WO2021063033A1

WO2021063033A1 - Energy consumption model training method for air conditioner and air conditioning system control method

Info

Publication number: WO2021063033A1
Application number: PCT/CN2020/097223
Authority: WO
Inventors: 汤潮; 郭琦; 袁德玉
Original assignee: 北京国双科技有限公司
Priority date: 2019-09-30
Filing date: 2020-06-19
Publication date: 2021-04-08
Also published as: CN112577161A; CN112577161B

Abstract

The present application relates to an energy consumption model training method for an air conditioning system, an air conditioning system control method and apparatus, a computer device, and a storage medium. The method comprises: respectively constructing energy consumption mechanism models of constituent devices according to input parameters and output parameters of the air conditioning system; training the energy consumption mechanism models on the basis of historical operation data; respectively constructing energy consumption models of the constituent devices; determining a control variable constraint and a state variable constraint of the air conditioning system; solving the optimum variables in the energy consumption models in a reasonable interval of a control variable by means of a genetic algorithm and a gradient optimization algorithm; and obtaining an air conditioning system control optimization policy.

Description

Air conditioning energy consumption model training method and air conditioning system control method

This application claims the priority of a Chinese patent application filed with the Chinese Patent Office, the application number is 201910941341.1, and the invention title "Air-conditioning energy consumption model training method and air-conditioning system control method" on September 30, 2019, the entire content of which is incorporated by reference In this application.

Technical field

This application relates to the technical field of air-conditioning systems, and specifically to a method for training an air-conditioning energy consumption model and a method for controlling the air-conditioning system.

Background technique

Air conditioning systems are currently widely used, and most of the scenarios such as public buildings, commercial buildings, and data centers use this system to achieve temperature control and dehumidification functions. Large-scale applications have also brought huge amounts of energy consumption. Taking data centers as an example, as of 2019, data center energy consumption accounted for more than 1.3% of global energy use, of which air conditioning systems accounted for more than 40%, while commercial Buildings and public buildings account for a higher proportion of energy consumption.

In order to reduce the energy consumption of the air-conditioning system and realize the energy-saving effect, there are currently a variety of energy-saving control methods for the air-conditioning system. Take the energy-saving control method for central air-conditioning system as an example. At present, BA (Building Automation System) has been popularized, and PID (closed-loop control system, closed-loop control)-based control methods are also widely used. Set the temperature difference between the inlet and outlet of the chilled water to control the frequency of the refrigerating pump, and set the temperature difference between the inlet and outlet temperature of the cooling water and the outdoor wet bulb temperature to control the frequency of the cooling tower. To a certain extent, the operating parameters of the equipment can be adjusted adaptively according to the environment.

However, the specific setting of the temperature difference and the chilled water outlet temperature in the above control process are actually set based on manual experience and cannot be adjusted adaptively according to the indoor and outdoor environments. For the building system, the environmental quantity will undoubtedly change. The temperature, humidity, load, etc. will all change with time. The artificial fixed value cannot be changed according to the perception of the environmental quantity, so it is impossible to achieve a better air-conditioning system. Energy-saving control.

Summary of the invention

Based on this, it is necessary to provide an air conditioning energy consumption model training method that can achieve good energy consumption prediction and an air conditioning system control method, device, computer device, and storage medium that can achieve good energy saving effects in response to the above technical problems.

An air conditioning energy consumption model training method, the method includes:

Determine the input parameters and output parameters of the components of the air conditioning system;

According to the input parameters and the output parameters, obtain the energy consumption mechanism model corresponding to the component equipment in the air conditioning system;

Obtain historical operating data of the air-conditioning system, train the energy consumption mechanism model according to the historical operating data, and obtain an energy consumption model of the components of the air-conditioning system.

In one of the embodiments, the obtaining the energy consumption mechanism model corresponding to the component equipment in the air conditioning system according to the input parameter and the output parameter includes:

Determine the mapping relationship between the input parameters and output parameters of the components of the air conditioning system according to the input parameters and the output parameters;

According to the mapping relationship, the energy consumption mechanism model relation formula of the component equipment in the air conditioning system is obtained, and the energy consumption mechanism model of the component equipment in the air conditioning system is constructed.

In one of the embodiments, the acquiring historical operating data of the air-conditioning system, training the energy consumption mechanism model according to the historical operating data, and obtaining the energy consumption model of the components of the air-conditioning system includes:

Acquiring historical operating data of the air-conditioning system;

Randomly divide the historical operation data into a training set part and a test set part;

Train the energy consumption mechanism model through the training set part, update the undetermined parameter values in the energy consumption mechanism model relational expression, and obtain the energy consumption mechanism model after training;

The energy consumption mechanism model after training is tested through the test set part, and when the test is passed, the energy consumption model of the component equipment in the air conditioning system is obtained.

In one of the embodiments, the obtaining historical operating data of the air-conditioning system includes:

Perform data integration and filling processing on the original historical operating data of the constituent equipment in the air conditioning system;

Based on the relevance of the operating data between the components of the air-conditioning system, remove abnormal data from the data after data integration and filling processing;

The historical operating data of the air-conditioning system in a stable operating state is extracted from the processed data after the abnormal data is eliminated.

In addition, this application also provides an air conditioning energy consumption model training device, which includes:

The parameter determination module is used to determine the input parameters and output parameters of the components of the air conditioning system;

An energy consumption mechanism model building module, configured to obtain an energy consumption mechanism model corresponding to the constituent devices in the air conditioning system according to the input parameters and the output parameters;

The energy consumption model building module is used to obtain historical operation data of the air conditioning system, and train the energy consumption mechanism model according to the historical operation data to obtain the energy consumption model of the components of the air conditioning system

The above air conditioning energy consumption model training method and device constructs the energy consumption mechanism model of each component according to the input and output parameters of the air conditioning system, and trains the energy consumption mechanism model based on historical operating data, and constructs the energy consumption of each component. model. In the whole process, on the one hand, the mechanism model is used to construct the energy consumption model to give full play to the global generalization ability of the mechanism model; on the other hand, the energy consumption model is constructed separately for the components of the entire air conditioning system, which is closer to the real energy consumption changes of the air conditioning system. Under the circumstances, the constructed energy consumption model can accurately realize the energy consumption prediction of the air-conditioning system.

An air conditioning system control method, the method includes:

Obtain the control quantity constraint and the state quantity constraint of the component equipment in the air-conditioning system, and determine a reasonable interval of the component equipment control quantity in the air-conditioning system according to the control quantity constraint and the state quantity constraint;

With the minimum energy consumption as the goal, the energy consumption model is optimized by genetic algorithm and gradient optimization algorithm, and the optimal control quantity combination of the air conditioning system is solved within the reasonable interval to obtain the air conditioning system control optimization strategy. The energy consumption model is an energy consumption model trained by the above air conditioning energy consumption model training method.

In one of the embodiments, the variable optimization of the energy consumption model by using a genetic algorithm and a gradient optimization algorithm includes:

For floating-point variables in the energy consumption model, perform variable optimization through genetic algorithm or gradient optimization algorithm;

For the integer variables in the energy consumption model, the number of integer variable combinations is obtained. When the number of integer variable combinations is less than the preset value that can be traversed, the variable optimization is performed through the gradient optimization algorithm. When the integer variable combination When the number is greater than the preset value and can be traversed, the genetic algorithm is used to optimize the variables.

In one of the embodiments, obtaining the control quantity constraint and the state quantity constraint of the component equipment in the air conditioning system, and determining a reasonable interval of the control quantity of the component equipment in the air conditioning system according to the control quantity constraint and the status quantity constraint includes:

Acquiring the environmental quantity constraint value and the state quantity constraint of the air-conditioning system and the current environmental value of the air-conditioning system;

If the current environmental value exceeds the environmental quantity constraint value, a factor is generated according to the extent to which the current environmental value exceeds the environmental quantity constraint value;

According to the current environmental value and the factor, setting a first range of the control amount of the component equipment in the air-conditioning system;

According to the state quantity constraint, obtain a reasonable operating range of the component equipment state quantity in the air-conditioning system;

According to the reasonable operating range of the state quantity, calculate the second range of the control quantity of the component equipment in the air-conditioning system;

The intersection of the first range of the control quantity and the second range of the control quantity is obtained, and a reasonable interval of the control quantity of the constituent equipment in the air conditioning system is obtained.

In one of the embodiments, the minimum energy consumption is the goal, the energy consumption model is optimized by genetic algorithm and gradient optimization algorithm, and the optimal control variable combination of the air conditioning system is solved within the reasonable interval. After obtaining the control optimization strategy of the air-conditioning system, it also includes:

Obtain the number of control parameters of the air conditioning system and the optional value of the parameter offset;

Establishing an orthogonal table according to the number and the optional value of the parameter offset;

The energy consumption model is updated iteratively on the basis of the non-replay extraction offset in the orthogonal table to the control value corresponding to the control optimization strategy of the air conditioning system.

In addition, the present application also provides a control device for an air conditioning system, the device including:

An interval determination module, configured to obtain the control quantity constraint and the state quantity constraint of the component equipment in the air-conditioning system, and determine a reasonable interval of the component equipment control quantity in the air-conditioning system according to the control quantity constraint and the state quantity constraint;

The control optimization module is used to optimize the variables of the energy consumption model through genetic algorithm and gradient optimization algorithm with the goal of minimum energy consumption, and solve the optimal control quantity combination of the air conditioning system within the reasonable interval to obtain the air conditioner The system control optimization strategy, wherein the energy consumption model is an energy consumption model trained by the above-mentioned air conditioning energy consumption model training device.

A computer device, including at least one processor, at least one memory, and a bus; wherein the processor and the memory communicate with each other through the bus; the processor is used to call a program in the memory Instruction to execute the above method.

A computer-readable storage medium, on which a computer program is stored, and when the computer program is executed by a processor, the steps of the above-mentioned method are realized.

The above-mentioned air conditioning system control method, device, computer equipment and storage medium construct the energy consumption mechanism model of each component according to the input and output parameters of the air conditioning system, and train the energy consumption mechanism model based on historical operating data to construct each component separately The energy consumption model of the equipment determines the control quantity constraints and state quantity constraints of the air conditioning system, and uses genetic algorithms and gradient optimization algorithms to solve the optimal variables in the energy consumption model within a reasonable interval of the control quantity, and obtains the control optimization strategy of the air conditioning system. In the whole process, within the effective boundary range of the environmental quantity constraint and the control quantity constraint, the energy consumption model of the equipment in the air-conditioning system is respectively solved for the optimal variables based on the genetic algorithm and the gradient optimization algorithm, and the energy in the current environment can be accurately obtained. The control strategy of the air-conditioning system with the least energy consumption achieves good energy-saving effects of the air-conditioning system.

Description of the drawings

Figure 1 is an application environment diagram of an air conditioning system control method in an embodiment;

FIG. 2 is a schematic flowchart of a training method for an air conditioning energy consumption model in an embodiment;

FIG. 3 is a schematic flowchart of a method for training an air conditioning energy consumption model in another embodiment;

Figure 4 is a schematic flow chart of an air conditioning system control method in an embodiment;

Figure 5 is a schematic diagram of an optimized algorithm decision tree;

Fig. 6 is a schematic flowchart of a control method of an air conditioning system in another embodiment;

Figure 7 is a schematic diagram of the processing flow of the boundary condition processing method;

Figure 8 is a schematic structural diagram of an air conditioning energy consumption model training device in an embodiment;

Figure 9 is a schematic diagram of the structure of an air conditioning system control device in an embodiment;

Fig. 10 is a diagram of the internal structure of a computer device in an embodiment.

Detailed ways

In order to make the purpose, technical solutions, and advantages of this application clearer, the following further describes the application in detail with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described here are only used to explain the application, and not used to limit the application.

The air conditioning system control method provided in this application can be applied to the application environment as shown in FIG. 1. In one of the application examples, the air conditioning system 102 communicates with the main control server 104 through the network. The main control server 104 determines the input parameters and output parameters of the components of the air-conditioning system, obtains the energy consumption mechanism model corresponding to the components of the air-conditioning system according to the input parameters and output parameters, obtains the historical operating data of the air-conditioning system, and compares them based on the historical operating data. The energy consumption mechanism model is trained to obtain the energy consumption model of the equipment in the air conditioning system. The main control server 104 may perform further processing on the air conditioning system 102 based on the accurately constructed energy consumption model.

In another application example, after the main control server 104 communicates with the air conditioning system 102 in the above manner to construct the energy consumption model of the air conditioning system, the main control server 104 also obtains the control quantity constraints and state quantity constraints of the components of the air conditioning system. According to the control quantity constraint and the state quantity constraint, determine the reasonable interval of the control quantity of the equipment in the air conditioning system, take the minimum energy consumption as the goal, use the genetic algorithm and the gradient optimization algorithm to optimize the energy consumption model, and solve the air conditioner within the reasonable interval The system optimal control quantity combination obtains the air conditioning system control optimization strategy. The main control server 104 outputs the air conditioning system control optimization strategy to the air conditioning system 102, and the air conditioning system 102 adjusts its operating parameters based on the control optimization strategy to achieve low energy operation. Among them, the terminal 102 may be, but is not limited to, a central air-conditioning system, an integrated air-conditioning system, an independent air-conditioning system, and the like. The main control server 104 may be implemented by an independent server or a server cluster composed of multiple servers.

In one embodiment, as shown in FIG. 2, a method for training an air conditioner energy consumption model is provided. Taking the method applied to the main control server 104 in FIG. 1 as an example for description, the method includes the following steps:

S100: Determine the input parameters and output parameters of the components of the air conditioning system.

The components of the air conditioning system include cooling towers, cooling pumps, chillers, refrigeration primary pumps, refrigeration secondary pumps, and air systems. Since the energy consumption of each device is related to its input parameters and output parameters, the input parameters and output parameters of these devices are first obtained here.

S200: Obtain the energy consumption mechanism model corresponding to the components of the air conditioning system according to the input parameters and output parameters.

In the field of energy consumption prediction, the current conventional method is to use a data-driven model to predict, that is, to use a machine learning model to fit historical operating data. The investment is low and the test error is generally low. It is acceptable, but this is not too great for the goal of energy optimization. Significantly, because the ultimate goal is not to predict the system energy consumption under the existing control strategy, but to predict the expected energy consumption after changing the control strategy. The main problems of the pure data-driven machine learning energy consumption model are reflected in: 1. Machine learning The model cannot give correct energy consumption predictions for parameter states that exceed the operating parameter space of the training sample; 2. In the machine learning model, the output changes due to changes in input may be different from business perception. To deal with the problem that the test sample feature space exceeds the training feature space, when the training sample cannot be effectively expanded, the most direct way is to use a mechanism/empirical model with global generalization capabilities. Energy consumption mechanism model, that is, combined with existing business knowledge or literature research results, a priori model formula is given. The formula contains a number of undetermined parameters, and historical data is used to fit the undetermined parameters to establish an energy consumption mechanism model; Compared with the machine learning model, the energy consumption mechanism model usually has better global generalization ability, that is, the performance of the energy consumption mechanism model is better for situations that have not occurred in the past.

The energy consumption prediction part realizes the function of predicting the energy consumption of the system through the environmental quantity and control variables, that is, obtaining the mapping relationship of system energy consumption=f (environmental quantity, control variable). Among them, the energy consumption mechanism model corresponding to the components of the air conditioning system includes the cooling tower energy consumption model, the cooling pump energy consumption model, the refrigeration pump energy consumption model and the main engine energy consumption model, as well as the main engine energy consumption model. Cooling water inlet temperature, and the cooling water inlet temperature is not directly controllable, so it is necessary to add a cooling water inlet temperature prediction model; the heat exchange of chilled water and cooling water are mainly affected by the load and the environment , There is little change in the process of adjusting the control quantity, so the heat exchange calculated from the system measurement value can be used as the input of the model. In simple terms, the above process of obtaining the energy consumption mechanism model can be understood as finding the mapping relationship (functional relationship) between the input parameters and output parameters of each component device, and using the mapping relationship to characterize the energy consumption mechanism model of each component device.

As shown in FIG. 3, in one of the embodiments, step S200 includes:

S220: Determine the mapping relationship between the input parameters and the output parameters of the components of the air conditioning system according to the input parameters and the output parameters.

S240: According to the mapping relationship, obtain the energy consumption mechanism model relation formula of the component equipment in the air conditioning system, and construct the energy consumption mechanism model of the component equipment in the air conditioning system.

Specifically, the cooling tower energy consumption mechanism model:

Cooling tower active power = f (cooling tower fan frequency, fan start and stop status)

Cooling pump energy consumption mechanism model:

Cooling pump active power = f (cooling pump frequency, pump start and stop state)

Energy consumption mechanism model of refrigeration primary pump:

The active power of the refrigerating primary pump = f (the frequency of the refrigerating primary pump, the start and stop state of the pump)

Main engine energy consumption mechanism model:

Active power of the host = f (start and stop status of the host, cooling water inlet temperature, chilled water supply temperature, chilled water heat exchange, cooling pump frequency, freezing primary pump frequency)

Cooling water inlet temperature prediction model:

Cooling water inlet temperature = f (outdoor temperature, outdoor humidity, cooling water heat exchange, cooling tower fan frequency, cooling pump frequency, start and stop status)

Energy consumption mechanism model of refrigeration secondary pump:

The active power of the secondary refrigeration pump = f (the frequency of the secondary refrigeration pump, the start and stop state of the pump)

Energy consumption mechanism model of wind system:

Wind system energy consumption mechanism model = f (chilled water supply temperature, secondary pump frequency, chilled water heat exchange)

Among them, the energy consumption mechanism model relationship used by the cooling tower, cooling pump, refrigerating primary pump, and refrigerating secondary pump:

P=(a*f ^b +c)*s

Among them, P is the power, f is the frequency, s is the switch state, and a, b, and c are undetermined parameters.

Host model:

p=p ₀ +k ₁ ·Q·(1+k ₂ (t _ci -12) ⁿ¹ )·(1-k ₃ ·(f _cwp -20))·(1-k ₄ ·(t _eo -8) ⁿ² )·(1-k ₅ ·(f _pchwp -20))

Where Q is the cooling capacity, t _ci is the cooling water inlet temperature, t _eo is the chilled water outlet temperature, f _cwp is the cooling pump frequency, f _pchwp is the freezing pump frequency, and the rest are undetermined parameters.

Cooling water inlet temperature model:

Among them, Q _c is the heat exchange amount of the cooling water, t _w is the outdoor wet bulb temperature, which is calculated from the outdoor temperature and humidity, f _ct is the cooling tower frequency, and a and b are undetermined parameters.

S300: Obtain historical operating data of the air-conditioning system, train the energy consumption mechanism model according to the historical operating data, and obtain the energy consumption model of the components of the air-conditioning system.

The historical operation data of the air-conditioning system includes the control quantity data, the state quantity data and the corresponding environmental quantity data in the historical operation. These historical data can be extracted from the operation log data of the air-conditioning system. It should be pointed out that the historical operating data contains historical operating data of each component of the air-conditioning system. The historical operating data of different component equipment is extracted and collected to obtain the historical operating data of the air-conditioning system. Using historical operating data as training data, the energy consumption mechanism model model is trained to obtain the energy consumption model of each component device. In simple terms, the training process can be understood as solving the undetermined parameters or constant values in the corresponding relationship formula of the above-mentioned energy consumption model through the training data. Optionally, the historical operation data can be divided into two parts: training set and test set. The energy consumption mechanism model is trained through the training set, and the model obtained after training is tested through the test set. After the test is passed, it is determined that the training is obtained. The model of is the energy consumption model of the composing equipment.

In the above air conditioning energy consumption model training method, the energy consumption mechanism model of each component is constructed according to the input parameters and output parameters of the air conditioning system, the energy consumption mechanism model is trained based on historical operating data, and the energy consumption model of each component is constructed separately. In the whole process, on the one hand, the mechanism model is used to construct the energy consumption model to give full play to the global generalization ability of the mechanism model; on the other hand, the energy consumption model is constructed separately for the components of the entire air conditioning system, which is closer to the real energy consumption changes of the air conditioning system. Under the circumstances, the constructed energy consumption model can accurately realize the energy consumption prediction of the air-conditioning system.

As shown in FIG. 3, in one of the embodiments, step S300 includes:

S320: Obtain historical operating data of the air conditioning system.

S340: Randomly divide the historical running data into a training set part and a test set part.

S360: Train the energy consumption mechanism model through the training set, update the undetermined parameter values in the energy consumption mechanism model relational formula, and obtain the energy consumption mechanism model after training.

S380: Test the energy consumption mechanism model after training through the test set. When the test is passed, the energy consumption model of the components in the air conditioning system is obtained.

The historical operating data of the air conditioner can be extracted from the operating log of the air-conditioning system. The historical operating data obtained is divided into two parts: a training set and a test set. The data in the training set is used as the training data, and the generated energy consumption The mechanism model relationship is used for training. This training process can be looped or carried until the energy consumption mechanism model relationship after training is summarized and the pending parameter values are accurately obtained, and the energy consumption mechanism model after training is obtained. Take cooling tower, cooling pump, refrigeration primary pump, and refrigeration secondary pump using energy consumption mechanism model relation: P=(a*f ^b +c)*s as an example, where a, b, c are undetermined parameters, passed The training part of the data is used to train the relational loop or iteratively, and finally get the specific values of a, b, c, and get the accurate relational expression of the energy consumption mechanism model, that is, the energy consumption model can be directly based on the energy consumption model in the follow-up. Calculation. After training, the energy consumption mechanism model is tested on the test set to detect whether the energy consumption mechanism model obtained is correct, and the effect of the model is tested. When the test passes, the energy consumption model of each component device is obtained. If it is not necessary, the model structure and model parameters are determined, and the relationship between model output and model input is verified, and whether it conforms to business cognition. If there are models or model parameters that violate business common sense, the model structure needs to be adjusted and retrained.

In one of the embodiments, obtaining historical operating data of the air conditioning system includes:

Perform data integration and filling processing on the original historical operating data of the components in the air conditioning system; based on the correlation of the operating data between the components in the air conditioning system, remove abnormal data from the data after data integration and filling processing; after removing the abnormal data from the processing The historical operating data of the air-conditioning system in a stable operating state is extracted from the data.

The original historical operating data directly obtains the unprocessed historical operating data, which can be directly imported from the air-conditioning system operating log data. The original data carries operating data from different components of the air conditioning system, and the operating data of different equipment is integrated and filled. Specifically, the original historical operating data includes cold source systems, dynamic loop systems, and power monitoring systems. The time stamps of different points in each system may be different. Operations such as rounding and aggregation are performed according to a certain granularity. The aggregated data is filled with vacant values. The original historical operating data may also contain abnormal data (error data). This type of abnormal data can be eliminated based on the correlation of the operating data between the components of the air conditioning system. Specifically, the filtering and elimination of abnormal data is mainly for cold The three data of source observation value, power monitoring data, and equipment switch status are verified. For a point, such as the frequency of the cooling tower fan, you can verify each other according to the frequency measurement value, power value, and start-stop status table. In the case of conflicts, the data can be considered valid, otherwise it will be eliminated. The data of the air-conditioning system in the stable operating state has reference significance. Therefore, in this embodiment, the historical operating data of the air-conditioning system in the stable operating state is extracted. Specifically, after the abnormal state filtering is performed, a reliable start-stop state of the equipment can be obtained. Considering that when the start-stop switching of the equipment occurs, the physical system needs a certain reaction time to reach a new steady state. The current state considered is the stable state. When the model is trained, it is necessary to filter the data identified as transient; set the time threshold to filter out the data for a certain period of time after the switch of the device start-stop state occurs.

In addition, as shown in FIG. 4, the present application also provides a control method for an air conditioning system, including:

S400: Obtain the control quantity constraint and the state quantity constraint of the components of the air conditioning system, and determine the reasonable interval of the control quantity of the component equipment in the air conditioning system according to the control quantity constraint and the state quantity constraint.

The control quantity constraint and the state quantity constraint can be regarded as the quantification/numerization of the boundary conditions in the subsequent optimization process of energy consumption model variables. Specifically, the boundary conditions of the air-conditioning system mainly include the terminal environment, the safe operation of the equipment and the reachability constraints of the internal state. Terminal environmental constraints: The final business function constraints of the air-conditioning system mainly include constraints on the temperature, humidity, fresh air volume, air pressure and other parameters of the air-conditioning terminal area. A major prerequisite for the energy-saving of the air-conditioning system is to meet the terminal's humidity, humidity and other indicators Requirements. Equipment safe operation constraints: Each equipment has a safety control range for its operating parameters. Energy saving must first be carried out under the premise of ensuring the safe operation of the equipment. Exceeding the range may cause equipment failure, such as pump frequency range, chiller water flow, water temperature range, and Some parameters change with time rate constraints, etc. Internal state reachability constraint: This refers to the constraint that keeps the components and models in the energy-saving optimization system in the correct operating state, that is, the premise and assumption in each modeling process can be met in actual operation, such as When the chilled water pump frequency is adjusted, it is usually assumed that the cooling capacity remains unchanged. The chilled return water temperature at the specified chilled outlet water temperature and chilled pump frequency is calculated by Q=cmΔt. When the chilled outlet water temperature is too high or the chilled pump frequency is too low, the actual operation of the system is frozen. The return water temperature is limited by the terminal heat transfer capacity and cannot reach the model calculation results at all, thus destroying the expected refrigeration capacity. The expected refrigerated return water temperature must also be limited to achieve the specified equipment heat absorption.

S500: With the goal of minimum energy consumption, the energy consumption model is optimized through genetic algorithm and gradient optimization algorithm, and the optimal control quantity combination of the air-conditioning system is solved within a reasonable interval, and the air-conditioning system control optimization strategy is obtained, based on the energy consumption model. The energy consumption model obtained by training the above air conditioning energy consumption model training method.

There are many optimization algorithms, such as evolutionary algorithms, swarm intelligence optimization algorithms, and simulated annealing algorithms. Simply put, the optimization algorithm is a relatively independent mathematical problem from the business logic. After the energy consumption model is established, the type of optimization algorithm to be used can be determined according to the specific parameters to be optimized. Here, genetic algorithm and gradient optimization algorithm are used to optimize the variables of the energy consumption model. With the minimum energy consumption as the goal, the optimal control quantity combination of air-conditioning is solved, and the control optimization strategy of the air-conditioning system is obtained.

In the above air conditioning system control method, the energy consumption mechanism model of each component is constructed according to the input parameters and output parameters of the air conditioning system, and the energy consumption mechanism model is trained based on historical operating data, and the energy consumption model of each component is constructed separately to determine the air conditioner The control quantity constraint and state quantity constraint of the system are used to solve the optimal variables in the energy consumption model through genetic algorithm and gradient optimization algorithm within a reasonable interval of the control quantity, and the control optimization strategy of the air conditioning system is obtained. In the whole process, within the effective boundary range of the environmental quantity constraint and the control quantity constraint, the energy consumption model of the equipment in the air-conditioning system is respectively solved for the optimal variables based on the genetic algorithm and the gradient optimization algorithm, and the energy in the current environment can be accurately obtained. The control strategy of the air-conditioning system with the least energy consumption realizes the good energy-saving effect of the air-conditioning system.

In one of the embodiments, the variable optimization of the energy consumption model through the genetic algorithm and the gradient optimization algorithm includes:

For floating-point variables in the energy consumption model, the genetic algorithm or gradient optimization algorithm is used to optimize the variables; for the integer variables in the energy consumption model, the number of integer variable combinations is obtained. When the number of integer variable combinations is less than the preset value, it can be traversed , Use the gradient optimization algorithm to optimize the variables. When the number of integer variable combinations is greater than the preset value that can be traversed, the genetic algorithm is used to optimize the variables.

Floating point variables can be simply understood as variable values that can have a decimal point, which can include frequency, temperature, pressure, and so on. An integer variable can be simply understood as a variable that can only be an integer value, which can include the switch state and the number of units that are turned on. The preset value is a threshold value set based on experience, which can be specifically set according to actual needs, and it can be specifically used to distinguish between "less" and "more". Specifically, when the variables to be optimized are floating-point variables, such as frequency, temperature, and pressure, select the gradient optimization algorithm or genetic algorithm; when the variables to be optimized include integer variables, such as the switch status of the device and the number of units to be turned on, the correct Support for mixed integer programming. When the number of combinations of integer variables is small and can be traversed, the mixed integer programming can be converted into multiple floating-point programming problems; when the number of combinations of integer variables is too large, the genetic algorithm can be used to directly program the mixed integers. Problem solving; usually there are dozens of large central air-conditioning equipment, and it is not feasible to traverse the switch combination directly, but after being restricted by linkage switches and deterministic business rules, the space of feasible equipment switch combinations will not be too large, traversing the switch combination It is still an effective way to use gradient-based optimization algorithms. Under non-differentiable or non-convex complex objective functions, the optimal or even close to optimal sub-optimal solution cannot be obtained, which significantly limits the choice of energy consumption model. On the premise that the optimization speed is acceptable, genetic algorithm optimization is preferred. For floating-point variables such as frequency, temperature, pressure, etc., the genetic algorithm can still be solved by integer programming. Because integer programming has a smaller search space, the measured optimization effect is often better than floating-point optimization.

In practical applications, the above two algorithms can be dealt with in the following ways: 1. Optimize algorithm coding to realize two types of optimization algorithms, namely genetic algorithm and the minimize method in the scipy package; 2. Establish an optimization algorithm selection decision tree The decision tree is shown in Figure 5; 3. Obtain the energy consumption model and the control parameters to be optimized; 4. Determine the control parameters and objective function of the current optimization problem, and select the corresponding optimization algorithm according to the corresponding situation.

As shown in FIG. 6, in one of the embodiments, step S400 includes:

S410: Obtain the environmental quantity constraint value and the state quantity constraint of the air-conditioning system and the current environmental value of the air-conditioning system.

The environmental constraint value is a preset parameter value, which can be read directly here. The current environmental value can be directly collected by a variety of sensors, such as the temperature and humidity of the current environment.

S420: If the current environmental value exceeds the environmental quantity constraint value, a factor is generated according to the extent to which the current environmental value exceeds the environmental quantity constraint value.

The factor can be generated as needed, and it only needs to meet the condition of "generating based on the degree of excess". Taking the environmental value as an example, assuming that a certain set temperature is 20 degrees and the current ambient temperature is 30 degrees, the generated factor It can be directly 3/2 (30/20); assuming that the current ambient temperature at another moment is 25 degrees, the corresponding factor can be directly 5/4 (25/4). It can be understood that this factor is only used to characterize the degree of excess. It is a relative value.

S430: According to the current environmental values and factors, set the first range of control variables of the components of the air conditioning system.

Taking the cooling capacity and the cooling water heat exchange control amount as an example, the cooling capacity and the cooling water heat exchange are calculated based on the current environment value, and the cooling capacity is updated by multiplying the cooling capacity and the cooling water heat exchange by the factor obtained in step S520. The heat exchange with the cooling water. The above processing indicates that the current cooling capacity cannot meet the needs of the terminal environment, and the cooling capacity and the cooling water heat exchange need to be adjusted. As far as the overall control of the device is concerned, based on the current environmental values and factors, reset the first range A of the control that constitutes the device.

S440: Obtain a reasonable operating range of the state quantities of the components in the air conditioning system according to the state quantity constraints.

Obtain the reasonable operating range of the pre-set air conditioning system components. Specifically, it is mainly necessary to obtain non-directly controllable state quantities, such as the return temperature of the chilled water and the reasonable operating range of the temperature of the cooling water in and out.

S450: According to the reasonable operating range of the state quantity, calculate the second range of the control quantity of the components in the air conditioning system.

According to the reasonable operating range of the state quantity, the second range B of the control quantity of each device is calculated. The calculation process can be obtained based on the conversion method of the conventional state quantity and the control quantity in the technical field of the air-conditioning system.

S460: Obtain the intersection of the first range of the control quantity and the second range of the control quantity, and obtain a reasonable interval for the control quantity of the equipment in the air conditioning system.

Calculate the intersection of the first range of the control quantity and the second range of the control quantity to obtain the intersection interval, that is, obtain the reasonable interval of the control quantity of the components in the air conditioning system.

Specifically, in practical applications, the entire reasonable interval for obtaining the control quantity based on the boundary conditions can be referred to the boundary condition processing block diagram shown in FIG. 7.

In one of the embodiments, with the goal of minimum energy consumption, the energy consumption model is optimized by genetic algorithm and gradient optimization algorithm, and the optimal control quantity combination of the air-conditioning system is solved within a reasonable interval, and the control optimization strategy of the air-conditioning system is obtained. ,Also includes:

Obtain the number of air-conditioning system control parameters and the optional value of the parameter offset; establish an orthogonal table according to the number and the optional value of the parameter offset; in the orthogonal table, there is no playback to extract the offset to the air-conditioning system control optimization The energy consumption model is updated iteratively on the control amount corresponding to the strategy.

In this embodiment, orthogonal table testing is introduced, and the sample space is expanded to further optimize the energy consumption model, so that the energy consumption model is closer to the real situation, and energy consumption calculation and prediction can be better. Specifically, the orthogonal table test refers to the addition of a certain offset based on the control value given by the algorithm, so that the air conditioning system can truly operate in different states. The main purpose of this is that in the early stage, the sample space distribution is relatively concentrated and single, which is not conducive to the establishment of a high-precision energy consumption model. By adding offsets, the sample space is expanded. The offset can be given randomly, but considering that the control parameters added with the offset must be run on the real device, and after the air-conditioning system switches the control, it takes some time to reach a steady state, and high-frequency control is not feasible, so The test with offset is very time-consuming, each test takes some time, and the time cost is very high. At this time, orthogonal table test is needed. The function of the orthogonal table test is to reduce the correlation of the test sample distribution as much as possible under the limited number of tests, and the correlation between the samples is reduced, that is, the distribution is more discrete, so that you can better explore the previous The situation that does not appear is conducive to the iterative update of the energy consumption model later.

The orthogonal table test specifically includes the following steps:

1. Determine the number n of parameters to be controlled and the optional values of parameter offset k _i , i=1, 2,...n; different control parameters, such as cooling tower frequency, chilled water outlet temperature, may be selected The number of offsets is different.

2. Establish an orthogonal table according to n and k _i , i=1, 2,...n.

3. Determine whether to introduce the orthogonal table test. If the model is already very accurate or you do not want to sacrifice the optimal control amount in exchange for more sample space, you can skip the orthogonal table test and skip step (4)(5); If the test is introduced, perform steps (4) (5).

4. After the optimization algorithm outputs the control combination, according to certain rules, there is no replacement extraction in the orthogonal table (if the orthogonal table has been traversed, then all are replaced, and a new round starts) one set of offsets , Added to the control value output by the algorithm.

5. Check whether the control quantity after the offset is added, whether it meets the constraints of the control quantity and the state quantity, if not, it will be corrected.

It should be understood that although the steps in the flowcharts of FIGS. 2, 3, 4 and 6 are displayed in sequence as indicated by the arrows, these steps are not necessarily executed in sequence in the order indicated by the arrows. Unless there is a clear description in this article, there is no strict order for the execution of these steps, and these steps can be executed in other orders. Moreover, at least part of the steps in Figures 2, 3, 4 and 6 may include multiple sub-steps or multiple stages. These sub-steps or stages are not necessarily executed at the same time, but can be performed at different times. For execution, the order of execution of these sub-steps or stages is not necessarily performed sequentially, but may be performed alternately or alternately with other steps or at least a part of the sub-steps or stages of other steps.

As shown in Figure 8, this application also provides an air conditioning energy consumption model training device, which includes:

The parameter determination module 100 is used to determine the input parameters and output parameters of the components of the air conditioning system;

The energy consumption mechanism model building module 200 is used to obtain the energy consumption mechanism model corresponding to the component equipment in the air conditioning system according to the input parameters and the output parameters;

The energy consumption model building module 300 is used to obtain historical operating data of the air-conditioning system, and to train the energy consumption mechanism model according to the historical operating data to obtain the energy consumption model of the components of the air-conditioning system.

The above air conditioning energy consumption model training device constructs the energy consumption mechanism model of each component device according to the input parameters and output parameters of the air conditioning system, trains the energy consumption mechanism model based on historical operating data, and constructs the energy consumption model of each component device. In the whole process, on the one hand, the mechanism model is used to construct the energy consumption model to give full play to the global generalization ability of the mechanism model; on the other hand, the energy consumption model is constructed separately for the components of the entire air conditioning system, which is closer to the real energy consumption changes of the air conditioning system. Under the circumstances, the constructed energy consumption model can accurately realize the energy consumption prediction of the air-conditioning system.

In one of the embodiments, the energy consumption mechanism model building module 200 is also used to determine the mapping relationship between the input parameters and output parameters of the components of the air conditioning system according to the input parameters and output parameters; according to the mapping relationship, the composition of the air conditioning system is obtained. The energy consumption mechanism model relation of the equipment is used to construct the energy consumption mechanism model of the composing equipment in the air conditioning system.

In one of the embodiments, the energy consumption model building module 300 is also used to obtain historical operating data of the air conditioning system; randomly divide the historical operating data into a training set part and a test set part; train the energy consumption mechanism model through the training set part , Update the undetermined parameter values in the relational expression of the energy consumption mechanism model to obtain the energy consumption mechanism model after training; test the energy consumption mechanism model after training through the test set part, and when the test passes, the energy consumption of the components in the air conditioning system is obtained. Consumption model.

In one of the embodiments, the energy consumption model building module 300 is also used to perform data integration and filling processing on the original historical operating data of the components in the air-conditioning system; based on the correlation of the operating data between the components in the air-conditioning system, the data is integrated Fill in the processed data for abnormal data elimination; extract the historical operating data of the air-conditioning system under stable operating conditions from the data after the abnormal data elimination processing.

As shown in Figure 9, the present application also provides an air conditioning system control device, which specifically includes:

The interval determination module 400 is used to obtain the control quantity constraint and the state quantity constraint of the component equipment in the air conditioning system, and determine the reasonable interval of the component equipment control quantity in the air conditioning system according to the control quantity constraint and the status quantity constraint;

The control optimization module 500 is used to optimize the variables of the energy consumption model through the genetic algorithm and the gradient optimization algorithm with the goal of minimum energy consumption, to solve the optimal control quantity combination of the air conditioning system within a reasonable interval, and to obtain the control optimization strategy of the air conditioning system. Wherein, the energy consumption model is an energy consumption model trained by the above-mentioned air conditioning energy consumption model training device.

The above-mentioned air conditioning system control device constructs the energy consumption mechanism model of each component device according to the input parameters and output parameters of the air conditioning system, trains the energy consumption mechanism model based on historical operating data, and constructs the energy consumption model of each component device to determine the air conditioner The control quantity constraint and state quantity constraint of the system are used to solve the optimal variables in the energy consumption model through genetic algorithm and gradient optimization algorithm within a reasonable interval of the control quantity, and the control optimization strategy of the air conditioning system is obtained. In the whole process, within the effective boundary range of the environmental quantity constraint and the control quantity constraint, the energy consumption model of the equipment in the air-conditioning system is respectively solved for the optimal variables based on the genetic algorithm and the gradient optimization algorithm, and the energy in the current environment can be accurately obtained. The control strategy of the air-conditioning system with the least energy consumption achieves good energy-saving effects of the air-conditioning system.

In one of the embodiments, the control optimization module 500 is also used to perform variable optimization through genetic algorithms or gradient optimization algorithms for floating-point variables in the energy consumption model; for integer variables in the energy consumption model, obtain the number of integer variable combinations. When the number of integer variable combinations is less than the preset value to be traversed, the variable optimization algorithm is used to optimize the variables; when the number of integer variable combinations is greater than the preset value to be traversed, the genetic algorithm is used to optimize the variables.

In one of the embodiments, the interval determination module 400 is also used to obtain the environmental quantity constraint value and state quantity constraint of the air conditioning system and the current environmental value of the air conditioning system; if the current environmental value exceeds the environmental quantity constraint value, according to the current environmental value The degree of excess of the environmental quantity constraint value generates a factor; according to the current environmental value and factor, the first range of the control quantity of the components in the air conditioning system is set; according to the state quantity constraint, the reasonable operating range of the component equipment in the air conditioning system is obtained; The reasonable operating range of the state quantity is calculated, and the second range of the control quantity of the components in the air conditioning system is calculated; the intersection of the first range of the control quantity and the second range of the control quantity is obtained, and the reasonable interval of the control quantity of the components in the air conditioning system is obtained.

In one of the embodiments, the above-mentioned air-conditioning system control device further includes an orthogonal table test module, which is used to obtain the number of air-conditioning system control parameters and the optional value of the parameter offset; according to the number and the optional value of the parameter offset , Establish an orthogonal table; extract the offset without playback in the orthogonal table to the control value corresponding to the control optimization strategy of the air conditioning system, and iteratively update the energy consumption model.

For the specific limitation of the air-conditioning system control device, please refer to the above-mentioned limitation on the air-conditioning system control method, which will not be repeated here. Each module in the above-mentioned air conditioning system control device can be implemented in whole or in part by software, hardware, and a combination thereof. The above-mentioned modules may be embedded in the form of hardware or independent of the processor in the computer equipment, or may be stored in the memory of the computer equipment in the form of software, so that the processor can call and execute the operations corresponding to the above-mentioned modules.

In one embodiment, a computer device is provided. The computer device may be a server, and its internal structure diagram may be as shown in FIG. 10. The computer equipment includes a processor, a memory, a network interface, and a database connected through a system bus. Among them, the processor of the computer device is used to provide calculation and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, a computer program, and a database. The internal memory provides an environment for the operation of the operating system and computer programs in the non-volatile storage medium. The database of the computer equipment is used to store data such as historical operation of the air-conditioning system. The network interface of the computer device is used to communicate with an external terminal through a network connection. The computer program is executed by the processor to realize an air conditioning system control method.

Those skilled in the art can understand that the structure shown in FIG. 10 is only a block diagram of a part of the structure related to the solution of the present application, and does not constitute a limitation on the computer device to which the solution of the present application is applied. The specific computer device may Including more or fewer parts than shown in the figure, or combining some parts, or having a different arrangement of parts.

In one embodiment, a computer device is provided, including a memory, a processor, and a computer program stored in the memory and capable of running on the processor, and the processor implements the following steps when the processor executes the computer program:

According to the input parameters and output parameters, the energy consumption mechanism model corresponding to the components of the air conditioning system is obtained;

Obtain historical operating data of the air-conditioning system, train the energy consumption mechanism model according to the historical operating data, and obtain the energy consumption model of the components of the air-conditioning system.

In an embodiment, the processor further implements the following steps when executing the computer program:

According to the input parameters and output parameters, determine the mapping relationship between the input parameters and output parameters of the components of the air conditioning system; according to the mapping relationship, obtain the energy consumption mechanism model relationship of the components of the air conditioning system, and construct the energy of the components of the air conditioning system. Model of consumption mechanism.

Obtain historical operating data of the air-conditioning system; randomly divide the historical operating data into a training set part and a test set part; train the energy consumption mechanism model through the training set part, update the undetermined parameter values in the energy consumption mechanism model relation, and get the post-training The energy consumption mechanism model of the air conditioning system; the energy consumption mechanism model after training is tested through the test set part. When the test is passed, the energy consumption model of the components in the air conditioning system is obtained.

Obtain the control quantity constraints and state quantity constraints of the components of the air-conditioning system, and determine the reasonable interval of the component equipment control quantities in the air-conditioning system according to the control quantity constraints and the state quantity constraints;

Taking the minimum energy consumption as the goal, the energy consumption model is optimized by genetic algorithm and gradient optimization algorithm, and the optimal control quantity combination of the air-conditioning system is solved within a reasonable interval, and the air-conditioning system control optimization strategy is obtained. The energy consumption model is based on The energy consumption model obtained by the training method of the above air conditioning energy consumption model training.

Obtain the environmental quantity constraint value and state quantity constraint of the air-conditioning system and the current environmental value of the air-conditioning system; if the current environmental value exceeds the environmental quantity constraint value, generate a factor according to the extent to which the current environmental value exceeds the environmental quantity constraint value; according to the current environment Values and factors, set the first range of the control variables of the components of the air conditioning system; obtain the reasonable operating range of the components of the equipment in the air conditioning system according to the state quantity constraints; calculate the control of the components of the equipment in the air conditioning system according to the reasonable operating range of the state quantities The second range of the control quantity; the intersection of the first range of the control quantity and the second range of the control quantity is obtained, and a reasonable interval of the control quantity of the components in the air conditioning system is obtained.

In one embodiment, a computer-readable storage medium is provided, on which a computer program is stored, and when the computer program is executed by a processor, the following steps are implemented:

In an embodiment, when the computer program is executed by the processor, the following steps are further implemented:

A person of ordinary skill in the art can understand that all or part of the processes in the above-mentioned embodiment methods can be implemented by instructing relevant hardware through a computer program. The computer program can be stored in a non-volatile computer readable storage medium. When the computer program is executed, it may include the processes of the above-mentioned method embodiments. Wherein, any reference to memory, storage, database or other media used in the embodiments provided in this application may include non-volatile and/or volatile memory. Non-volatile memory may include read only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM), or flash memory. Volatile memory may include random access memory (RAM) or external cache memory. As an illustration and not a limitation, RAM is available in many forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double data rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), synchronous chain Channel (Synchlink) DRAM (SLDRAM), memory bus (Rambus) direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM), etc.

The technical features of the above embodiments can be combined arbitrarily. In order to make the description concise, all possible combinations of the technical features in the above embodiments are not described. However, as long as there is no contradiction in the combination of these technical features, they should be It is considered as the range described in this specification.

The above embodiments only express several implementation manners of the present application, and the description is relatively specific and detailed, but it should not be understood as a limitation on the scope of the invention patent. It should be pointed out that for those of ordinary skill in the art, without departing from the concept of this application, several modifications and improvements can be made, and these all fall within the protection scope of this application. Therefore, the scope of protection of the patent of this application shall be subject to the appended claims.

Industrial applicability

The solutions provided in the embodiments of this application can be applied to the field of air-conditioning systems. In the embodiments of this application, firstly, the input parameters and output parameters of the components of the air-conditioning system are determined; according to the input parameters and output parameters, the corresponding components of the air-conditioning system are obtained. The energy consumption mechanism model of the air conditioning system; obtain the historical operating data of the air conditioning system, train the energy consumption mechanism model according to the historical operating data, and construct the energy consumption model of the components of the air conditioning system. Then, obtain the control quantity constraints and state quantity constraints of the components of the air conditioning system, and determine the reasonable range of the control quantity of the components in the air conditioning system according to the control quantity constraints and state quantity constraints; take the minimum energy consumption as the goal, through genetic algorithm and gradient The optimization algorithm optimizes the variables of the energy consumption model constructed above, solves the optimal control quantity combination of the air conditioning system within a reasonable interval, and obtains the control optimization strategy of the air conditioning system. Realize the good energy-saving effect of the air-conditioning system.

Claims

An air conditioning energy consumption model training method, the method includes:

Determine the input parameters and output parameters of the components of the air conditioning system;

According to the input parameters and the output parameters, obtain the energy consumption mechanism model corresponding to the component equipment in the air conditioning system;

Obtain historical operating data of the air-conditioning system, train the energy consumption mechanism model according to the historical operating data, and construct an energy consumption model of the components of the air-conditioning system.
The method according to claim 1, wherein said obtaining the energy consumption mechanism model corresponding to the component equipment in the air conditioning system according to the input parameter and the output parameter comprises:

Determine the mapping relationship between the input parameters and output parameters of the components of the air conditioning system according to the input parameters and the output parameters;

According to the mapping relationship, the energy consumption mechanism model relation formula of the component equipment in the air conditioning system is obtained, and the energy consumption mechanism model of the component equipment in the air conditioning system is constructed.
The method according to claim 2, wherein the acquiring historical operating data of the air-conditioning system, training the energy consumption mechanism model according to the historical operating data, and obtaining the energy consumption of the components of the air-conditioning system The model includes:

Acquiring historical operating data of the air-conditioning system;

Randomly divide the historical operation data into a training set part and a test set part;

Train the energy consumption mechanism model through the training set part, update the undetermined parameter values in the energy consumption mechanism model relational expression, and obtain the energy consumption mechanism model after training;

The energy consumption mechanism model after training is tested through the test set part, and when the test is passed, the energy consumption model of the component equipment in the air conditioning system is obtained.
The method according to claim 1, wherein said obtaining historical operating data of the air conditioning system comprises:

Perform data integration and filling processing on the original historical operating data of the constituent equipment in the air conditioning system;

Based on the relevance of the operating data between the components of the air-conditioning system, remove abnormal data from the data after data integration and filling processing;

The historical operating data of the air-conditioning system in a stable operating state is extracted from the processed data after the abnormal data is eliminated.
An air conditioning system control method, the method includes:

Obtain the control quantity constraint and the state quantity constraint of the component equipment in the air-conditioning system, and determine a reasonable interval of the component equipment control quantity in the air-conditioning system according to the control quantity constraint and the state quantity constraint;

With the minimum energy consumption as the goal, the energy consumption model is optimized by genetic algorithm and gradient optimization algorithm, and the optimal control quantity combination of the air conditioning system is solved within the reasonable interval to obtain the air conditioning system control optimization strategy. The energy consumption model is an energy consumption model trained by the method of any one of claims 1-4.
The method according to claim 5, wherein said performing variable optimization on said energy consumption model through genetic algorithm and gradient optimization algorithm comprises:

For floating-point variables in the energy consumption model, perform variable optimization through genetic algorithm or gradient optimization algorithm;

For the integer variables in the energy consumption model, the number of integer variable combinations is obtained. When the number of integer variable combinations is less than the preset value that can be traversed, the variable optimization is performed through the gradient optimization algorithm. When the integer variable combination When the number is greater than the preset value and can be traversed, the genetic algorithm is used to optimize the variables.
The method according to claim 5, wherein said obtaining the control quantity constraint and the state quantity constraint of the component equipment in the air-conditioning system, and determining the control quantity of the component equipment in the air-conditioning system according to the control quantity constraint and the state quantity constraint Reasonable ranges for quantities include:

Acquiring the environmental quantity constraint value and the state quantity constraint of the air-conditioning system and the current environmental value of the air-conditioning system;

If the current environmental value exceeds the environmental quantity constraint value, a factor is generated according to the extent to which the current environmental value exceeds the environmental quantity constraint value;

According to the current environmental value and the factor, setting a first range of the control amount of the component equipment in the air-conditioning system;

According to the state quantity constraint, obtain a reasonable operating range of the component equipment state quantity in the air-conditioning system;

According to the reasonable operating range of the state quantity, calculate the second range of the control quantity of the component equipment in the air-conditioning system;

The intersection of the first range of the control quantity and the second range of the control quantity is obtained, and a reasonable interval of the control quantity of the constituent equipment in the air conditioning system is obtained.
The method according to claim 5, wherein the minimum energy consumption is the goal, the energy consumption model is optimized by genetic algorithm and gradient optimization algorithm, and the air-conditioning system minimum value is solved within the reasonable interval. After obtaining the optimal control quantity combination, after obtaining the air conditioning system control optimization strategy, it also includes:

Obtain the number of control parameters of the air conditioning system and the optional value of the parameter offset;

Establishing an orthogonal table according to the number and the optional value of the parameter offset;

The energy consumption model is updated iteratively on the basis of the non-replay extraction offset in the orthogonal table to the control value corresponding to the control optimization strategy of the air conditioning system.
A computer device, including at least one processor, at least one memory, and a bus; wherein the processor and the memory communicate with each other through the bus; the processor is used to call a program in the memory Instructions to perform the method according to any one of claims 1 to 8.
A computer-readable storage medium having a computer program stored thereon, and when the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 8 are realized.