A medical chiller IoT edge control method

Abstract

The present invention discloses an Internet of Things edge control method for a medical chiller, comprising: obtaining chiller operating parameters, cooling tower operating parameters, and environmental parameters of different types of areas in a hospital, inputting the chiller operating parameters, cooling tower operating parameters, and environmental parameters of the different types of areas into a cooling load prediction model to obtain future cooling load prediction values ​​of the different types of areas; inputting the chiller operating parameters, cooling tower operating parameters, and future cooling load prediction values ​​of the different types of areas into a deep reinforcement learning model to obtain adjustment strategies for chillers and cooling towers in different types of areas; dynamically adjusting cooling water and chilled water distribution paths and their distribution priorities according to different types of areas, future cooling load prediction values, and adjustment strategies for chillers and cooling towers; and controlling the operation of chillers and cooling towers in corresponding areas through adjustment strategies for chillers and cooling towers in different types of areas, cooling water and chilled water regional distribution paths, and their distribution priorities.

Description

Technical Field

[0001] The present invention relates to the field of Internet of Things edge control technology, and in particular to an Internet of Things edge control method for a medical chiller. Background Art

[0002] The central air-conditioning system of a hospital is a key infrastructure used to ensure the temperature, humidity and air quality requirements of wards, operating rooms, ICUs, drug storage areas and medical equipment rooms. The refrigeration function of the air-conditioning system is generally achieved through a chiller system. The chiller system mainly includes: a chiller and a cooling tower. In the cooling unit, cooling is achieved by circulating chilled water. After the system is started, the compressor starts working, the chilled water cycle starts, the water pump starts, and the chilled water and cooling water start to circulate; the chilled water is cooled to the set temperature through the evaporator, and the chilled water is transported to the air-conditioning terminal equipment to absorb indoor heat; the chilled water releases the heat absorbed from the chilled water in the condenser, and the cooling water takes the heat away; the cooling water transports the heat to the cooling tower and dissipates the heat into the atmosphere through contact with the air; the chilled water completes the absorption of indoor heat, and the cooling water completes the discharge of heat, and the entire cycle continues.

[0003] Due to the emergency switching in hospitals, such as emergency surgery in the operating room and expansion of the ICU, the cooling load demand in a certain area suddenly increases. The chiller (chiller and cooling tower) needs to cool down quickly to ensure that the ambient temperature of the operation room or ward reaches the required standard. In normal times, the chiller and cooling tower work together to ensure constant ambient temperature. However, in an emergency, the load response of the chiller (compressor and evaporator) is usually faster because the compressor can quickly adjust the frequency or start and stop; however, the cooling tower is limited by the inertia of the fan, the inertia of the cooling water cycle, and the response speed of the heat exchange equipment, and the adjustment speed is slow, resulting in a difference in the response speed of the two, and it is impossible to achieve rapid cooling, which may lead to an increase in unilateral energy consumption. In addition, the existing technology only controls at the equipment level when controlling. Due to the complex pipe network of the hospital's chiller, there is no joint control based on the actual priority.

[0004] The purpose of this invention is to design an IoT edge control method for a medical chiller in view of the above-mentioned problems in the prior art. Summary of the invention

[0005] In view of this, the purpose of the present invention is to propose an IoT edge control method for a medical chiller, which can solve the above-mentioned problems.

[0006] The present invention provides an Internet of Things edge control method for a medical chiller, comprising:

[0007] The chiller operating parameters, cooling tower operating parameters, and environmental parameters of different types of areas in the hospital are obtained, and the chiller operating parameters, cooling tower operating parameters, and environmental parameters of different types of areas are input into the cooling load prediction model to obtain the future cooling load prediction values ​​of different types of areas;

[0008] The chiller operating parameters, cooling tower operating parameters, and future cooling load forecasts for different types of areas are input into the deep reinforcement learning model to obtain adjustment strategies for chillers and cooling towers in different types of areas.

[0009] Dynamically adjust cooling water and chilled water distribution paths and their distribution priorities according to different types of areas, future cooling load forecast values, and adjustment strategies of chillers and cooling towers;

[0010] The operation of chillers and cooling towers in the corresponding areas is controlled through the adjustment strategies of chillers and cooling towers in different types of areas, the cooling water and chilled water area allocation paths and their allocation priorities.

[0011] Further, the acquisition of chiller operating parameters, cooling tower operating parameters, and environmental parameters of different types of areas in the hospital includes:

[0012] Obtain the evaporator temperature, condenser temperature, compressor frequency, and chilled water flow of the chiller;

[0013] Get the cooling water inlet / outlet temperature, fan speed, and cooling water flow rate of the cooling tower;

[0014] Get indoor temperature and humidity, outdoor temperature and humidity.

[0015] Furthermore, the cooling load prediction model is obtained through the following steps:

[0016] Collect the chiller operating parameters, cooling tower operating parameters, and environmental parameters at several different time points, calibrate the actual cooling load data at the corresponding time points, and use them as training samples for the cooling load prediction model;

[0017] A cooling load prediction model is constructed through the long short-term memory network model. The operating parameters of the cooling unit, cooling tower and environmental parameters at past time points are used as the input of the cooling load prediction model, and the actual cooling load data at future time points are used as the output of the cooling load prediction model to train the cooling load prediction model.

[0018] Furthermore, the deep reinforcement learning model is obtained through the following steps:

[0019] The future cooling load forecast value, environmental parameters, chiller operating parameters, and cooling tower operating parameters are used as state variables, and the compressor frequency adjustment amount and chilled water flow adjustment amount of the chiller, the fan speed adjustment amount and cooling water flow rate adjustment amount of the cooling tower are used as action variables to build a deep reinforcement learning model.

[0020] A reward function is constructed based on the regional cooling load demand satisfaction rate and total energy consumption. When the reward value of the reward function is higher than the reward threshold, it is marked as a positive reward signal, otherwise it is marked as a negative reward signal. The deep reinforcement learning model is optimized according to the reward signal.

[0021] Furthermore, the reward function constructed by the regional cooling load demand satisfaction rate and the total energy consumption includes:

[0022] The cooling load demand satisfaction rate is calculated by using the actual cooling load data and the future cooling load forecast value. The calculation formula is as follows:

[0023] ;

[0024] The total energy consumption is calculated by adding the real-time power of the compressor, the real-time power of the evaporator, the real-time power of the condenser, and the real-time power of the fan;

[0025] The reward function is obtained by weighting the cooling load demand satisfaction rate and the total energy consumption. The calculation formula is as follows:

[0026] ，

[0027] in, is the corresponding weight coefficient.

[0028] Furthermore, the dynamically adjusting the cooling water and chilled water distribution paths and their distribution priorities according to different types of regions, future cooling load forecast values, and adjustment strategies of chillers and cooling towers includes:

[0029] Construct a chilled water flow model based on chilled water flow nodes and paths, and calculate the cost of each chilled water flow path;

[0030] Construct a cooling water flow model based on cooling water flow nodes and paths, and calculate the cost of each cooling water flow path;

[0031] Based on the chilled water flow path overhead and the cooling water flow path overhead, calculate the comprehensive flow path overhead for each area, determine the chilled water and cooling water area allocation priority based on the comprehensive flow path overhead, and select the allocation path based on the chilled water flow path overhead and the cooling water flow path overhead.

[0032] Further, the constructing of the chilled water flow model according to the chilled water flow nodes and paths includes:

[0033] The chilled water flow model is constructed by taking each area as the target node, the chiller as the source node, and the chilled water pipes connecting the target node and the source node as the chilled water path.

[0034] Further, the calculation of the cost of each chilled water flow path includes:

[0035] defining regional priority factors according to the regional types of different types of regions;

[0036] Define the chilled water load demand factor based on the future cooling load forecast;

[0037] According to the frequency adjustment time of the compressor of the chiller, the frequency adjustment time of the compressor is obtained;

[0038] The priority factor, chilled water load demand factor, and compressor frequency adjustment time are combined to calculate the chilled water flow path cost. The calculation formula is as follows:

[0039] .

[0040] Further, the constructing of the cooling water flow model according to the cooling water flow nodes and paths includes:

[0041] The condenser of the chiller is taken as the target node, the cooling tower is taken as the source node, and the cooling water pipe connecting the target node and the source node is taken as the cooling water path to construct the cooling water flow model.

[0042] Further, the calculation of the cost of each cooling water flow path includes:

[0043] Define the cooling efficiency factor based on the cooling water temperature and the condenser demand temperature;

[0044] The current cooling water load capacity is defined based on the ratio of the current load to the maximum capacity of the cooling tower;

[0045] According to the time required for cooling tower fan adjustment, the fan start-up adjustment time is obtained;

[0046] The cooling water flow path cost is calculated by combining the cooling efficiency factor, the current cooling water load capacity, and the fan start-up adjustment time. The calculation formula is as follows:

[0047] .

[0048] Beneficial effects of the present invention:

[0049] First, through the cooling load prediction model, the changing trend of cooling load demand can be perceived in advance, and the fan speed and cooling water flow rate of the cooling tower can be adjusted in advance to make up for the response lag and solve the problem of slow response of the cooling tower. At the same time, the predicted future cooling load forecast value can be used to solve the coordination problem between the chiller and the cooling tower and the complex problem of the hospital pipe network.

[0050] Second, through the deep reinforcement learning model, a coordinated control strategy for the chiller and cooling tower is established. In an emergency, the compressor frequency of the chiller is adjusted first, and the cooling tower fan and water pump are started in advance to ensure the overall response and coordination of the system. The reinforcement learning model can dynamically adjust the working mode of the chiller and cooling tower according to the cooling load demand at different time points to avoid unilateral energy consumption increase.

[0051] Third, by building models for the chilled water loop and the cooling water loop respectively, in the chilled water and cooling water systems, the available resources (chilled water, cooling water) are limited and cannot fully meet the maximum demand of all regions at the same time. It is necessary to ensure that resources are supplied to higher priority regions first, so as to optimize the overall operation cost. After determining the priority, for each region, further select the specific chilled water and cooling water path with the lowest path cost to ensure that even for the high-priority region, the path used is still the best in the local area. BRIEF DESCRIPTION OF THE DRAWINGS

[0052] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings in the following description are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0053] Figure 1 It is a method flow chart of an embodiment. DETAILED DESCRIPTION

[0054] To facilitate understanding by those skilled in the art, the structure of the present invention is now further described in detail with reference to the embodiments and the accompanying drawings. It should be understood that the steps mentioned in the present embodiment, unless otherwise specified in their order, can be adjusted in order according to actual needs, and can even be executed simultaneously or partially simultaneously.

[0055] Example

[0056] like Figure 1 As shown, an embodiment of the present invention provides an IoT edge control method for a medical chiller, including:

[0057] S1 obtains the chiller operating parameters, cooling tower operating parameters, and environmental parameters of different types of areas in the hospital, inputs the chiller operating parameters, cooling tower operating parameters, and environmental parameters of different types of areas into the cooling load prediction model, and obtains the future cooling load prediction values ​​of different types of areas;

[0058] S101 obtains the evaporator temperature, condenser temperature, compressor frequency, and chilled water flow rate of the chiller;

[0059] S102 obtains the cooling water inlet / outlet temperature, fan speed, and cooling water flow rate of the cooling tower;

[0060] S103 obtains indoor temperature and humidity, and outdoor temperature and humidity.

[0061] In this step, a temperature sensor is installed at the chilled water inlet of the evaporator to obtain the evaporator temperature; a temperature sensor is installed at the chilled water outlet of the condenser to measure the temperature of the chilled water after cooling in the condenser; the current frequency of the compressor operation is obtained through the inverter used by the compressor; a dedicated chilled water flow meter is installed in the chilled water pipeline to dynamically monitor the chilled water flow; temperature sensors are installed at the inlet and outlet of the cooling water main to obtain the cooling water inlet / outlet temperature; the fan is usually driven by a variable frequency drive, and its speed can be read from the inverter; a flow meter is installed in the cooling water pipeline to obtain the cooling water flow rate; the number of people indoors can be obtained through an infrared sensor or camera installed indoors.

[0062] Furthermore, the cooling load prediction model is obtained through the following steps:

[0063] Collect the chiller operating parameters, cooling tower operating parameters, and environmental parameters at several different time points, calibrate the actual cooling load data at the corresponding time points, and use them as training samples for the cooling load prediction model;

[0064] A cooling load prediction model is constructed through the long short-term memory network model. The operating parameters of the cooling unit, cooling tower and environmental parameters at past time points are used as the input of the cooling load prediction model, and the actual cooling load data at future time points are used as the output of the cooling load prediction model to train the cooling load prediction model.

[0065] In this step, the actual cooling load data can be measured by energy metering equipment (such as cooling meter or flow meter). The cooling load is usually calculated by the chilled water flow and temperature difference: ， Where Q is the cooling load, m is the chilled water flow, c is the specific heat capacity, Δ T=The temperature difference between the inlet and outlet of chilled water. Evaporator temperature: reflects the heat absorption capacity of chilled water and directly affects the cooling load. Condenser temperature: reflects the heat release capacity of chilled water and is closely related to the efficiency of the cooling tower. Compressor frequency: affects the cooling capacity of the chiller. The higher the frequency, the stronger the cooling capacity. Chilled water flow rate: directly determines the size of the cooling load. The larger the flow rate, the higher the cooling load. These parameters can reflect the operating status and cooling capacity of the chiller and are the core influencing factors for predicting the cooling load. Cooling water inlet / outlet temperature: affects the heat exchange efficiency of the condenser, and thus affects the cooling capacity of the chiller. Fan speed: determines the cooling capacity of the cooling tower. The higher the speed, the higher the cooling efficiency. Cooling water flow rate: reflects the circulation capacity of cooling water. The larger the flow rate, the higher the cooling efficiency. The operating parameters of the cooling tower determine the heat dissipation efficiency of the condenser and indirectly affect the cooling load capacity of the chiller. Indoor temperature and humidity: reflects the heat load demand inside the building and directly affects the cooling load. Outdoor temperature and humidity: affects the cooling efficiency of the cooling tower (for example, the lower the wet bulb temperature, the higher the cooling efficiency), and thus affects the operating status of the chiller. Environmental parameters are external factors affecting cooling load, which can reflect the heat load demand of the building and the working efficiency of the chiller.

[0066] The Long Short-Term Memory (LSTM) network model is a deep learning model suitable for processing time series data, which can capture the time dependency and long-term trend of data. The operating parameters of the chiller and cooling tower, as well as the environmental parameters, will change over time and have a direct impact on the future cooling load. The Long Short-Term Memory (LSTM) network model can well learn these time series characteristics and predict future cooling loads.

[0067] Through the cooling load prediction model, the changing trend of cooling load demand can be perceived in advance, and the fan speed and cooling water flow rate of the cooling tower can be adjusted in advance to compensate for the response lag and solve the problem of slow response of the cooling tower. At the same time, the predicted future cooling load forecast value can be used to solve the coordination problem between the chiller and the cooling tower and the complex problem of the hospital pipe network.

[0068] S2 inputs the chiller operating parameters, cooling tower operating parameters, and future cooling load forecast values ​​of different types of areas into the deep reinforcement learning model to obtain the adjustment strategies of chillers and cooling towers in different types of areas;

[0069] Furthermore, the deep reinforcement learning model is obtained through the following steps:

[0070] The future cooling load forecast value, environmental parameters, chiller operating parameters, and cooling tower operating parameters are used as state variables, and the compressor frequency adjustment amount and chilled water flow adjustment amount of the chiller, the fan speed adjustment amount and cooling water flow rate adjustment amount of the cooling tower are used as action variables to build a deep reinforcement learning model.

[0071] A reward function is constructed based on the regional cooling load demand satisfaction rate and total energy consumption. When the reward value of the reward function is higher than the reward threshold, it is marked as a positive reward signal, otherwise it is marked as a negative reward signal. The deep reinforcement learning model is optimized according to the reward signal.

[0072] In this step, these state variables (future cooling load forecast, environmental parameters, chiller operating parameters, cooling tower operating parameters) fully cover the internal state and external environment of the entire system, which is an important basis for formulating the optimal adjustment strategy. The action variables directly correspond to the adjustable control parameters of the equipment, so that the output of the reinforcement learning model can be mapped to the physical execution layer to form a closed-loop control.

[0073] The response speed of the chiller (such as the compressor) is fast, while the response speed of the cooling tower is slow. In traditional methods, the control strategies of the two are often independent, resulting in an uncoordinated overall response of the system. The following problems may occur: Unilateral energy consumption increases: The chiller responds quickly to the cooling load demand, but the cooling tower is not adjusted in time, resulting in reduced condenser heat dissipation efficiency, increased compressor load, and increased energy consumption. Local overload: The delayed response of the cooling tower may cause the condenser temperature to rise, affecting the operating stability of the chiller. Therefore, through the deep reinforcement learning model, a collaborative control strategy for the chiller and the cooling tower is established. In an emergency, the compressor frequency of the chiller is adjusted first, and the cooling tower fan and water pump are started in advance to ensure the overall response and coordination of the system. The reinforcement learning model can dynamically adjust the working mode of the chiller and the cooling tower according to the cooling load demand at different time points to avoid unilateral energy consumption increases.

[0074] Furthermore, the reward function is constructed by the regional cooling load demand satisfaction rate and the total energy consumption, including:

[0075] The cooling load demand satisfaction rate is calculated by using the actual cooling load data and the future cooling load forecast value. The calculation formula is as follows:

[0076] ;

[0077] The total energy consumption is calculated by adding the real-time power of the compressor, the real-time power of the evaporator, the real-time power of the condenser, and the real-time power of the fan;

[0078] The reward function is obtained by weighting the cooling load demand satisfaction rate and the total energy consumption. The calculation formula is as follows:

[0079] ，

[0080] in, is the corresponding weight coefficient.

[0081] In this step, the reward function directly includes the system goal: the cooling load demand satisfaction rate reflects the satisfaction of the cooling load demand, and the total energy consumption reflects the system energy efficiency. The reward function guides the reinforcement learning model to achieve a comprehensive balance between cooling load demand and energy efficiency, while supporting weighted optimization. Whether the cooling load demand can be met is the core goal of chiller control, especially in hospital scenarios, where certain areas (such as operating rooms and ICUs) have high real-time requirements for cooling load demand. The actual cooling load data can be measured by energy metering equipment (such as cooling meters or flow meters). The cooling load is usually calculated by the chilled water flow and temperature difference: ， Where Q is the cooling load, m is the chilled water flow, c is the specific heat capacity, Δ T is the temperature difference between the inlet and outlet of the chilled water. The future cooling load forecast value is obtained by predicting in step S1. Energy consumption optimization is the core goal of ensuring low-cost operation, especially for hospital refrigeration systems with high energy consumption, and energy-saving optimization is of great significance. The actual power of the compressor, evaporator, condenser and fan is measured in real time through the built-in power metering module. The weight coefficient can be dynamically adjusted according to the demand priority.

[0082] S3 dynamically adjusts the cooling water and chilled water distribution paths and their distribution priorities according to different types of areas, future cooling load forecast values, and the adjustment strategies of chillers and cooling towers;

[0083] S301 constructs a chilled water flow model according to the chilled water flow nodes and paths, and calculates the cost of each chilled water flow path;

[0084] S3011 constructs a chilled water flow model by taking each area as a target node, a chiller as a source node, and a chilled water pipe connecting the target node and the source node as a chilled water path.

[0085] S3012 defines a regional priority factor according to the regional type of different types of regions;

[0086] S3013 defines a chilled water load demand factor based on future cooling load forecast values;

[0087] S3014 obtains the compressor frequency adjustment time according to the compressor frequency adjustment time of the chiller;

[0088] S3015 combines the priority factor, chilled water load demand factor, and compressor frequency adjustment time to calculate the chilled water flow path cost. The calculation formula is as follows:

[0089] .

[0090] In this step, the chiller is the main source of chilled water and is responsible for providing cooling for the entire chilled water system. Therefore, using the chiller as the source node of the chilled water flow is consistent with the actual physical structure of the chilled water system. Each area (such as the operating room, ICU, and general ward) is the final user of chilled water, and its cooling load demand determines the distribution and flow path of chilled water. Therefore, using the area as the target node can accurately describe the distribution target of chilled water. The chilled water flow system can be abstracted as a network graph model, in which: the source node is the chiller, which is responsible for providing chilled water; the target node is each area, which is responsible for receiving chilled water; the path is the chilled water pipeline, which is responsible for transmitting chilled water.

[0091] The regional priority factor reflects the priority of different types of regions for cooling load. For example, the demand for cooling load in operating rooms and ICUs is more urgent and has a higher priority. In the calculation of path cost, the higher the priority factor, the smaller the path cost, which means that the path needs to be allocated chilled water resources first. By introducing the priority factor, it can ensure that high-priority regions get priority in chilled water allocation.

[0092] The chilled water load demand factor reflects the actual demand for chilled water in a region, which is usually determined by the future cooling load forecast. The higher the cooling load demand, the greater the load demand factor. In the path cost calculation, the larger the load demand factor, the smaller the path cost, indicating that chilled water is more likely to flow to areas with higher load demand. By introducing the load demand factor, it is possible to achieve a reasonable allocation of chilled water resources and avoid resource waste or insufficient allocation.

[0093] The compressor frequency adjustment time reflects the speed at which the chiller responds to the cooling load demand. The longer the adjustment time, the slower the path response. In the path cost calculation, the longer the adjustment time, the greater the path cost, indicating that the path has a slower response speed. By introducing the adjustment time, the path with a faster response speed can be selected first, improving the system's rapid response capability.

[0094] S302 constructs a cooling water flow model according to cooling water flow nodes and paths, and calculates the cost of each cooling water flow path;

[0095] S3021 takes the condenser of the chiller as the target node, the cooling tower as the source node, and the cooling water pipe connecting the target node and the source node as the cooling water path to construct a cooling water flow model.

[0096] S3022 defines a cooling efficiency factor based on the cooling water temperature and the condenser demand temperature;

[0097] S3023 defines the current cooling water load capacity according to the ratio of the current load to the maximum capacity of the cooling tower;

[0098] S3024 obtains the fan start adjustment time according to the time required for the cooling tower fan adjustment;

[0099] S3025 calculates the cooling water flow path cost by combining the cooling efficiency factor, the current cooling water load capacity, and the fan start-up adjustment time. The calculation formula is as follows:

[0100] .

[0101] In this step, the cooling tower is the main cooling equipment for cooling water, which is responsible for providing low-temperature cooling water to the condenser. Therefore, the cooling tower is the source of cooling water flow. The condenser is the main user of cooling water, and its performance directly affects the operating efficiency of the chiller. Therefore, the condenser is the target node of cooling water flow. The cooling water system can be abstracted as a network graph model, in which: the source node is the cooling tower, which is responsible for providing cooling water; the target node is the condenser, which is responsible for receiving cooling water; the path is the cooling water pipeline, which is responsible for transmitting cooling water.

[0102] The cooling efficiency factor reflects the heat exchange efficiency of the cooling water and is usually determined by the difference between the cooling water temperature and the condenser demand temperature. The lower the cooling water temperature, the higher the cooling efficiency; when the cooling water temperature is close to the condenser demand temperature, the cooling efficiency is low. The higher the cooling efficiency factor, the smaller the path overhead, indicating that the path can more efficiently meet the cooling needs of the condenser. By introducing the cooling efficiency factor, the path with high cooling efficiency can be preferentially selected to reduce system energy consumption.

[0103] The current cooling water load capacity reflects the load pressure of the cooling tower, which is usually determined by the ratio of the current load to the maximum capacity of the cooling tower. The higher the load, the closer the cooling capacity of the cooling tower is to the upper limit, and the greater the load pressure of the cooling water path. The larger the current cooling water load capacity, the smaller the path overhead, indicating that the path can carry more cooling water flow. By introducing load capacity, it is possible to avoid excessive concentration of cooling water resources on certain paths and prevent the cooling tower from overloading.

[0104] The fan start adjustment time reflects the speed at which the cooling tower fan responds to cooling demand. The longer the adjustment time, the slower the path response. The longer the fan start adjustment time, the greater the path overhead, indicating that the path has a slower response speed. By introducing the fan adjustment time, the path with a faster response speed can be preferentially selected, thereby improving the system's rapid response capability.

[0105] S303 calculates the comprehensive flow path cost of each area based on the chilled water flow path cost and the cooling water flow path cost, determines the chilled water and cooling water area allocation priority based on the comprehensive flow path cost, and selects the allocation path based on the chilled water flow path cost and the cooling water flow path cost.

[0106] In this step, in the chilled water and cooling water systems, the available resources (chilled water, cooling water) are limited and cannot fully meet the maximum demand of all regions at the same time. It is necessary to ensure that resources are supplied to higher priority regions first, so that the overall operation cost is optimized. The flow of chilled water and cooling water is a mutually coupled system. The smaller the comprehensive flow path cost, the higher the allocation priority. Low cost means that the resource allocation to this path is the most efficient in the overall system. High-priority regions obtain more chilled water flow, and at the same time, they are preferentially equipped with cold water flow with higher cooling efficiency, and the allocation ratio of low-priority regions is reduced. The following example is used to illustrate: Chilled water flow path cost: Region 1: 3, Region 2: 5, Region 3: 2; Cooling water flow path cost: Region 1 (the region where the condenser is located): 7, Region 2 (the region where the condenser is located): 4, Region 3 (the region where the condenser is located): 3; Assuming that both weights are 0.5, the comprehensive flow path cost: Region 1: 5, Region 2: 4.5, Region 3: 2.5; Sort by priority: Region 3, Region 2, Region 1.

[0107] Due to the particularity of hospitals, the chilled water paths and cooling water paths in operating rooms, ICUs and wards in hospitals generally have multi-circuit designs, which is a necessary condition for meeting the high performance requirements and safe operation of hospitals. After determining the priority, for each area, the chilled water and cooling water flow paths with the lowest path overhead are further selected. The flow rate and flow direction can be adjusted through electric valves and water pumps to ensure that even in high-priority areas, the path used is still the best in the local area.

[0108] S4 controls the operation of chillers and cooling towers in the corresponding areas through the adjustment strategies of chillers and cooling towers in different types of areas, the cooling water and chilled water regional allocation paths and their allocation priorities.

[0109] In this step, the adjustment strategies of the chiller and cooling tower obtained in the previous step are used to adjust the compressor frequency of the chiller, the chilled water flow rate, the fan speed of the cooling tower, and the cooling water flow rate. In the scenario where the demand for chilled water and cooling water is insufficient, the chilled water flow rate adjustment and the cooling water flow rate adjustment are allocated according to the regional priority order, and priority is given to areas with high priority demand.

[0110] It will be appreciated by those skilled in the art that embodiments of the present invention may be provided as methods, systems, or computer program products. Therefore, the present invention may take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware. Furthermore, the present invention may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0111] The present invention is described with reference to flowcharts and / or block diagrams of methods, devices (systems), and computer program products according to embodiments of the present invention. It should be understood that each process and / or block in the flowchart and / or block diagram, as well as the combination of processes and / or blocks in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0112] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0113] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0114] It should be noted that in the claims, any reference signs placed between brackets shall not be construed as limiting the claims. The word "comprising" does not exclude the presence of components or steps not listed in the claim. The word "a" or "an" preceding a component does not exclude the presence of a plurality of such components. The invention may be implemented by means of hardware comprising several different components and by means of a suitably programmed computer. In a unit claim enumerating several means, several of these means may be embodied by the same item of hardware. The use of the words first, second, and third etc. does not indicate any order. These words may be interpreted as names.

[0115] Although the preferred embodiments of the present invention have been described, those skilled in the art may make additional changes and modifications to these embodiments once they have learned the basic creative concept. Therefore, the appended claims are intended to be interpreted as including the preferred embodiments and all changes and modifications that fall within the scope of the present invention.

[0116] Obviously, those skilled in the art may make various changes and modifications to the present invention without departing from the spirit and scope of the present invention. Thus, if these modifications and variations of the present invention fall within the scope of the claims of the present invention and their equivalents, the present invention is also intended to include these modifications and variations.

[0117] In the present invention, unless otherwise clearly specified and limited, the terms "installed", "connected", "connected", "fixed" and the like should be understood in a broad sense, for example, it can be a fixed connection, a detachable connection, or an integral connection; it can be a mechanical connection or an electrical connection; it can be a direct connection or an indirect connection through an intermediate medium, it can be the internal connection of two elements or the interaction relationship between two elements. For ordinary technicians in this field, the specific meanings of the above terms in the present invention can be understood according to specific circumstances.

[0118] In the description of this specification, the description with reference to the terms "one embodiment", "some embodiments", "example", "specific example", or "some examples" etc. means that the specific features, structures, materials or characteristics described in conjunction with the embodiment or example are included in at least one embodiment or example of the present invention. In this specification, the schematic representation of the above terms should not be understood as necessarily referring to the same embodiment or example. Moreover, the specific features, structures, materials or characteristics described may be combined in any one or more embodiments or examples in a suitable manner. In addition, those skilled in the art may combine and combine the different embodiments or examples described in this specification and the features of the different embodiments or examples, unless they are contradictory.

Claims

1. A medical chiller IoT edge control method, characterized in that: include: The chiller operating parameters, cooling tower operating parameters, and environmental parameters of different types of areas in the hospital are obtained, and the chiller operating parameters, cooling tower operating parameters, and environmental parameters of different types of areas are input into the cooling load prediction model to obtain the future cooling load prediction values ​​of different types of areas; The chiller operating parameters, cooling tower operating parameters, and future cooling load forecasts for different types of areas are input into the deep reinforcement learning model to obtain adjustment strategies for chillers and cooling towers in different types of areas. Dynamically adjust cooling water and chilled water distribution paths and their distribution priorities according to different types of areas, future cooling load forecast values, and adjustment strategies of chillers and cooling towers; Control the operation of chillers and cooling towers in the corresponding areas through the adjustment strategies of chillers and cooling towers in different types of areas, the regional allocation paths of cooling water and chilled water, and their allocation priorities; The method of dynamically adjusting the cooling water and chilled water distribution paths and their distribution priorities according to different types of regions, future cooling load forecast values, and adjustment strategies of chillers and cooling towers includes: Construct a chilled water flow model based on chilled water flow nodes and paths, and calculate the cost of each chilled water flow path; Construct a cooling water flow model based on cooling water flow nodes and paths, and calculate the cost of each cooling water flow path; Calculate the comprehensive flow path cost of each area according to the chilled water flow path cost and the cooling water flow path cost, determine the chilled water and cooling water area allocation priority according to the comprehensive flow path cost, and select the allocation path according to the chilled water flow path cost and the cooling water flow path cost; The calculation of the cost of each chilled water flow path includes: defining regional priority factors according to the regional types of different types of regions; Define the chilled water load demand factor based on the future cooling load forecast; According to the frequency adjustment time of the compressor of the chiller, the frequency adjustment time of the compressor is obtained; The chilled water flow path cost is calculated by combining the regional priority factor, chilled water load demand factor, and compressor frequency adjustment time. The calculation formula is as follows: ； The calculation of the cost of each cooling water flow path includes: Define the cooling efficiency factor based on the cooling water temperature and the condenser demand temperature; The current cooling water load capacity is defined based on the ratio of the current load to the maximum capacity of the cooling tower; According to the time required for cooling tower fan adjustment, the fan start-up adjustment time is obtained; The cooling water flow path cost is calculated by combining the cooling efficiency factor, the current cooling water load capacity, and the fan start-up adjustment time. The calculation formula is as follows: 。 2. According to claim 1, a medical chiller IoT edge control method is characterized in that: The acquisition of chiller operating parameters, cooling tower operating parameters, and environmental parameters in different types of areas of the hospital includes: Obtain the evaporator temperature, condenser temperature, compressor frequency, and chilled water flow of the chiller; Get the cooling water inlet / outlet temperature, fan speed, and cooling water flow rate of the cooling tower; Get indoor temperature and humidity, outdoor temperature and humidity.

3. The method for edge control of the Internet of Things for medical chillers according to claim 2, characterized in that: The cooling load prediction model is obtained through the following steps: Collect the chiller operating parameters, cooling tower operating parameters, and environmental parameters at several different time points, calibrate the actual cooling load data at the corresponding time points, and use them as training samples for the cooling load prediction model; A cooling load prediction model is constructed through the long short-term memory network model. The chiller operating parameters, cooling tower operating parameters, and environmental parameters at past time points are used as the input of the cooling load prediction model, and the actual cooling load data at future time points are used as the output of the cooling load prediction model to train the cooling load prediction model.

4. The method for edge control of the medical chiller Internet of Things according to claim 3, characterized in that: The deep reinforcement learning model is obtained through the following steps: The future cooling load forecast value, environmental parameters, chiller operating parameters, and cooling tower operating parameters are used as state variables, and the compressor frequency adjustment amount and chilled water flow adjustment amount of the chiller, the fan speed adjustment amount and cooling water flow rate adjustment amount of the cooling tower are used as action variables to build a deep reinforcement learning model. A reward function is constructed based on the regional cooling load demand satisfaction rate and total energy consumption. When the reward value of the reward function is higher than the reward threshold, it is marked as a positive reward signal, otherwise it is marked as a negative reward signal. The deep reinforcement learning model is optimized according to the reward signal.

5. A medical chiller IoT edge control method according to claim 4, characterized in that: The reward function constructed by using the regional cooling load demand satisfaction rate and the total energy consumption includes: The cooling load demand satisfaction rate is calculated by using the actual cooling load data and the future cooling load forecast value. The calculation formula is as follows: ； The total energy consumption is calculated by adding the real-time power of the compressor, the real-time power of the evaporator, the real-time power of the condenser, and the real-time power of the fan; The reward function is obtained by weighting the cooling load demand satisfaction rate and the total energy consumption. The calculation formula is as follows: ， in, is the corresponding weight coefficient.

6. The method for edge control of the Internet of Things for medical chillers according to claim 1, characterized in that: The constructing of the chilled water flow model according to the chilled water flow nodes and paths comprises: The chilled water flow model is constructed by taking each area as the target node, the chiller as the source node, and the chilled water pipes connecting the target node and the source node as the chilled water path.

7. The method for edge control of the medical chiller Internet of Things according to claim 1, characterized in that: The construction of the cooling water flow model according to the cooling water flow nodes and paths includes: The condenser of the chiller is taken as the target node, the cooling tower is taken as the source node, and the cooling water pipe connecting the target node and the source node is taken as the cooling water path to construct the cooling water flow model.

Images