Air conditioner control system and air conditioner

Abstract

This invention discloses an air conditioning control system and an air conditioner. An environmental parameter acquisition module obtains environmental parameters and the actual operating parameters of the air conditioner. A model invocation module inputs the environmental parameters into a pre-trained energy consumption optimization model, which outputs the target operating parameters for the air conditioner. A PID control module receives the target operating parameters at its first input terminal and the actual operating parameters at its second input terminal, and outputs control commands. These commands control the air conditioner's operation, ensuring that the actual operating parameters reach the target operating parameters. When the air conditioner operates at the target parameters, its energy consumption is optimized. Therefore, this invention's air conditioning control system and air conditioner, combining an energy consumption optimization model and PID control, provide stable control of the air conditioner, achieving optimal energy consumption with stable and accurate control, thus solving the technical problems of poor control stability and high energy consumption in existing technologies.

Description

Technical Field

[0001] This invention relates to the field of air conditioning technology, and more particularly to air conditioning control systems and air conditioners. Background Technology

[0002] Central air conditioning systems are one of the largest sources of energy consumption in commercial buildings. Energy consumption is a key performance indicator for users, directly impacting their operating costs.

[0003] Central air conditioning control schemes can be divided into two main categories: traditional and intelligent control methods.

[0004] Most current intelligent control systems incorporate neural networks, which involve a large amount of computation and have relatively poor adjustment stability, thus requiring improvement.

[0005] Current air conditioning controls are energy-intensive and unstable, lacking a smart and stable energy-saving control strategy. Summary of the Invention

[0006] This invention proposes an air conditioning control system that solves the technical problems of poor control stability and high energy consumption in the prior art.

[0007] To achieve the above objectives, the present invention adopts the following technical solution:

[0008] This invention provides an air conditioning control system, comprising:

[0009] The parameter acquisition module is configured to acquire environmental parameters and the actual operating parameters of the air conditioner.

[0010] The model calling module is configured to input the environmental parameters into a pre-trained energy consumption optimization model, and the energy consumption optimization model outputs the target operating parameters of the air conditioner.

[0011] The PID control module is configured such that the target operating parameters are input at its first input terminal, the actual operating parameters are input at its second input terminal, and a control command is output at its output terminal to control the operation of the air conditioner.

[0012] In some embodiments of this application, the environmental parameters include: indoor ambient temperature and outdoor ambient temperature;

[0013] The target operating parameters include: target chilled water outlet temperature, target chilled water return temperature, and target cooling water flow rate;

[0014] The actual operating parameters include: actual chilled water outlet temperature, actual chilled water return temperature, and actual cooling water flow rate.

[0015] In some embodiments of this application, the PID control module includes:

[0016] The first PID control unit has a first input terminal for the target outlet temperature of chilled water, a second input terminal for the actual outlet temperature of chilled water, and an output terminal for the control command of chilled water outlet temperature.

[0017] The second PID control unit has a first input terminal for the target return water temperature of chilled water, a second input terminal for the actual return water temperature of chilled water, and an output terminal for the control command of the return water temperature of chilled water.

[0018] The third PID control unit has a first input terminal for the target flow rate of cooling water, a second input terminal for the actual flow rate of cooling water, and an output terminal for the cooling water flow control command.

[0019] In some embodiments of this application, the energy consumption optimization model is an agent model based on the TD3 algorithm.

[0020] In some embodiments of this application, the air conditioning control system further includes:

[0021] The model building module includes:

[0022] Air conditioning model building unit, which is used to build an air conditioning model for simulating air conditioning;

[0023] An energy consumption model establishment unit is used to establish an energy consumption optimization model based on the TD3 algorithm; the input parameters of the energy consumption optimization model are the environmental parameters of the air conditioning model, and the output parameters of the energy consumption optimization model are the target operating parameters of the air conditioning model.

[0024] A PID model building unit is used to build a PID control model. The first input terminal of the PID control model is the output parameters of the energy consumption optimization model, the second input terminal is the actual operating parameters of the air conditioning model, and the output terminal outputs control commands to the air conditioning model.

[0025] The air conditioning control system also includes:

[0026] The model training module uses the TD3 algorithm to train the energy consumption optimization model.

[0027] In some embodiments of this application, when training an energy consumption optimization model using the TD3 algorithm, the model updates its own hyperparameters according to the reward function.

[0028] In some embodiments of this application, the reward function r of the TD3 algorithm t for:

[0029]

[0030] Where cost is the real-time energy consumption of the air conditioner, λ is the temperature penalty coefficient, and T inIndoor ambient temperature, and T These are the upper and lower temperature limits.

[0031] In some embodiments of this application, training an energy consumption optimization model using the TD3 algorithm specifically includes:

[0032] Calculate the average reward value obtained in each n rounds of training;

[0033] Calculate the difference between the currently calculated average and the previously calculated average;

[0034] If the difference is within the set range, the training ends, and the trained energy consumption optimization model is obtained.

[0035] In some embodiments of this application, an air conditioning model, a PID control model, and an energy consumption optimization model are created in MATLAB, and the energy consumption optimization model is simulated and trained.

[0036] The present invention provides an air conditioner, including the aforementioned air conditioner control system.

[0037] The technical solution of this invention has the following advantages over the prior art: The air conditioning control system and air conditioner of this invention acquire environmental parameters and the actual operating parameters of the air conditioner through a parameter acquisition module. The model calling module inputs the environmental parameters into a pre-trained energy consumption optimization model, which outputs the target operating parameters of the air conditioner. The first input terminal of the PID control module inputs the target operating parameters, the second input terminal inputs the actual operating parameters of the air conditioner, and the output terminal outputs control commands. These control commands control the operation of the air conditioner, ensuring that the actual operating parameters reach the target operating parameters. When the air conditioner operates at the target operating parameters, its energy consumption is optimized. Therefore, the air conditioning control system and air conditioner of this invention, combining the energy consumption optimization model and PID control, provide stable control of the air conditioner, resulting in optimal energy consumption, stable control, and high accuracy, thus solving the technical problems of poor control stability and high energy consumption in the prior art.

[0038] Other features and advantages of the present invention will become clearer after reading the detailed embodiments of the invention in conjunction with the accompanying drawings. Attached Figure Description

[0039] To more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings used in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0040] Figure 1This is a schematic diagram of the structure of one embodiment of the air conditioning control system of the present invention;

[0041] Figure 2 A flowchart illustrating one embodiment of the steps performed by the air conditioning control system of the present invention;

[0042] Figure 3 A schematic diagram of one embodiment of a chiller unit;

[0043] Figure 4 A schematic diagram of one embodiment of a PID control module;

[0044] Figure 5 A schematic diagram of the structure of one embodiment of the model building module;

[0045] Figure 6 This is a connection diagram of the energy consumption optimization model, PID control model, and air conditioning model.

[0046] Figure 7 This is a schematic diagram of the TD3 agent structure;

[0047] Figure 8 A flowchart illustrating one embodiment of training an energy consumption optimization model using the TD3 algorithm;

[0048] Figure 9 A flowchart illustrating another embodiment of training an energy consumption optimization model using the TD3 algorithm;

[0049] Figure 10 This is a flowchart of another embodiment of training an energy consumption optimization model using the TD3 algorithm.

[0050] Figure label:

[0051] 11. Compressor; 12. Condenser; 13. Electronic expansion valve; 14. Evaporator;

[0052] 15. Chilled water pump; 16. Terminal equipment; 17. Cooling water pump; 18. Cooling tower. Detailed Implementation

[0053] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0054] In the description of this application, it should be understood that the terms "center", "upper", "lower", "front", "rear", "left", "right", "vertical", "horizontal", "top", "bottom", "inner", "outer", etc., indicate the orientation or positional relationship based on the orientation or positional relationship shown in the accompanying drawings. They are only for the convenience of describing this application and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation. Therefore, they should not be construed as limitations on this application.

[0055] The terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of technical features indicated. Therefore, a feature defined as "first" or "second" may explicitly or implicitly include one or more of that feature. In the description of this application, unless otherwise stated, "a plurality of" means two or more.

[0056] In the description of this application, it should be noted that, unless otherwise expressly specified and limited, the terms "installation," "connection," and "linking" should be interpreted broadly. For example, they can refer to a fixed connection, a detachable connection, or an integral connection; they can refer to a mechanical connection or an electrical connection; they can refer to a direct connection or an indirect connection through an intermediate medium; and they can refer to the internal connection between two components. Those skilled in the art can understand the specific meaning of the above terms in this application based on the specific circumstances.

[0057] In this invention, unless otherwise explicitly specified and limited, "above" or "below" the second feature can include direct contact between the first and second features, or contact between the first and second features through another feature between them. Furthermore, "above," "over," and "on top" of the second feature includes the first feature directly above or diagonally above the second feature, or simply indicates that the first feature is at a higher horizontal level than the second feature. "Below," "below," and "under" the second feature includes the first feature directly below or diagonally below the second feature, or simply indicates that the first feature is at a lower horizontal level than the second feature.

[0058] The following disclosure provides many different embodiments or examples for implementing various structures of the invention. To simplify the disclosure, specific examples of components and arrangements are described below. These are merely examples and are not intended to limit the invention. Furthermore, reference numerals and / or letters may be repeated in different examples; such repetition is for simplification and clarity and does not in itself indicate a relationship between the various embodiments and / or arrangements discussed. In addition, examples of various specific processes and materials are provided in this invention, but those skilled in the art will recognize the application of other processes and / or the use of other materials.

[0059] Air conditioners execute refrigeration and heating cycles using a compressor, condenser, expansion valve, and evaporator. These cycles are controlled by a controller, which manages the refrigerant flow and the opening of the expansion valve. The refrigeration and heating cycles involve a series of processes including compression, condensation, expansion, and evaporation, ultimately supplying refrigerant to the conditioned and heat-exchanged air.

[0060] The compressor compresses refrigerant gas under high temperature and pressure and discharges the compressed refrigerant gas. The discharged refrigerant gas flows into the condenser. The condenser condenses the compressed refrigerant into a liquid phase, and the heat is released to the surrounding environment through the condensation process.

[0061] The expansion valve expands the high-temperature, high-pressure liquid refrigerant condensed in the condenser into a low-pressure liquid refrigerant. The evaporator evaporates the expanded refrigerant in the expansion valve, returning the low-temperature, low-pressure refrigerant gas to the compressor. The evaporator achieves its cooling effect by utilizing the latent heat of refrigerant evaporation to exchange heat with the material being cooled. Throughout the cycle, the air conditioner regulates the temperature of the indoor space.

[0062] An air conditioner outdoor unit refers to the part of the refrigeration cycle that includes the compressor and the outdoor heat exchanger. An air conditioner indoor unit includes the indoor heat exchanger, and an expansion valve can be provided in either the outdoor or indoor unit.

[0063] The indoor and outdoor heat exchangers function as either condensers or evaporators. When the indoor heat exchanger is used as a condenser, the air conditioner functions as a heater in heating mode; when the indoor heat exchanger is used as an evaporator, the air conditioner functions as a cooler in cooling mode.

[0064] The air conditioning control system in this embodiment includes a parameter acquisition module, a model calling module, a PID control module, etc., see [link / reference]. Figure 1 As shown.

[0065] The parameter acquisition module is configured to acquire environmental parameters and the actual operating parameters of the air conditioner.

[0066] The model calling module is configured to input the environmental parameters obtained by the parameter acquisition module into the pre-trained energy consumption optimization model, and the energy consumption optimization model outputs the target operating parameters of the air conditioner.

[0067] The PID control module is configured such that the target operating parameters are input at its first input terminal, the actual operating parameters of the air conditioner are input at its second input terminal, and control commands are output at its output terminal to control the operation of the air conditioner.

[0068] The system uses sensors to acquire environmental parameters and the actual operating parameters of the air conditioner, and then transmits them to the parameter acquisition module.

[0069] The energy consumption optimization model takes environmental parameters as input and outputs target operating parameters that optimize air conditioning energy consumption based on these parameters. When the air conditioner operates at these target parameters, its energy consumption will be optimal.

[0070] The PID control module uses the PID control algorithm. The PID control module calculates the deviation between the target operating parameters and the actual operating parameters of the air conditioner. Based on continuously eliminating the deviation between the two, it outputs control commands to control the operation of the air conditioner, thereby making the actual operating parameters reach the target operating parameters and achieving optimal energy consumption of the air conditioner.

[0071] The PID control algorithm provides a relatively stable output and can compensate for fluctuations in the energy consumption optimization model, thus achieving stable control of the air conditioner.

[0072] For the air conditioning control system, please refer to the following steps. Figure 2 As shown.

[0073] Step S11: The parameter acquisition module acquires environmental parameters.

[0074] Step S12: The model calling module inputs environmental parameters into the pre-trained energy consumption optimization model, and the energy consumption optimization model outputs the target operating parameters of the air conditioner.

[0075] Step S13: The parameter acquisition module acquires the actual operating parameters of the air conditioner.

[0076] Step S14: The first input terminal of the PID control module receives the target operating parameters, the second input terminal receives the actual operating parameters, and the output terminal outputs control commands. The control commands control the operation of the air conditioner so that the actual operating parameters of the air conditioner reach the target operating parameters, thereby achieving optimal energy consumption of the air conditioner.

[0077] The air conditioning control system in this embodiment acquires environmental parameters and the actual operating parameters of the air conditioner through a parameter acquisition module. A model invocation module inputs the environmental parameters into a pre-trained energy consumption optimization model, which outputs the target operating parameters for the air conditioner. The PID control module receives the target operating parameters at its first input and the actual operating parameters at its second input, and outputs control commands. These commands control the air conditioner's operation, ensuring that the actual operating parameters reach the target parameters. When the air conditioner operates at the target parameters, its energy consumption is optimized. Therefore, the air conditioning control system in this embodiment, combining the energy consumption optimization model and PID control, provides stable control of the air conditioner, achieving optimal energy consumption with stable and accurate control, thus solving the technical problems of poor control stability and high energy consumption in existing technologies.

[0078] In some embodiments of this application, "air conditioning" refers to a central air conditioning system, specifically a chiller unit. See [link to relevant documentation]. Figure 3 As shown, the compressor 11, condenser 12, electronic expansion valve 13, and evaporator 14 form a refrigerant circulation system, which performs the air conditioning cooling cycle and heating cycle.

[0079] The water flowing through the evaporator 14 exchanges heat with the refrigerant, and the chilled water is sent to the indoor terminal equipment 16 by the chilled water pump 15 to achieve cooling of the indoor space or equipment.

[0080] The water flowing through the condenser 12 exchanges heat with the refrigerant, and the cooling water is sent to the cooling tower 18 by the cooling water pump 17. The water in the cooling tower 18 exchanges heat with the air.

[0081] In some embodiments of this application, environmental parameters include: indoor ambient temperature (the indoor ambient temperature where the terminal device is located) and outdoor ambient temperature.

[0082] The target operating parameters include: target chilled water outlet temperature, target chilled water return temperature, and target cooling water flow rate.

[0083] Actual operating parameters include: actual chilled water outlet temperature, actual chilled water return temperature, and actual cooling water flow rate.

[0084] Indoor and outdoor ambient temperatures can comprehensively characterize the environmental parameters of an air conditioner.

[0085] Chilled water outlet temperature, chilled water return temperature, and cooling water flow rate can comprehensively characterize the operating parameters of the chiller unit, enabling precise control of the chiller unit.

[0086] Multiple temperature sensors are used to collect indoor ambient temperature, outdoor ambient temperature, actual chilled water outlet temperature, and actual chilled water return temperature; a flow sensor is used to collect the actual cooling water flow rate, and then the collected information is transmitted to the parameter acquisition module.

[0087] The parameter acquisition module acquires the indoor ambient temperature, outdoor ambient temperature, and the actual outlet temperature, return temperature, and flow rate of the chilled water in the air conditioner.

[0088] The energy consumption optimization model takes indoor and outdoor ambient temperatures as inputs and outputs target chilled water outlet temperature, target chilled water return temperature, and target cooling water flow rate.

[0089] The PID control module outputs control commands based on the target outlet temperature of chilled water, the target return temperature of chilled water, the target flow rate of cooling water, and the actual outlet temperature of chilled water, the actual return temperature of chilled water, and the actual flow rate of cooling water in the air conditioner. This controls the operation of the air conditioner so that the actual outlet temperature of chilled water, the actual return temperature of chilled water, and the actual flow rate of cooling water reach the target outlet temperature of chilled water, the target return temperature of chilled water, and the target flow rate of cooling water, respectively, thereby optimizing the energy consumption of the air conditioner.

[0090] In some embodiments of this application, the control commands are directed to the compressor, the electronic expansion valve, and the cooling water pump, so that the actual outlet temperature of the chilled water reaches the target outlet temperature of the chilled water, the actual return temperature of the chilled water reaches the target return temperature of the chilled water, and the actual flow rate of the cooling water reaches the target flow rate of the cooling water.

[0091] In some embodiments of this application, to facilitate the implementation of PID control and improve control efficiency, the PID control module includes a first PID control unit, a second PID control unit, a third PID control unit, etc. (See also...) Figure 4 As shown.

[0092] The first PID control unit receives the target chilled water outlet temperature at its first input terminal and the actual chilled water outlet temperature at its second input terminal. Its output terminal outputs a chilled water outlet temperature control command. This command controls the operation of the air conditioner, ensuring that the actual chilled water outlet temperature reaches the target chilled water outlet temperature.

[0093] The second PID control unit has a first input terminal for the target chilled water return temperature, a second input terminal for the actual chilled water return temperature, and an output terminal for a chilled water return temperature control command. The air conditioner is controlled by this chilled water return temperature control command to ensure that the actual chilled water return temperature reaches the target chilled water return temperature.

[0094] The third PID control unit receives the target cooling water flow rate at its first input terminal, the actual cooling water flow rate at its second input terminal, and outputs a cooling water flow rate control command at its output terminal. This command controls the air conditioner's operation, ensuring the actual cooling water flow rate reaches the target flow rate.

[0095] In some embodiments of this application, the energy consumption optimization model is an agent model based on the TD3 algorithm. The energy consumption optimization model is trained using the TD3 algorithm.

[0096] The TD3 algorithm is a deep reinforcement learning algorithm that combines deep deterministic policy gradient algorithm and dual Q-learning. The TD3 algorithm exhibits good learning stability and is suitable for handling problems with continuous action spaces, thus improving both the stability and efficiency of the algorithm.

[0097] The agent model based on the TD3 algorithm has good stability and high training efficiency.

[0098] In some embodiments of this application, the air conditioning control system further includes a model building module and a model training module. The model building module includes an air conditioning model building unit, an energy consumption model building unit, a PID model building unit, etc., see [link to relevant documentation]. Figure 5 As shown.

[0099] The air conditioning model building unit is used to build an air conditioning model for simulating air conditioning.

[0100] The energy consumption model building unit is used to construct an energy consumption optimization model based on the TD3 algorithm. The input parameters of the energy consumption optimization model are the environmental parameters of the air conditioning model, and the output parameters of the energy consumption optimization model are the target operating parameters of the air conditioning model.

[0101] The PID model building unit is used to build the PID control model. The first input of the PID control model receives the output parameters of the energy consumption optimization model, and the second input receives the actual operating parameters of the air conditioning model. Its output sends control commands to the air conditioning model. (See [link / reference]). Figure 6 As shown.

[0102] The model training module uses the TD3 algorithm to train an energy consumption optimization model.

[0103] By establishing an air conditioning model and a PID control model, and training the energy consumption optimization model together, the number of training rounds of the TD3 algorithm can be reduced, the stability of the output of the energy consumption optimization model can be improved, and the output action can be prevented from deviating excessively from the reference value.

[0104] The trained energy consumption optimization model can output reasonable target operating parameters to optimize the energy consumption of the air conditioner.

[0105] In some embodiments of this application, when training an energy consumption optimization model using the TD3 algorithm, the hyperparameters are updated according to the reward function to accelerate convergence, improve training speed, and train an accurate energy consumption optimization model.

[0106] In terms of modeling, the operation of the central air conditioning system maintains the required temperature in each area based on the current temperature and external environmental disturbances. The temperature of the area in the next time step is determined only by the current state of the system, environmental disturbances, and the air input from the air conditioning system. Therefore, it can be described by a Markov process in deep reinforcement learning, which is represented by a quadruple {s, a, p, r}, where s is the environmental state (environmental parameter) that the TD3 agent needs to be familiar with, a is the action that the air conditioning system can take, p is the probability, and r is the reward obtained after taking the action.

[0107] The total energy consumption of an air conditioner is the sum of the power consumption of the refrigeration system and the power consumption of the water pump. The energy consumption of the water pump is related to the temperature difference between the inlet and outlet of the chilled water. Considering the energy consumption pattern of the refrigeration system, the factors affecting the energy consumption of the air conditioning unit are divided into two parts: one part is related to temperature, namely the inlet and outlet temperatures of the chilled water and cooling water; the other part is related to flow rate, namely the flow rates of the chilled water and cooling water. Therefore, the actions output by the TD3 intelligent agent are the chilled water outlet temperature, the chilled water return temperature, and the cooling water flow rate.

[0108] In the process of reducing air conditioning energy consumption, comfort requirements were taken into account to ensure that the system's outlet air temperature was within the thermal comfort range. Therefore, the outdoor ambient temperature and indoor temperature were used as the state space of the deep reinforcement learning algorithm.

[0109] The TD3-PID algorithm proposed in this application is integrated with central air conditioning as follows: Figure 7 As shown, the action value output by the TD3 agent is input into the PID algorithm, and then the PID algorithm outputs the control action (control command) for the central air conditioning system.

[0110] The TD3 algorithm comprises six neural networks, each consisting of an input layer, hidden layers, and an output layer. The agent is trained through iterative updates within these networks. The TD3 algorithm updates the output of the evaluation network by using the minimum value of the target Q output from two evaluation networks, thus avoiding overestimation of actions and improving training robustness.

[0111] The advantage of combining TD3 with PID algorithm is that it reduces the number of training rounds of TD3 algorithm and improves the stability of control command output, avoiding excessive deviation of the output action from the reference value.

[0112] In terms of specific training, the cooling area of ​​the central air conditioning system is used as the environment. The temperature of the cooling area (indoor temperature) and the outdoor temperature are input as state variables into the TD3 algorithm agent. Simultaneously, the reward value corresponding to the previous action is provided. The agent updates its hyperparameters based on the state and reward value from the environment and returns the new action to the environment. After multiple rounds of algorithm iteration, the training of the agent is completed. The trained agent is then implanted into the air conditioning controller to reduce the energy consumption of the air conditioning system while ensuring indoor thermal comfort.

[0113] Regarding the hyperparameter settings of the algorithm, this application takes into account the energy consumption of the air conditioner and the user experience, and transforms them into the design of the algorithm reward function, which evaluates the user's physical sensation through temperature values.

[0114] Minimizing energy consumption and maintaining an ideal temperature to ensure comfort are contradictory goals. Since the reward function is an important basis for training the TD3 algorithm to update its own parameters and evaluate the quality of actions, this application adopts the method of designing a reward function to balance energy consumption and comfort.

[0115] In some embodiments of this application, the reward function r of the TD3 algorithm t for:

[0116]

[0117] Where cost is the cost(s) t ,a t ), where λ is the real-time energy consumption of the air conditioner, and T is the temperature penalty coefficient. in Indoor ambient temperature, and T These are the upper and lower limits of temperature to maintain indoor comfort.

[0118] To minimize the total energy consumption of the central air conditioning system, a negative sign is added before the real-time energy consumption figure. It is worth noting that the comfort evaluation in this reward function uses only temperature as a criterion, reducing the complexity of algorithm training and improving its efficiency.

[0119] In some embodiments of this application, the training of the energy consumption optimization model using the TD3 algorithm specifically includes the following steps, see below. Figure 8 As shown.

[0120] Step S21: Calculate the average reward value obtained in each n rounds of training.

[0121] In each training round, the reward function calculates a reward value.

[0122] After n rounds of training, n reward values ​​are obtained, and the average of these n reward values ​​is calculated.

[0123] Step S22: Calculate the difference between the currently calculated average and the previously calculated average.

[0124] Step S23: Determine whether the difference is within the set range.

[0125] If the difference is not within the set range, the next round of training will continue.

[0126] If the difference is within the set range, it means that the difference is very small. Then, proceed to step S24: training ends, and the trained energy consumption optimization model is obtained.

[0127] For example, n is 50. Every 50 training rounds, the average reward value of the 50 rounds is calculated. The average reward value is then compared with the average reward value of the previous 50 rounds. If the difference between the two is within the set range, it means that the TD3 algorithm has converged and the training is over.

[0128] By designing steps S21 to S24, the difference between the average reward value of the current n rounds and the average reward value of the previous n rounds is calculated. If the difference is within the set range, it indicates that the TD3 algorithm has converged, the judgment is accurate, and the training is terminated in time to avoid excessive training time.

[0129] In some embodiments of this application, when training the energy consumption optimization model using the TD3 algorithm, the following steps are performed, see [link to relevant documentation]. Figure 9 As shown.

[0130] Step S31: Count the number of training rounds.

[0131] Step S32: Determine whether the target number of rounds has been reached.

[0132] If the target number of training rounds has not been reached, continue training.

[0133] If the number of training rounds reaches the target number of rounds, then proceed to step S33: training ends, and the trained energy consumption optimization model is obtained.

[0134] By designing steps S31 to S33, training can be terminated promptly when the target number of training rounds is reached, thus avoiding excessively long training times.

[0135] In some embodiments of this application, an air conditioning model, a PID control model, and an energy consumption optimization model are created in MATLAB, and the energy consumption optimization model is simulated and trained.

[0136] Using MATLAB for simulation training is easy to implement and inexpensive.

[0137] This application creatively proposes a PC-based programming and training scheme for the TD3 algorithm to regulate central air conditioning. The TD3 algorithm can be programmed and trained using only a single MATLAB software, eliminating the need to directly apply the TD3 algorithm to a real-world environment and reducing trial-and-error costs.

[0138] This application creatively proposes a training scheme for the TD3-PID algorithm in the field of central air conditioning. This application uses only MATLAB to implement the simulation of the central air conditioning system and the training and interaction of the TD3-PID algorithm. The specific process is as follows: Figure 10 As shown.

[0139] Step S41: Begin.

[0140] Step S42: Create the model in Simulink in MATLAB.

[0141] Create an air conditioning model and a PID control model (PID control strategy) in Simulink in MATLAB.

[0142] Step S43: Write the TD3 algorithm in the m file.

[0143] That is, to construct an energy consumption optimization model.

[0144] Step S44: Set the simulation time and step size.

[0145] Step S45: Simulink interacts with the m-file.

[0146] Step S46: Determine whether the target number of rounds has been reached or the algorithm has converged.

[0147] If not, return to step S45.

[0148] If so, proceed to step S47: End.

[0149] An air conditioning model and PID control strategy (PID control model) are created in Simulink within MATLAB. The TD3 algorithm code is then written in an m-file, and the simulation start and end times and step size are set, using a clock signal for simulation. After the program runs, the algorithm training interface displayed in MATLAB is observed. The convergence of the TD3-PID algorithm is evaluated by combining the output Q-value, the reward value obtained in each round, and the average reward value over 50 rounds. After several consecutive iterations, the simulation training of the TD3 agent is completed, resulting in a trained TD3 agent (energy consumption optimization model).

[0150] Finally, the parameters of the deep neural network inside the trained TD3 agent are extracted and embedded into the control strategy (air conditioning controller), which enables the actual deployment of the deep reinforcement learning algorithm and realizes intelligent control in actual engineering, achieving the goal of balancing air conditioning energy consumption and user comfort.

[0151] The air conditioning control system of this application exhibits stable output through PID control, which can compensate for fluctuations in the neural network. This application proposes a scheme for controlling a central air conditioning system using a combination of PID control and a single deep reinforcement learning TD3 algorithm (i.e., TD3-PID). This scheme effectively regulates the dynamically nonlinear central air conditioning system while shortening the training time of the TD3 algorithm and improving output stability. Furthermore, this application creatively proposes a PC-based programming and training scheme for the TD3 algorithm to regulate the central air conditioning system. The TD3 algorithm can be programmed and trained using only a single MATLAB software, eliminating the need to directly apply the TD3 algorithm to a real-world environment and reducing trial-and-error costs. Regarding the hyperparameter settings of the algorithm, this application considers air conditioning energy consumption and user experience, transforming these into the design of an algorithm reward function, using temperature values ​​to assess the user's comfort.

[0152] This application has three key technical advantages:

[0153] First, this application adopts a TD3-PID scheme that combines a single TD3 algorithm with PID control, reducing the training difficulty and output stability of the TD3 algorithm. It boasts high training efficiency and strong practicality.

[0154] Secondly, this application proposes to use only MATLAB to implement the simulation of the central air conditioning system, the writing of the TD3 algorithm, and the interaction of the air conditioning system algorithm.

[0155] Third, the hyperparameter configuration in this application is primarily based on temperature, and the reward function settings take into account air conditioning energy consumption and user experience. This avoids considering too many parameters and improves computational efficiency.

[0156] Based on the design of the air conditioning control system described above, this embodiment also proposes an air conditioner, including the aforementioned air conditioning control system.

[0157] By designing the aforementioned air conditioning control system in the air conditioner, the air conditioner can be stably controlled, resulting in optimal energy consumption, stable control, and high accuracy. This solves the technical problems of poor control stability and high energy consumption in the prior art.

[0158] The air conditioner in this embodiment takes into account both energy consumption and user comfort, and can reduce air conditioner energy consumption while ensuring user experience.

[0159] In some embodiments of this application, the air conditioning control system can be integrated onto the air conditioning controller.

[0160] In the description of the above embodiments, specific features, structures, materials, or characteristics may be combined in any suitable manner in one or more embodiments or examples.

[0161] The above are merely specific embodiments 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. An air conditioning control system, characterized in that, include: The parameter acquisition module is configured to acquire environmental parameters and the actual operating parameters of the air conditioner; the model invocation module is configured to input the environmental parameters into a pre-trained energy consumption optimization model, and the energy consumption optimization model outputs the target operating parameters of the air conditioner. The PID control module is configured such that the target operating parameters are input at its first input terminal, the actual operating parameters are input at its second input terminal, and a control command is output at its output terminal to control the operation of the air conditioner.

2. The air conditioning control system according to claim 1, characterized in that: The environmental parameters include: indoor ambient temperature and outdoor ambient temperature; The target operating parameters include: target chilled water outlet temperature, target chilled water return temperature, and target cooling water flow rate; The actual operating parameters include: actual chilled water outlet temperature, actual chilled water return temperature, and actual cooling water flow rate.

3. The air conditioning control system according to claim 2, characterized in that: The PID control module includes: The first PID control unit has a first input terminal for the target outlet temperature of chilled water, a second input terminal for the actual outlet temperature of chilled water, and an output terminal for the control command of chilled water outlet temperature. The second PID control unit has a first input terminal for the target return water temperature of chilled water, a second input terminal for the actual return water temperature of chilled water, and an output terminal for the control command of the return water temperature of chilled water. The third PID control unit has a first input terminal for the target flow rate of cooling water, a second input terminal for the actual flow rate of cooling water, and an output terminal for the cooling water flow control command.

4. The air conditioning control system according to claim 1, characterized in that: The energy consumption optimization model is an agent model based on the TD3 algorithm.

5. The air conditioning control system according to any one of claims 1 to 4, characterized in that: The air conditioning control system also includes: The model building module includes: Air conditioning model building unit, which is used to build an air conditioning model for simulating air conditioning; An energy consumption model establishment unit is used to establish an energy consumption optimization model based on the TD3 algorithm; the input parameters of the energy consumption optimization model are the environmental parameters of the air conditioning model, and the output parameters of the energy consumption optimization model are the target operating parameters of the air conditioning model. A PID model building unit is used to build a PID control model. The first input terminal of the PID control model is the output parameters of the energy consumption optimization model, the second input terminal is the actual operating parameters of the air conditioning model, and the output terminal outputs control commands to the air conditioning model. The air conditioning control system also includes: The model training module uses the TD3 algorithm to train the energy consumption optimization model.

6. The air conditioning control system according to claim 5, characterized in that: When training an energy consumption optimization model using the TD3 algorithm, the model updates its own hyperparameters based on the reward function.

7. The air conditioning control system according to claim 6, characterized in that: The reward function r of the TD3 algorithm t for: Where cost is the real-time energy consumption of the air conditioner, λ is the temperature penalty coefficient, and T in Indoor ambient temperature, and T These are the upper and lower temperature limits.

8. The air conditioning control system according to claim 5, characterized in that: When training an energy consumption optimization model using the TD3 algorithm, the specific steps include: Calculate the average reward value obtained in each n rounds of training; Calculate the difference between the currently calculated average and the previously calculated average; If the difference is within the set range, the training ends, and the trained energy consumption optimization model is obtained.

9. The air conditioning control system according to claim 5, characterized in that: An air conditioning model, a PID control model, and an energy consumption optimization model were created in MATLAB, and the energy consumption optimization model was simulated and trained.

10. An air conditioner, characterized in that, include: The air conditioning control system as described in any one of claims 1 to 9.

Images