Dynamic optimization method for energy consumption of central air conditioner of building

Abstract

The invention provides a building central air conditioner energy consumption dynamic optimization method which comprises the following steps: S1, for any water chilling unit, acquiring real-time operation data, and inputting the real-time operation data and meteorological data in a future target time period into a cold demand prediction model to obtain a system cold load demand quantity; s2, constructing an objective function by taking the minimum total energy consumption or the minimum operation cost of the system under the predicted system cooling load demand as a limiting condition, and constructing a system optimization model in combination with meteorological data, refrigerator performance parameters, cooling tower operation constraint conditions and water pump equipment operation state constraint conditions; s3, solving the system optimization model by adopting an optimization algorithm to obtain the optimal starting number of the water chilling units, the optimal load rate of each water chilling unit and a cooling tower fan operation strategy in one dispatching cycle or multiple dispatching cycles in the future; and S4, generating a control instruction, and issuing and executing the control instruction. The method has the beneficial effects that systematic energy conservation and consumption reduction can be realized, global collaborative optimization is realized, and high automation and intelligence are realized.

Description

Technical Field

[0001] This invention relates to the technical field of building energy management and HVAC systems, and more specifically, to a method for dynamic optimization of energy consumption in building central air conditioning. Background Technology

[0002] In building energy consumption, central air conditioning systems, especially their core component chillers, are the main energy-consuming units. To achieve energy conservation and consumption reduction, existing control methods generally rely on building automation systems (BAS) or building energy management systems (BMS). These traditional systems mostly use rule-based control based on fixed thresholds (such as setting fixed supply and return water temperatures, pressure differences, or flow rates) or simple PID control logic to control the start-up, shutdown, and operation of chillers, pumps, and cooling towers.

[0003] However, these static, experience-based control strategies have significant limitations. First, they lack foresight, failing to anticipate future weather changes and dynamic fluctuations in building loads. This results in a reactive rather than proactive system response, easily leading to response delays or energy waste. Second, traditional control models often focus on the local optimization or steady-state operation of individual devices, lacking coordinated scheduling and overall energy efficiency optimization of the entire cooling system (including multiple chillers, pumps, cooling towers, and cold storage devices). This leads to mismatches between system components, making it difficult to consistently operate at the highest overall energy efficiency.

[0004] Furthermore, existing optimization methods are insufficient in characterizing the time-varying and nonlinear features of chiller unit performance, and their control models struggle to accurately reflect the true energy efficiency characteristics of the equipment during actual operation. Simultaneously, the formulation of control strategies is often disconnected from dynamic electricity price signals, failing to fully utilize the peak-shaving and valley-filling potential of devices such as cold storage to further reduce operating costs.

[0005] Therefore, there is an urgent need in this field for a solution that can overcome the above-mentioned defects, namely, a dynamic optimization method for building central air conditioning energy consumption that can systematically save energy and reduce consumption, achieve global collaborative optimization, and realize a high degree of automation and intelligence. Summary of the Invention

[0006] The technical problem to be solved by this invention is how to achieve systematic energy saving and consumption reduction, global collaborative optimization, and high automation and intelligence. In order to overcome the defects of the above-mentioned existing technologies (or related technologies), this invention provides a method for dynamic optimization of energy consumption of building central air conditioning. This invention provides a method for dynamic optimization of energy consumption of building central air conditioning, comprising the following steps: Step S1: For any central air conditioning chiller unit in the target building, obtain the real-time operating data of the chiller unit, and input the real-time operating data and the meteorological data for the future target time period into the pre-trained cooling load prediction model to obtain the system cooling load demand for the future target time period. Step S2: The objective function is constructed with the minimum total energy consumption or minimum operating cost of the system under the predicted cooling load demand as the constraint condition. The system optimization model is constructed by combining the meteorological data, chiller performance parameters, cooling tower operation constraints and water pump equipment operation status constraints. Step S3: Use an optimization algorithm to solve the system optimization model to obtain the optimal number of chiller units to be started, the optimal load rate of each chiller unit, and the cooling tower fan operation strategy in one or more future scheduling cycles. Step S4: Based on the optimal number of units to be turned on, the optimal load rate, and the cooling tower fan operation strategy, control commands are generated and sent to the chiller unit and cooling tower control system for dynamic optimization operation.

[0007] Compared with existing technologies, the present invention provides a method for dynamic optimization of energy consumption in building central air conditioning systems, which has the following advantages: This invention combines real-time operating data of chiller units with a cooling load prediction model and system optimization algorithms. Under the premise of meeting the predicted cooling load demand of the system, it takes the lowest total energy consumption or lowest operating cost of the system as the optimization objective, realizing dynamic decision-making and coordinated control for the whole system. It can significantly reduce the energy consumption and operating costs of central air conditioning systems. By introducing meteorological data and equipment performance prediction results for future target time periods, the central air conditioning system is transformed from traditional passive response control to proactive forward optimization control. It can adjust the operating strategy in advance to cope with changes in operating conditions, thereby effectively avoiding energy waste.

[0008] Furthermore, in step S3, the present invention uniformly optimizes the optimal number of chiller units to be turned on, the optimal load rate of a single chiller unit, and the cooling tower fan operation strategy. This enables key equipment such as chiller units, water pumps, cooling towers, and cold storage devices to operate collaboratively under the same optimization framework. This avoids the problem of limited overall energy efficiency caused by optimizing only a single device or local link in the prior art. It improves the overall operating efficiency at the system level. The entire optimization process, from data acquisition and load prediction to control command generation, is automated, significantly reducing the reliance on human experience and intervention, and improving the intelligence level and management efficiency of the system operation.

[0009] In one possible implementation, the real-time operating data obtained in step S1 includes the current operating status parameters of the chiller unit and system environmental parameters. The current operating status parameters include at least the chiller unit's on / off status, partial load rate, cooling water inlet temperature, cooling water supply and return temperature difference, current system cooling load, and cooling tower operating status. The system environmental parameters include the current outdoor dry-bulb temperature and the current outdoor wet-bulb temperature. The meteorological data includes the outdoor air temperature and outdoor humidity for the target time period in the weather forecast. In step S1, the current operating status parameters, the system environmental parameters, and the meteorological data are input into the cooling load prediction model to obtain the system cooling load demand.

[0010] Compared with existing technologies, the above technical solution can clearly identify the key input features required by the cooling load forecasting model, providing a comprehensive and high-quality data foundation for model prediction, thereby effectively improving the accuracy of load forecasting and performance forecasting, and ensuring the scientificity and reliability of subsequent system optimization decisions.

[0011] In one possible implementation, in step S1, the cooling load prediction model is constructed using a long short-term memory neural network model or a time series convolutional neural network model, and is used to predict the system cooling load demand for a future target time period.

[0012] Compared with existing technologies, the above-mentioned technical solution can effectively characterize the long-term trend and short-term fluctuation characteristics of meteorological data changes by using deep learning models specifically designed for time-series data modeling. It has higher prediction accuracy than traditional statistical forecasting methods and provides more reliable forward-looking information for the generation of subsequent control commands.

[0013] In one possible implementation, the cooling load prediction model uses the historical cooling load data of the chiller unit, the real-time operating data, and the meteorological data as input features to learn the mapping relationship between cooling load changes, operating conditions, and meteorological conditions, and to predict the cooling load demand of the system within a future target time period.

[0014] Compared with existing technologies, the above-mentioned technical solution can accurately characterize the nonlinear characteristics of the performance parameters of chiller units as they change with operating conditions. It can not only obtain accurate performance prediction results, but also improve the model's adaptability to complex operating conditions, thereby providing reliable performance constraints for system optimization models.

[0015] In one possible implementation, the system optimization model constructed in step S2 takes the minimum total energy consumption or minimum operating cost of the system as the objective function while meeting the system's cooling load demand, and combines the meteorological data, the performance parameters of the chiller unit, the operating constraints of the cooling tower, and the operating status constraints of the water pump equipment for modeling.

[0016] In one possible implementation, the system optimization model constructed in step S2 is solved using a mixed-integer linear programming model or a mixed-integer nonlinear programming model to obtain the optimal operating scheme of the chiller unit within one or more future scheduling cycles.

[0017] Compared with existing technologies, the above-mentioned technical solutions can effectively handle complex optimization problems involving both discrete and continuous variables, such as chiller start-up and shutdown decisions, load allocation, and operation strategy selection. Under the premise of meeting equipment operation constraints and system physical constraints, theoretically optimal or near-optimal operation schemes can be obtained, thereby maximizing energy-saving potential.

[0018] In one possible implementation, the building central air conditioning energy consumption dynamic optimization method adopts a rolling optimization mechanism, which periodically executes steps S1 to S4 at preset time intervals. At the beginning of each rolling cycle, the system cooling load demand is re-predicted based on the latest acquired real-time operating data and meteorological data, and the solution results of the system optimization model are updated to dynamically adjust the control commands.

[0019] Compared with existing technologies, the above-mentioned technical solution can continuously absorb the latest real-time operating data, dynamically correct prediction errors and sudden operating conditions, and keep the system in the optimal or near-optimal state under the current operating conditions, significantly improving the stability and robustness of the system operation.

[0020] In one possible implementation, the cooling tower fan operation strategy obtained in step S3 includes at least one of the following: simultaneous supply and storage mode, separate cold storage mode, separate cold release mode, and combined cold storage and cold release mode.

[0021] Compared with existing technologies, the above-mentioned technical solution can intelligently switch between different cold storage and release modes according to load changes and electricity price characteristics. It can store cold during off-peak hours and release cold during peak hours, thereby effectively shaving off peaks and filling valleys, reducing operating costs, and improving the system's scheduling flexibility for complex load changes. Attached Figure Description

[0022] Figure 1 This is a flowchart of the steps of the present invention; Figure 2This is a schematic diagram of the architecture of the chiller unit and cooling tower control system of the present invention; Figure 3 This is a schematic diagram of the cooling demand prediction model of the present invention. Detailed Implementation

[0023] First, those skilled in the art should understand that these embodiments are merely illustrative of the technical principles of the present invention and are not intended to limit the scope of protection of the present invention. Those skilled in the art can make adjustments as needed to adapt to specific application scenarios.

[0024] The present invention will now be described in further detail with reference to the accompanying drawings and specific embodiments.

[0025] See Figure 1 This invention discloses a method for dynamic optimization of energy consumption of building central air conditioning, comprising the following steps: Step S1: For any central air conditioning chiller unit in the target building, obtain the real-time operating data of the chiller unit, and input the real-time operating data and the meteorological data for the future target time period into the pre-trained cooling load prediction model to obtain the system cooling load demand for the future target time period. Step S2: The objective function is constructed with the minimum total energy consumption or minimum operating cost of the system under the predicted cooling load demand as the constraint condition. Combined with meteorological data, chiller performance parameters, cooling tower operation constraints and water pump equipment operation status constraints, the system optimization model is constructed. Step S3: Use an optimization algorithm to solve the system optimization model to obtain the optimal number of chiller units to be started, the optimal load rate of each chiller unit, and the cooling tower fan operation strategy in one or more future scheduling cycles. Step S4: Based on the optimal number of units to be turned on, the optimal load rate, and the cooling tower fan operation strategy, control commands are generated and sent to the chiller unit and cooling tower control system for dynamic optimization operation.

[0026] In this embodiment of the invention, the real-time operating data includes the current operating status parameters of the chiller unit and system environmental parameters. The current operating status parameters include at least the chiller unit's start-up status, partial load rate, cooling water inlet temperature, cooling water supply and return temperature difference, current system cooling load, and cooling tower operating status. The system environmental parameters include the current outdoor dry-bulb temperature and the current outdoor wet-bulb temperature. The meteorological data includes the outdoor air temperature and outdoor humidity for the target time period in the weather forecast. In step S1, the current operating status parameters, system environmental parameters, and meteorological data are input into the cooling load prediction model to obtain the system cooling load demand.

[0027] In this embodiment of the invention, see Figure 2 , Figure 2 The diagram shows the specific architecture of the chiller unit and cooling tower control system, including a cold source side system, a chilled water storage system, a chilled water system, detection and actuation components, and an optimization control system. The cold source side system includes n cooling towers for supplying chilled water to the building load side and cooling the chiller unit's condensate return water; a cooling water pump P1 for driving the cooling water circulation between the cooling towers and the chiller unit; cooling water supply pipes (green) and return pipes (red) forming the cooling water circulation loop; and temperature measuring points for collecting the cooling water inlet and outlet temperatures. The chilled water storage system includes components for low-load operation. The system includes a cold storage device (TES) that stores cold air during periods of low electricity prices and releases it during periods of high load; cold storage valves V4, V5, and V6 for controlling the cold storage and release operation modes; and n release pumps to drive the cold storage device to participate in the system's cooling supply. This cold storage and release system can switch between simultaneous supply and storage, standalone cold storage, standalone cold release, and combined operation modes. The chilled water system includes chilled water supply lines (blue) and return lines (orange) for supplying cooling to end loads and recovering chilled water; chilled water pump P2 for driving the chilled water circulation between the chiller units, the cold storage device, and the load side; and a distributor for connecting multiple chiller units in parallel. The system includes a water collector, regulating valves CV1-CV3 for adjusting chilled water flow and switching operating conditions, and electric valves V1-V3. The detection and actuation elements include flow meters for collecting chilled or cooling water flow from each branch, temperature sensors for collecting the supply and return temperatures of cooling and chilled water, and valves and pumps that are the main actuators for system optimization control. The optimization control system includes a cooling demand prediction module, a system optimization decision module, and a control command generation module. This optimization control system is communicatively connected to the chiller, pumps, cooling tower, valves, and cold storage device to receive real-time operating data and meteorological data, and to issue control commands to each actuator.

[0028] In this embodiment of the invention, see Figure 3 , Figure 3 The diagram shows the specific structure of the cooling demand prediction model. Based on its data input layer, parallel feature extraction layer, long short-term memory neural network, and feature classification layer, step S1 specifically includes: Step S11, Data Acquisition and Preprocessing: Real-time operating data of the chiller unit is collected through the sensor and BAS system interface. To ensure data quality, data cleaning and data standardization preprocessing are required. Data cleaning uses a sliding window statistical method (such as the Z-score method) to identify and remove abnormal data points. For missing data, linear interpolation or a K-nearest neighbor (KNN) based filling method is used for repair. Data standardization preprocessing involves Z-score standardization of multi-source heterogeneous data (such as cooling water inlet temperature and partial load rate of the chiller unit) to convert it into a distribution with a mean of 0 and a standard deviation of 1. Step S12: Based on LSTM temperature prediction, the preprocessed real-time operating data is... The input is fed into a Long Short-Term Memory (LSTM) neural network, which captures long-term dependencies through the following gating mechanism: Forgotten Gate: ; Input Gate: ; ; Cell status update: ; Output gate: ; ; in, For the sigmoid function, , , , These are the weight coefficients used in network training. for The hidden state at all times for The hidden state at all times for Input at any time , , , The bias parameters are used for training the network, and the network output is a predicted sequence of outdoor dry and wet bulb temperatures for the next H hours. ; Step S13, based on GPR performance coefficient prediction, real-time running data (mainly partial load rate) will be used. and cooling water inlet temperature The input is given to the Gaussian Process Regression (GPR) model, which treats the performance coefficient COP as a Gaussian process. ; in, Represents the Gaussian distribution coefficient. Represents the mean function, The covariance function is represented by the radial basis function (RBF), which is preferred. This represents the predicted partial load factor. This indicates the predicted cooling water inlet temperature.

[0029] In this embodiment of the invention, step S2 specifically includes: Step S21, define the objective function. Taking the lowest system operating cost as an example, the objective function is constructed as follows: ; in, express Electricity price at any given time; Indicates the first Taiwan chiller unit Power consumption at any given moment; Indicates that the water pump is in Power consumption at any given moment; Indicates that the cooling tower is Power consumption at any given moment; Indicates the time step; Indicates the total time step size for prediction; Indicates the prediction time domain; Step S22: Construct system constraints. The system optimization model must satisfy the following core physical and operational constraints: Cooling load balance constraints: ; in, Indicates the first Taiwan chiller unit Cooling capacity at any given time; Indicates the cooling capacity of the cold storage device; Indicates the cold storage capacity of the cold storage device; This indicates the system's cooling load requirement; Equipment operating constraints: ; in, Indicates the lower limit of the load rate; Indicates the first Taiwan chiller unit A binary variable representing the start / stop status at any given time; Indicates the upper limit of the load rate; Indicates the current load rate; Minimum continuous running time constraint: ; This constraint mandates that once the chiller unit is started, it must run continuously for at least [duration missing]. For at least 24 hours, once shut down, the machine must remain shut down continuously for at least [number] hours. Hours; Dynamic constraints of cold storage devices: ; ; in, Indicates that the cold storage device is in The relative cooling capacity at any given time; Indicates that the cold storage device is in The relative cooling capacity at any given time; This indicates the cold storage efficiency of the cold storage device; This indicates the cooling efficiency of the cold storage device; Indicates the total cold storage capacity; This indicates the lower limit of the relative cold storage capacity of the cold storage device; This indicates the upper limit of the relative cold storage capacity of the cold storage device.

[0030] In this embodiment of the invention, in step S3, the mixed-integer linear programming (MILP) or mixed-integer nonlinear programming (MINLP) model, which consists of the objective function and constraints, is input into a commercial solver (such as Gurobi or CPLEX) for solving. The solver uses algorithms such as branch-and-bound to search for the globally optimal or near-optimal sequence of decision variables. .

[0031] In this embodiment of the invention, the cooling tower fan operation strategy obtained in step S3 includes at least one of the following: simultaneous supply and storage mode, separate cold storage mode, separate cold release mode, and combined cold storage and cold release mode. In step S3, the optimal number of chillers to be operated at each moment is precisely calculated as the optimal number of chiller units to be turned on. The optimal load rate of each chiller unit is to allocate a suitable load to each operating chiller unit so that it can operate in the high-efficiency zone as much as possible. The cold storage and cold release strategy of the cold storage device is to clarify which mode the cold storage device should be in at each moment, such as simultaneous supply and storage, separate cold storage, separate cold release, or combined cold storage and cold release mode, as well as the specific cold storage / cold release power.

[0032] In this embodiment of the invention, step S4 specifically includes: Step S41: Instruction issued, the first time step in the solution result is used. The control strategy is converted into specific control commands and sent to the field controller for execution via industry standard protocols such as OPC UA or BACnet; Step S42, rolling optimization, the system enters the next control cycle. Then, repeat steps S11 to S41, update the input of the system optimization model with the latest real-time running data, and resolve the optimization problem, thereby incorporating the latest system state feedback into the decision-making process, forming a closed-loop feedback control system with strong anti-interference capability.

[0033] In this embodiment of the invention, a cooling-to-standard duration is introduced for startup optimization. Before the building is put into use each day, the system estimates the cooling-to-standard duration for that day. Subsequently, the optimal startup time Determined by the following formula: ; in, Indicates the building's expected opening time. This indicates a safety buffer time. This method ensures that the ambient temperature is within acceptable limits when personnel arrive, while avoiding energy waste caused by premature startup.

[0034] In the description of this invention, the references to "one embodiment," "some embodiments," "in this embodiment," "specific example," or "some examples," etc., indicate that a specific feature, structure, material, or characteristic described in connection with that embodiment or example is included in at least one embodiment or example of the invention. In this specification, the illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples. Moreover, without contradiction, those skilled in the art can combine and integrate the different embodiments or examples described in this specification, as well as the features of different embodiments or examples.

[0035] The above description is merely a specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the technical scope disclosed in the present invention should be included within the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be determined by the scope of the claims.

Claims

1. A method for dynamic optimization of energy consumption of building central air conditioning, characterized in that, Includes the following steps: Step S1: For any central air conditioning chiller unit in the target building, obtain the real-time operating data of the chiller unit, and input the real-time operating data and the meteorological data for the future target time period into the pre-trained cooling load prediction model to obtain the system cooling load demand for the future target time period. Step S2: The objective function is constructed with the minimum total energy consumption or minimum operating cost of the system under the predicted cooling load demand as the constraint condition. The system optimization model is constructed by combining the meteorological data, chiller performance parameters, cooling tower operation constraints and water pump equipment operation status constraints. Step S3: Use an optimization algorithm to solve the system optimization model to obtain the optimal number of chiller units to be started, the optimal load rate of each chiller unit, and the cooling tower fan operation strategy in one or more future scheduling cycles. Step S4: Based on the optimal number of units to be turned on, the optimal load rate, and the cooling tower fan operation strategy, control commands are generated and sent to the chiller unit and cooling tower control system for dynamic optimization operation.

2. The method for dynamic optimization of building central air conditioning energy consumption according to claim 1, characterized in that, The real-time operating data obtained in step S1 includes the current operating status parameters of the chiller unit and system environmental parameters. The current operating status parameters include at least the chiller unit's on-state, partial load rate, cooling water inlet temperature, cooling water supply and return temperature difference, current system cooling load, and cooling tower operating status. The system environmental parameters include the current outdoor dry-bulb temperature and the current outdoor wet-bulb temperature. The meteorological data includes the outdoor air temperature and outdoor humidity for the target time period in the weather forecast. In step S1, the current operating status parameters, the system environmental parameters, and the meteorological data are input into the cooling load prediction model to obtain the system cooling load demand.

3. The method for dynamic optimization of building central air conditioning energy consumption according to claim 1, characterized in that, In step S1, the cooling load prediction model is constructed using a long short-term memory neural network model or a time series convolutional neural network model, and is used to predict the system cooling load demand for a future target time period.

4. The method for dynamic optimization of building central air conditioning energy consumption according to claim 1, characterized in that, The cooling load prediction model uses the historical cooling load data of the chiller unit, the real-time operating data, and the meteorological data as input features to learn the mapping relationship between cooling load changes, operating conditions, and meteorological conditions, and to predict the cooling load demand of the system within a future target time period.

5. The method for dynamic optimization of building central air conditioning energy consumption according to claim 1, characterized in that, The system optimization model constructed in step S2 takes the minimum total energy consumption or minimum operating cost of the system as the objective function under the premise of meeting the system's cooling load demand, and combines the meteorological data, the performance parameters of the chiller unit, the operating constraints of the cooling tower, and the operating status constraints of the water pump equipment for modeling.

6. The method for dynamic optimization of building central air conditioning energy consumption according to claim 1, characterized in that, The system optimization model constructed in step S2 is solved using a mixed-integer linear programming model or a mixed-integer nonlinear programming model to obtain the optimal operating scheme of the chiller unit in one or more future scheduling cycles.

7. The method for dynamic optimization of building central air conditioning energy consumption according to claim 1, characterized in that, The building central air conditioning energy consumption dynamic optimization method adopts a rolling optimization mechanism, which periodically executes steps S1 to S4 at preset time intervals. At the beginning of each rolling cycle, the system cooling load demand is re-predicted based on the latest acquired real-time operating data and meteorological data, and the solution results of the system optimization model are updated to dynamically adjust the control commands.

8. The method for dynamic optimization of building central air conditioning energy consumption according to claim 1, characterized in that, The cooling tower fan operation strategy obtained in step S3 includes at least one of the following: simultaneous supply and storage mode, separate cold storage mode, separate cold release mode, and combined cold storage and cold release mode.

Images