may also be used to convert AC power from the energy grid 414 to DC power that can be stored in the battery unit 302. The power output of the battery unit 302 is P bat It is shown as P bat This can be positive when the battery unit 302 is supplying power to the inverter 410 (i.e., the battery unit 302 is discharging), or negative when the battery unit 302 is receiving power from the power inverter 410 (i.e., the battery unit 302 is charging).

[0047] In some examples, the power inverter 410 receives a DC power output from the battery unit 302 and converts the DC power output into an AC power output that can be supplied to the powered CEF component 402. The power inverter 410 may use a local oscillator to synchronize the frequency of the AC power output with the frequency of the energy grid 414 (e.g., 50 Hz or 60 Hz) and may limit the voltage of the AC power output so that it does not exceed the grid voltage. In some embodiments, the power inverter 410 is a resonant inverter that includes or uses an LC circuit to remove harmonics from a simple square wave in order to achieve a sine wave that matches the frequency of the energy grid 414. In various embodiments, the power inverter 410 may operate with or without a transformer, using a high-frequency transformer or a low-frequency transformer. A low-frequency transformer may directly convert the DC output from the battery unit 302 into an AC output that can be supplied to the powered CEF component 402. The high-frequency transformer may employ a multi-stage process, which involves converting the DC output to high-frequency AC, then back to DC, and finally to an AC output supplied to the powered CEF component 402.

[0048] The power output of PV panel 308 is P PV It is shown as follows: Power output P of PV panel 308 PV The energy P generated by the PV panel 308 may be stored in the battery unit 302 and / or used to power the powered CEF component 402. In some embodiments, the energy P generated by the PV panel 308 may be stored in the battery unit 302 and / or used to power the powered CEF component 402. PV The system measures and provides an indicator of PV power to the predictive CEF controller 304. For example, the PV panel 308 is configured to provide an indicator of PV power percentage (i.e., PV%) to the predictive CEF controller 304. The PV power percentage may represent the percentage of the maximum PV power currently being used by the PV panel 308.

[0049] The power junction 412 is the point where the power-type CEF component 402, the energy grid 414, the PV panel 308, and the power inverter 410 are electrically connected. The power supplied from the power inverter 410 to the power junction 412 is P bat as shown. P bat can be positive when the power inverter 410 is providing power to the power junction 412 (i.e., the battery unit 302 is discharging), and can be negative when the power inverter 410 is receiving power from the power junction 412 (i.e., the battery unit 302 is charging). The power supplied from the energy grid 414 to the power junction 412 is P grid as shown, and the power supplied from the PV panel 308 to the power junction 412 is P PV as shown. P bat , P PV , and P grid are combined at the power junction 412 to form P total (i.e., P total = P grid + P bat + P PV ). P total can be defined as the power provided from the power junction 412 to the power-type CEF component 402. In some examples, P total is greater than P grid . For example, when the battery unit 302 is discharging, P bat can be positive, which is added to the grid power P bat and the PV power P PV when they are combined to form P grid to form P total . In other examples, P grid is less than P PV . For example, when the battery unit 302 is charging, P total can be negative, which is the combination of P grid , P bat , P bat , P PV , and P grid to form Ptotal When forming, grid power P grid and PV power P PV It is subtracted from.

[0050] The predictive CEF controller 304 may be configured to control the powered CEF components 402 and the power inverter 410. In some embodiments, the predictive CEF controller 304 controls the battery power setpoint P sp,bat It generates and provides it to the power inverter 410. Battery power setpoint P sp,bat This is the battery power setpoint P sp,bat To achieve this, the power inverter 410 is (P sp,bat (If negative) Use the power available at power junction 412 to charge battery unit 302, or (P sp,bat (If positive) the battery unit 302 is discharged to supply power to the power junction 412, which may include a positive or negative power value (e.g., kW).

[0051] In some embodiments, the predictive CEF controller 304 generates and provides control signals to the powered CEF component 402. The predictive CEF controller 304 may use multi-stage optimization techniques to generate control signals. For example, the predictive CEF controller 304 may include an economic controller configured to determine the optimal amount of energy to be consumed by the powered CEF component 402 at each time step during the optimization period. The optimal amount of energy to be consumed may minimize a cost function that takes into account the cost of the energy consumed by the CEF 300. The cost of energy may be based on the time-varying energy price from the electric utility 418. In some embodiments, the predictive CEF controller 304 determines the optimal amount of energy to purchase from the energy grid 414 (i.e., the grid power setpoint P) at each of a plurality of time steps. sp,grid ) and the optimal amount of power to store or release from the battery unit 302 (i.e., battery power setpoint P sp,batThe predictive CEF controller 304 can monitor the actual power consumption of the powered CEF component 402 and use the actual power consumption as a feedback signal when generating the optimal power setpoint.

[0052] The predictive CEF controller 304 sets a temperature setpoint (e.g., zone temperature setpoint T) that achieves the optimal amount of power consumption at each time step. sp,zone , chilled water temperature set point T sp,chw The predictive CEF controller 304 may include a tracking controller configured to generate (etc.). In some embodiments, the predictive CEF controller 304 uses an equipment model for the powered CEF component 402 to determine the amount of heating or cooling that can be generated by the CEF component 402 based on optimal power consumption. Based on the power setpoint and / or temperature setpoint, the predictive CEF controller 304 determines the building zone temperature T zone To predict how it will change, a zone temperature model can be used in combination with weather forecasts from weather service 416.

[0053] In some embodiments, the predictive CEF controller 304 generates control signals for the powered CEF components 402 using temperature setpoints. The control signals may include on / off commands, speed setpoints for the fans of the cooling tower 404, power setpoints for the compressors of the chiller 406, chilled water temperature setpoints for the chiller 406, pressure or flow rate setpoints for the pumps 408, or other types of setpoints for individual devices of the powered CEF components 402. In other embodiments, the control signals may include temperature setpoints generated by the predictive CEF controller 304 (e.g., zone temperature setpoint T sp,zone , chilled water temperature set point T sp,chw This may include, for example. The temperature setpoint may be provided to a powered CEF component 402 or a local controller of the powered CEF component 402 that operates to achieve the temperature setpoint. For example, the local controller of chiller 406 receives the chilled water temperature T from the chilled water temperature sensor. chw Measurement values, and / or zone temperature T from zone temperature sensor. zoneThe measured values ​​can be received. The local controller can use a feedback control process (e.g., PID, ESC, MPC, etc.) to increase or decrease the amount of cooling provided by the chiller 406 and drive the measured temperature(s) to a temperature setpoint(s). A similar feedback control process can be used to control the cooling tower 404 and / or pump 408. The multi-stage optimization performed by the predictive CEF controller 304 is described in more detail with reference to Figure 5.

[0054] Predictive CEF Controller Referring now to Figure 5, a block diagram illustrating a predictive CEF controller 304 in more detail according to an exemplary embodiment is shown. The predictive CEF controller 304 is shown to include a communication interface 502 and a processing circuit 504. The communication interface 502 can facilitate communication between the controller 304 and an external system or device. For example, the communication interface 502 can receive zone temperature T from a zone temperature sensor 516. zone The communication interface 502 can receive measurements of the power consumption of the power-driven CEF component 402. In some embodiments, the communication interface 502 receives measurements of the state of charge (SOC) of the battery unit 302, which may be provided as a percentage of the maximum battery capacity (i.e., battery %). The communication interface 502 can receive weather forecasts from the weather service 416 and predicted energy costs and demand costs from the electric utility 418. In some embodiments, the predictive CEF controller 304 uses the communication interface 502 to provide control signals to the power-driven CEF component 402 and the power inverter 410.

[0055] The communication interface 502 may include wired or wireless communication interfaces (e.g., jacks, antennas, transmitters, receivers, transceivers, wired terminals, etc.) for data communication with external systems or devices. In various embodiments, communication may be direct (e.g., local wired or wireless communication) or via a communication network (e.g., a WAN, the Internet, a cellular network, etc.). For example, the communication interface 502 may include Ethernet cards and ports for transmitting and receiving data over an Ethernet-based communication link or network. In another example, the communication interface 502 may include Wi-Fi transceivers for communication over a wireless communication network or cellular or mobile phone communication transceivers.

[0056] The processing circuit 504 is shown to include a processor 506 and a memory 508. The processor 506 may be a general-purpose or specific-purpose processor, an application-specific integrated circuit (ASIC), one or more field-programmable gate arrays (FPGAs), a group of processing components, or other suitable processing components. The processor 506 is configured to execute computer code or instructions stored in the memory 508 or received from other computer-readable media (e.g., a CD-ROM, network storage, remote server, etc.).

[0057] Memory 508 may include one or more devices (e.g., memory units, memory devices, storage devices, etc.) for storing data and / or computer code to complete and / or facilitate the various processes described herein. Memory 508 may include random access memory (RAM), read-only memory (ROM), hard drive storage, temporary storage, non-volatile memory, flash memory, optical memory, or any other suitable memory for storing software objects and / or computer instructions. Memory 508 may include database components, object code components, script components, or any other type of information structure to support the various activities and information structures described herein. Memory 508 may be communicably connected to the processor 506 via processing circuit 504 and may include computer code for executing one or more processes described herein (e.g., by the processor 506). When the processor 506 executes instructions stored in memory 508 to complete the various activities described herein, the processor 506 generally configures the controller 304 (and more specifically, the processing circuit 504) to complete such activities.

[0058] Referring further to Figure 5, the predictive CEF controller 304 is shown to include an economy controller 510, a tracking controller 512, and an equipment controller 514. Controllers 510-514 may be configured to perform a multi-state optimization process for generating control signals to the power inverter 410 and the powered CEF components 402. Briefly, the economy controller 510 optimizes the predictive cost function to determine the optimal amount of electricity to purchase from the energy grid 414 at each time step of the optimization period (i.e., the grid power setpoint P sp,grid ), the optimal amount of energy to store or release from the battery unit 302 (i.e., the battery power setpoint P sp,bat ), and / or the optimal amount of power to be consumed by the powered CEF component 402 (i.e., CEF power setpoint P sp,totalThe tracking controller 512 can determine the optimal power setpoint P. sp,grid , P sp,bat , and / or P sp,total Use this to determine the optimal temperature setpoint (for example, zone temperature setpoint T sp,zone , chilled water temperature set point T sp,chw (etc.) and the optimal battery charging or discharging speed (i.e., Bat C / D The device controller 514 can determine the optimal temperature setpoint T sp,zone , or T sp,chw Using the actual (e.g., measured) temperature T zone and / or T chw Control signals can be generated for the powered CEF component 402 to drive it to a setpoint (for example, using a feedback control technique). Each of the controllers 510-514 is described in detail below.

[0059] Economic Controller The economic controller 510 optimizes the predicted cost function to determine the optimal amount of electricity to purchase from the energy grid 414 at each time step of the optimization period (i.e., the grid power setpoint P sp,grid ), the optimal amount of energy to store or release from the battery unit 302 (i.e., the battery power setpoint P sp,bat ), and / or the optimal amount of power to be consumed by the powered CEF component 402 (i.e., CEF power setpoint P sp,total ) can be configured to determine. An example of a forecast cost function that can be optimized by the economic controller 510 is shown in the following equation.

number

[0060] The first and second terms of the predicted cost function J represent the cost of electricity consumed by the power equation CEF component 402 over the optimization period. The parameter C at each time step k ec The value of (k) can be defined by the energy cost information provided by the electric utility 418. In some embodiments, the cost of electricity changes as a function of time, and this is different C at different time steps k. ec The variable P yields the value of (k). chiller (k) and P HRC (k) is a decision variable that can be optimized by the economic controller 510. In some embodiments, the total power consumption P of the power-type CEF component 402 at time step k is... total (k) is P chiller (k) and P HRC (k) is equal to the sum of (i.e., P total (k=P) chiller (k) + P HRC (k)). Therefore, in some embodiments, the first two terms of the prediction cost function are summation [Number] can be replaced with.

[0061] The third term of the prediction cost function J represents the cost of the fuel (e.g., natural gas) consumed by the CEF300 over the duration of the optimization period. The value of C gas (k) at each time step k can be defined by the energy cost information provided by the natural gas utility. In some embodiments, the cost of the gas varies as a function of time, which results in different values of C gas (k) at different time steps k. The variable F gas (k) is a decision variable that can be optimized by the economic controller 510.

[0062] The fourth term of the prediction cost function J represents the demand charge. The demand charge is an additional charge imposed by some utility providers based on the maximum power consumption during the corresponding demand charge period. For example, the demand charge rate C DC can be specified in dollars per unit of power (e.g., $ / kW), and the demand charge can be calculated by multiplying the peak power usage (e.g., kW) during the demand charge period. In the prediction cost function J, the demand charge rate C DC can be defined by the demand cost information received from the electric utility 418. The variable P grid (k) is a decision variable that can be optimized by the economic controller 510 to reduce the peak power usage max(P grid (k)) that occurs during the demand charge period. Load shifting can enable the economic controller 510 to smooth out momentary spikes in the electrical demand of the CEF300 by storing energy in the battery unit 302 when the power consumption of the power-type CEF component 402 is low. The stored energy is used to draw peak power P gridcan be discharged from the battery unit 302 to reduce the demand charge thereby incurred.

[0063] The last term of the prediction cost function J represents the amount of cost savings due to the use of the battery unit 302. Unlike the previous terms of the cost function J, the last term is subtracted from the total cost. The value of the parameter C ec (k) at each time step k can be defined by the energy cost information provided by the electric utility 418. In some embodiments, the cost of electricity varies as a function of time, which results in different values of C ec (k) at different time steps k. The variable P bat (k) is a decision variable that can be optimized by the economic controller 510. A positive value of P bat (k) indicates that the battery unit 302 is discharging, and a negative value of P bat (k) indicates that the battery unit 302 is charging. The power P bat (k) discharged from the battery unit 302 can be used to satisfy some or all of the total power consumption P total (k) of the power-type CEF component 402, which reduces the amount of power P grid (k) purchased from the energy grid 414 (i.e., P grid (k)=P total (k)-P bat (k)-P PV (k)). However, when the battery unit 302 is charged, a negative value of P grid (k) is brought about that is added to the total amount of power P bat (k) purchased from the energy grid 414.

[0064] In some embodiments, the power P PV provided by the PV panel 308 is not included in the prediction cost function J because generating PV power incurs no cost. However, the power P PV generated by the PV panel 308 can be used to satisfy the total power consumption P total(k) can be partially or entirely satisfied, which is the amount of electricity P purchased from the energy grid 414. grid (k) Reduce (i.e., P grid (k=P) total (k)-P bat (k)-P PV (k)). The amount of PV energy P generated during any time step k. PV However, this can be predicted by the economic controller 510. Several techniques for predicting the amount of PV power generated by PV panels are described in U.S. Patent Application No. 15 / 247,869. (U.S. Patent Application Publication No. 2017 / 0104449) U.S. Patent Application No. 15 / 247,844 (U.S. Patent Application Publication No. 2017 / 0104337) , and U.S. Patent Application No. 15 / 247,788 (U.S. Patent Application Publication No. 2017 / 0102675) These patent applications are described in [the relevant document]. Each of these patent applications has a filing date of August 25, 2016, and the entire disclosure of each of these patent applications is incorporated herein by reference.

[0065] The economic controller 510 can optimize the predicted cost function J over the duration of the optimization period to determine the optimal value of the decision variable at each time step during the optimization period. In some embodiments, the optimization period has a duration of approximately one day, and each time step is approximately 15 minutes. However, the duration of the optimization period and time steps may vary in other embodiments and may be adjusted by the user. Advantageously, the economic controller 510 can use the battery unit 302 to implement load shifting by drawing electricity from the energy grid 414 when energy prices are low and / or when the power consumed by the power-driven CEF component 402 is low. The electricity is stored in the battery unit 302 and can be discharged later when energy prices are high and / or when the power consumption of the power-driven CEF component 402 is high. This allows the economic controller 510 to reduce the cost of electricity consumed by the CEF 300, smooth out instantaneous spikes in the electricity demand of the CEF 300, and thereby reduce the resulting demand charges.

[0066] The economic controller 510 may be configured to impose constraints on the optimization of the predicted cost function J. In some embodiments, the constraint is the temperature T of the building zones served by the CEF 300. zone This includes the constraints. The economic controller 510 controls the actual or predicted temperature T zone The minimum temperature boundary T min and the maximum temperature boundary T max Between (that is, T min ≦T zone ≦T max ) can be configured to always be maintained. Parameter T min and T max This can vary over time to define different temperature ranges at different times (e.g., occupied temperature range, unoccupied temperature range, daytime temperature range, nighttime temperature range, etc.).

[0067] To ensure that the zone temperature constraints are met, the economic controller 510 determines the building zone temperature T as a function of the decision variables optimized by the economic controller 510. zone This can be modeled. In some embodiments, the economic controller 510 models the temperature of the building zones using a heat transfer model. For example, the heating or cooling dynamics of the building zones can be described by the energy balance.

number

number

number

number

number

number

[0068] The previous energy balance combined all mass and air properties of the building zones into a single zone temperature. Other heat transfer models that can be used by the economic controller 510 include the following air and mass zone models.

number

[0069] The previous equation combines all the mass characteristics of the building zones into a single zone mass. Other heat transfer models that can be used by the economic controller 510 include the following air, shallow mass, and deep mass zone models.

number

[0070] In some embodiments, the economic controller 510 uses weather forecasts from the weather service 416 to determine the ambient air temperature T at each time step of the optimization period. a and / or disturbances

number

number

number

[0071] In some embodiments, the economic controller 510 provides a power setpoint P sp,grid and P sp,bat As a function of the amount of heating or cooling applied to a building zone by CEF300,

number

number

[0072] In some embodiments, the economic controller 510 controls the zone temperature T as shown in the following equation. zone and zone temperature setpoint T sp,zone As a function of the amount of heating or cooling applied to the building zone by the CEF300 (i.e.,

number

number

[0073] The model used by the economic controller 510 is the amount of heating or cooling provided by the CEF300.

number

[0074] In some embodiments, the economic controller 510 controls the amount of heating or cooling provided by the CEF300.

number

[0075] The economic controller 510 controls operation v as shown in the following equation air The amount of heating or cooling provided by CEF300

number

number

number

[0076] In some embodiments,

number

number

number

number

number

number

[0077] Zone temperature T zone In addition to the constraints on the above, the economic controller 510 can impose constraints on the charge state (SOC) and charge / discharge rate of the battery unit 302. In some embodiments, the economic controller 510 generates and imposes the following power constraints on the predicted cost function J. P bat ≤P rated -P bat ≤P rated In the formula, P bat P is the amount of electricity released from the battery unit 302. rated This is the rated battery power of the battery unit 302 (for example, the maximum speed at which the battery unit 302 can be charged or discharged). These power constraints apply when the battery unit 302 is rated P ratedWe guarantee that the battery will not be charged or discharged at a speed exceeding its maximum possible charge / discharge rate.

[0078] In some embodiments, the economic controller 510 generates and imposes one or more capacity constraints on the predicted cost function J. The capacity constraints are the battery power P charged or discharged during each time step. bat This can be used to relate the capacity and SOC of the battery unit 302. The capacity constraints can ensure that the capacity of the battery unit 302 is maintained within acceptable lower and upper limits at each time step of the optimization period. In some embodiments, the economy controller 510 generates the following capacity constraints: C a (k)-P bat (k)Δt≦C rated C a (k)-P bat (k)Δt≧0 In the formula, C a (k) is the available battery capacity (e.g., kWh) at the start of time step k, and P bat (k) is the rate at which the battery unit 302 is discharged during time step k (e.g., kW), Δt is the duration of each time step, and C rated This is the maximum rated capacity (e.g., kWh) of the battery unit 302. Item P bat (k)Δt represents the change in battery capacity during time step k. These capacity constraints apply when the capacity of battery unit 302 is zero and when it is at its maximum rated capacity C. rated Ensure that it is maintained between them.

[0079] In some embodiments, the economic controller 510 generates and imposes one or more capacity constraints on the operation of the powered CEF component 402. For example, the powered CEF component 402 has a maximum power consumption P total,max It may have a corresponding maximum operating point (e.g., maximum pump speed, maximum cooling capacity, etc.). The economic controller 510 provides power P to the powered CEF component 402 as shown in the following equation. total Set to zero and maximum power consumption Ptotal,max It can be configured to generate constraints that limit the relationship between the two. 0≦P total ≤P total,max P total =P sp,grid +P sp,bat In the formula, the total power P supplied to the power-type CEF component 402. total This is the grid power setpoint P sp,grid and battery power setpoint P sp,bat It is the sum of the two.

[0080] In some embodiments, the economic controller 510 generates and imposes one or more capacity constraints on the operation of one or more subplants of the CEF300. For example, heating may be provided by the heater subplant 202, and cooling may be provided by the chiller subplant 206. The operation of the heater subplant 202 and the chiller subplant 206 may be defined by subplant curves for each of the heater subplant 202 and the chiller subplant 206. Each subplant curve may define the resource production of the subplant (e.g., refrigerated tonnage, kW heating, etc.) as a function of one or more resources (e.g., electricity, natural gas, water, etc.) consumed by the subplant. Some examples of subplant curves that may be used by the economic controller 510 are provided in U.S. Patent Application No. 14 / 634,609, filed February 27, 2015. (U.S. Patent Application Publication No. 2015 / 0316902) This will be explained in more detail.

[0081] The economic controller 510 may be configured to use subplant curves to identify the maximum heating amount that can be provided by the heater subplant 202 and the maximum cooling amount that can be provided by the chiller subplant 206. The economic controller 510 can generate and impose constraints that limit the amount of heating provided by the heater subplant 202 between zero and the maximum heating amount. Similarly, the economic controller 510 can generate and impose constraints that limit the amount of cooling provided by the chiller subplant 206 between zero and the maximum cooling amount.

[0082] The economic controller 510 optimizes the predicted cost function J according to the constraints, and determines the decision variable P total , P chiller , P HRC F gas , P grid , and P bat The optimal value of can be determined, where P total =P bat +P grid +P PV In some embodiments, the economic controller 510 is P total , P bat , and / or P grid The optimal value is used to generate a power setpoint for the tracking controller 512. The power setpoint is the battery power setpoint P for each time step k during the optimization period. sp,bat , grid power setpoint P sp,grid , and / or CEF power setpoint P sp,total This may include the following: The economic controller 510 can provide power setpoints to the tracking controller 512.

[0083] Tracking controller The tracking controller 512 controls the optimal power setpoint P generated by the economic controller 510. sp,grid , P sp,bat , and / or P sp,total Use this to determine the optimal temperature setpoint (for example, zone temperature setpoint T sp,zone , chilled water temperature set point T sp,chw (etc.) and the optimal battery charging or discharging speed (i.e., Bat C / D ) can be determined. In some embodiments, the tracking controller 512 can determine the power setpoint P of the CEF300. sp,total Zone temperature setpoint T is expected to achieve this. sp,zone , and / or chilled water temperature setpoint T sp,chw In other words, the tracking controller 512 provides the CEF300 with the optimal amount of power P determined by the economic controller 510. total Zone temperature setpoint T that consumes sp,zone and / or chilled water temperature setpoint T sp,chw It can generate.

[0084] In some embodiments, the tracking controller 512 uses a power consumption model to determine the power consumption of the CEF300 based on the zone temperature T. zone and zone temperature setpoint T sp,zone To associate with, for example, the tracking controller 512 uses the model of the equipment controller 514 to determine the zone temperature T zone and zone temperature setpoint T sp,zone The control operation performed by the equipment controller 514 can be determined as a function of this. An example of such a zone regulation controller model is shown in the following equation. v air =f3(T zone , T sp,zone ) In the formula, v air This is the velocity (i.e., control action) of the airflow into the building zone.

[0085] The tracking controller 512 controls the zone temperature T zone and zone temperature setpoint T sp,zone As a function of the power consumption P of the CEF300, total This can be defined. An example of such a model is shown in the following equation. P total =f4(T zone , T sp,zone )

[0086] The function f4 can be identified from the data. For example, the tracking controller 512 is P total and T zone Collect measurements of T sp,zone The corresponding value can be identified. The tracking controller 512 can identify P total , T zone , and T sp,zone The collected values ​​can be used as training data to perform a system identification process and determine a function f4 that defines the relationships between such variables.

[0087] The tracking controller 512 uses a similar model to measure the total power consumption P of the CEF300. totaland the chilled water temperature setpoint T sp,chw The relationship between the two can be determined. For example, the tracking controller 512 determines the zone temperature T zone and chilled water temperature set point T sp,chw As a function of the power consumption P of the CEF300, total This can be defined. An example of such a model is shown in the following equation. P total =f5(T zone , T sp,chw )

[0088] The function f5 can be identified from the data. For example, the tracking controller 512 is P total and T zone Collect measurements of T sp,chw The corresponding value can be identified. The tracking controller 512 can identify P total , T zone , and T sp,chw The collected values ​​can be used as training data to perform a system identification process and determine the function f5 that defines the relationships between such variables.

[0089] Tracking controller 512 is P total , T sp,zone , and T sp,chw Using the relationship between T sp,zone , and T sp,chw The value of can be determined. For example, the tracking controller 512 can determine P as input from the economic controller 510. total The value (i.e., P sp,total ) can receive, T sp,zone and T sp,chw The corresponding value can be used to make a determination. The tracking controller 512 uses T sp,zone , and T sp,chw The value can be provided as an output to the device controller 514.

[0090] In some embodiments, the tracking controller 512 controls the battery power setpoint P sp,bat Use the optimal speed to charge or discharge the battery unit 302. C / DDetermine the following: For example, the battery power setpoint P sp,bat This can define a power value (kW) that can be converted by the tracking controller 512 into a control signal for the power inverter 410 and / or the equipment controller 514. In other embodiments, the battery power setpoint P sp,bat This is supplied directly to the power inverter 410, and battery power P bat It is used by the power inverter 410 to control it.

[0091] Device controller The device controller 514 sets the optimal temperature setpoint T generated by the tracking controller 512. sp,zone or T sp,chw This can be used to generate a control signal for the powered CEF component 402. The control signal generated by the instrument controller 514 is used to control the actual (e.g., measured) temperature T zone and / or T chw The motor can be driven to a set point. The instrument controller 514 can generate control signals for the powered CEF component 402 using any of a variety of control techniques. For example, the instrument controller 514 can generate control signals for the powered CEF component 402 using a state-based algorithm, an extreme value search control (ESC) algorithm, a proportional-integral (PI) control algorithm, a proportional-integral-derivative (PID) control algorithm, a model predictive control (MPC) algorithm, or other feedback control algorithm.

[0092] The control signals may include on / off commands, speed setpoints for the cooling tower 404 fan, power setpoints for the chiller 406 compressor, chilled water temperature setpoints for the chiller 406, pressure setpoints or flow rate setpoints for the pump 408, or other types of setpoints for individual devices of the powered CEF component 402. In other embodiments, the control signals may include temperature setpoints (e.g., zone temperature setpoint T) generated by the predictive CEF controller 304. sp,zone , chilled water temperature set point T sp,chwThis may include, for example. The temperature setpoint may be provided to a powered CEF component 402 or a local controller of the powered CEF component 402 that operates to achieve the temperature setpoint. For example, the local controller of chiller 406 receives the chilled water temperature T from the chilled water temperature sensor. chw Measurement values, and / or zone temperature T from zone temperature sensor. zone Measurement values ​​can be received.

[0093] In some embodiments, the equipment controller 514 is configured to provide control signals to the power inverter 410. The control signals provided to the power inverter 410 are directed to the battery power setpoint P sp,bat , and / or optimal charge / discharge rate Bat C / D It may include the following: The device controller 514 controls the battery power setpoint P sp,bat To achieve this, the power inverter 410 may be configured to operate. For example, the equipment controller 514 may set the battery power setpoint P sp,bat Accordingly, the power inverter 410 can either charge the battery unit 302 or discharge the battery unit 302.

[0094] Referring now to Figure 6, several embodiments of the user interface 600 that may be generated by the predictive CEF controller 304 are shown. As discussed above, the economic controller 510 calculates each power consumption value (e.g., P) including grid power and / or battery power at each time step of the optimization period. chiller , P HRC It can be configured to determine parts such as grid power and / or battery power. The user interface 600 can be used to communicate to the user the relative portion of each power consumption value, including grid power and / or battery power.

[0095] Interface 600 illustrates a dispatch chart. The upper half of the dispatch chart corresponds to cooling, and the lower half corresponds to heating. The midpoint line between the upper and lower halves corresponds to zero load / power for both halves. Positive cooling values ​​are shown as displacements above the midpoint line, and positive heating values ​​are shown as displacements below the midpoint line. Lines 602 and 612 represent the required cooling load and required heating load, respectively, at each time step of the optimization period. Lines 604 and 614 represent the charge levels of the batteries used to power the cooling equipment (e.g., chiller subplant) and heating equipment (e.g., heater subplant) over the duration of the optimization period.

[0096] As discussed above, the economic controller 510 may be configured to determine the optimal power setpoint for each time step of the optimization period. The results of the optimization performed by the economic controller 510 may be represented in a dispatch chart. For example, the dispatch chart is shown to include vertical columns for each time step of the optimization period. Each column may contain one or more bars representing the power setpoint determined by the economic controller 510 for the corresponding time step. The color of each bar indicates the type of power setpoint. For example, the gray bars 608 and 618 (shown as white bars in Figure 6) represent grid power setpoints (e.g., P sp,grid ) may indicate, while the green bars 606 and 616 (shown as shaded bars in Figure 6) indicate the battery power setpoint (e.g., P sp,bat This can indicate the magnitude of the corresponding power setpoint in that time step. The height of each bar indicates the magnitude of the corresponding power setpoint in that time step.

[0097] A green bar 606 positioned above the required cooling line 602 indicates that the cooling equipment battery is being charged (i.e., excess energy used to charge the battery), while a green bar 606 positioned below the required cooling line 602 indicates that the cooling equipment battery is being discharged (i.e., battery power used to meet part of the required cooling load). The charge level of the cooling equipment battery increases when the cooling equipment battery is being charged and decreases when the cooling equipment battery is being discharged.

[0098] Similarly, a green bar 616 positioned below the required heating wire 612 indicates that the heating equipment battery is charged (i.e., excess energy used to charge the battery), while a green bar 616 positioned above the required heating wire 612 indicates that the heating equipment battery is discharged (i.e., battery power used to meet part of the required heating load). The charge level of the heating equipment battery increases when the heating equipment battery is charged and decreases when the heating equipment battery is discharged.

[0099] Air-cooled chiller with battery unit and predictive control Referring here to Figures 7 and 8, an air-cooled chiller 700 is shown, according to several embodiments, comprising a battery unit 702 and a predictive chiller controller 704. The chiller 700 may be configured to provide a chilled fluid (e.g., chilled water 718) to a cooling load 734 via chilled water pipes 714. The cooling load 734 may include, for example, a building zone, a supply air stream flowing through an air duct, airflow within an air processing unit or rooftop unit, fluid flowing through a heat exchanger, a refrigerator or freezer, a condenser or evaporator, a cooling coil, or any other type of system, device, or space requiring cooling. In some embodiments, a pump 732 circulates the chilled fluid to the cooling load 734 via a chilled fluid circuit 738. The chilled fluid can absorb heat from the cooling load 734, thereby providing cooling to the cooling load 734 and warming the chilled fluid. The warm fluid (shown in Figure 7 as return water 716) may return to the chiller 700 via a return water pipe 712.

[0100] Chiller 700 is shown to include a condenser 722, a compressor 720, an evaporator 724, an expansion device 726, and a fan 730. The compressor 720 may be configured to circulate the refrigerant between the condenser 722 and the evaporator 724 via a refrigerant circuit 736. The compressor 720 operates to compress the refrigerant to a high-pressure, high-temperature state. The compressed refrigerant flows through the condenser 722, which transfers heat from the refrigerant in the refrigerant circuit 736 to the airflow 728. The fan 730 can be used to push the airflow 728 through or over the condenser 722 to provide cooling to the refrigerant in the condenser 722. The cooled refrigerant then flows through the expansion device 726, which expands the refrigerant to a low-temperature, low-pressure state. The expanded refrigerant flows through the evaporator 724, which transfers heat from the cold fluid in the cold fluid circuit 738 to the refrigerant in the refrigerant circuit 736.

[0101] In some embodiments, the chiller 700 includes one or more photovoltaic (PV) panels 708. A PV panel 708 may include an assembly of photovoltaic cells. Photovoltaic cells are configured to convert solar energy (i.e., sunlight) into electricity using a photovoltaic material such as monocrystalline silicon, polycrystalline silicon, amorphous silicon, cadmium telluride, copper indium gallenium / sulfide selenide, or other materials that exhibit a photovoltaic effect. In some embodiments, the photovoltaic cells are contained within a packaged assembly that forms the PV panel 708. Each PV panel 708 may include a plurality of linked photovoltaic cells. The PV panels 708 can be combined to form a photovoltaic array.

[0102] In some embodiments, the PV panel 708 is configured to maximize solar energy collection. For example, the chiller 700 may include a solar tracker (e.g., a GPS tracker, a sunshine sensor, etc.) that adjusts the angle of the PV panel 708 so that it is directly facing the sun throughout the day. The solar tracker may allow the PV panel 708 to receive direct sunlight for a longer period of the day, which may increase the total amount of energy produced by the PV panel 708. In some embodiments, the chiller 700 includes a collection of mirrors, lenses, or sun concentrators configured to guide and / or concentrate sunlight onto the PV panel 708. The energy generated by the PV panel 708 may be stored in a battery unit 702 and / or used to power various components of the chiller 700.

[0103] In some embodiments, the battery unit 702 includes one or more battery cells 706. The battery cells 706 are configured to store and release electrical energy (i.e., electricity). In some embodiments, the battery unit 702 is charged using electricity from an external energy grid (e.g., provided by an electric utility). The electricity stored in the battery unit 702 can be released to power one or more powered components of the chiller 700 (e.g., a fan 730, a compressor 720, a pump 732, etc.). Advantageously, the battery unit 702 allows the chiller 700 to shift its electrical load over time by drawing electricity from the energy grid and charging the battery unit 702 when energy prices are low, and releasing the stored electricity when energy prices are high. In some embodiments, the battery unit 702 has sufficient energy capacity to power the chiller 700 for approximately 4-6 hours when operating at maximum capacity, so that it can be utilized during periods of high energy cost and charged during periods of low energy cost.

[0104] As shown in Figure 8, the chiller 700 may include a fuel cell 802. In some embodiments, the fuel cell 802 is a fuel cell configured to generate electrical energy using chemical reactions. For example, the fuel cell 802 may convert the chemical energy of hydrogen and an oxidizer (e.g., oxygen) into electricity through a pair of oxidation-reduction reactions. In other embodiments, the fuel cell 802 is a hydrocarbon fuel cell that uses one or more of diesel, methanol, natural gas, etc., to generate electricity. The fuel cell 802 may be controlled to generate electricity to augment grid energy or other energy sources, to supplement battery discharge during high-energy-cost periods, and to generate electricity (e.g., during high-energy-cost periods) to charge the battery. The fuel cell may require fuel replacement (e.g., hydrogen supply), which may be purchased periodically and added to the chiller 700. In embodiments in which the chiller 700 includes a fuel cell 802, the control and optimization processes described herein are configured to take into account the contribution of the fuel cell 802 and the cost of operating the fuel cell 802 when generating control outputs for various components of the chiller 700, including the fuel cell 802. For example, optimization performed by the predictive chiller controller 704 may determine whether to operate the fuel cell 802 to generate electricity for each time step of the optimization period.

[0105] In some embodiments, the predictive chiller controller 704 performs an optimization process to determine whether to charge or discharge the battery unit 702 during each of several time steps that occur during the optimization period. The predictive chiller controller 704 may use weather and price data 710 to predict the amount of heating / cooling and the cost of electricity required during each of the several time steps. The predictive chiller controller 704 can optimize an objective function that takes into account the cost of electricity purchased from the energy grid over the duration of the optimization period. During each time step, the predictive chiller controller 704 can determine the amount of electricity to purchase from the energy grid and the amount of electricity to store or release from the battery unit 702. The objective function and optimization performed by the predictive chiller controller 704 are described in more detail with reference to Figures 9-10.

[0106] Predictive chiller control system Referring now to Figure 9, block diagrams of predictive chiller control systems 900 according to several embodiments are shown. Some of the components shown in the control system 900 may be part of the chiller 700. For example, the chiller 700 may include a powered chiller component 902, a battery unit 702, a predictive chiller controller 704, a power inverter 910, and a power junction 912. The powered chiller component 902 may include any components of the chiller 700 that consume power (e.g., electricity) during operation. For example, the powered chiller component 902 is shown to include a cooling fan 730, a compressor 720, and a pump 732.

[0107] The power inverter 910 may be configured to convert electrical power between direct current (DC) and alternating current (AC). For example, the battery unit 702 may be configured to store and output DC power, while the energy grid 914 and the powered chiller component 902 may be configured to consume and supply AC power. The power inverter 910 may be used to convert the DC power from the battery unit 702 into a sinusoidal AC output synchronized to the grid frequency of the energy grid 914 and / or the powered chiller component 902. The power inverter 910 may also be used to convert the AC power from the energy grid 914 into DC power that can be stored in the battery unit 702. The power output of the battery unit 702 is P bat It is shown as P bat This can be positive when the battery unit 702 is supplying power to the inverter 910 (i.e., the battery unit 702 is discharging), and negative when the battery unit 702 is receiving power from the power inverter 910 (i.e., the battery unit 702 is charging).

[0108] In some examples, the power inverter 910 receives a DC power output from the battery unit 702 and converts the DC power output into an AC power output that can be supplied to the powered chiller component 902. The power inverter 910 may use a local oscillator to synchronize the frequency of the AC power output with the frequency of the energy grid 914 (e.g., 50 Hz or 60 Hz) and may limit the voltage of the AC power output so that it does not exceed the grid voltage. In some embodiments, the power inverter 910 is a resonant inverter that includes or uses an LC circuit to remove harmonics from a simple square wave in order to achieve a sine wave that matches the frequency of the energy grid 914. In various embodiments, the power inverter 910 may operate with or without a transformer, using a high-frequency transformer or a low-frequency transformer. A low-frequency transformer may directly convert the DC output from the battery unit 702 into an AC output that can be supplied to the powered chiller component 902. The high-frequency transformer may employ a multi-stage process, which involves converting the DC output to high-frequency AC, then back to DC, and finally to an AC output supplied to the powered chiller component 902.

[0109] The power output of the PV panel 708 is P PV It is shown as follows: Power output P of PV panel 708 PV The energy can be stored in the battery unit 702 and / or used to supply power to the powered chiller component 902. In some embodiments, the amount of energy P generated by the PV panel 708 is PV The system measures and provides an indicator of PV power to the predictive chiller controller 704. For example, the PV panel 708 is shown to provide an indicator of PV power percentage (i.e., PV%) to the predictive chiller controller 704. The PV power percentage may represent the percentage of the maximum PV power that the PV panel 708 is currently operating on.

[0110] The power junction 912 is the point where the powered chiller components 902, the energy grid 914, the PV panels 708, and the power inverter 910 are electrically connected. The power supplied from the power inverter 910 to the power junction 912 is P bat It is shown as P bat This can be positive when the power inverter 910 is supplying power to the power junction 912 (i.e., the battery unit 702 is discharging), and negative when the power inverter 910 is receiving power from the power junction 912 (i.e., the battery unit 702 is charging). The power supplied from the energy grid 914 to the power junction 912 is P grid As shown, the power supplied from PV panel 708 to power junction 912 is P PV It is shown as P bat , P PV , and P grid These are combined at power junction 912, P total (that is, P total =P grid +P bat +P PV ) forms P total P can be defined as the power supplied from the power junction 912 to the powered chiller component 902. In some examples, P total P grid It is larger than that. For example, when battery unit 702 is discharging, P bat This can be positive, and this is P bat and P PV P grid Combined with P total When forming, grid power P grid and PV power P PV It is added to P. total P grid It may be less than P. For example, when the battery unit 702 is charging, bat It can be negative, and this is P bat , P PV , and P grid These are combined into Ptotal When forming, grid power P grid and PV power P PV It is subtracted from.

[0111] The predictive chiller controller 704 may be configured to control the powered chiller components 902 and the power inverter 910. In some embodiments, the predictive chiller controller 704 controls the battery power setpoint P sp,bat It generates and provides it to the power inverter 910. Battery power setpoint P sp,bat This is the battery power setpoint P sp,bat To achieve this, the power inverter 910 uses the power available in the power junction 912 or the discharge battery unit 702 (P sp,bat If positive, charge the battery unit 702 (P sp,bat If negative, it may include a positive or negative power value (e.g., kW) that causes power junction 912 to supply power.

[0112] In some embodiments, the predictive chiller controller 704 generates and provides control signals to the powered chiller component 902. The predictive chiller controller 704 may use multi-stage optimization techniques to generate control signals. For example, the predictive chiller controller 704 may include an economic controller configured to determine the optimal amount of energy to be consumed by the powered chiller component 902 at each time step during the optimization period. The optimal amount of energy to be consumed may minimize a cost function that takes into account the cost of energy consumed by the chiller 700. The cost of energy may be based on the time-varying energy price from the electric utility 918. In some embodiments, the predictive chiller controller 704 determines the optimal amount of energy to purchase from the energy grid 914 (i.e., the grid power setpoint P) at each of a plurality of time steps. sp,grid ) and the optimal amount of energy to store or release from the battery unit 702 (i.e., the battery power setpoint P sp,batThe predictive chiller controller 704 can monitor the actual power consumption of the powered chiller component 902 and use the actual power consumption as a feedback signal when generating the optimal power setpoint.

[0113] The predictive chiller controller 704 sets a temperature setpoint (e.g., air temperature setpoint T) that achieves the optimal power consumption at each time step. sp,air , chilled water temperature set point T sp,water The predictive chiller controller may include a tracking controller configured to generate (etc.). In some embodiments, the predictive chiller controller 704 uses an equipment model for the powered chiller component 902 to determine the amount of heating or cooling that can be generated by the chiller component 902 based on the optimal power consumption. The predictive chiller controller 704 uses a temperature model to determine the chilled water temperature T based on the power setpoint. water It is possible to predict how it will change.

[0114] In some embodiments, the predictive chiller controller 704 generates control signals for the powered chiller components 902 using temperature setpoints. These control signals may include on / off commands, speed setpoints for the fan 730, power setpoints for the compressor 720, chilled water temperature setpoints for the chiller 700, pressure or flow rate setpoints for the pump 732, or other types of setpoints for individual devices of the powered chiller components 902. In other embodiments, the control signals may include temperature setpoints generated by the predictive chiller controller 704 (e.g., air temperature setpoint T sp,air , chilled water temperature set point T sp,water This may include, for example. The temperature setpoint may be provided to the powered chiller component 902 or the local controller of the powered chiller component 902, which operates to achieve the temperature setpoint. For example, the local controller of the fan 730 receives the chilled water temperature T from the chilled water temperature sensor. water Measurement values ​​and / or air temperature T from the air temperature sensor airThe local controller can receive a measurement of the temperature of the airflow 728 (i.e., the temperature of the airflow 728). The local controller can use a feedback control process (e.g., PID, ESC, MPC, etc.) to increase or decrease the amount of air supplied by the fan 730 and drive the measured temperature(s) up to a temperature setpoint. A similar feedback control process can be used for the compressor 720 and / or pump 732. The multi-stage optimization performed by the predictive chiller controller 704 will be described in more detail with reference to Figure 10.

[0115] Predictive Chiller Controller Referring now to Figure 10, a block diagram illustrating a predictive chiller controller 704 in more detail according to an exemplary embodiment is shown. The predictive chiller controller 704 is shown to include a communication interface 1002 and a processing circuit 1004. The communication interface 1002 can facilitate communication between the controller 704 and an external system or device. For example, the communication interface 1002 can receive air temperature T from a temperature sensor 1016. air and cold water temperature T water The communication interface 1002 can receive measurements of the power consumption of the powered chiller component 902. In some embodiments, the communication interface 1002 receives measurements of the state of charge (SOC) of the battery unit 702, which may be provided as a percentage of the maximum battery capacity (i.e., battery %). The communication interface 1002 can receive weather forecasts from the weather service 916 and predicted energy costs and demand costs from the electric utility 918. In some embodiments, the predictive chiller controller 704 uses the communication interface 1002 to provide control signals to the powered chiller component 902 and the power inverter 910.

[0116] The communication interface 1002 may include wired or wireless communication interfaces (e.g., jacks, antennas, transmitters, receivers, transceivers, wired terminals, etc.) for data communication with external systems or devices. In various embodiments, communication may be direct (e.g., local wired or wireless communication) or via a communication network (e.g., a WAN, the Internet, a cellular network, etc.). For example, the communication interface 1002 may include Ethernet cards and ports for transmitting and receiving data over an Ethernet-based communication link or network. In another example, the communication interface 1002 may include Wi-Fi transceivers for communication over a wireless communication network or cellular or mobile phone communication transceivers.

[0117] The processing circuit 1004 is shown to include a processor 1006 and a memory 1008. The processor 1006 may be a general-purpose or specific-purpose processor, an application-specific integrated circuit (ASIC), one or more field-programmable gate arrays (FPGAs), a group of processing components, or other suitable electronic processing components. The processor 1006 is configured to execute computer code or instructions stored in the memory 1008 or received from other computer-readable media (e.g., CD-ROM, network storage, remote server, etc.).

[0118] Memory 1008 may include one or more devices (e.g., memory units, memory devices, storage devices, etc.) for storing data and / or computer code to complete and / or facilitate the various processes described herein. Memory 1008 may include random access memory (RAM), read-only memory (ROM), hard drive storage, temporary storage, non-volatile memory, flash memory, optical memory, or any other suitable memory for storing software objects and / or computer instructions. Memory 1008 may include database components, object code components, script components, or any other type of information structure to support the various activities and information structures described herein. Memory 1008 may be communicably connected to processor 1006 via processing circuit 1004 and may include computer code for executing one or more processes described herein (e.g., by processor 1006). When the processor 1006 executes instructions stored in memory 1008 to complete the various activities described herein, the processor 1006 generally configures the controller 704 (and more specifically, the processing circuit 1004) to complete such activities.

[0119] Referring further to Figure 10, the predictive chiller controller 704 is shown to include an economy controller 1010, a tracking controller 1012, and an equipment controller 1014. Controllers 1010-1014 may be configured to perform a multi-state optimization process to generate control signals for the power inverter 910 and the powered chiller components 902. Briefly, the economy controller 1010 optimizes the predictive cost function to determine the optimal amount of electricity to purchase from the energy grid 914 at each time step of the optimization period (i.e., the grid power setpoint P sp,grid ), the optimal amount of energy to store or release from the battery unit 702 (i.e., the battery power setpoint P sp,bat ), and / or the optimal amount of power to be consumed by the powered chiller component 902 (i.e., chiller power setpoint Psp,total The tracking controller 1012 can determine the optimal power setpoint P. sp,grid , P sp,bat , and / or P sp,total Use this to determine the optimal temperature setpoint (for example, air setpoint T sp,air , chilled water temperature set point T sp,water (etc.), as well as the optimal battery charging or discharging speed (i.e., Bat C / D The device controller 1014 can determine the optimal temperature setpoint T. sp,air or T sp,water Using the actual (e.g., measured) temperature T air and / or T water Control signals can be generated for the powered chiller component 902 to drive it to a setpoint (for example, using a feedback control technique). Each of the controllers 1010 to 1014 is described in detail below.

[0120] Economic Controller The economic controller 1010 optimizes the predicted cost function to determine the optimal amount of electricity to purchase from the energy grid 914 at each time step of the optimization period (i.e., the grid power setpoint P sp,grid ), the optimal amount of energy to store or release from the battery unit 702 (i.e., the battery power setpoint P sp,bat ), and / or the optimal amount of power consumed by the powered chiller component 902 (i.e., chiller power setpoint P sp,total ) can be configured to determine. An example of a forecast cost function that can be optimized by the economic controller 1010 is shown in the following equation.

number

[0121] The first, second, and third terms of the predicted cost function J represent the cost of electricity consumed by the powered chiller component 902 over the duration of the optimization period. The parameter C at each time step k ec The value of (k) may be defined by the energy cost information provided by the electrical utility 918. In some embodiments, the cost of electricity varies as a function of time, and this is different C at different time steps k. ec The variable P yields the value of (k). fan (k), P comp (k), and P pump (k) is a decision variable that can be optimized by the economic controller 1010. In some embodiments, the total power consumption P of the powered chiller components 902 at time step k is... total (k) is P fan (k), P comp (k), and P pump (k) is equal to the sum of (i.e., P total (k=P) fan (k) + P comp (k) + P pump (k) Therefore, in some embodiments, the first three terms of the prediction cost function are sum

number

[0122] The fourth term of the forecast cost function J represents demand charges. Demand charges are additional charges imposed by several utility providers based on the maximum power consumption during the relevant demand charging period. For example, the demand charging rate C DC The demand charge rate C may be specified in US dollars per unit of electricity (e.g., $ / kW) and can be calculated by multiplying it by the peak electricity usage (e.g., kW) during the demand charge period. In the forecast cost function J, the demand charge rate C DC This can be defined by demand cost information received from the electric utility 918. Variable P grid (k) is the peak power usage max(P) that occurs during the demand-charging period. grid (k)) is a decision variable that can be optimized by the economic controller 1010 to reduce load shift. Load shift can enable the economic controller 1010 to smooth out instantaneous spikes in the electricity demand of the chiller 700 by storing energy in the battery unit 702 when the power consumption of the power chiller component 902 is low. The stored energy can be used to draw peak power P from the energy grid 914 when the power consumption of the power chiller component 902 is high. grid To reduce this and thereby decrease the resulting demand charges, it may be discharged from the battery unit 702.

[0123] The last term of the predicted cost function J represents the cost savings resulting from the use of battery unit 702. Unlike the preceding terms of the cost function J, the last term is subtracted from the total cost. Parameter C at each time step k ec The value of (k) may be defined by the energy cost information provided by the electrical utility 918. In some embodiments, the cost of electricity varies as a function of time, and this is different C at different time steps k. ec The variable P yields the value of (k). bat (k) is a decision variable that can be optimized by the economic controller 1010. batA positive value for (k) indicates that the battery unit 702 is discharging, and P bat A negative value for (k) indicates that the battery unit 702 is charging. Power P is the power emitted from the battery unit 702. bat (k) is the total power consumption P of the powered chiller component 902. total (k) can be used to satisfy some or all of the amount of electricity P purchased from the energy grid 914. grid (k) Reduce (i.e., P grid (k=P) total (k)-P bat (k)-P PV (k)). However, when the battery unit 702 is charged, the power P purchased from the energy grid 914 grid P is added to the total amount of (k) bat A negative value for (k) is obtained.

[0124] In some embodiments, the power P provided by the PV panel 708 PV The power generated by the PV panel 708 does not incur any cost and is therefore not included in the predicted cost function J. However, the power P generated by the PV panel 708 is not included in the predicted cost function J. PV Using this, the total power consumption P of the powered chiller component 902 total (k) can be partially or entirely satisfied, which is the amount of electricity P purchased from the energy grid 914. grid (k) Reduce (i.e., P grid (k=P) total (k)-P bat (k)-P PV (k)). The amount of PV energy P generated during any time step k. PV This can be predicted by the economic controller 1010. Several techniques for predicting the amount of PV power generated by PV panels are described in U.S. Patent Application No. 15 / 247,869. (U.S. Patent Application Publication No. 2017 / 0104449) U.S. Patent Application No. 15 / 247,844 (U.S. Patent Application Publication No. 2017 / 0104337) , and U.S. Patent Application No. 15 / 247,788 (U.S. Patent Application Publication No. 2017 / 0102675)These patent applications are described in [the relevant document]. Each of these patent applications has a filing date of August 25, 2016, and the entire disclosure of each of these patent applications is incorporated herein by reference.

[0125] The economic controller 1010 can optimize the predicted cost function J over the duration of the optimization period to determine the optimal value of the decision variable at each time step during the optimization period. In some embodiments, the optimization period has a duration of approximately one day, and each time step is approximately 15 minutes. However, the duration of the optimization period and time steps may vary in other embodiments and may be adjusted by the user. Advantageously, the economic controller 1010 can use the battery unit 702 to implement load shifting by drawing electricity from the energy grid 914 when energy prices are low and / or when the power consumed by the powered chiller component 902 is low. The electricity is stored in the battery unit 702 and can be discharged later when energy prices are high and / or the power consumption of the powered chiller component 902 is high. This allows the economic controller 1010 to reduce the cost of electricity consumed by the chiller 700, smooth out instantaneous spikes in the electricity demand of the chiller 700, and thereby reduce the resulting demand charges.

[0126] The economic controller 1010 may be configured to impose constraints on the optimization of the predicted cost function J. In some embodiments, the constraint is the temperature T of the chilled water produced by the chiller 700. water This includes constraints on the minimum temperature boundary T. The economic controller 1010 has a minimum temperature boundary T. min and the maximum temperature boundary T max Between (that is, T min ≦T water ≦T max Actual or predicted temperature T water It can be configured to always maintain the parameter T. min and T max This can vary over time to define different temperature ranges at different points in time.

[0127] Water temperature T water In addition to the constraints on the above, the economic controller 1010 can impose constraints on the charge state (SOC) and charge / discharge rate of the battery unit 702. In some embodiments, the economic controller 1010 generates and imposes the following power constraints on the predicted cost function J. P bat ≤P rated -P bat ≤P rated In the formula, P bat This is the amount of power released from the battery unit 702, and P rated This is the rated battery power of the battery unit 702 (e.g., the maximum speed at which the battery unit 702 can be charged or discharged). These power constraints apply when the battery unit 702 is rated P rated We guarantee that the battery will not be charged or discharged at a speed exceeding its maximum possible charge / discharge rate.

[0128] In some embodiments, the economic controller 1010 generates and imposes one or more capacity constraints on the predicted cost function J. The capacity constraints are the battery power P charged or discharged during each time step. bat This can be used to relate the capacity and SOC of the battery unit 702. The capacity constraints can ensure that the capacity of the battery unit 702 is maintained within acceptable lower and upper limits at each time step of the optimization period. In some embodiments, the economy controller 1010 generates the following capacity constraints: C a (k)-P bat (k)Δt≦C rated C a (k)-P bat (k)Δt≧0 In the formula, C a (k) is the available battery capacity (e.g., kWh) at the start of time step k, and P bat (k) is the rate at which the battery unit 702 is discharged during time step k (e.g., kW), Δt is the duration of each time step, and Crated This is the maximum rated capacity (e.g., kWh) of the battery unit 702. Item P bat (k)Δt represents the change in battery capacity during time step k. These capacity constraints apply when the capacity of battery unit 702 is zero and when it is at its maximum rated capacity C. rated Ensure that it is maintained between them.

[0129] In some embodiments, the economic controller 1010 generates and imposes one or more capacity constraints on the operation of the powered chiller component 902. For example, the powered chiller component 902 has a maximum power consumption P total,max It may have a corresponding maximum operating point (e.g., maximum pump speed, maximum cooling capacity, etc.). The economic controller 1010 supplies power P to the powered chiller component 902 as shown in the following equation. total Set to zero and maximum power consumption P total,max It can be configured to generate constraints that limit the relationship between the two. 0≦P total ≤P total,max P total =P sp,grid +P sp,bat In the formula, the total power P supplied to the powered chiller component 902 total This is the grid power setpoint P sp,grid and battery power setpoint P sp,bat It is the sum of the two.

[0130] The economic controller 1010 optimizes the predicted cost function J according to the constraints, and determines the decision variable P total , P fan , P comp , P pump , P grid , and P bat The optimal value of can be determined, and in the formula, P total =P bat +P grid +P PV In some embodiments, the economic controller 1010 is P total , P bat , and / or P gridThe optimal value is used to generate power setpoints for the tracking controller 1012. The power setpoints are the battery power setpoints P for each time step k during the optimization period. sp,bat , grid power setpoint P sp,grid , and / or chiller power setpoint P sp,total This may include the following: The economic controller 1010 can provide power setpoints to the tracking controller 1012.

[0131] Tracking controller The tracking controller 1012 sets the optimal power setpoint P generated by the economic controller 1010. sp,grid , P sp,bat , and / or P sp,total Use this to determine the optimal temperature setpoint (for example, air temperature setpoint T sp,air , chilled water temperature set point T sp,water (etc.) and the optimal battery charging or discharging speed (i.e., Bat C / D ) can be determined. In some embodiments, the tracking controller 1012 can determine the power setpoint P of the chiller 700. sp,total The air temperature setpoint T is expected to achieve this. sp,air and / or chilled water temperature setpoint T sp,water It generates the air temperature setpoint T. sp,air and / or chilled water temperature setpoint T sp,water This can generate power P determined by the economic controller 1010 to the chiller 700. total Consume the optimal amount.

[0132] In some embodiments, the tracking controller 1012 controls the battery power setpoint P sp,bat Use the optimal charging or discharging speed for battery unit 702. C / D Determine the following: For example, the battery power setpoint P sp,bat This can define a power value (kW) that can be converted by the tracking controller 1012 into a control signal for the power inverter 910 and / or the equipment controller 1014. In other embodiments, the battery power setpoint P sp,batThis is supplied directly to the power inverter 910, and battery power P bat It is used by the power inverter 910 to control it.

[0133] Device controller The device controller 1014 sets the optimal temperature setpoint T generated by the tracking controller 1012. sp,air or T sp,water This can be used to generate a control signal for the powered chiller component 902. The control signal generated by the equipment controller 1014 is the actual (e.g., measured) temperature T air and / or T water The powered chiller component 902 can be driven to a set point. The equipment controller 1014 can generate control signals for the powered chiller component 902 using any of a variety of control techniques. For example, the equipment controller 1014 can generate control signals for the powered chiller component 902 using a state-based algorithm, an extreme value search control (ESC) algorithm, a proportional-integral (PI) control algorithm, a proportional-integral-derivative (PID) control algorithm, a model predictive control (MPC) algorithm, or other feedback control algorithm.

[0134] The control signals may include on / off commands, speed setpoints for the fan 730, power setpoints for the compressor 720, pressure or flow rate setpoints for the pump 732, or other types of setpoints for individual devices of the powered chiller components 902. In other embodiments, the control signals may include temperature setpoints (e.g., air temperature setpoint T) generated by the predictive chiller controller 704. sp,air , chilled water temperature set point T sp,water This may include, for example. The temperature setpoint may be provided to the powered chiller component 902 or the local controller of the powered chiller component 902, which operates to achieve the temperature setpoint. For example, the local controller for the fan 730 receives the chilled water temperature T from the chilled water temperature sensor. water Measurement values ​​and / or air temperature T from the air temperature sensor airThe measured temperature can be received, and the speed of the fan 730 can be adjusted to drive the measured temperature to the temperature setpoint.

[0135] In some embodiments, the equipment controller 1014 is configured to provide control signals to the power inverter 910. The control signals provided to the power inverter 910 are directed to the battery power setpoint P sp,bat , and / or optimal charge / discharge rate Bat C / D It may include the following: The device controller 1014 sets the battery power setpoint P sp,bat To achieve this, the power inverter 910 may be configured to operate. For example, the equipment controller 1014 may set the battery power setpoint P sp,bat Accordingly, the power inverter 910 can either charge the battery unit 702 or discharge the battery unit 702.

[0136] Pump unit with battery and predictive control unit Referring here to Figures 11-12, a pump unit 1100 having a battery unit 1102 and a predictive pump controller 1104 is shown according to several embodiments. The pump unit 1100 may be configured to circulate fluid through an HVAC device 1134 via a fluid circuit 1138. The HVAC device 1134 may include, for example, a heating or cooling coil, an air handling unit, a rooftop unit, a heat exchanger, a refrigerator or freezer, a condenser or evaporator, a cooling tower, or any other type of system or device that receives fluid in the HVAC system. In some embodiments, the pump 1132 receives fluid (e.g., inlet water 1116) via an inlet water pipe 1112 and outputs fluid (e.g., outlet water 1118) via an outlet water pipe 1114.

[0137] In some embodiments, the battery unit 1102 includes one or more battery cells 1106. The battery cells 1106 are configured to store and release electrical energy (i.e., electricity). In some embodiments, the battery unit 1102 is charged using electricity from an external energy grid (e.g., provided by an electric utility). The electricity stored in the battery unit 1102 can be released to power one or more powered components of the pump unit 1100 (e.g., pump 1132). Advantageously, the battery unit 1102 allows the pump unit 1100 to draw electricity from the energy grid and charge the battery unit 1102 when energy prices are low, and release the stored electricity when energy prices are high, thereby shifting the electrical load of the pump unit 1100 over time. In some embodiments, the battery unit 1102 has sufficient energy capacity to power the pump unit 1100 for approximately 4-6 hours when operating at maximum capacity, so that it can be utilized during periods of high energy costs and charged during periods of low energy costs.

[0138] As shown in Figure 12, the pump unit 1100 may include a fuel cell 1202. In some embodiments, the fuel cell 1202 is a fuel cell configured to generate electrical energy using chemical reactions. For example, the fuel cell 1202 may convert the chemical energy of hydrogen and an oxidizer (e.g., oxygen) into electricity through a pair of oxidation-reduction reactions. In other embodiments, the fuel cell 1202 is a hydrocarbon fuel cell that uses one or more of the following to generate electricity: diesel, methanol, natural gas, etc. The fuel cell 1202 may be controlled to generate electricity to boost grid energy or other energy sources, to supplement battery discharge during high-energy-cost periods, or to generate electricity (e.g., during high-energy-cost periods) to charge the battery. The fuel cell may require fuel replacement (e.g., hydrogen supply), which may be purchased periodically and added to the pump unit 1100. In embodiments in which the pump unit 1100 includes a fuel cell 1202, the control and optimization processes described herein are configured to take into account the power contribution of the fuel cell 1202 and the cost of operating the fuel cell 1202 when generating control outputs for various components of the pump unit 1100 including the fuel cell 1202. For example, optimization performed by the predictive pump controller 1104 may determine whether to operate the fuel cell 1202 to generate electricity for each time step of the optimization period.

[0139] In some embodiments, the predictive pump controller 1104 performs an optimization process to determine whether to charge or discharge the battery unit 1102 during each of several time steps that occur during the optimization period. The predictive pump controller 1104 may use weather and price data 1110 to predict the amount of heating / cooling and the cost of electricity required during each of the several time steps. The predictive pump controller 1104 can optimize an objective function that takes into account the cost of electricity purchased from the energy grid over the duration of the optimization period. During each time step, the predictive pump controller 1104 can determine the amount of electricity to purchase from the energy grid and the amount of electricity to store or release from the battery unit 1102. The objective function and optimization performed by the predictive pump controller 1104 are described in more detail with reference to Figures 13-14.

[0140] Predictive pump control system Referring now to Figure 13, block diagrams of predictive pump control systems 1300 according to several embodiments are shown. Some of the components shown in the control system 1300 may be part of a pump unit 1100. For example, the pump unit 1100 may include a pump 1132, a battery unit 1102, a predictive pump controller 1104, a power inverter 1310, and a power junction 1312.

[0141] The power inverter 1310 may be configured to convert electrical power between direct current (DC) and alternating current (AC). For example, the battery unit 1102 may be configured to store and output DC power, while the energy grid 1314 and pump 1132 may be configured to consume and supply AC power. The power inverter 1310 may be used to convert the DC power from the battery unit 1102 into a sinusoidal AC output synchronized to the grid frequency of the energy grid 1314 and / or pump 1132. The power inverter 1310 may also be used to convert the AC power from the energy grid 1314 into DC power that can be stored in the battery unit 1102. The power output of the battery unit 1102 is P bat It is shown as P bat This value can be positive when the battery unit 1102 is supplying power to the power inverter 1310 (i.e., the battery unit 1102 is discharging), or negative when the battery unit 1102 is receiving power from the power inverter 1310 (i.e., the battery unit 1102 is charging).

[0142] In some examples, the power inverter 1310 receives a DC power output from the battery unit 1102 and converts the DC power output into an AC power output that can be supplied to the pump 1132. The power inverter 1310 may use a local oscillator to synchronize the frequency of the AC power output with the frequency of the energy grid 1314 (e.g., 50 Hz or 60 Hz) and may limit the voltage of the AC power output so that it does not exceed the grid voltage. In some embodiments, the power inverter 1310 is a resonant inverter that includes or uses an LC circuit to remove harmonics from a simple square wave in order to achieve a sine wave that matches the frequency of the energy grid 1314. In various embodiments, the power inverter 1310 may operate with or without a transformer, using a high-frequency transformer or a low-frequency transformer. A low-frequency transformer may directly convert the DC output from the battery unit 1102 into an AC output that can be supplied to the pump 1132. The high-frequency transformer may employ a multi-stage process, which involves converting the DC output to high-frequency AC, then back to DC, and finally to an AC output supplied to the pump 1132.

[0143] The power junction 1312 is the point where the pump 1132, the energy grid 1314, and the power inverter 1310 are electrically connected. The power supplied from the power inverter 1310 to the power junction 1312 is P bat It is shown as P bat This can be positive when the power inverter 1310 is supplying power to the power junction 1312 (i.e., the battery unit 1102 is discharging), or negative when the power inverter 1310 is receiving power from the power junction 1312 (i.e., the battery unit 1102 is charging). The power supplied from the energy grid 1314 to the power junction 1312 is P grid It is shown as P bat and P grid These are combined at power junction 1312, P total (that is, P total =P grid +P bat) forms P total P can be defined as the power supplied from power junction 1312 to pump 1132. In some examples, P total P grid It is larger than. For example, when the battery unit 1102 is discharging, P bat This can be positive, and this is P bat and P grid and are combined to form P total When forming, grid power P grid It is added to P. total P grid It may be less than. For example, when the battery unit 1102 is charging, P bat It can be negative, and this is P bat and P grid and are combined to form P total When forming, grid power P grid It is subtracted from.

[0144] The predictive pump controller 1104 may be configured to control the pump 1132 and the power inverter 1310. In some embodiments, the predictive pump controller 1104 controls the battery power setpoint P sp,bat It generates and provides it to the power inverter 1310. Battery power setpoint P sp,bat This may include a positive or negative power value (e.g., kW), which corresponds to the battery power setpoint P. sp,bat In order to achieve this, (P sp,bat (If negative) the power inverter 1310 is instructed to charge the battery unit 1102 using the power available at the power junction 1312, or (P sp,bat If the condition is positive, the battery unit 1102 is discharged to supply power to the power junction 1312.

[0145] In some embodiments, the predictive pump controller 1104 generates and provides control signals to the pump 1132. The predictive pump controller 1104 may use multi-stage optimization techniques to generate control signals. For example, the predictive pump controller 1104 may include an economic controller configured to determine the optimal amount of energy to be consumed by the pump 1132 at each time step during the optimization period. The optimal amount of energy to be consumed may minimize a cost function that takes into account the cost of energy consumed by the pump unit 1100. The cost of energy may be based on the time-varying energy price from the electric utility 1318. In some embodiments, the predictive pump controller 1104 determines the optimal amount of energy to purchase from the energy grid 1314 at each of a plurality of time steps (i.e., the grid power setpoint P sp,grid ) and the optimal amount of energy to store or release from the battery unit 1102 (i.e., battery power setpoint P sp,bat The predictive pump controller 1104 can monitor the actual power consumption of the pump 1132 and use the actual power consumption as a feedback signal when generating the optimal power setpoint.

[0146] The predictive pump controller 1104 sets a flow setpoint that achieves the optimal power consumption at each time step. sp and differential pressure setpoint DP sp This may include a tracking controller configured to generate the following. In some embodiments, the predictive pump controller 1104 uses an instrument model for the pump 1132 to determine the amount of fluid flow and / or differential pressure generated by the pump 1132 based on optimal power consumption.

[0147] In some embodiments, the predictive pump controller 1104 sets the flow setpoint Flow sp and differential pressure setpoint DP spThe control signals for the pump 1132 are generated using the Flow control signal. The control signals may include on / off commands, speed setpoints, or other types of setpoints that affect the operation of the pump 1132. In other embodiments, the control signals may include flow setpoints generated by the predictive pump controller 1104. sp and differential pressure setpoint DP sp This may include the following. The setpoint may be provided to the pump 1132 or the local controller of the pump 1132 that operates to achieve the setpoint. For example, the local controller for the pump 1132 may receive measurements of the differential pressure DP across the pump 1132 from one or more pressure sensors, and / or measurements of the fluid flow caused by the pump 1132 from one or more flow sensors. The local controller may use a feedback control process (e.g., PID, ESC, MPC, etc.) to increase or decrease the speed of the pump 1132 and drive the measured fluid flow rate and / or differential pressure to the setpoint. The multi-stage optimization performed by the predictive pump controller 1104 is described in more detail with reference to Figure 14.

[0148] Predictive pump controller Referring here to Figure 14, a block diagram illustrating a predictive pump controller 1104 in more detail according to an exemplary embodiment is shown. The predictive pump controller 1104 is shown to include a communication interface 1402 and a processing circuit 1404. The communication interface 1402 can facilitate communication between the controller 1104 and an external system or device. For example, the communication interface 1402 may receive measurements of fluid flow from a flow sensor 1416, measurements of differential pressure DP across the pump 1132 from a pressure sensor 1418, and measurements of power consumption of the pump 1132. In some embodiments, the communication interface 1402 receives measurements of the state of charge (SOC) of the battery unit 1102, which may be provided as a percentage of the maximum battery capacity (i.e., battery %). The communication interface 1402 may receive weather forecasts from a weather service 916 and predicted energy costs and demand costs from an electrical utility 1318. In some embodiments, the predictive pump controller 1104 provides control signal pumps 1132 and power inverters 1310 using a communication interface 1402.

[0149] The communication interface 1402 may include wired or wireless communication interfaces (e.g., jacks, antennas, transmitters, receivers, transceivers, wired terminals, etc.) for data communication with external systems or devices. In various embodiments, communication may be direct (e.g., local wired or wireless communication) or via a communication network (e.g., a WAN, the Internet, a cellular network, etc.). For example, the communication interface 1402 may include Ethernet cards and ports for transmitting and receiving data over an Ethernet-based communication link or network. In another example, the communication interface 1402 may include Wi-Fi transceivers for communication over a wireless communication network or cellular or mobile phone communication transceivers.

[0150] The processing circuit 1404 is shown to include a processor 1406 and a memory 1408. The processor 1406 may be a general-purpose or specific-purpose processor, an application-specific integrated circuit (ASIC), one or more field-programmable gate arrays (FPGAs), a group of processing components, or other suitable electronic processing components. The processor 1406 is configured to execute computer code or instructions stored in the memory 1408 or received from other computer-readable media (e.g., a CD-ROM, network storage, remote server, etc.).

[0151] Memory 1408 may include one or more devices (e.g., memory units, memory devices, storage devices, etc.) for storing data and / or computer code to complete and / or facilitate the various processes described herein. Memory 1408 may include random access memory (RAM), read-only memory (ROM), hard drive storage, temporary storage, non-volatile memory, flash memory, optical memory, or any other suitable memory for storing software objects and / or computer instructions. Memory 1408 may include database components, object code components, script components, or any other type of information structure to support the various activities and information structures described herein. Memory 1408 may be communicably connected to processor 1406 via processing circuit 1404 and may include computer code for executing one or more processes described herein (e.g., by processor 1406). When the processor 1406 executes instructions stored in memory 1408 to complete the various activities described herein, the processor 1406 generally configures the controller 1104 (and more specifically, the processing circuit 1404) to complete such activities.

[0152] Referring further to Figure 14, the predictive pump controller 1104 is shown to include an economy controller 1410, a tracking controller 1412, and an equipment controller 1414. Controllers 1410-1414 may be configured to perform a multi-state optimization process to generate control signals for the power inverter 1310 and the pump 1132. Briefly, the economy controller 1410 optimizes the predictive cost function to determine the optimal amount of electricity to purchase from the energy grid 1314 at each time step of the optimization period (i.e., the grid power setpoint P sp,grid ), the optimal amount of energy to store or release from the battery unit 1102 (i.e., the battery power setpoint P sp,bat ), and / or the optimal amount of power to be consumed by pump 1132 (i.e., pump power setpoint P sp,pump The tracking controller 1412 can determine the optimal power setpoint P. sp,grid , P sp,bat , and / or P sp,pump Use the optimal flow setpoint Flow sp, Pressure setting point DP sp, and the optimal battery charge / discharge speed (i.e., Bat C / D The device controller 1414 can determine the optimal setpoint Flow. sp and / or DP sp This can be used to generate control signals for pump 1132 that drive the actual (e.g., measured) flow rate and / or pressure DP to a set point (e.g., using a feedback control technique). Each of the controllers 1410-1414 is described in detail below.

[0153] Economic Controller The economic controller 1410 optimizes the predicted cost function to determine the optimal amount of electricity to purchase from the energy grid 1314 at each time step of the optimization period (i.e., the grid power setpoint P). sp,grid ), the optimal amount of energy to store or release from the battery unit 1102 (i.e., the battery power setpoint P sp,bat), and / or the optimal amount of power to be consumed by pump 1132 (i.e., pump power setpoint P sp,pump ) can be configured to determine. An example of a forecast cost function that can be optimized by the economic controller 1410 is shown in the following equation.

number

[0154] The first term of the predicted cost function J represents the cost of electricity consumed by pump 1132 over the duration of the optimization period. The parameter C at each time step k ec The value of (k) can be defined by the energy cost information provided by the electrical utility 1318. In some embodiments, the cost of electricity changes as a function of time, and this is different C at different time steps k. ec The variable P yields the value of (k). pump (k) is a decision variable that can be optimized by the economic controller 1410.

[0155] The second term of the forecast cost function J represents demand charges. Demand charges are additional charges imposed by several utility providers based on the maximum power consumption during the relevant demand charging period. For example, the demand charging rate C DC The demand charge rate C may be specified in US dollars per unit of electricity (e.g., $ / kW) and can be calculated by multiplying it by the peak electricity usage (e.g., kW) during the demand charge period. In the forecast cost function J, the demand charge rate C DC This can be defined by demand cost information received from the electric utility 1318. Variable P grid (k) is the peak power usage max(P) that occurs during the demand-charging period. grid (k)) is a decision variable that can be optimized by the economic controller 1410 to reduce load shift. Load shift can allow the economic controller 1410 to smooth out instantaneous spikes in the electricity demand of pump unit 1100 by storing energy in battery unit 1102 when the power consumption of pump 1132 is low. The stored energy can be used to draw peak power from energy grid 1314 when the power consumption of pump 1132 is high. grid To reduce the consumption and thereby decrease the resulting demand charges, the battery unit 1102 may release the emissions.

[0156] The last term of the predicted cost function J represents the cost savings resulting from the use of battery unit 1102. Unlike the preceding terms of the cost function J, the last term is subtracted from the total cost. Parameter C at each time step k ec The value of (k) can be defined by the energy cost information provided by the electrical utility 1318. In some embodiments, the cost of electricity changes as a function of time, and this is different C at different time steps k. ec The variable P yields the value of (k). bat (k) is a decision variable that can be optimized by the economic controller 1410. bat A positive value for (k) indicates that the battery unit 1102 is discharging, and P batA negative value for (k) indicates that the battery unit 1102 is charging. Power P is the power emitted from the battery unit 1102. bat (k) is the total power consumption P of pump 1132. total (k) may be used to satisfy some or all of the amount of electricity P purchased from the energy grid 1314. grid (k) Reduce (i.e., P grid (k=P) total (k)-P bat (k)). However, when the battery unit 1102 is charged, the total amount of energy P purchased from the energy grid 1314 is grid P is added to (k) bat A negative value for (k) is obtained.

[0157] The economic controller 1410 can optimize the predicted cost function J over the duration of the optimization period to determine the optimal value of the decision variable at each time step during the optimization period. In some embodiments, the optimization period has a duration of approximately one day, and each time step is approximately 15 minutes. However, the duration of the optimization period and time steps may vary in other embodiments and may be adjusted by the user. Advantageously, the economic controller 1410 can use the battery unit 1102 to implement load shifting by drawing electricity from the energy grid 1314 when energy prices are low and / or when the power consumed by the pump 1132 is low. The electricity is stored in the battery unit 1102 and can be released later when energy prices are high and / or the power consumption of the pump 1132 is high. This allows the economic controller 1410 to reduce the cost of electricity consumed by the pump unit 1100, smooth out instantaneous spikes in the electricity demand of the pump unit 1100, and reduce the resulting demand charges.

[0158] The economic controller 1410 may be configured to impose constraints on the optimization of the predicted cost function J. In some embodiments, the constraints include constraints on the flow rate and / or differential pressure DP produced by the pump 1132. The economic controller 1410 also imposes constraints on the minimum flow boundary flow. min and the maximum flow boundary max Between (that is, Flow min ≤Flow≦Flow max ) can be configured to always maintain the actual or predicted flow rate. Parameter Flow min and Flow max This can vary over time to define different flow ranges at different times. Similarly, the economic controller 1410 has a minimum pressure boundary DP. min and maximum pressure boundary DP max Between (i.e., DP min ≤DP ≤DP max ) can be configured to always maintain the actual or predicted pressure DP. Parameter DP min and DP max This can vary over time to define different flow rate ranges at different times.

[0159] In addition to constraints on fluid flow rate and differential pressure DP, the economic controller 1410 can impose constraints on the state of charge (SOC) and charge / discharge rate of the battery unit 1102. In some embodiments, the economic controller 1410 generates and imposes the following power constraints on the predicted cost function J. P bat ≤P rated -P bat ≤P rated In the formula, P bat P is the amount of power released from the battery unit 1102. rated This is the rated battery power of the battery unit 1102 (e.g., the maximum speed at which the battery unit 1102 can be charged or discharged). These power constraints apply when the battery unit 1102 is rated P rated We guarantee that the battery will not be charged or discharged at a speed exceeding its maximum possible charge / discharge rate.

[0160] In some embodiments, the economic controller 1410 generates and imposes one or more capacity constraints on the predicted cost function J. The capacity constraints are the battery power P charged or discharged during each time step. bat This can be used to relate the capacity and SOC of the battery unit 1102. The capacity constraints can ensure that the capacity of the battery unit 1102 is maintained within acceptable lower and upper limits at each time step of the optimization period. In some embodiments, the economy controller 1410 generates the following capacity constraints: C a (k)-P bat (k)Δt≦C rated C a (k)-P bat (k)Δt≧0 In the formula, C a (k) is the available battery capacity (e.g., kWh) at the start of time step k, and P bat (k) is the rate at which the battery unit 1102 is discharged during time step k (e.g., kW), Δt is the duration of each time step, and C rated This is the maximum rated capacity (e.g., kWh) of the battery unit 1102. Item P bat (k)Δt represents the change in battery capacity during time step k. These capacity constraints apply when the capacity of battery unit 1102 is zero and when it is at its maximum rated capacity C. rated Ensure that it is maintained between them.

[0161] In some embodiments, the economic controller 1410 generates and imposes one or more capacity constraints on the operation of the pump 1132. For example, the pump 1132 has a maximum power consumption P pump,max It may have a corresponding maximum operating point (e.g., maximum pump speed, maximum differential pressure, etc.). The economic controller 1410 supplies power P to the pump 1132 as shown in the following equation. pump Set to zero and maximum power consumption P pump,max It can be configured to generate constraints that limit the relationship between the two. 0≦Ppump ≤P pump,max P pump =P sp,grid +P sp,bat In the formula, the total power P supplied to pump 1132 pump This is the grid power setpoint P sp,grid and battery power setpoint P sp,bat It is the sum of the two.

[0162] The economic controller 1410 optimizes the constrained predicted cost function J to determine the decision variable P pump , P grid , and P bat The optimal value of can be determined, where P pump =P bat +P grid In some embodiments, the economic controller 1410 is P pump , P bat , and / or P grid Using the optimal value for , power setpoints are generated for the tracking controller 1412. The power setpoints are the battery power setpoints P for each time step k during the optimization period. sp,bat , grid power setpoint P sp,grid , and / or pump power setpoint P sp,pump This may include the following: The economic controller 1410 can provide power setpoints to the tracking controller 1412.

[0163] Tracking controller The tracking controller 1412 controls the optimal power setpoint P generated by the economic controller 1410. sp,grid , P sp,bat , and / or P sp,pump Use the optimal flow setpoint Flow sp, Optimal pressure setting point DP sp, and the optimal battery charging or discharging speed (i.e., Bat C / D ) can be determined. In some embodiments, the tracking controller 1412 can determine the power setpoint P for the pump 1132. sp,pump Flow setting point that is expected to achieve spand / or pressure setpoint DP sp In other words, the tracking controller 1412 generates the optimal amount of power P determined by the economic controller 1410. pump Flow setting point for supplying to pump 1132 sp and / or pressure setpoint DP sp It can generate.

[0164] In some embodiments, the tracking controller 1412 controls the battery power setpoint P sp,bat Use the optimal charging or discharging speed for the battery unit 1102. C / D Determine the following: For example, the battery power setpoint P sp,bat This can define a power value (kW) that can be converted by the tracking controller 1412 into a control signal for the power inverter 1310 and / or the equipment controller 1414. In other embodiments, the battery power setpoint P sp,bat This is supplied directly to the power inverter 1310, and battery power P bat It is used by the power inverter 1310 to control it.

[0165] Device controller The device controller 1414 controls the optimal flow setpoint generated by the tracking controller 1412. sp and / or pressure setpoint DP sp The control signals for pump 1132 can be generated using the instrument controller 1414. The control signals generated by the instrument controller 1414 can drive the actual (e.g., measured) flow rate and pressure DP to set points. The instrument controller 1414 can generate control signals for pump 1132 using any of a variety of control techniques. For example, the instrument controller 1414 can generate control signals for pump 1132 using a state-based algorithm, an extreme value search control (ESC) algorithm, a proportional-integral (PI) control algorithm, a proportional-integral-derivative (PID) control algorithm, a model predictive control (MPC) algorithm, or other feedback control algorithm.

[0166] The control signals may include on / off commands, speed commands for pump 1132, power commands for pump 1132, or other types of operation commands for pump 1132. In other embodiments, the control signals may include flow setpoints generated by the predictive pump controller 1104. sp and / or pressure setpoint DP sp This may include the following. The setpoint may be provided to the pump 1132 or a local controller of the pump 1132 that operates to achieve the setpoint. For example, the local controller for the pump 1132 may receive a measured fluid flow rate from the flow sensor 1416 and / or a measured differential pressure DP from the pressure sensor 1418, and may adjust the speed of the pump 1132 to drive the measured flow rate and / or pressure to the setpoint.

[0167] In some embodiments, the equipment controller 1414 is configured to provide control signals to the power inverter 1310. The control signals provided to the power inverter 1310 are directed to the battery power setpoint P sp,bat and / or optimal charge / discharge rate Bat C / D This may include the following. The device controller 1414 controls the battery power setpoint P sp,bat To achieve this, the power inverter 1310 may be configured to operate. For example, the equipment controller 1414 may set the battery power setpoint P sp,bat Accordingly, the power inverter 1310 can either charge the battery unit 1102 or discharge the battery unit 1102.

[0168] Cooling tower equipped with battery unit and predictive control unit Referring here to Figure 15, several embodiments of a cooling tower system 1500 are shown. The system 1500 is shown to include a cooling tower 1512 and a battery unit 1502 having a predictive cooling tower controller 1504. The cooling tower 1512 may be configured to provide cooling to a cooling load 1522. The cooling load 1522 may include, for example, a building zone, a supply air stream flowing through an air duct, airflow in an air processing unit or rooftop unit, fluid flowing through a heat exchanger, a refrigerator or freezer, a condenser or evaporator, a cooling coil, or any other type of system, device, or space requiring cooling. In some embodiments, a pump 1516 circulates a chilled fluid to the cooling load 1522 via a cooling tower circuit 1532. The chilled fluid can absorb heat from the cooling load 1522, thereby providing cooling to the cooling load 1522 and warming the chilled fluid.

[0169] The cooling tower 1512 may be configured to cool the water in the cooling tower circuit 1532 by transferring heat from the water to the outside air. The cooling tower 1512 may include a fan 1514 that causes cooling air to flow through the cooling tower 1512. The cooling tower 1512 places the cold air into a heat exchange relationship with the warmer water, thereby transferring heat from the warmer water to the colder air. The cooling tower circuit 1532 is shown and described as circulating water, but it should be understood that any type of coolant or working fluid (e.g., water, glycol, CO2, etc.) may be used in the cooling tower circuit 1532.

[0170] Referring further to Figure 15, the system 1500 is shown to include a battery unit 1502. In some embodiments, the battery unit 1502 includes one or more photovoltaic (PV) panels 1508. A PV panel 1508 may include an assembly of photovoltaic cells. Photovoltaic cells are configured to convert solar energy (i.e., sunlight) into electricity using photovoltaic materials such as monocrystalline silicon, polycrystalline silicon, amorphous silicon, cadmium telluride, copper indium gallenium / sulfide selenide, or other materials that exhibit a photovoltaic effect. In some embodiments, the photovoltaic cells are contained within a packaged assembly that forms the PV panel 1508. Each PV panel 1508 may include a plurality of linked photovoltaic cells. The PV panels 1508 can be combined to form a photovoltaic array.

[0171] In some embodiments, the PV panel 1508 is configured to maximize solar energy collection. For example, the battery unit 1502 may include a solar tracker (e.g., a GPS tracker, a sunshine sensor, etc.) that adjusts the angle of the PV panel 1508 so that it is directly facing the sun throughout the day. The solar tracker may allow the PV panel 1508 to receive direct sunlight for a longer period of the day, potentially increasing the total amount of energy produced by the PV panel 1508. In some embodiments, the battery unit 1502 includes a collection of mirrors, lenses, or sun concentrators configured to guide and / or concentrate sunlight onto the PV panel 1508. The energy generated by the PV panel 1508 may be stored in a battery cell 1506 and / or used to power various components of the cooling tower 1512.

[0172] In some embodiments, the battery unit 1502 includes one or more battery cells 1506. The battery cells 1506 are configured to store and release electrical energy (i.e., electricity). In some embodiments, the battery unit 1502 is charged using electricity from an external energy grid (e.g., provided by an electric utility). The electricity stored in the battery unit 1502 can be released to power one or more powered components of the cooling tower 1512 (e.g., a fan 1514, a pump 1516, etc.). Advantageously, the battery unit 1502 allows the cooling tower 1512 to shift its electrical load over time by drawing electricity from the energy grid and charging the battery unit 1502 when energy prices are low, and releasing the stored electricity when energy prices are high. In some embodiments, the battery unit 1502 has sufficient energy capacity to power the cooling tower 1512 for approximately 4 to 6 hours when operating at maximum capacity, so that it can be utilized during high-energy-cost periods and charged during low-energy-cost periods.

[0173] In some embodiments, the predictive cooling tower controller 1504 performs an optimization process to determine whether to charge or discharge the battery unit 1502 during each of several time steps that occur during the optimization period. The predictive cooling tower controller 1504 may use weather and price data 1510 to predict the amount of heating / cooling and the cost of electricity required during each of the several time steps. The predictive cooling tower controller 1504 can optimize an objective function that takes into account the cost of electricity purchased from the energy grid over the duration of the optimization period. In some embodiments, the objective function also takes into account the cost of operating the various components of the cooling tower 1512 (e.g., the cost of natural gas used to fuel the boiler). During each time step, the predictive cooling tower controller 1504 can determine the amount of electricity to purchase from the energy grid and the amount of electricity to store or release from the battery unit 1502. The objective function and optimization performed by the predictive cooling tower controller 1504 are described in more detail with reference to Figures 16-17.

[0174] Predictive cooling tower control system Referring now to Figure 16, block diagrams of predictive cooling tower control systems 1600 according to several embodiments are shown. Some of the components shown in the control system 1600 may be part of a cooling tower 1512. For example, the cooling tower 1512 may include a powered cooling tower component 1602, a battery unit 1502, a fuel cell 1691, a predictive cooling tower controller 1504, a power inverter 1610, and a power junction 1612. The powered cooling tower component 1602 may include any component of the cooling tower 1512 that consumes power (e.g., electricity) during operation. For example, the powered cooling tower component 1602 is shown to include a cooling fan 1514 and a pump 1516.

[0175] The fuel cell 1691 is a fuel cell configured to generate electrical energy using chemical reactions. For example, the fuel cell 1691 may convert the chemical energy of hydrogen and an oxidizer (e.g., oxygen) into electricity through a pair of oxidation-reduction reactions. In other embodiments, the fuel cell 1691 is a hydrocarbon fuel cell that uses one or more of the following to generate electricity: diesel, methanol, natural gas, etc. The fuel cell 1691 may be controlled to generate electricity to augment grid energy or other energy sources, to supplement battery discharge during high-energy-cost periods, or to generate electricity (e.g., during high-energy-cost periods) to charge batteries. The fuel cell may require fuel replacement (e.g., hydrogen supply), which may be purchased periodically. In embodiments in which the cooling tower 1512 includes the fuel cell 1691, the control and optimization processes described herein are configured to take into account the contribution of the fuel cell 1691 and the cost of operating the fuel cell 1691 when generating control outputs for various components of the cooling tower 1512 including the fuel cell 1691. For example, the optimization performed by the predictive cooling tower controller 1504 may determine whether to operate the fuel cell 1691 to generate electricity for each time step of the optimization period.

[0176] The power inverter 1610 may be configured to convert electrical power between direct current (DC) and alternating current (AC). For example, the battery unit 1502 may be configured to store and output DC power, while the energy grid 1614 and the powered cooling tower component 1602 may be configured to consume and supply AC power. The power inverter 1610 may be used to convert DC power from the battery unit 1502 into a sinusoidal AC output synchronized to the grid frequency of the energy grid 1614 and / or the powered cooling tower component 1602. The power inverter 1610 may also be used to convert AC power from the energy grid 1614 into DC power that can be stored in the battery unit 1502. The power output of the battery unit 1502 is P bat It is shown as P batThis value can be positive when the battery unit 1502 is supplying power to the inverter 1610 (i.e., the battery unit 1502 is discharging), or negative when the battery unit 1502 is receiving power from the power inverter 1610 (i.e., the battery unit 1502 is charging).

[0177] In some examples, the power inverter 1610 receives a DC power output from the battery unit 1502 and converts the DC power output into an AC power output that can be provided to the powered cooling tower component 1602. The power inverter 1610 may use a local oscillator to synchronize the frequency of the AC power output with the frequency of the energy grid 1614 (e.g., 50 Hz or 60 Hz) and may limit the voltage of the AC power output so that it does not exceed the grid voltage. In some embodiments, the power inverter 1610 is a resonant inverter that includes or uses an LC circuit to remove harmonics from a simple square wave in order to achieve a sine wave that matches the frequency of the energy grid 1614. In various embodiments, the power inverter 1610 may operate with or without a transformer, using a high-frequency transformer or a low-frequency transformer. A low-frequency transformer may directly convert the DC output from the battery unit 1502 into an AC output that can be provided to the powered cooling tower component 1602. The high-frequency transformer may employ a multi-stage process, which involves converting the DC output to high-frequency AC, then back to DC, and finally to an AC output supplied to the powered cooling tower component 1602.

[0178] The power output of PV panel 1508 is P PV It is shown as follows: Power output P of PV panel 1508 PV The energy can be stored in the battery unit 1502 and / or used to supply power to the powered cooling tower component 1602. In some embodiments, the amount of energy P generated by the PV panel 1508 is PVThe system measures and provides an indicator of PV power to the predictive cooling tower controller 1504. For example, a PV panel 1508 is shown that provides the predictive cooling tower controller 1504 with an indicator of PV power percentage (i.e., PV%). The PV power percentage may represent the percentage of the maximum PV power that the PV panel 1508 is currently operating at.

[0179] The power junction 1612 is the point where the powered cooling tower components 1602, the energy grid 1614, the PV panels 1508, and the power inverter 1610 are electrically connected. The power supplied from the power inverter 1610 to the power junction 1612 is P bat It is shown as P bat This can be positive when the power inverter 1610 is supplying power to the power junction 1612 (i.e., the battery unit 1502 is discharging), or negative when the power inverter 1610 is receiving power from the power junction 1612 (i.e., the battery unit 1502 is charging). The power supplied from the energy grid 1614 to the power junction 1612 is P grid As shown, the power supplied from PV panel 1508 to power junction 1612 is P PV As shown, the power supplied from fuel cell 1891 is P FC It is shown as P bat , P PV , P FC and P grid These are combined at power junction 1612, P total (that is, P total =P grid +P bat +P PV +P FC ) forms P total P can be defined as the power supplied from the power junction 1612 to the powered cooling tower component 1602. In some examples, P total P grid It is larger than. For example, when the battery unit 1502 is discharging, P bat This can be positive, and this is Pbat and P PV P grid Combined with P total When forming, grid power P grid and PV power P PV It is added to P. total P grid It may be less than P. For example, when the battery unit 1502 is charging, bat It can be negative, and this is P bat , P PV , and P grid These are combined into P total When forming, grid power P grid and PV power P PV It is subtracted from.

[0180] The predictive cooling tower controller 1504 may be configured to control the powered cooling tower components 1602 and the power inverter 1610. In some embodiments, the predictive cooling tower controller 1504 controls the battery power setpoint P sp,bat It generates and provides it to the power inverter 1610. Battery power setpoint P sp,bat This may include a positive or negative power value (e.g., kW), which corresponds to the battery power setpoint P. sp,bat In order to achieve this, (P sp,bat (If negative) the power inverter 1610 is instructed to charge the battery unit 1502 using the power available at the power junction 1612, or (P sp,bat If the condition is positive, the battery unit 1502 is discharged to supply power to the power junction 1612. In some embodiments, the predictive cooling tower controller 1504 sets the fuel cell setpoint P sp,FC It generates and provides it to the fuel cell 1691.

[0181] In some embodiments, the predictive cooling tower controller 1504 generates and provides control signals to the powered cooling tower component 1602. The predictive cooling tower controller 1504 may use multi-stage optimization techniques to generate control signals. For example, the predictive cooling tower controller 1504 may include an economic controller configured to determine the optimal amount of electricity to be consumed by the powered cooling tower component 1602 at each time step during the optimization period. The optimal amount of electricity to be consumed may minimize a cost function that accounts for the cost of energy consumed by the cooling tower 1512. The cost of energy may be based on the time-varying energy price from the electric utility 1618. In some embodiments, the predictive cooling tower controller 1504 determines the optimal amount of electricity to purchase from the energy grid 1614 (i.e., the grid power setpoint P) at each of a plurality of time steps. sp,grid ) and the optimal amount of energy to store or release from the battery unit 1502 (i.e., battery power setpoint P sp,bat The predictive cooling tower controller 1504 can monitor the actual power consumption of the powered cooling tower components 1602 and use the actual power consumption as a feedback signal when generating the optimal power setpoint.

[0182] The predictive cooling tower controller 1504 may include a tracking controller configured to generate temperature setpoints that achieve optimal power consumption at each time step. The temperature setpoints are, for example, collection unit water temperature setpoint T. sp,sump (i.e., the water temperature set point in the collection unit 1518) and / or the condenser water temperature set point T sp,cond (i.e., a temperature setpoint for the hot water returning to the cooling tower 1512) may be included. In some embodiments, the predictive cooling tower controller 1504 uses an equipment model for the powered cooling tower component 1602 to determine the amount of cooling that can be generated by the cooling tower 1512 based on the optimal power consumption.

[0183] In some embodiments, the predictive cooling tower controller 1504 generates control signals for the powered cooling tower components 1602 using temperature setpoints. These control signals may include on / off commands, speed setpoints for the fan 1514, differential pressure or flow rate setpoints for the pump 1516, or other types of setpoints for individual devices of the powered cooling tower components 1602. In other embodiments, the control signals may include temperature setpoints generated by the predictive cooling tower controller 1504 (e.g., collection unit water temperature setpoint T). sp,sump , condenser water temperature set point T sp,cond This may include, for example. The temperature setpoint may be provided to the powered cooling tower component 1602 or the local controller of the powered cooling tower component 1602, which operates to achieve the temperature setpoint. For example, the local controller of fan 1514 receives the collection unit water temperature T from the collection unit water temperature sensor. cump Measurements of and / or condenser water temperature T from the condenser water temperature sensor cond The measured values ​​can be received. The local controller can use a feedback control process (e.g., PID, ESC, MPC, etc.) to increase or decrease the speed of fan 1514 and drive the measured temperature up to the temperature setpoint. A similar feedback control process can be used to control pump 1516. The multi-stage optimization performed by the predictive cooling tower controller 1504 is described in more detail with reference to Figure 17.

[0184] Predictive Cooling Tower Controller Referring now to Figure 17, a block diagram illustrating a predictive cooling tower controller 1504 in more detail according to an exemplary embodiment is shown. The predictive cooling tower controller 1504 is shown to include a communication interface 1702 and a processing circuit 1704. The communication interface 1702 can facilitate communication between the controller 1504 and an external system or device. For example, the communication interface 1702 can collect water temperature T from a temperature sensor 1716. sump and condenser water temperature T condThe communication interface 1702 can receive measurements of the power consumption of the powered cooling tower component 1602. In some embodiments, the communication interface 1702 receives measurements of the state of charge (SOC) of the battery unit 1502, which may be provided as a percentage of the maximum battery capacity (i.e., battery %). The communication interface 1702 can receive weather forecasts from the weather service 1616 and predicted energy costs and demand costs from the electric utility 1618. In some embodiments, the predictive cooling tower controller 1504 uses the communication interface 1702 to provide control signals to the powered cooling tower component 1602 and the power inverter 1610.

[0185] The communication interface 1702 may include wired or wireless communication interfaces (e.g., jacks, antennas, transmitters, receivers, transceivers, wired terminals, etc.) for data communication with external systems or devices. In various embodiments, communication may be direct (e.g., local wired or wireless communication) or via a communication network (e.g., a WAN, the Internet, a cellular network, etc.). For example, the communication interface 1702 may include Ethernet cards and ports for transmitting and receiving data over an Ethernet-based communication link or network. In another example, the communication interface 1702 may include Wi-Fi transceivers for communication over a wireless communication network or cellular or mobile phone communication transceivers.

[0186] The processing circuit 1704 is shown to include a processor 1706 and a memory 1708. The processor 1706 may be a general-purpose or specific-purpose processor, an application-specific integrated circuit (ASIC), one or more field-programmable gate arrays (FPGAs), a group of processing components, or other suitable electronic processing components. The processor 1706 is configured to execute computer code or instructions stored in the memory 1708 or received from other computer-readable media (e.g., CD-ROM, network storage, remote server, etc.).

[0187] Memory 1708 may include one or more devices (e.g., memory units, memory devices, storage devices, etc.) for storing data and / or computer code to complete and / or facilitate the various processes described herein. Memory 1708 may include random access memory (RAM), read-only memory (ROM), hard drive storage, temporary storage, non-volatile memory, flash memory, optical memory, or any other suitable memory for storing software objects and / or computer instructions. Memory 1708 may include database components, object code components, script components, or any other type of information structure to support the various activities and information structures described herein. Memory 1708 may be communicably connected to processor 1706 via processing circuit 1704 and may include computer code for executing one or more processes described herein (e.g., by processor 1706). When the processor 1706 executes instructions stored in memory 1708 to complete the various activities described herein, the processor 1706 generally configures the controller 1504 (and more specifically, the processing circuit 1704) to complete such activities.

[0188] Referring further to Figure 17, the predictive cooling tower controller 1504 is shown to include an economic controller 1710, a tracking controller 1712, and an equipment controller 1714. Controllers 1710-1714 may be configured to perform a multi-state optimization process to generate control signals for the power inverter 1610 and the powered cooling tower components 1602. Briefly, the economic controller 1710 optimizes the predictive cost function to determine the optimal amount of electricity to purchase from the energy grid 1614 at each time step of the optimization period (i.e., the grid power setpoint P sp,grid ), the optimal amount of energy to store or release from the battery unit 1502 (i.e., the battery power setpoint P sp,bat), and / or the optimal amount of power to be consumed by the powered cooling tower component 1602 (i.e., the cooling tower power setpoint P sp,total The tracking controller 1712 can determine the optimal power setpoint P. sp,grid , P sp,bat , and / or P sp,total Use this to determine the optimal temperature setpoint (for example, the water temperature setpoint T in the collection unit). sp,sump , condenser water temperature set point T sp,cond (etc.) and the optimal battery charging or discharging speed (i.e., Bat C / D The device controller 1714 can determine the optimal temperature setpoint T sp,zone , or T sp,chw Using the actual (e.g., measured) temperature T zone and / or T chw Control signals can be generated for the powered cooling tower component 1602 to drive it to a setpoint (for example, using a feedback control technique). Each of the controllers 1710 to 1714 is described in detail below.

[0189] Economic Controller The economic controller 1710 optimizes the predicted cost function to determine the optimal amount of electricity to purchase from the energy grid 1614 at each time step of the optimization period (i.e., the grid power setpoint P). sp,grid ), the optimal amount of energy to store or release from the battery unit 1502 (i.e., the battery power setpoint P sp,bat ), and / or the optimal amount of power to be consumed by the powered cooling tower component 1602 (i.e., the cooling tower power setpoint P sp,total ) can be configured to determine. An example of a forecast cost function that can be optimized by the economic controller 1710 is shown in the following equation.

number

[0190] The first and second terms of the predicted cost function J represent the cost of electricity consumed by the powered cooling tower component 1602 over the optimization period. The parameter C at each time step k ec The value of (k) can be defined by the energy cost information provided by the electric utility 1618. In some embodiments, the cost of electricity varies as a function of time, and this is different C at different time steps k. ec The variable P yields the value of (k). fan (k) and P cond (k) is a decision variable that can be optimized by the economic controller 1710. In some embodiments, the total power consumption P of the powered cooling tower components 1602 at time step k. total (k) is P fan (k) and P pump (k) is equal to the sum of (i.e., P total (k=P) fan (k) + P pump (k)). Therefore, in some embodiments, the first two terms of the prediction cost function are summation

number

[0191] The third term of the forecast cost function J represents demand charges. Demand charges are additional charges imposed by certain utilities based on maximum power consumption during the relevant demand charging period. For example, the demand charging rate C DC The demand charge rate C may be specified in US dollars per unit of electricity (e.g., $ / kW) and can be calculated by multiplying it by the peak electricity usage (e.g., kW) during the demand charge period. In the forecast cost function J, the demand charge rate C DC This can be defined by demand cost information received from the electric utility 1618. Variable P grid (k) is the peak power usage max(P) that occurs during the demand-charging period. grid (k)) is a decision variable that can be optimized by the economic controller 1710 to reduce load shift. Load shift can be used by the economic controller 1710 to smooth out instantaneous spikes in the electricity demand of the cooling tower 1512 by storing energy in the battery unit 1502 when the power consumption of the powered cooling tower component 1602 is low. The stored energy can be used to draw peak power P from the energy grid 1614 when the power consumption of the powered cooling tower component 1602 is high. grid To reduce the amount of electricity used and thereby decrease the resulting demand charges, it may be discharged from the battery unit 1502.

[0192] The last term of the predicted cost function J represents the cost savings resulting from the use of battery unit 1502. Unlike the preceding terms of the cost function J, the last term is subtracted from the total cost. Parameter C at each time step k ec The value of (k) can be defined by the energy cost information provided by the electric utility 1618. In some embodiments, the cost of electricity varies as a function of time, and this is different C at different time steps k. ec The variable P yields the value of (k). bat (k) is a decision variable that can be optimized by the economic controller 1710. batA positive value for (k) indicates that the battery unit 1502 is discharging, and P bat A negative value for (k) indicates that the battery unit 1502 is charging. Power P is the power emitted from the battery unit 1502. bat (k) is the total power consumption P of the powered cooling tower components 1602. total (k) can be used to satisfy some or all of the amount of electricity P purchased from the energy grid 1614. grid (k) (that is, P grid (k=P) total (k)-P bat (k)-P PV (k)) is reduced. However, when the battery unit 1502 is charged, the total amount of energy P purchased from the energy grid 1614 is reduced. grid P is added to (k) bat A negative value for (k) is obtained.

[0193] In some embodiments, the power P provided by the PV panel 1508 PV The power generated by PV panels does not incur any costs and is therefore not included in the predicted cost function J. However, the power P generated by PV panels 1508 is not included in the predicted cost function J. PV Using this, the total power consumption P of the powered cooling tower component 1602 total (k) can be partially or entirely satisfied, which is the amount of electricity P purchased from the energy grid 1614. grid (k) Reduce (i.e., P grid (k=P) total (k)-P bat (k)-P PV (k)). The amount of PV energy P generated during any time step k. PV This can be predicted by the economic controller 1710. Several techniques for predicting the amount of PV power generated by PV panels are described in U.S. Patent Application No. 15 / 247,869. (U.S. Patent Application Publication No. 2017 / 0104449) U.S. Patent Application No. 15 / 247,844 (U.S. Patent Application Publication No. 2017 / 0104337) , and U.S. Patent Application No. 15 / 247,788 (U.S. Patent Application Publication No. 2017 / 0102675)These patent applications are described in [the relevant document]. Each of these patent applications has a filing date of August 25, 2016, and the entire disclosure of each of these patent applications is incorporated herein by reference.

[0194] The economic controller 1710 can optimize the predicted cost function J over the duration of the optimization period to determine the optimal value of the decision variable at each time step during the optimization period. In some embodiments, the optimization period has a duration of approximately one day, and each time step is approximately 15 minutes. However, the duration of the optimization period and time steps may vary in other embodiments and may be adjusted by the user. Advantageously, the economic controller 1710 can use the battery unit 1502 to implement load shifting by drawing electricity from the energy grid 1614 when energy prices are low and / or when the power consumed by the powered cooling tower component 1602 is low. The electricity is stored in the battery unit 1502 and can be discharged later when energy prices are high and / or when the power consumption of the powered cooling tower component 1602 is high. This allows the economic controller 1710 to reduce the cost of electricity consumed by the cooling tower 1512, smooth out instantaneous spikes in the electricity demand of the cooling tower 1512, and reduce the resulting demand charges.

[0195] The economic controller 1710 may be configured to impose constraints on the optimization of the predicted cost function J. In some embodiments, the constraint is the temperature T of the collected water produced by the cooling tower 1512. sump This includes constraints on the actual or predicted temperature T. The economic controller 1710 uses the actual or predicted temperature T. zump The minimum temperature boundary T min and the maximum temperature boundary T max Between (that is, T min ≦T sump ≦T max ) can be configured to maintain at all times. Similarly, the economic controller 1710 can control the actual or predicted temperature T cond The minimum temperature boundary T min and the maximum temperature boundary Tmax Between (that is, T min ≦T cond ≦T max ) can be configured to always be maintained. Parameter T min and T max This can vary over time to define different temperature ranges at different points in time.

[0196] To ensure that the temperature constraint is met, the economic controller 1710 determines the temperature T as a function of the decision variable optimized by the economic controller 1710. sump and T cond This can be modeled. Several techniques for developing a temperature model and relating temperature to the decision variable of the predictive cost function J are described in U.S. Patent No. 9,436,179 granted on September 6, 2016, and U.S. Patent Application No. 14 / 694,633 filed on April 23, 2015. (U.S. Patent Application Publication No. 2016 / 0313751) , and U.S. Patent Application No. 15 / 199,910, filed on June 30, 2016. (U.S. Patent Application Publication No. 2018 / 0004171) This is described in more detail therein. The entire disclosures of each of these patents and patent applications are incorporated herein by reference.

[0197] temperature T sump and T cond In addition to the constraints on the above, the economic controller 1710 can impose constraints on the charge state (SOC) and charge / discharge rate of the battery unit 1502. In some embodiments, the economic controller 1710 generates and imposes the following power constraints on the predicted cost function J. P bat ≤P rated -P bat ≤P rated In the formula, P bat This is the amount of electricity released from the battery unit 1502, and P rated This is the rated battery power of the battery unit 1502 (e.g., the maximum speed at which the battery unit 1502 can be charged or discharged). These power constraints apply when the battery unit 1502 is rated P ratedWe guarantee that the battery will not be charged or discharged at a speed exceeding its maximum possible charge / discharge rate.

[0198] In some embodiments, the economic controller 1710 generates and imposes one or more capacity constraints on the predicted cost function J. The capacity constraints are the battery power P charged or discharged during each time step. bat This can be used to relate the capacity and SOC of the battery unit 1502. The capacity constraints can ensure that the capacity of the battery unit 1502 is maintained within acceptable lower and upper limits at each time step of the optimization period. In some embodiments, the economy controller 1710 generates the following capacity constraints: C a (k)-P bat (k)Δt≦C rated C a (k)-P bat (k)Δt≧0 In the formula, C a (k) is the available battery capacity (e.g., kWh) at the start of time step k, and P bat (k) is the rate at which the battery unit 1502 is discharged during time step k (e.g., kW), Δt is the duration of each time step, and C rated This is the maximum rated capacity (e.g., kWh) of the battery unit 1502. Item P bat (k)Δt represents the change in battery capacity during time step k. These capacity constraints apply when the capacity of battery unit 1502 is zero and when it is at its maximum rated capacity C. rated Ensure that it is maintained between them.

[0199] In some embodiments, the economic controller 1710 generates and imposes one or more capacity constraints on the operation of the powered cooling tower component 1602. For example, the powered cooling tower component 1602 has a maximum power consumption P total,max It may have a corresponding maximum operating point (e.g., maximum pump speed, maximum cooling capacity, etc.). The economic controller 1710 supplies power P to the powered cooling tower component 1602 as shown in the following equation. totalSet to zero and maximum power consumption P total,max It can be configured to generate constraints that limit the relationship between the two. 0≦P total ≤P total,max P total =P sp,grid +P sp,bat In the formula, the total power P supplied to the powered cooling tower component 1602 total This is the grid power setpoint P sp,grid and battery power setpoint P sp,bat It is the sum of the two.

[0200] The economic controller 1710 optimizes the predicted cost function J according to the constraints, and determines the decision variable P total , P fan , P pump , P grid , and P bat The optimal value of can be determined, where P total =P bat +P grid +P PV In some embodiments, the economic controller 1710 is P total , P bat , and / or P grid The optimal value is used to generate power setpoints for the tracking controller 1712. The power setpoints are the battery power setpoints P for each time step k during the optimization period. sp,bat , grid power setpoint P sp,grid , and / or cooling tower power setpoint P sp,total This may include the following: The economic controller 1710 can provide power setpoints to the tracking controller 1712.

[0201] Tracking controller The tracking controller 1712 controls the optimal power setpoint P generated by the economic controller 1710. sp,grid , P sp,bat , and / or P sp,total Use this to determine the optimal temperature setpoint (for example, the water temperature setpoint T in the collection unit). sp,sump , condenser water temperature set point T sp,cond(etc.) and the optimal battery charging or discharging speed (i.e., Bat C / D ) can be determined. In some embodiments, the tracking controller 1712 can determine the power setpoint P for the cooling tower 1512. sp,total The water temperature setpoint T of the collection unit is expected to achieve this. sp,sump and / or condenser water temperature setpoint T sp,cond In other words, the tracking controller 1712 outputs the optimal amount of power P determined by the economic controller 1710 to the cooling tower 1512. total The water temperature set point T in the collection unit consumes the water. sp,sump and / or condenser water temperature setpoint T sp,cond It can generate.

[0202] In some embodiments, the tracking controller 1712 uses a power consumption model to determine the power consumption of the cooling tower 1512, and the water temperature of the collection unit T sump and collection unit water temperature setpoint T sp,sump To associate with, for example, the tracking controller 1712 uses the model of the equipment controller 1714 to determine the water temperature of the collection unit T sump and collection unit water temperature setpoint T sp,sump The control operation performed by the equipment controller 1714 can be determined as a function of this. An example of such a zone regulation controller model is shown in the following equation. P total =f4(T sump , T sp,sump )

[0203] The function f4 can be identified from the data. For example, the tracking controller 1712 is P total and T sump Collect measurements of T sp,sump The corresponding value can be identified. The tracking controller 1712 is P total , T sump , and T sp,sump The collected values ​​can be used as training data to perform a system identification process and determine the function f4 that defines the relationships between such variables.

[0204] The tracking controller 1712 uses a similar model to determine the total power consumption P of the cooling tower 1512. total and condenser water temperature setpoint T sp,cond The relationship between the two can be determined. For example, the tracking controller 1712 can determine the condenser water temperature T cond and condenser water temperature setpoint T sp,cond The power consumption P of cooling tower 1512 is expressed as a function of total This can be defined. An example of such a model is shown in the following equation. P total =f5(T cond , T sp,cond )

[0205] The function f5 can be identified from the data. For example, the tracking controller 1712 is P total and T cond Collect measurements of T sp,cond The corresponding value can be identified. The tracking controller 1712 is P total , T cond , and T sp,cond The collected values ​​can be used as training data to perform a system identification process and determine the function f5 that defines the relationships between such variables.

[0206] The tracking controller 1712 is P total , T sp,sump , and T sp,cond Using the relationship between T sp,sump , and T sp,cond The value of can be determined. For example, the tracking controller 1712 takes P as input from the economic controller 1710. total The value (i.e., P sp,total ) can receive, T sp,sump and T sp,cond The corresponding value can be used to make a determination. The tracking controller 1712 is T sp,sump , and T sp,cond The value can be provided as an output to the device controller 1714.

[0207] In some embodiments, the tracking controller 1712 controls the battery power setpoint Psp,bat Use sp,bat to determine the optimal speed Bat for charging or discharging the battery unit 1502. C / D For example, the battery power set point P sp,bat may define a power value (kW) that can be converted by the tracking controller 1712 into a control signal for the power inverter 1610 and / or the equipment controller 1714. In other embodiments, the battery power set point P sp,bat is provided directly to the power inverter 1610 and is used by the power inverter 1610 to control the battery power P bat .

[0208] Equipment controller The equipment controller 1714 can generate a control signal for the mechanical cooling tower component 1602 using the optimal temperature set point T sp,sump or T sp,cond generated by the tracking controller 1712. The control signal generated by the equipment controller 1714 can drive the actual (e.g., measured) temperature T sump and / or T cond to the set point. The equipment controller 1714 can use any of a variety of control techniques to generate a control signal for the mechanical cooling tower component 1602. For example, the equipment controller 1714 can use a state-based algorithm, an extremum seeking control (ESC) algorithm, a proportional integral (PI) control algorithm, a proportional integral derivative (PID) control algorithm, a model predictive control (MPC) algorithm, or other feedback control algorithms to generate a control signal for the mechanical cooling tower component 1602.

[0209] The control signal may include an on / off command, a speed set point for the fan 1514, a pressure set point or a flow rate set point for the pump 1516, or other types of set points for individual devices of the mechanical cooling tower component 1602. In other embodiments, the control signal is a temperature set point generated by the predictive cooling tower controller 1504 (e.g., the collected water temperature set point T sp,sump , the condenser water temperature set point T sp,condmay include, etc. The temperature set point can be provided to the powered cooling tower component 1602 or the local controller of the powered cooling tower component 1602 that operates to achieve the temperature set point. For example, the local controller for the fan 1514 receives the measured value of the water temperature T sump of the collected water and / or the measured value of the condenser water temperature T cond and can adjust the speed of the fan 1514 to drive the measured temperature to the set point.

[0210] In some embodiments, the device controller 1714 is configured to provide a control signal to the power inverter 1610. The control signal provided to the power inverter 1610 may include the battery power set point P sp,bat , and / or the optimal charge / discharge rate Bat C / D . The device controller 1714 can be configured to operate the power inverter 1610 to achieve the battery power set point P sp,bat . For example, the device controller 1714 can charge or discharge the battery unit 1502 to the power inverter 1610 according to the battery power set point P sp,bat .

[0211] Valve unit with battery and predictive control unit Referring here to Figures 18-19, a valve unit 1800 having a battery unit 1802 and a predictive valve controller 1804 is shown according to several embodiments. The valve unit 1800 may be configured to control a valve 1832 via a valve actuator 1834. The valve 1832 may be a fluid control valve configured to control the flow rate of fluid from an inlet pipe 1812 to an outlet pipe 1814. The actuator 1834 may include a motor or other powered component configured to adjust the position of the valve 1832. In some embodiments, the valve unit 1800 is configured to control the flow of fluid through an HVAC device 1836 via a fluid circuit 1838. The HVAC device 1836 may include, for example, a heating or cooling coil, an air handling unit, a rooftop unit, a heat exchanger, a refrigerator or freezer, a condenser or evaporator, a cooling tower, or any other type of system or device that receives fluid in the HVAC system.

[0212] In some embodiments, the battery unit 1802 includes one or more battery cells 1806. The battery cells 1806 are configured to store and release electrical energy (i.e., electricity). In some embodiments, the battery unit 1802 is charged using electricity from an external energy grid (e.g., provided by an electric utility). The electricity stored in the battery unit 1802 can be released to power one or more powered components of the valve unit 1800 (e.g., actuator 1834). Advantageously, the battery unit 1802 allows the valve unit 1800 to time-shift its electrical load by drawing electricity from the energy grid and charging the battery unit 1802 when energy prices are low, and releasing the stored electricity when energy prices are high. In some embodiments, the battery unit 1802 has sufficient energy capacity to power the valve unit 1800 for approximately 4-6 hours when operating at maximum capacity, so that it can be utilized during periods of high energy cost and charged during periods of low energy cost.

[0213] As shown in Figure 19, the valve unit 1800 may include a fuel cell 1901. In some embodiments, the fuel cell 1901 is a fuel cell configured to generate electrical energy using chemical reactions. For example, the fuel cell 1901 may convert the chemical energy of hydrogen and an oxidizer (e.g., oxygen) into electricity through a pair of oxidation-reduction reactions. In other embodiments, the fuel cell 1901 is a hydrocarbon fuel cell that uses one or more of the following to generate electricity: diesel, methanol, natural gas, etc. The fuel cell 1901 may be controlled to generate electricity to boost grid energy or other energy sources, to supplement battery discharge during high-energy-cost periods, or to generate electricity (e.g., during high-energy-cost periods) to charge the battery. The fuel cell may require fuel replacement (e.g., hydrogen supply), which may be purchased periodically and added to the valve unit 1100. In embodiments in which the valve unit 1100 includes a fuel cell 1901, the control and optimization processes described herein are configured to take into account the contribution of the fuel cell 1901 and the cost of operating the fuel cell 1901 when generating control outputs for various components of the valve unit 1100 including the fuel cell 1901. For example, optimization performed by the predictive valve controller 1804 may determine whether to operate the fuel cell 1901 to generate electricity for each time step of the optimization period.

[0214] In some embodiments, the predictive valve controller 1804 performs an optimization process to determine whether to charge or discharge the battery unit 1802 during each of several time steps that occur during the optimization period. The predictive valve controller 1804 may use weather and price data 1810 to predict the amount of heating / cooling and the cost of electricity required during each of the several time steps. The predictive valve controller 1804 can optimize an objective function that takes into account the cost of electricity purchased from the energy grid over the optimization period. During each time step, the predictive valve controller 1804 can determine the amount of electricity to purchase from the energy grid and the amount of electricity to store or discharge from the battery unit 1802. The objective function and optimization performed by the predictive valve controller 1804 are described in more detail with reference to Figures 20-21.

[0215] Predictive valve control system Referring now to Figure 20, block diagrams of predictive valve control systems 2000 according to several embodiments are shown. Some of the components shown in the control system 2000 may be part of a valve unit 1800. For example, the valve unit 1800 may include an actuator 1834, a battery unit 1802, a predictive valve controller 1804, a power inverter 2010, and a power junction 2012.

[0216] The power inverter 2010 may be configured to convert electrical power between direct current (DC) and alternating current (AC). For example, the battery unit 1802 may be configured to store and output DC power, while the energy grid 2014 and actuator 1834 may be configured to consume and supply AC power. The power inverter 2010 may be used to convert DC power from the battery unit 1802 into a sinusoidal AC output synchronized to the grid frequency of the energy grid 2014 and / or actuator 1834. The power inverter 2010 may also be used to convert AC power from the energy grid 2014 into DC power that can be stored in the battery unit 1802. The power output of the battery unit 1802 is P bat It is shown as P bat This value can be positive when the battery unit 1802 is supplying power to the power inverter 2010 (i.e., the battery unit 1802 is discharging), or negative when the battery unit 1802 is receiving power from the power inverter 2010 (i.e., the battery unit 1802 is charging).

[0217] In some examples, the power inverter 2010 receives a DC power output from the battery unit 1802 and converts the DC power output into an AC power output that can be supplied to the actuator 1834. The power inverter 2010 may use a local oscillator to synchronize the frequency of the AC power output with the frequency of the energy grid 2014 (e.g., 50 Hz or 60 Hz) and may limit the voltage of the AC power output so that it does not exceed the grid voltage. In some embodiments, the power inverter 2010 is a resonant inverter that includes or uses an LC circuit to remove harmonics from a simple square wave in order to achieve a sine wave that matches the frequency of the energy grid 2014. In various embodiments, the power inverter 2010 may operate with or without a transformer, using a high-frequency transformer or a low-frequency transformer. A low-frequency transformer may directly convert the DC output from the battery unit 1802 into an AC output that can be supplied to the actuator 1834. The high-frequency transformer may employ a multi-stage process, which involves converting the DC output to high-frequency AC, then back to DC, and finally to an AC output supplied to the actuator 1834.

[0218] Power junction 2012 is the point where actuator 1834, energy grid 2014, and power inverter 2010 are electrically connected. The power supplied from power inverter 2010 to power junction 2012 is P bat It is shown as P bat This can be positive if the power inverter 2010 is supplying power to the power junction 2012 (i.e., the battery unit 1802 is discharging), or negative if the power inverter 2010 is receiving power from the power junction 2012 (i.e., the battery unit 1802 is charging). The power supplied from the energy grid 2014 to the power junction 2012 is P grid It is shown as P bat and P grid These were combined in Power Junction 2012, P total (that is, P total =P grid+P bat ) forms P total P can be defined as the power supplied from power junction 2012 to actuator 1834. In some examples, P total P grid It is larger than that. For example, when battery unit 1802 is discharging, P bat This can be positive, and this is P bat and P grid and are combined to form P total When forming, grid power P grid It is added to P. total P grid It may be less than P. For example, when the battery unit 1802 is charging, bat It can be negative, and this is P bat and P grid and are combined to form P total When forming, grid power P grid It is subtracted from.

[0219] The predictive valve controller 1804 may be configured to control the actuator 1834 and the power inverter 2010. In some embodiments, the predictive valve controller 1804 controls the battery power setpoint P sp,bat It generates and provides it to the power inverter 2010. Battery power setpoint P sp,bat This may include a positive or negative power value (e.g., kW), which corresponds to the battery power setpoint P. sp,bat In order to achieve this, (P sp,bat (If negative) The power inverter 2010 is instructed to charge the battery unit 1802 using the power available at the power junction 2012. or (P sp,bat (If positive) Discharge battery unit 1802 to supply power to power junction 2012.

[0220] In some embodiments, the predictive valve controller 1804 generates and provides control signals to the actuator 1834. The predictive valve controller 1804 may use multi-stage optimization techniques to generate control signals. For example, the predictive valve controller 1804 may include an economic controller configured to determine the optimal amount of energy to be consumed by the actuator 1834 at each time step during the optimization period. The optimal amount of energy to be consumed may minimize a cost function that accounts for the cost of the energy consumed by the valve unit 1800. The cost of energy may be based on the time-varying energy price from the electric utility 2018. In some embodiments, the predictive valve controller 1804 determines the optimal amount of energy to purchase from the energy grid 2014 (i.e., the grid power setpoint P) at each of a plurality of time steps. sp,grid ) and the optimal amount of energy to store or release from the battery unit 1802 (i.e., battery power setpoint P sp,bat The predictive valve controller 1804 can monitor the actual power consumption of the actuator 1834 and use the actual power consumption as a feedback signal when generating the optimal power setpoint.

[0221] The predictive valve controller 1804 may include a tracking controller configured to generate positioning points for the actuator 1834 that achieve the optimal power consumption at each time step. In some embodiments, the predictive valve controller 1804 uses an instrument model for the actuator 1834 to determine the position of the actuator 1834 corresponding to the optimal power consumption.

[0222] In some embodiments, the predictive valve controller 1804 generates a control signal for the actuator 1834 using a position setpoint. The control signal may include an on / off command, a position command, a voltage signal, or other types of setpoints that affect the operation of the actuator 1834. In other embodiments, the control signal may include a position setpoint generated by the predictive valve controller 1804. The setpoint may be provided to the actuator 1834 or a local controller for the actuator 1834 that operates to achieve the setpoint. For example, the local controller for the actuator 1834 may receive measured values ​​of the valve position from one or more position sensors. The local controller may use a feedback control process (e.g., PID, ESC, MPC, etc.) to adjust the position of the actuator 1834 and / or the valve 1832 and drive the measured position to the setpoint. The multi-stage optimization performed by the predictive valve controller 1804 is described in more detail with reference to Figure 21.

[0223] Predictive valve controller Referring here to Figure 21, a block diagram illustrating a predictive valve controller 1804 in more detail according to an exemplary embodiment is shown. The predictive valve controller 1804 is shown to include a communication interface 2102 and a processing circuit 2104. The communication interface 2102 can facilitate communication between the controller 1804 and an external system or device. For example, the communication interface 2102 may receive measurements of valve position from a position sensor 2118 and measurements of power consumption of the actuator 1834. In some embodiments, the communication interface 2102 receives measurements of the state of charge (SOC) of the battery unit 1802, which may be provided as a percentage of the maximum battery capacity (i.e., battery %). The communication interface 2102 may receive weather forecasts from a weather service 916 and predicted energy costs and demand costs from an electrical utility 2018. In some embodiments, the predictive valve controller 1804 uses the communication interface 2102 to provide control signals to the actuator 1834 and the power inverter 2010.

[0224] The communication interface 2102 may include wired or wireless communication interfaces (e.g., jacks, antennas, transmitters, receivers, transceivers, wired terminals, etc.) for data communication with external systems or devices. In various embodiments, communication may be direct (e.g., local wired or wireless communication) or via a communication network (e.g., a WAN, the Internet, a cellular network, etc.). For example, the communication interface 2102 may include Ethernet cards and ports for transmitting and receiving data over an Ethernet-based communication link or network. In another example, the communication interface 2102 may include Wi-Fi transceivers for communication over a wireless communication network or cellular or mobile phone communication transceivers.

[0225] The processing circuit 2104 is shown to include a processor 2106 and a memory 2108. The processor 2106 may be a general-purpose or specific-purpose processor, an application-specific integrated circuit (ASIC), one or more field-programmable gate arrays (FPGAs), a group of processing components, or other suitable electronic processing components. The processor 2106 is configured to execute computer code or instructions stored in the memory 2108 or received from other computer-readable media (e.g., CD-ROM, network storage, remote server, etc.).

[0226] Memory 2108 may include one or more devices (e.g., memory units, memory devices, storage devices, etc.) for storing data and / or computer code to complete and / or facilitate the various processes described herein. Memory 2108 may include random access memory (RAM), read-only memory (ROM), hard drive storage, temporary storage, non-volatile memory, flash memory, optical memory, or any other suitable memory for storing software objects and / or computer instructions. Memory 2108 may include database components, object code components, script components, or any other type of information structure to support the various activities and information structures described herein. Memory 2108 may be communicably connected to processor 2106 via processing circuit 2104 and may include computer code for executing one or more processes described herein (e.g., by processor 2106). When the processor 2106 executes instructions stored in memory 2108 to complete the various activities described herein, the processor 2106 generally configures the controller 1804 (and more specifically, the processing circuit 2104) to complete such activities.

[0227] Referring further to Figure 21, the predictive valve controller 1804 is shown to include an economy controller 2110, a tracking controller 2112, and an equipment controller 2114. Controllers 2110-2114 may be configured to perform a multi-state optimization process to generate control signals for the power inverter 2010 and actuator 1834. Briefly, the economy controller 2110 optimizes the predictive cost function to determine the optimal amount of electricity to purchase from the energy grid 2014 at each time step of the optimization period (i.e., the grid power setpoint P sp,grid ), the optimal amount of energy to store or release from the battery unit 1802 (i.e., the battery power setpoint P sp,bat ), and / or the optimal amount of power to be consumed by actuator 1834 (i.e., pump power setpoint P sp,act The tracking controller 2112 can determine the optimal power setpoint P. sp,grid , P sp,bat , and / or P sp,act Using the optimal position setting point Pos for valve 1832 sp and the optimal battery charge / discharge speed (i.e., Bat C / D The device controller 2114 can determine the optimal position setting point Pos sp This can be used to generate control signals for actuator 1834 that drive the actual (e.g., measured) position to a setpoint (e.g., using a feedback control technique). Each of the controllers 2110-2114 is described in detail below.

[0228] Economic Controller The economic controller 2110 optimizes the predicted cost function to determine the optimal amount of electricity to purchase from the energy grid 2014 at each time step of the optimization period (i.e., the grid power setpoint P). sp,grid ), the optimal amount of energy to store or release from the battery unit 1802 (i.e., the battery power setpoint P sp,bat ), and / or the optimal amount of power to be consumed by actuator 1834 (i.e., actuator power setpoint Psp,act ) can be configured to determine. An example of a forecast cost function that can be optimized by the economic controller 2110 is shown in the following equation.

number

[0229] The first term of the predicted cost function J represents the cost of the electricity consumed by actuator 1834 over the duration of the optimization period. The parameter C at each time step k ec The value of (k) can be defined by the energy cost information provided by the Electric Utility 2018. In some embodiments, the cost of electricity changes as a function of time, and this is different C at different time steps k. ec The variable P yields the value of (k). act (k) is a decision variable that can be optimized by the economic controller 2110.

[0230] The second term of the prediction cost function J represents the demand charge. The demand charge is an additional charge imposed by some utility providers based on the maximum power consumption during the corresponding demand charge period. For example, the demand charge rate C DC can be specified in dollars per unit of power (e.g., $ / kW), and the demand charge can be calculated by multiplying the peak power usage (e.g., kW) during the demand charge period. In the prediction cost function J, the demand charge rate C DC can be defined by the demand cost information received from the electric utility 2018. The variable P grid (k) is a decision variable that can be optimized by the economic controller 2110 to reduce the peak power usage max(P grid (k)) that occurs during the demand charge period. Load shifting can enable the economic controller 2110 to smooth out momentary spikes in the electrical demand of the valve unit 1800 by storing energy in the battery unit 1802 when the power consumption of the actuator 1834 is low. The stored energy can be released from the battery unit 1802 to reduce the peak power draw P grid from the energy grid 2014 when the power consumption of the actuator 1834 is high, thereby reducing the resulting demand charge.

[0231] The last term of the prediction cost function J represents the cost savings resulting from the use of the battery unit 1802. Unlike the previous terms of the cost function J, the last term is subtracted from the total cost. The value of the parameter C ec (k) at each time step k can be defined by the energy cost information provided by the electric utility 2018. In some embodiments, the cost of electricity varies as a function of time, which results in different values of C ec (k) at different time steps k. The variable P bat (k) is a decision variable that can be optimized by the economic controller 2110. A positive value of P bat (k) indicates that the battery unit 1802 is discharging, and P batA negative value for (k) indicates that the battery unit 1802 is charging. Power P is the power emitted from the battery unit 1802. bat (k) is the total power consumption P of the actuator 1834. total (k) can be used to satisfy some or all of the following: the amount of electricity P purchased from the energy grid 2014. grid (k) Reduce (i.e., P grid (k=P) total (k)-P bat (k)). However, when battery unit 1802 is charged, the total amount of energy P purchased from the energy grid 2014 grid P is added to (k) bat A negative value for (k) is obtained.

[0232] The economic controller 2110 can optimize the predicted cost function J over the duration of the optimization period to determine the optimal value of the decision variable at each time step during the optimization period. In some embodiments, the optimization period has a duration of approximately one day, and each time step is approximately 15 minutes. However, the duration of the optimization period and time steps may vary in other embodiments and may be adjusted by the user. Advantageously, the economic controller 2110 can use the battery unit 1802 to implement load shifting by drawing electricity from the energy grid 2014 when energy prices are low and / or when the power consumed by the powered actuator 1834 is low. The electricity is stored in the battery unit 1802 and can be discharged later when energy prices are high and / or when the power consumption of the actuator 1834 is high. This allows the economic controller 2110 to reduce the cost of electricity consumed by the valve unit 1800, smooth out instantaneous spikes in the electricity demand of the valve unit 1800, and reduce the resulting demand charges.

[0233] The economic controller 2110 may be configured to impose constraints on the optimization of the predicted cost function J. In some embodiments, the constraints include constraints on the position of the actuator 1834. The economic controller 2110 uses the actual or predicted position as a minimum position boundary Pos min and maximum position boundary Pos max Between (that is, Pos min ≦Pos≦Pos max ) can be configured to always be maintained. Parameter Pos min and Pos max This can vary over time to define different positional ranges at different times.

[0234] In addition to constraints on the position of valve 1832, the economy controller 2110 can impose constraints on the state of charge (SOC) and charge / discharge rate of battery unit 1802. In some embodiments, the economy controller 2110 generates and imposes the following power constraints on the predicted cost function J. P bat ≤P rated -P bat ≤P rated In the formula, P bat This is the amount of energy released from the battery unit 1802, and P rated This is the rated battery power of the battery unit 1802 (e.g., the maximum speed at which the battery unit 1802 can be charged or discharged). These power constraints apply when the battery unit 1802 is rated P rated We guarantee that the battery will not be charged or discharged at a speed exceeding its maximum possible charge / discharge rate.

[0235] In some embodiments, the economic controller 2110 generates and imposes one or more capacity constraints on the predicted cost function J. The capacity constraints are the battery power P charged or discharged during each time step. batThis can be used to relate the capacity and SOC of the battery unit 1802. The capacity constraints can ensure that the capacity of the battery unit 1802 is maintained within acceptable lower and upper limits at each time step of the optimization period. In some embodiments, the economy controller 2110 generates the following capacity constraints: C a (k)-P bat (k)Δt≦C rated C a (k)-P bat (k)Δt≧0 In the formula, C a (k) is the available battery capacity (e.g., kWh) at the start of time step k, and P bat (k) is the rate at which the battery unit 1802 is discharged during time step k (e.g., kW), Δt is the duration of each time step, and C rated This is the maximum rated capacity (e.g., kWh) of the battery unit 1802. Item P bat (k)Δt represents the change in battery capacity during time step k. These capacity constraints apply when the capacity of battery unit 1802 is zero and when it is at its maximum rated capacity C. rated Ensure that it is maintained between them.

[0236] In some embodiments, the economic controller 2110 generates and imposes one or more capacity constraints on the operation of the actuator 1834. For example, the actuator 1834 has a maximum power consumption P act,max It may have a corresponding maximum operating point (e.g., maximum operating speed, maximum position, etc.). The economic controller 2110 supplies power P to the actuator 1834 as shown in the following equation. act Set to zero and maximum power consumption P act,max It can be configured to generate constraints that limit the relationship between the two. 0≦P act ≤P act,max P act =P sp,grid +P sp,bat In the formula, the total power P supplied to actuator 1834act This is the grid power setpoint P sp,grid and battery power setpoint P sp,bat It is the sum of the two.

[0237] The economic controller 2110 optimizes the constrained predicted cost function J to determine the decision variable P act , P grid , and P bat The optimal value of can be determined, where P act =P bat +P grid In some embodiments, the economic controller 2110 is P act , P bat , and / or P grid Using the optimal value for , power setpoints are generated for the tracking controller 2112. The power setpoints are the battery power setpoints P for each time step k during the optimization period. sp,bat , grid power setpoint P sp,grid , and / or actuator power setpoint P sp,act This may include the following: The economic controller 2110 can provide power setpoints to the tracking controller 2112.

[0238] Tracking controller The tracking controller 2112 controls the optimal power setpoint P generated by the economic controller 2110. sp,grid , P sp,bat , and / or P sp,act Use the optimal positioning point Pos sp and the optimal battery charge / discharge rate (i.e., Bat C / D ) can be determined. In some embodiments, the tracking controller 2112 can determine the power setpoint P for the actuator 1834. sp,act Position setting point Pos sp It generates the optimal amount of power P determined by the economic controller 2110. In other words, the tracking controller 2112 generates the optimal amount of power P determined by the economic controller 2110. act Position setting point Pos that consumes power to actuator 1834 sp It can generate.

[0239] In some embodiments, the tracking controller 2112 controls the battery power setpoint P sp,bat Use the optimal charging or discharging speed for battery unit 1802. C / D Determine the following: For example, the battery power setpoint P sp,bat This can define a power value (kW) that can be converted by the tracking controller 2112 into a control signal for the power inverter 2010 and / or the equipment controller 2114. In other embodiments, the battery power setpoint P sp,bat This is directly supplied to the Power Inverter 2010, and battery power P bat It is used by the power inverter 2010 to control it.

[0240] Device controller The device controller 2114 sets the optimal positioning point Pos generated by the tracking controller 2112. sp The control signal for actuator 1834 can be generated using the device controller 2114. The control signal generated by the device controller 2114 can drive the actual (e.g., measured) position of the valve 1832 setpoint. The device controller 2114 can generate the control signal for actuator 1834 using any of a variety of control techniques. For example, the device controller 2114 can generate the control signal for actuator 1834 using a state-based algorithm, an extreme value search control (ESC) algorithm, a proportional-integral (PI) control algorithm, a proportional-integral-derivative (PID) control algorithm, a model predictive control (MPC) algorithm, or other feedback control algorithm.

[0241] The control signal may include an on / off command, a position command, a voltage signal, or other types of setpoints that affect the operation of the actuator 1834. In other embodiments, the control signal may include a position setpoint generated by the predictive valve controller 1804. The setpoint may be provided to the actuator 1834 or a local controller for the actuator 1834 that operates to achieve the setpoint. For example, the local controller for the actuator 1834 may receive measured values ​​of the valve position from one or more position sensors. The local controller may use a feedback control process (e.g., PID, ESC, MPC, etc.) to adjust the position of the actuator 1834 and / or the valve 1832 and drive the measured position to the setpoint.

[0242] In some embodiments, the equipment controller 2114 is configured to provide control signals to the power inverter 2010. The control signals provided to the power inverter 2010 control the battery power setpoint P sp,bat , and / or optimal charge / discharge rate Bat C / D It may include the following: The device controller 2114 controls the battery power setpoint P sp,bat To achieve this, the power inverter 2010 may be configured to operate. For example, the equipment controller 2114 sets the battery power setpoint P sp,bat Accordingly, the power inverter 2010 can either charge the battery unit 1802 or discharge the battery unit 1802.

[0243] Fuel cell optimization Referring here to Figure 22, a flowchart of process 2200 is shown relating to the optimization of the operation of fuel cells integrated with building equipment (e.g., fuel cell 808 in Figure 8, fuel cell 202 in Figure 12, fuel cell 1691 in Figure 16, or fuel cell 1901 in Figure 19) or included in modular energy units (e.g., see Figures 23-24, discussed below). Process 2200 may be performed by various predictive controllers or power management circuits described herein, e.g., predictive chiller controller 704, predictive pump controller 1104, predictive cooling tower controller 1504, etc. In other embodiments, the optimization process in Figure 22 may be performed on a standalone fuel cell and may be performed by a controller of the standalone fuel cell. In other embodiments, process 2200 may be performed by cloud-based optimization resources, as shown in Figures 29-30 and discussed below.

[0244] In step 2202, fuel cells integrated with building equipment (e.g., fuel cell 808 in Figure 8, fuel cell 202 in Figure 12, fuel cell 1691 in Figure 16, or fuel cell 1901 in Figure 19, as considered above) are monitored. For example, data on power generation, fuel level, fuel utilization rate, on / off status, etc., may be collected to monitor the fuel cells. For example, step 2202 may include collecting a time series of fuel cell data showing one or more fuel cell-related variables for each time step in the time series. Step 2202 may result in the collection and aggregation of a set of fuel cell training data showing past usage, fuel consumption, etc., of the fuel cells. In some embodiments, data on building equipment supplied by the fuel cells, building conditions affected by the operation of the building equipment, or other variables that may affect the demand for fuel cells (e.g., weather, building schedule, etc.) may be collected as part of monitoring the fuel cells in step 2202.

[0245] In step 2204, the future usage of the fuel cells is predicted. The future usage of the fuel cells can be predicted based on monitoring of the fuel cells from step 2202, for example, based on a model trained on time-series data collected by monitoring the fuel cells in step 2202. For example, the data collected in step 2202 can be used to train a neural network that predicts the future usage of the fuel cells based on the set of inputs represented in the data collected in step 2202. For example, weather forecast data, building setting points, time of day, current usage, etc., can be used to predict the future usage of the fuel cells. The future usage of the fuel cells can be predicted in terms of fuel consumption (e.g., amount of hydrogen, percentage of fuel cell capacity, etc.).

[0246] In step 2206, fuel prices are tracked. The price of fuel for a fuel cell (e.g., additional hydrogen to refuel a hydrogen fuel cell) is tracked to observe fluctuations in fuel prices over time. Tracking fuel prices may include identifying average prices, modeling fluctuations in fuel cell prices, adaptively improving fuel price forecasts, identifying periods of low prices, identifying periods of high prices, etc. Step 2206 may also include monitoring multiple fuel sources to identify the lowest cost option for fuel for a fuel cell. In some embodiments, step 2206 may include identifying preferred hydrogen production methods (e.g., hydrogen produced from water electrolysis by wind power rather than from fossil fuels) and generating hydrogen source suitability for cleaner production.

[0247] In step 2208, an optimization is performed to minimize the cost of purchasing fuel for fuel cells, subject to fuel storage limits or cost constraints or penalties. For example, a fuel storage tank may be located near the fuel cells (e.g., in a building) and may have a fixed capacity (i.e., the maximum amount of fuel it can hold). In such embodiments, constraints are defined to prevent the optimization from reaching a solution that results in overfilling of the storage tank. As another example, in some scenarios (e.g., leased storage area, resource consumption required to maintain the storage tank within a favorable temperature range), storing more fuel may incur higher costs for the user.

[0248] The optimization in step 2208 can be performed using the tracked fuel prices from step 2206 and the predicted future fuel cell usage from step 2204. For example, a predicted cost function can be formulated that shows the total cost of purchasing fuel, where the purchase time and the amount of fuel to be acquired are the decision variables for optimization, subject to constraints (e.g., balance equations, models) that predictively balance fuel consumption constraints and fuel storage constraints to ensure that enough fuel is available to meet the predicted fuel cell usage. The total cost that still satisfies those constraints and requirements may be the output of the optimization. The optimization strategy and formulation can use similar techniques as described in detail above.

[0249] In step 2210, additional fuel is ordered or otherwise obtained according to the results of the optimization in step 2208. For example, the optimization may indicate purchasing a specific amount of fuel from a specific fuel provider at a specific time, and step 2210 may include automatically executing that transaction or controlling distribution equipment or devices to obtain the fuel at that time. For example, step 2210 may include generating a notification to a technician (e.g., via the user's mobile phone or other device) instructing the technician to obtain the optimal amount of fuel at a specific future time. As another example, step 2210 may include automatically moving an autonomous vehicle to a provider fuel distribution site to obtain the optimized amount of fuel at an identified time. The fuel can then be added to a fuel cell or stored for later use by the fuel cell. This can optimize the operational costs of integrating the fuel cell into building equipment or other building systems.

[0250] Fuel cells can be incorporated into various devices for the execution of process 2200. For example, the various devices described above may be adapted to include one or more fuel cells. As another example, fuel cells are included in U.S. Patent Application No. 15 / 963,860, filed on February 25, 2020. (U.S. Patent Application Publication No. 2018 / 0340704) The AHU and RTU equipment described herein, or U.S. Patent Application No. 16 / 314,277, filed on June 29, 2017, and incorporated herein by reference. (U.S. Patent Application Publication No. 2019 / 0235453) This can be added to the VRF equipment listed.

[0251] Modular energy unit Referring here to Figures 23-24, diagrams of the modular energy unit 2300 according to exemplary embodiments are shown. The modular energy unit 2300 is a encapsulated, integrated product that, in various embodiments, can be deployed in a building or other site and electrically connected between the building and the utility grid to provide substantial energy cost savings, reduced carbon emissions, and reduced reliance on third-party energy grids. The modular energy unit 2300 may be similar in size to other large building equipment (e.g., rooftop units, chillers, industrial boilers, etc.). In some examples, the modular energy unit 2300 may have physical dimensions similar to or smaller than a shipping container suitable for transport by truck on standard roads. Thus, the modular energy unit 2300 can be easily deployed to provide the various benefits discussed below without requiring significant on-site construction or customization. For example, as discussed in detail below, the modular energy unit 2300 can provide a plug-and-play service for achieving net-zero carbon emissions for a building.

[0252] As shown in Figures 23-24, the modular energy unit 2300 includes a housing 2302. The housing 2302 may be similar in size to a shipping container, as mentioned above, for transport by truck on standard roads. In some embodiments, the housing 2302 is mobile (e.g., on a trailer with wheels). The modular energy unit 2300 is shown to include a photovoltaic cell 2304, a wind turbine 2306, a fuel cell 2308, a nuclear microreactor 2310, a gravity energy system 2312, and a battery 2314 inside or on top of the housing 2302. Various embodiments of these elements may be included. For example, in many examples, the nuclear microreactor 2310 and / or the gravity energy system 2312 are omitted. In another example, the fuel cell 2308 may be omitted. In yet another example, in some cases only one of the wind turbine 2306 and the photovoltaic cell 2304 is included. All such variations are within the scope of this disclosure. The modular energy unit 2300 is also shown to include a power management circuit 2314 connected to a wireless communication device (indicated as a cellular modem 2316), a utility grid electrical inlet 2318 for connection to a utility grid 2324, and an electrical outlet 2320 for connection to the building electrical system 2322.

[0253] Therefore, the modular energy unit 2300 includes multiple energy source and storage solutions with complementary properties that facilitate the reliable provision of electrical energy to meet building loads in various scenarios. The photovoltaic cell 2304 is mounted on the modular energy unit 2300 (for example, on the roof 2303 of the housing 2302) and configured to convert light (solar radiation) into electrical energy. The photovoltaic cell 2304 is conductively coupled to the power management circuit 2314 and supplies electricity from the photovoltaic cell 2304 to the power management circuit 2314. The photovoltaic cell 2304 can have a spread substantially identical to the roof 2303 of the housing, for example, covering an area larger than 90% of the surface area of ​​the roof 2303. The photovoltaic cell 2304 is thereby configured to provide the power management circuit 2314 with free, renewable, zero-carbon electrical energy without requiring the installation of additional solar panels or support structures separate from the modular energy unit 2300. In some embodiments, for example, in scenarios where a building or premises include complementary photovoltaic cells (e.g., previously installed elsewhere in the building or premises), the additional photovoltaic cells can be electrically connected to the power management circuit 2314, enabling the power management circuit 2314 to manage the electricity from such cells.

[0254] The modular energy unit 2300 is also shown to include a wind turbine 2306 extending from the roof 2303 of the housing 2302. The wind turbine 2306 may be configured to be easily erected without technical expertise (e.g., using simple tools and following simple indicators) when delivering the modular energy unit 2300 to its destination in a building or premises. The wind turbine 2306 is configured to convert wind energy into electricity through the rotation of a wind-driven turbine. The wind turbine 2306 is conductively connected to a power management circuit 2314 and provides electrical energy to the power management circuit 2314. Although one wind turbine 2306 is shown in the embodiment, various embodiments may include multiple wind turbines 2306. In some embodiments, for example, in scenarios where the building or premises include a supplementary wind turbine (e.g., a pre-installed wind energy harvesting device), such a supplementary wind turbine may also be connected to the power management circuit 2314 so that the power management circuit 2314 can receive and manage electricity from the supplementary wind turbine.

[0255] Therefore, the photovoltaic cell 2304 and the wind turbine 2306 provide the modular energy unit 2300 with the ability to convert environmental conditions (wind, sunlight) into electricity. Other embodiments may include a variety of other environmental energy collection devices and / or combinations thereof. The environmental energy collection devices (e.g., the photovoltaic cell 2304 and the wind turbine 2306) provide free, zero-carbon electricity during times when favorable environmental conditions are present. As will be discussed in the following sections, the modular energy unit may also include an energy source configured to provide base load or supplemental energy generation for use during times when environmental conditions do not provide sufficient energy to the photovoltaic cell 2304 and the wind turbine 2306 (e.g., a low-wind night).

[0256] The modular energy unit 2300 is shown to include a base-load or auxiliary energy generating device, shown as a nuclear microreactor 2312 and a fuel cell 2308. The nuclear microreactor 2312 generates electric power using energy from a nuclear fission or fusion reaction, for example, by converting the heat from such a reaction into electricity using a power conversion cycle. The nuclear microreactor 2312 may use enriched uranium fuel, e.g., highly purified, low-enriched uranium. The nuclear microreactor 2312 may be configured to be controllable to provide varying levels of electricity at varying times. In some examples, the nuclear microreactor 2312 may provide base-load energy complemented by other energy sources, providing a reliable backup energy source when other wind / solar energy, energy grids, and other energy sources are unavailable. Thus, the nuclear microreactor 2312 is very low-risk, can be made small enough to fit inside the housing 2302 of the modular energy unit 2300 and operate within the housing 2302, and can operate for decades, although it requires handling of fuel materials. Therefore, the nuclear microreactor 2312 is configured to provide electrical energy to the power management circuit with zero carbon emissions. In many embodiments, the nuclear reactor 2312 is omitted.

[0257] The fuel cell 2308 is configured to generate electrical energy through a chemical reaction, for example, a redox reaction between hydrogen and oxygen in one embodiment where the fuel cell 2308 is a hydrogen fuel cell. The fuel cell 2308 can be controlled to initiate the chemical reaction and convert the chemical fuel into electrical energy on demand, while being substantially stable to remain dormant when other power sources are in use. The fuel cell 2308 may be a hydrogen fuel cell or some other type of fuel cell (e.g., hydrocarbons). In embodiments where the fuel cell 2308 is a hydrogen fuel cell, the fuel cell 2308 can operate to generate electricity without carbon emissions, while emitting only water. The fuel cell 2308 may be located within the housing 2300 to provide easy access for refueling the fuel cell 2308 (e.g., via a port on the outer wall of the housing 2302). In some embodiments, the fuel cell 2308 is configured to transmit a signal via the cellular modem 2316 indicating that the fuel cell 2308 should be refilled, for example, based on process 2200 in Figure 22. The fuel cell 2308 is conductively connected to a power management circuit 2314 and supplies electricity to the power management circuit 2314. As will be discussed below, the power management circuit 2314 can control the fuel cell 2308 between on and off states, to different power levels, etc. Thus, the modular energy unit 2300 includes an electricity generation device that is independent of environmental conditions.

[0258] The modular energy unit 2300 also includes a battery 2313 and a rechargeable energy storage device, indicated as a gravity energy storage device 2310. As will be discussed below, the rechargeable energy storage device can release energy to a power management circuit 2314 as needed and store the energy supplied to the rechargeable energy storage device by the power management circuit 2314. In various embodiments, other types of rechargeable energy storage devices can be used.

[0259] As shown in Figures 23-24, the rechargeable energy storage device includes a gravity energy system 2310. The gravity energy system 2310 is configured to convert electricity into gravitational potential energy (for energy storage) and to convert gravitational potential energy into electricity (for releasing the stored energy). The gravity energy system 2310 can provide long-term energy storage without degradation. In some embodiments, the gravity energy system 2310 is configured to move one or more high-mass objects (e.g., one or more blocks of high-density metal) on a cable connected to one or more winches, which is configured to consume electrical energy to raise the high-mass objects through the cable and to extract electrical energy when lowering the high-mass objects. In other embodiments, the gravity energy system 2310 is configured to convert gravitational potential energy into electricity by pumping a fluid (e.g., water, high-density fluid) upward, consuming and storing electricity, and allowing the fluid to flow downward through one or more turbines, as shown and described by reference in Figure 25, for example. In some embodiments, the gravity energy system 2310 is configured to collect rainwater in a rainwater tank at a high altitude or elevation (e.g., a building rooftop, water tower, mountain, or hill) and to convert the gravitational energy in the stored rainwater into electricity as the rainwater flows downward, as described with reference to Figure 25. Figure 23 shows the gravity energy system 2310 housed within a housing 2303. In other embodiments, the gravity energy system 2310 may include an extension above the roof 2303 to increase the amount of gravitational potential energy that can be stored by the gravity energy system 2310. The gravity energy system 2310 is electrically connected to the power management circuit 2314 to receive electrical energy from or to the power management circuit 2314 and to receive control signals from the power management circuit 2314.

[0260] The modular energy unit 2300 is also shown to include a battery 2313. The battery 2313 is conductively connected to a power management circuit 2314 and is configured to store electrical energy from the power management circuit 2314 and release energy to the power management circuit 2314. The charging or discharging rate of the battery 2313 can be controlled by the power management circuit 2314. The battery 2313 may be located in one or more battery packs containing multiple battery cells (e.g., 8, 50, 100, 1000, etc.). For example, the battery 2313 may be a lithium-ion battery.

[0261] In some embodiments, the modular energy unit 2300 also includes heating and / or cooling devices configured to maintain the interior of the housing 2300 at or near a suitable temperature for the operation of its components, for example, at or near an efficient temperature for the battery 2313, the nuclear microreactor 2312, and the fuel cell 2308.

[0262] The modular energy unit 2300 is also shown to include a utility grid inlet 2318 configured to connect to the utility grid 2324 in order to provide a conductive path between the utility grid 2324 and the power management circuit 2314. The utility grid inlet 2318 may primarily allow the power management circuit to receive electricity from the utility grid 2324. In some embodiments and some scenarios, the utility grid inlet 2318 may also allow the power management circuit 2314 to return electricity to the utility grid 2324 (for example, by selling the electricity to a utility company and participating in a frequency response or other incentive program).

[0263] The modular energy unit 2300 is also shown to include an electrical energy outlet 2320. The electrical energy outlet 2320 is configured to electrically connect a power management circuit 2314 to a building electrical system 2322. The building electrical system 2322 includes building wiring configured to transmit electricity from the electrical energy outlet 2320 to various electrical devices within the building, such as outlets, lighting, and appliances. The power management circuit 2314 can detect the demand for electricity from the building electrical system 232 through its connection to the electrical energy outlet 2320.

[0264] The power management circuit 2314 is configured to manage the various flows of electricity provided by the components of the modular energy unit 2300 and to manage the flow of electricity to the building electrical system 2322 and energy storage devices (e.g., batteries 2313 and gravity energy system 2310). As will be discussed in detail below with reference to later figures, the power management circuit 2314 is configured to optimally allocate demand across the components of the modular energy unit 2300 and utility grid 2324 in order to meet the building's demands while minimizing energy costs, carbon emissions, or a combination thereof. The power management circuit 2314 may locally include computing components (e.g., memory devices and processing circuits) configured to perform optimization locally, as described below, or it may communicate with remote computing resources (e.g., cloud servers) via the cellular modem 2316 to facilitate optimal control of the modular energy unit.

[0265] Therefore, the modular energy unit 2300 is configured to manage the amount of electricity used from the utility grid 2324 and to be electrically positioned between the utility grid 2324 and the building electrical system 2322 to complement such electricity. Thus, the modular energy unit 2300 can be easily installed at a single point in the building without requiring the coordination or installation of devices within and throughout the building, as is the case with other building energy management systems. Thus, the modular energy unit 2300 is an easy-to-install, modular, integrated, plug-and-play solution for reducing energy costs and reducing or eliminating carbon emissions associated with the operation of the building.

[0266] Modular energy unit with rainwater and groundwater harvesting capabilities Referring here to Figure 25, a schematic illustrative diagram of an embodiment of a modular energy unit 2500 including a gravity energy system using rainwater and groundwater functions is shown according to an exemplary embodiment. The modular energy unit 2500 is shown to include a water-based energy system 2502, one or more wind turbines 2306, a photovoltaic cell 2304, and a gravity energy system implemented as a power management circuit (shown as a power accumulator 2501). In various embodiments, the modular energy unit 2500 may include various elements of the modular energy unit 2300 shown in Figures 23-24.

[0267] The water-based energy system 2502 is shown to include an upper water tank 2504 and a lower water tank 2506. The upper water tank 2504 is positioned to be exposed to rain (e.g., on a rooftop, on the roof 2303 of the housing 2302, connected to a downspout, ditch, etc.) and includes one or more openings configured to receive rainwater from the environment. A filter may be included to prevent debris and other contaminants from entering the upper water tank 2504. The lower water tank 2506 is physically positioned below the lower water tank (i.e., closer to the center of the Earth) such that the units of water in the upper water tank 2504 have a higher gravitational potential energy than the units of water in the lower water tank 2506. In some implementations, both the upper water tank 2504 and the lower water tank 2506 are positioned on a housing of an integrated shipping container-sized modular energy unit. In other embodiments, the upper water tank 2504 is located on the roof of a building supplied by a modular energy unit, while the lower water tank is located on the ground, underground, or at subsurface level, thus maximizing the gravitational potential between the upper and lower water tanks, taking into account the existing building structure. For example, some older industrial buildings still contain unused rooftop tanks that were previously decommissioned due to other technological advancements but can be reused for use in a water-based energy system 2502 without making significant structural changes to the building. The lower tank 2506 and the upper tank 2504 are connected by pipes, tubes, hoses, etc., to allow fluid flow between them.

[0268] As shown in Figure 25, the water-based energy system 2502 includes a pump 2508 and a motor 2510 positioned in the flow path between the lower tank 2506 and the upper tank 2504. The motor 2510 is configured to drive the pump 2508 to draw water from the lower tank 2506 to the upper tank 2504. The motor 2510 consumes electrical energy and uses the pump 2508 to convert the electrical energy into gravitational potential energy of the water drawn up by the pump 2508. The motor 2510 is controllable by a power accumulator 2501, for example, when surplus power is generated by the wind turbine 2306 and the photovoltaic cell 2304. A first controllable valve 2512 is positioned along the pipe between the lower water tank 2506 and the upper water tank 2504, for example, to prevent gravity from drawing water back through the pump when the pump is not operating. The first controllable valve 2512 is controllable by a valve controller 2513.

[0269] The water-based energy system 2502 is also shown to include a turbine 2514 connected along a flow path between an upper tank 2504 and a lower tank 2506. A second controllable valve 2516 is positioned between the upper tank 2504 and the turbine 2514 and is controllable by a valve control 2513. When the second controllable valve 2516 is controlled to open, gravity draws water from the upper tank 2504, through the turbine 2514, into the lower water tank 2506. The turbine 2514 may be positioned in close proximity to the lower water tank 2506. The gravitational potential energy of the water is converted into kinetic energy through gravitational acceleration as the water flows from the upper tank to the turbine 2514. The turbine 2514 then converts its kinetic energy into rotational kinetic energy, which is then converted into electricity via a generator 2518. Thereafter, the gravitational potential energy of the water in the upper tank 2504 is converted into electricity, which is supplied to a power accumulator 2501. The power accumulator 2501 can, for example, or according to various other optimization techniques described herein, cause the valve control 2513 to open valve 2516 when the power accumulator 2501 requires additional electricity to meet the building's demands.

[0270] By remaining open to rainwater, the upper tank 2502 can receive water without requiring the motor 2510 to operate, and therefore requires no electrical input. Thus, especially in rainy seasons and climates, the water-based energy system 2502 can provide a considerable amount of free electricity (i.e., extracted from the environment). To avoid overfilling of the lower water tank, the lower water tank may include an overflow function configured to drain into groundwater (or a drainage, sewer, or other system) when the lower water tank reaches its capacity.

[0271] Therefore, the modular energy unit 2500 can provide substantially constant power generation or power generation that always meets time-varying demand. For example, especially during periods of sunny and / or strong winds, solar and wind power are converted into electricity to meet demand, and gravitational potential energy is stored in the water-based energy system 2502. Then, during periods of cloudy or non-strong winds, that gravitational potential energy can be converted back into electricity via the turbine 2514. In addition, since periods of low sunlight can correspond to periods of high rainfall, the rainfall capture feature of the upper tank 2504 is particularly well suited to complement the photovoltaic power generation of the photovoltaic cell 2304. Thus, the modular energy unit 2500 utilizes multiple energy source and storage solutions to provide reliable, consistent, zero-carbon emission power.

[0272] Optimized control of modular energy units Referring now to Figure 26, a flowchart of a process 2600 for optimally controlling a modular energy unit (e.g., the modular energy unit 2300 in Figure 23) is shown according to an exemplary embodiment. Process 2600 may be performed, for example, by the power management circuit 2314 in Figure 24 and / or via cloud-based optimization resources (e.g., servers, memory, and processing circuits, etc.) that can communicate with the power management circuit 2314 via a cellular modem 2316. In some adaptations of Figure 26, process 2600 may be performed by the power accumulator 2501 in Figure 25.

[0273] In step 2602 of process 2600, the electricity demand on the modular energy unit is predicted. For example, the prediction of the electricity demand on the modular energy unit may include the electricity demand to be provided at each time step of the optimization period (e.g., instantaneous power in kilowatts, energy in joules or kilowatt-hours over small time increments). The electricity demand on the modular energy unit may be predicted using one or more gray-box models, neural network models, or several other modeling approaches. As an example, the electricity demand for a modular energy unit is predicted in U.S. Patent Application No. 14 / 717,593, filed May 20, 2015. (U.S. Patent Application Publication No. 2015 / 0316907) This can be predicted using the load predictor features described in (the entire disclosure of which is incorporated herein by reference).

[0274] In step 2604, the power generated from available wind turbines and / or solar power generators is predicted based on, for example, weather forecasts, data on the capacity and historical performance of the wind turbine 2306 and photovoltaic cell 2304, and / or other relevant data. For example, a technique for predicting the amount of renewable energy to be generated (e.g., PV power generated by PV panels) is described in U.S. Patent Application No. 15 / 247,869. (U.S. Patent Application Publication No. 2017 / 0104449) U.S. Patent Application No. 15 / 247,844 (U.S. Patent Application Publication No. 2017 / 0104337) , and U.S. Patent Application No. 15 / 247,788 (U.S. Patent Application Publication No. 2017 / 0102675) These are described in [the relevant document], each of which has a filing date of August 25, 2016, and is incorporated herein by reference.

[0275] In step 2606, battery capacity, gravity energy storage capacity, fuel cell fill level, and any other variables that may affect the amount of available energy capacity or storage capacity in the components of the modular energy unit 2300 are tracked. By tracking these values, and by considering in real time the available power and storage indicators in the various components of the modular energy unit 2300, it is possible to predict them in the future.

[0276] In step 2808, utility rates and carbon emissions data are obtained from the utility grid. For example, electricity from the utility grid may be subject to time-of-use pricing, such that pricing changes over time. Other pricing structures, incentive programs, penalties, etc., that may be related to energy from the utility grid, such as those described elsewhere in this specification, may also be identified in step 2808. In addition, in some scenarios, the utility grid may make available information indicating carbon emissions related to grid power at a particular point in time (e.g., time-varying power:carbon or carbon:power ratio, tons of CO2 per kWh), which may vary due to the utility grid receiving power from multiple sources that emit carbon at different rates (e.g., if natural gas plants, wind power plants, and solar power plants are connected to the energy grid).

[0277] In step 2610, using the data from steps 2602, 2604, 2606, and 2608, control decisions for the components of the modular energy unit are generated, in particular, so that the control decisions are predicted to reduce costs and / or carbon emissions associated with meeting the electricity demand of the modular energy unit. Figures 31-36 provide flowcharts of the process that can be adapted for use in generating control decisions for the components of the modular energy unit in various embodiments. As another example, the approach used by the predictive CEF controller described above may be adapted, for example, to handle the power contributions and constraints of the components of the modular energy unit 2300.

[0278] As one such example, the optimization problem may be formulated (for example, using an objective function) which includes decision variables representing the amount of energy or electrical energy to discharge from or charge battery 2313, the amount of energy or electrical energy to store in or extract from the gravity energy system 2310, the amount of energy or electrical energy to generate in fuel cell 2308, and / or the amount of energy or electrical energy to obtain from the utility grid. The optimization problem may include an objective function that represents the cost or penalty associated with carbon emissions, in addition to the total economic cost resulting from the selection of a particular decision variable (for example, the cost of purchasing electricity from the utility grid and the cost of fuel for fuel cell 2308). The cost or penalty associated with carbon emissions may be implemented by adding the cost of purchasing carbon offsets equal to the amount of carbon expected to be emitted over the optimization period, with respect to the values ​​of the decision variables selected (for example, to achieve net-zero emissions) or the cost of carbon credits associated with the expected emissions. Therefore, carbon emissions targets, as other technical targets for the modular energy unit 2300, can be quantified in shared units (e.g., dollars). An optimization problem may be performed to minimize an objective function subject to constraints on the capacity or other limitations of the various components of the modular energy unit 2300 and to determine the value of a decision variable that ensures the building's electricity demand is met.

[0279] In step 2612, the components of the modular energy unit are operated according to control decisions. For example, the fuel cell 2308, the gravity storage system 2310, and the battery 2313 can be controlled using the values ​​of the decision variables described above, which result from optimization. The power management circuit 2314 is configured to guide electricity between the various components of the modular energy unit 2300 (for example, from the photovoltaic cell 2304 to one of the building's outlets 2320 or the gravity energy system 2310 or the battery 2313) according to the values ​​of the decision variables. Thereafter, the modular energy unit 2300 is controlled to operate in an optimal manner that reduces both the cost of purchasing energy to operate the building and the carbon emissions associated with meeting the building's electricity demands.

[0280] Deployment of carbon neutrality in buildings using modular energy units Referring to Figure 27, flowcharts of the process for deploying modular energy units to achieve net-zero carbon emissions for a building are shown according to several embodiments. In step 2702, the modular energy units are manufactured in a production facility. In some embodiments, the production facility itself is carbon neutral (e.g., powered by green energy sources). In some embodiments, the modular energy units are manufactured from recycled materials (e.g., 80% or more recycled materials). The modular energy units manufactured in step 27 may be stock units having the same design, configuration, etc., and thus provide a simple and efficient process for providing modular energy units compared to alternative approaches that choose to install separate energy devices in a building and then attempt to integrate the devices on-site at a specific building.

[0281] In step 2704, one of the modular energy units from the production facility is transported to the building via road. It should be understood that steps 2704-2714 may be carried out separately for any or all of the modular energy units manufactured in step 2702. As described above, the modular energy units may have physical dimensions suitable for transport by semi-track via road, preferably without special handling as an oversized load or similar consideration.

[0282] In step 2706, the modular energy unit is electrically installed between the building and the utility grid so that it can manage demand on the utility grid. Step 2706 may include basic electrical wiring steps that can be easily achieved by a typical electrician without special training or expertise in energy systems. Step 2706 may simply include placing the housing of the modular energy system in a desired location, e.g., next to the building. In some scenarios, the modular energy system is installed on the roof of the building (e.g., to maximize exposure to sunlight). Step 2706 may include some simple mechanical adjustments to the modular data center to mount a wind turbine on the housing, e.g., as described above, but preferably no special expertise is required to achieve a suitable installation.

[0283] In step 2708, building data indicating the building's energy load is collected. For example, a modular data center can operate during the initial learning phase to collect data on building load, solar power availability, wind power availability, etc. During this data collection phase, the modular data center can operate according to a default control scheme, perform learning procedures (e.g., automated testing of system capabilities), and continue to meet the building's energy demands.

[0284] In step 2710, a predictive controller for a specific modular energy unit is trained based on the collected data. Thus, the predictive controller for the modular energy unit can be software-customized based on the demands of a specific building served by the modular energy unit and other relevant data (e.g., data indicating the availability of solar and wind energy). Therefore, in the example of process 2700, the modular data center may be supplied from the production facility with software customization based on the training data obtained in step 2710, without hardware customization. The predictive controller utilizes a gray-box system identification approach (e.g., U.S. Patent Application No. 16 / 447,724, filed June 20, 2019, incorporated herein by reference). (U.S. Patent Application Publication No. 2020 / 0218991) The predictive controller logic may be trained based on machine learning approaches that train neural network models (as described in [reference]), or other approaches that are suitable for a particular implementation of the predictive controller logic. Step 2710 may be performed locally on a modular energy unit or on remote cloud-based optimization resources.

[0285] In step 2712, the modular energy unit is controlled using a trained predictive controller and collected data to reduce or eliminate the use of carbon-emitting energy sources. For example, the predictive controller may optimally utilize zero-emission energy sources included in the modular energy unit (e.g., wind and solar power in the modular energy unit, hydrogen fuel cells, etc.) over other energy sources to supply the building and charge batteries 2313 and other energy storage devices. The predictive controller may also shift utility purchases to times when the utility grid is primarily supplied by renewable energy (e.g., solar, wind, geothermal, hydroelectric, etc.). Thus, the predictive controller can reduce the use of carbon-emitting energy sources and move the building's energy consumption toward zero carbon emissions. If carbon emissions are completely eliminated, process 2700 can be terminated in step 2712.

[0286] If some reduced levels of carbon emissions remain (for example, due to continued reliance on carbon emission production in the energy grid under certain conditions), process 2700 proceeds to step 2714, where any remaining carbon emissions are automatically offset using one or more carbon capture processes. Using the predictive controller and the data collected thereby, the remaining carbon emissions (e.g., in units of CO2) that can be used to initiate and run a desired carbon offset program can be estimated. In some embodiments, step 2714 includes automatically purchasing carbon offsets from third-party carbon offset providers, e.g., verified carbon offset providers engaged in carbon sequestration through forest conservation, afforestation efforts, etc. Carbon offset acquisition and management can be integrated with cloud-based resources that also host the predictive controller of the modular energy unit. In some embodiments, step 2714 includes automatically operating carbon capture / recovery technologies to capture a corresponding amount of carbon from the atmosphere. Various other approaches are possible for running a carbon offset program as an automated response to data generated by the modular energy unit. The reliance on carbon capture, carbon sequestration, and carbon offsetting to achieve net-zero emissions is substantially reduced by the installation and use of modular energy units, as included in Step 2714, which ensures that the carbon neutrality target is achieved through the implementation of Step 2700.

[0287] Modular energy unit equipped with building setting point optimization device Referring now to Figure 28, a block diagram of a system including another embodiment of the modular energy unit by several embodiments is shown. In the embodiment of Figure 28, the modular energy unit is configured to optimize building setpoints to influence the building's energy demand, thereby increasing the modular energy unit's ability to reduce energy costs and emissions. For example, time-shifted energy consumption may allow the modular energy unit to match consumption to periods of high availability of green energy while reducing demand during periods when the modular energy unit needs to rely on carbon-emitting energy sources. The example in Figure 28 enables such advantages.

[0288] As shown in Figure 28, the modular energy unit 2800 may be able to communicate with the building device 2802. The building device 2802 may include one or more thermostats of the building and / or HVAC equipment (e.g., air treatment units, chillers, VAV boxes, variable refrigerant flow system indoor and / or outdoor units), or other equipment within the building. In the example shown, the building device 2802 preferably includes a thermostat configured to control the HVAC equipment to drive the indoor air temperature toward a temperature setpoint. In such a case, changing the temperature setpoint can affect the building's resource demands, particularly by the building's HVAC equipment. Therefore, as described in the following sections, the modular energy unit 2800 is able to communicate with the building device 2802 to provide the building device 2802 with a time-varying temperature setpoint (or other setpoints, e.g., for airflow, humidity, lighting, etc.) in order to coordinate the building's operation with the energy operation of the modular energy unit 2800.

[0289] Figure 28 shows a building setpoint optimization device 2308 included in the modular energy unit 2800. The building setpoint optimization device 2308 is constructed with circuits (e.g., memory and processing components) configured to optimize the building setpoint in conjunction with the power management circuit 2314. For example, in some embodiments, the power management circuit 2314 provides the building setpoint optimization device 2308 with an index for effective use-time pricing of the electricity that can be provided by the modular energy unit 2800 (i.e., electricity costs resulting from the use of green energy generators 2304 / 2306, fuel cells 2308, and energy storage devices 2310 / 2313, in addition to purchases from the utility grid 2324). As another example, the power management circuit 2314 may provide the building setpoint optimization device 2308 with an effective carbon-to-power ratio of the electricity supplied from the modular energy unit 2800 to the building electrical system 2322. Optimized operation of the power management circuit can result in effective usage-time pricing or carbon-to-power ratios that are significantly lower than the rates set by the utility company. The building setting point optimization device 2308 is, for example, U.S. Patent Application No. 15 / 199,909, filed June 30, 2016. (U.S. Patent Application Publication No. 2018 / 0004172) U.S. Patent Application No. 13 / 802,154, filed on March 13, 2013. (U.S. Patent No. 9,852,481) U.S. Patent Application No. 16 / 687,122, filed on November 18, 2019. (U.S. Patent Application Publication No. 2021 / 0148592) , and / or U.S. Patent Application No. 16 / 598,539 filed on October 10, 2019 (U.S. Patent Application Publication No. 2020 / 0041158) The effective usage-time pricing can be used as input to a building setpoint optimization process for reducing costs and / or carbon emissions, using the optimization process described in (the entire disclosure of which is incorporated herein by reference).

[0290] Next, the building setpoint optimization device 2308 distributes the optimized setpoints to the building device 2802. The building device 2802 then operates according to the optimized setpoints, for example, by operating HVAC equipment to drive the building temperature up to the indoor air temperature setpoint generated by the building setpoint optimization device 2308. Thus, in the embodiment of Figure 28, the modular energy unit provides coordinated optimization of both building demand and energy production from multiple sources to meet optimized demand.

[0291] Networked modular energy units Referring now to Figure 29, a block diagram is shown illustrating a network 2900 of modular energy units 2901 and other elements according to several embodiments. Specifically, Figure 29 shows multiple modular energy units 2900 provided for use in a campus building 2902 and a modular data center 2904. The multiple modular energy units 2901 are able to communicate with each other via a local mesh network (e.g., a mesh Wi-Fi network) (e.g., arranged in a daisy chain), and one of the modular energy units 2901 is connected to a cloud-based optimization resource 2908 via a cellular network 2906. The cloud-based optimization resource 2908 is shown to be interoperable with a carbon isolation service 2910 (e.g., via an application programming interface).

[0292] Multiple modular energy units 2900 may be configured, for example, according to one of the examples in Figures 22-28. In this example, the modular energy units include a wireless communication interface (e.g., Wi-Fi networking hardware) or a wired communication interface (e.g., an Ethernet port) to enable direct communication between the modular energy units 2901. The modular energy units 2900 can be connected to each other, for example, in a daisy-chain architecture or a loop or ring architecture. The architecture shown in Figure 29 allows the cloud-based optimization resource 2908 to communicate with a single modular energy unit 2900 separately, rather than all of the modular energy units 2900, thereby simplifying communication and reducing potential access points to cybersecurity threats. The cellular network 2906 may include security features, for example, using a relaxed network identity-defining networking paradigm or other security communication protocols. As described in Patent No. 16 / 680,881 (filed November 12, 2019, and incorporated herein by reference), approaches for operating under intermittent connectivity to a cloud computing system may be implemented in some embodiments on or for a modular energy unit 2900.

[0293] The modular energy units 2901 work together to meet the energy needs of the on-site buildings 2902 and the modular data center 2904. The on-site buildings 2902 can include one or more buildings (e.g., two, three, five, ten, twenty, etc.). Figure 29 illustrates how multiple modular energy units can be provided together when suitable for larger energy needs that exceed the capacity of a single modular energy unit 2901. Multiple modular energy units may be electrically connected in series or parallel as an aggregate between the on-site buildings 2902 and the energy grid, or they may be connected to different buildings of the on-site buildings 2902, for example. In the example in Figure 29, the multiple modular energy units are electrically connected and configured to share energy among themselves in order to make the most of the available energy storage and generation capacity of the modular energy unit 2901.

[0294] The modular data center 2904 is configured to provide computing resources (servers, data storage media, etc.) on a premises to facilitate the tasks of people working or studying on the premises, and therefore includes computing components (e.g., servers, etc.) and HVAC equipment for controlling the temperature of the computing components. The modular data center 2904 may have a similar physical footprint to the modular energy unit 2901 and therefore can be easily distributed and installed together with the modular energy unit 2901. Figure 29 illustrates how the modular data center 2901 may be installed together with the modular energy unit 2901 that provides the energy load for the modular data center, and thus provides both computing components and the energy storage and generation functions necessary to power those computing components in a cost-effective, low-carbon, or zero-carbon manner. In some embodiments, the processes described in this example are performed by servers in the modular data center 2904, while in some embodiments they are performed by cloud-based optimization resources 2908.

[0295] The cloud-based optimization resource 2908 is configured to generate optimal control decisions for the modular energy units 2901, including coordinating the operation of multiple modular energy units 2901. For example, the cloud-based optimization resource 2908 can allocate energy storage or release across various energy storage devices of multiple modular energy units 2901. The cloud-based optimization resource can perform any of the optimization approaches described herein.

[0296] The cloud-based optimization resource 2908 can also estimate the carbon emissions associated with the operation of the on-site building 2902 and the modular data center 2904. The cloud-based optimization resource 2908 can communicate with carbon segregation (offsetting, capture, etc.) services and automatically obtain, for example, a carbon offset equal to the carbon emissions associated with the operation of the on-site building 2902. Thus, the network 2900 can, for example, automatically reduce energy costs and actual carbon emissions while reaching carbon neutrality. In another example, the cloud-based optimization resource 2908 can communicate with a carbon credit market and automatically purchase the carbon credits necessary to permit (e.g., authorize in consideration of government regulations) the carbon emissions resulting from the operation of the on-site building 2902 and the modular data center 2904, thereby automatically ensuring compliance with applicable regulatory limits on carbon emissions.

[0297] Referring here to Figure 30, several embodiments of the network 3000 are shown. The network 3000 is configured similarly to Figure 29 and includes modular energy units 2901 installed for use in the on-site buildings 2902, connected to the cloud-optimized resources 2908 via a cellular network 2906. The network 3000 further includes a building management system 3004 for use in the on-site buildings 2902, in particular by monitoring building data, building equipment, building setpoints, etc. The building management system 3004 may include controllers, network devices, sensors, etc., placed within the building to facilitate the control of building equipment. The building management system 3004 can communicate with the cloud-optimized resources 2908, for example, via the cellular network 2906 or via some other communication path (e.g., a wired internet connection).

[0298] In the example shown in Figure 30, the cloud optimization resource 2908 provides coordinated predictive control of modular energy units 2901 and building equipment through a building management system, e.g., active setpoint management. For example, the cloud optimization resource 2908 may formulate an optimization problem that includes, as decision variables, building setpoints (e.g., indoor air temperature setpoints), on / off decisions for building equipment, the amount of energy to store or release from the energy storage system of each modular energy unit 2901, and the amount of energy to be generated in the fuel cells of the modular energy units 2901, all of which can be determined in an integrated manner as a result of a single minimization process. The optimization problem may include, for example, a predicted temperature or temperature setpoint within a comfortable temperature range, e.g., U.S. Patent Application No. 16 / 943,955, filed July 30, 2020, incorporated herein by reference. (U.S. Patent Application Publication No. 2020 / 0355391) As described, the optimization problem may be subject to a building temperature constraint that requires staying within a defined range using one or more neural networks. The optimization problem may also be subject to load balancing constraints that ensure all energy stored, generated, purchased, or consumed is taken into account by the results of the optimization problem. The optimization problem may also be subject to constraints based on maximum allowable carbon emissions, or penalties based on carbon emissions.

[0299] In some embodiments, the optimization problem seeks to minimize an objective function that includes a first term corresponding to the economic costs of the building's operation (e.g., the cost of purchasing energy from the grid, the cost of equipment wear and tear) and a second term corresponding to the internalization of costs related to carbon emissions (e.g., the cost of purchasing carbon credits, the cost of purchasing carbon offsets, a weighted penalty term, etc.). By solving the optimization problem, the cloud-based optimization resource 2908 can generate control decisions for both the modular energy unit 2901 and the building management system 3002, thus enabling a high level of cost and carbon savings through the coordinated operation of the modular energy unit 2901 and building equipment.

[0300] Predictive control through carbon emission optimization Referring here to Figure 31, a flowchart of process 3100 for predictive control by carbon emission optimization according to an exemplary embodiment is shown. Process 3100 may be performed, for example, by a cloud-based optimization resource 2908. Process 3100 may also be performed by any of the predictive controllers described above (e.g., predictive CEF controller 304, predictive chiller controller 704, predictive pump controller 1104, predictive cooling tower controller 1504, or predictive valve controller 1804). Process 3100 is incorporated herein by reference, for example, by U.S. Patent Application No. 16 / 598,539, filed October 10, 2019. (U.S. Patent Application Publication No. 2020 / 0041158) As described in [reference], the process may be performed by or for a smart thermostat. Process 3100 is also, for example, U.S. Patent Application No. 15 / 199,909, filed June 30, 2016. (U.S. Patent Application Publication No. 2018 / 0004172) U.S. Patent Application No. 13 / 802,154, filed on March 13, 2013. (U.S. Patent No. 9,852,481) , or U.S. Patent Application No. 16 / 687,122 filed on November 18, 2019 (U.S. Patent Application Publication No. 2021 / 0148592) This may be done by a building management system or a controller for building equipment, as described in (the entire disclosure of which is incorporated herein by reference).

[0301] In step 3102, a time-varying indicator of carbon emissions per unit of energy or electricity, such as the electricity:carbon ratio (i.e., the average carbon per unit of electricity provided by the grid) or the marginal operating emission rate (MOER) (i.e., carbon per marginal unit of electricity, as described below), is received from the utility grid. That is, in the scenario of step 3102, the utility grid operator provides an estimate of the amount of carbon emitted to produce each unit of energy or electricity provided to customers of the utility grid. The carbon:electricity ratio or carbon:energy ratio may vary over time because renewable energy may contribute different percentages of total grid energy under different environmental conditions, different time periods, etc. In addition, the power sources that generate marginal energy units also change over time because different renewable energy sources and fossil fuel consuming plants may be operating at different times or under different total demands for the grid, and as a result, the carbon emissions associated with marginal energy consumption (i.e., whether or not the next energy unit is consumed) also change over time.

[0302] The emission rate associated with marginal energy consumption is referred to herein as the marginal operating emission rate (MOER), and in some embodiments, it may be distributed by a utility grid operator to its customers, for example, at a frequency of once every five minutes. To illustrate some examples, Figure 31B shows a graph of MOER over time in different seasons. The first graph 3150 shows the MOER over time for one week in February. In the example shown, the MOER may fluctuate between approximately 1000 lb / MWh and 0 lb / MWh. In the first graph 3150, the MOER is zero for parts of several days, for example, during the middle part of a day when photovoltaic energy production is sufficient to meet the grid's demand. At night, under cloudy conditions, or during periods of high demand, the grid relies on fossil fuel-based energy (coal, natural gas, etc.) to meet demand, and as a result, when such plants come into operation to meet the marginal demand of the utility grid, the MOER jumps to a higher value. The second graph 3152, showing the MOER for one week in July, shows that MOER can deviate from a standard pattern or fluctuation, indicating that higher MOERs can occur even during the daytime (when solar power is available) under periods of high demand. The first graph 3150 and the second graph 3152 show that there is an opportunity to reduce marginal emissions by shifting consumption over time to periods with lower MOERs, which can be achieved by process 3100 as described herein. The third graph 3154, showing the MOER for one week in August, shows that during periods of high, constant demand (e.g., during heat waves when air conditioning is running continuously and creating high demand on the utility grid), carbon emission sources remain constantly operational to meet marginal energy demands, so that MOER remains substantially constant throughout such periods.

[0303] In step 3102, assuming that such data is available from the utility grid, a time-varying indicator of carbon emissions (e.g., MOER) is received from the utility grid. In some scenarios, the utility grid may also provide forecasts or expected MOER or carbon:electricity ratio or carbon:energy ratio for future periods.

[0304] In scenarios where such information is not directly available from the utility grid (or other third part), the forecast controller may be configured to generate estimates of MOER or carbon:energy ratio or carbon:power ratio itself, as illustrated in steps 3104-3108 of process 3100. In step 3104, data is collected that identifies the available power on the energy grid, i.e., different energy sources and general information related to the production of energy sources supplied to the energy grid. This information is usually available even if detailed estimates of carbon emissions or real-time MOER are not shared by the utility company. Step 3104 may include collecting this data and building models of various energy sources on the utility grid. Step 3106 includes obtaining meteorological and time-of-day data (e.g., meteorological data for different time periods across the forecast range). In step 3108, the data from steps 3104 and 3106 are used to estimate the time-varying values ​​(e.g., mean, MOER) of carbon emissions per unit energy or power received from the energy grid. Step 3106 may include simulating the energy grid based on the best available information to perform a modeling approach that generates estimates and predictions of the carbon:power ratio or carbon:energy ratio or MOER over a prediction range. In some embodiments, a probabilistic optimization process is performed in which multiple scenarios with different time series values ​​of MOER are generated, and then, for example, U.S. Patent Application No. 16 / 115,290 filed on March 14, 2019. (U.S. Patent Application Publication No. 2019 / 0079473)As described in (the entire disclosure is incorporated here by reference) regarding the utility rate, it is used to optimize the overall objective across all multiple scenarios.

[0305] In step 3110, an objective function is generated that calculates the total carbon emissions, total marginal carbon emissions, or effective carbon-to-energy ratio over the forecast range, based on the forecast building load. The forecast building load is, for example, U.S. Patent Application No. 16 / 418,715, filed on May 21, 2019. (U.S. Patent Application Publication No. 2020 / 0371482) The system identification and gray-box modeling approach described in (the entire disclosure of which is incorporated herein by reference) can be used to model a building setpoint, e.g., a building temperature setpoint, and other building-related variables (such as indoor air temperature, outdoor air temperature, etc.). As another example, step 3110 can be modeled as a function of a building setpoint, e.g., a building temperature setpoint, and other building-related variables (such as indoor air temperature, outdoor air temperature, etc.) using a system identification and gray-box modeling approach such as that described in (the entire disclosure of which is incorporated herein by reference). (U.S. Patent Application Publication No. 2015 / 0316907) U.S. Patent Application No. 16 / 115,290, filed on August 28, 2018. (U.S. Patent Application Publication No. 2019 / 0079473) , or U.S. Application No. 15 / 199,910 filed on June 30, 2016. (U.S. Patent Application Publication No. 2018 / 0004171) The objective function used (all incorporated herein by reference) may be adapted by replacing the variable indicating the utility rate with a time-varying value of MOER, carbon:energy ratio, or carbon:power ratio. The objective function may also calculate the total emissions value by multiplying the carbon:energy ratio by the predicted or target energy consumption of building equipment. For example, the objective function may be:

number

[0306] In step 3112, time-varying setpoints are generated for a building to optimize an objective function subject to one or more constraints. For example, temperature setpoints for each time step across the optimization range may be generated in step 3112. As another example, energy consumption targets for building equipment may be generated in step 3112. Various details of such embodiments are provided in the applications cited above and incorporated herein by reference. Optimizing the objective function may include performing a gradient descent or other minimization process that seeks to achieve the smallest possible total carbon emissions while satisfying one or more constraints (e.g., a constraint to ensure occupant comfort).

[0307] In step 3114, building equipment is operated according to optimized setpoints. Since step 3112 uses a time-varying MOER or carbon:power ratio or carbon:energy ratio as input, step 3114 may include shifting building equipment temporally to low-carbon periods and away from high-carbon emission periods. For example, the building may be pre-cooled or pre-heated during low-carbon periods (e.g., cooled below a preferred temperature setpoint and heated above a preferred temperature setpoint) to reduce or eliminate the operation of cooling equipment (such as chillers) during high-carbon periods. Process 3100 can thereby reduce carbon emissions associated with the building's energy consumption. In other embodiments, process 3100 is implemented as a planning tool and used to generate reports, analyses, predicted carbon savings, predicted cost savings, recommendations, etc., resulting from implementing the optimization strategy of process 3100, instead of, or in addition to, controlling building equipment, as shown in the example in Figure 31.

[0308] Referring here to Figure 32, flowcharts of process 3200 for optimizing the operation of building equipment while internalizing the cost of carbon emissions are shown in several embodiments. Process 3200 can be performed in various embodiments by the same various controllers / processors as process 3100 described above.

[0309] In Step 3202, pricing for carbon offsets or carbon credits is established. Carbon offsets refer to the market for carbon sequestration and carbon capture services, such as reforestation or deforestation services, through which polluters can pay third parties to carry out activities to remove carbon from the atmosphere. If carbon offsets fully cover the emissions of a technology or entity (e.g., tons of CO2 emitted = tons of CO2 equivalent sequestrated), the technology or entity is called a carbon neutral or net-zero carbon emitter. Carbon credits refer to a regulated market, active in some jurisdictions, where companies cannot generate emissions exceeding a defined amount of freely transferable carbon credits traded on the market. In either case, units of carbon emissions (e.g., tons of CO2) may be associated with the economic cost of offsetting or obtaining regulatory permission for those emissions.

[0310] Step 3202 may include, for example, providing interoperability between a digital marketplace for carbon offsets or credits and a predictive controller for building equipment via one or more APIs and internet connectivity. Step 3202 may include monitoring pricing for carbon offsets or credits and building one or more models to predict future prices. In other embodiments, step 3202 may include obtaining data indicating the price that building owners have prepaid or contracted to pay for carbon offsets or credits. 【03...

Claims

1. A cascade control system for adjusting and controlling carbon emissions associated with building equipment operating in a distributed manner across multiple subsystems, A first controller configured to generate carbon emission targets for each of several subsystems based on time-varying values ​​of one or more emission rates related to electricity from the utility grid, Multiple second controllers and Equipped with, Each of the aforementioned plurality of second controllers is, It corresponds to one of the aforementioned subsystems, To generate control decisions for building equipment of the corresponding subsystem, Using the control decision, the building equipment of the corresponding subsystem is operated and It is configured to do the following: The control decision is configured to cause the building equipment to achieve the carbon emission target for the corresponding subsystem. The control decision is a cascade control system based on the time-varying values ​​of one or more emission rates.

2. The cascade control system according to claim 1, wherein the first controller generates the carbon emission target based on the time-varying values ​​of one or more emission rates, and one or more devices of the building equipment consume the power.

3. The cascade control system according to claim 1, wherein the first controller generates the carbon emissions target using a predictive control process, the predictive control process takes into account the aggregate carbon emissions of the plurality of subsystems that are expected to result from the carbon emissions target, and further takes into account the comfort of occupants of one or more buildings serviced by the building equipment based on several reductions that are expected to occur in order to meet the carbon emissions target.

4. The cascade control system according to claim 3, wherein the first controller generates the carbon emission target using a multi-objective optimization process having multiple objectives, including a carbon objective based on the aggregate carbon emissions of the plurality of subsystems and a comfort objective.

5. The cascade control system according to claim 4, wherein the aforementioned multiple objectives further include the cost of purchasing resources consumed by the building equipment.

6. The cascade control system according to claim 1, wherein the first controller generates the carbon emission target based on both the carbon emissions associated with the plurality of subsystems and other carbon emissions that are not controllable by the cascade control system.

7. The cascade control system according to claim 6, wherein the other carbon emissions are attributable to the transport of goods or people.

8. The cascade control system according to claim 1, wherein the first controller is configured to generate the carbon emissions target based on a budget or goal for total emissions over a certain period of time.

9. A cascade control system for adjusting and controlling carbon emissions associated with building equipment operating in a distributed manner across multiple subsystems, The first controller generates carbon emission targets for each of several subsystems based on time-varying values ​​of one or more emission rates related to power from the utility grid, In each of the multiple second controllers corresponding to one of the multiple subsystems, the carbon emission target is received from the first controller. Each of the plurality of second controllers generates a control decision for the building equipment of the corresponding subsystem, Each of the plurality of second controllers uses the control decision to operate the building equipment of the corresponding subsystem. Includes, The control decision is configured to cause the building equipment to achieve the carbon emission target for the corresponding subsystem. The control determination is a cascade control method based on the time-varying values ​​of one or more emission rates.

10. The cascade control method according to claim 9, wherein the first controller generates the carbon emission target based on the time-varying values ​​of one or more emission rates, and one or more devices of the building equipment consume the power.

11. The cascade control method according to claim 9, wherein the first controller generates the carbon emissions target using a predictive control process, the predictive control process takes into account the aggregate carbon emissions of the plurality of subsystems that are expected to result from the carbon emissions target, and further takes into account the comfort of occupants of one or more buildings serviced by the building equipment based on several reductions that are expected to occur in order to meet the carbon emissions target.

12. The cascade control method according to claim 11, wherein the first controller generates the carbon emission target using a multi-objective optimization process having multiple objectives, including a carbon objective based on the aggregate carbon emissions of the plurality of subsystems and a comfort objective.

13. The cascade control method according to claim 12, wherein the aforementioned multiple objectives further include the cost of purchasing resources consumed by the building equipment.

14. The cascade control method according to claim 9, wherein the first controller generates the carbon emission target based on both the carbon emissions associated with the plurality of subsystems and other carbon emissions that are not controllable by the cascade control system.

15. The cascade control method according to claim 14, wherein the other carbon emissions are attributable to the transport of goods or people.

16. The cascade control method according to claim 9, wherein the first controller is configured to generate the carbon emissions target based on a budget or goal for total emissions over a certain period of time.

17. One or more non-temporary computer-readable media storing instructions, wherein when the instructions are executed by one or more processors, the one or more processors... The first controller generates carbon emission targets for each of several subsystems based on time-varying values ​​of one or more emission rates related to power from the utility grid, In each of the multiple second controllers corresponding to one of the multiple subsystems, the carbon emission target is received from the first controller. Each of the plurality of second controllers generates a control decision for the building equipment of the corresponding subsystem, Each of the plurality of second controllers uses the control decision to operate the building equipment of the corresponding subsystem. Perform an action that includes the following: The control decision is configured to cause the building equipment to achieve the carbon emission target for the corresponding subsystem. The control decision is based on one or more non-temporary computer-readable media, which are the time-varying values ​​of the one or more emission rates.

18. The first controller generates the carbon emission target based on time-varying values ​​of one or more emission rates, and one or more devices of the building equipment consume the power, one or more non-temporary computer-readable media according to claim 17.

19. One or more non-temporary computer-readable media according to claim 17, wherein the first controller generates the carbon emissions target using a predictive control process, the predictive control process takes into account the aggregate carbon emissions of the plurality of subsystems that are expected to result from the carbon emissions target, and further takes into account the comfort of the occupants of one or more buildings serviced by the building equipment based on several reductions that are expected to occur in order to meet the carbon emissions target.

20. One or more non-temporary computer-readable media according to claim 17, wherein the first controller generates the carbon emissions target based on both the carbon emissions associated with the plurality of subsystems and other carbon emissions that are not controllable by the first controller or the plurality of second controllers.

Images