WO2024073937A1

WO2024073937A1 - Method for assessing air conditioner control algorithm, and apparatus, device and storage medium

Info

Publication number: WO2024073937A1
Application number: PCT/CN2022/137863
Authority: WO
Inventors: 陈峥; 韩星
Original assignee: 佳都科技集团股份有限公司; 广州新科佳都科技有限公司
Priority date: 2022-10-08
Filing date: 2022-12-09
Publication date: 2024-04-11
Also published as: CN115930376A

Abstract

A method for assessing an air conditioner control algorithm, and an apparatus, a device and a storage medium. The method comprises: upon the end of each time step, performing model parameter updating of a system model according to first data, wherein the first data comprises at least one of an actual air conditioner operation parameter, weather information, an indoor environment state, information regarding the number of people indoors, an air conditioner energy consumption value and an air conditioner cold energy output value of the current time step (S101); performing simulation processing on the first data and a simulation result of the previous time step by means of the system model and according to a plurality of preset candidate control algorithms, so as to obtain first simulation results corresponding to the respective candidate control algorithms (S102); upon the end of each cycle, calculating a performance index of the current cycle according to accumulated first data and/or first simulation results (S103); and obtaining assessment results of the respective candidate control algorithms in the current cycle according to an accumulated performance index of the current cycle (S104). By means of the method, the apparatus, the device and the storage medium, the problem of existing air conditioner control algorithm assessment results being inaccurate can be solved, thereby improving the accuracy of an air conditioner control algorithm assessment result.

Description

Air conditioning control algorithm evaluation method, device, equipment and storage medium

Technical Field

The embodiments of the present application relate to the technical field of air conditioning system control, and in particular, to an air conditioning control algorithm evaluation method, device, equipment and storage medium.

Background technique

With the development of social economy, the energy consumption of buildings has increased year by year, accounting for about 40% of the global energy demand. In my country, building energy consumption accounts for more than 30% of the total social energy consumption. At the same time, air conditioning and heating systems account for about half of the total building energy consumption, and the proportion has been increasing in recent years. The energy-saving compliance rate of public buildings is less than 10%, so how to reduce the energy consumption of air conditioning systems is the primary task of building energy conservation.

Generally, air conditioning systems use new and more energy-efficient control algorithms and/or models to control air conditioning operation in order to reduce the energy consumption of the air conditioning system. However, it is often impossible to accurately evaluate whether the new control algorithms and/or models can achieve the expected energy-saving effect. In the past, the energy-saving effects of new and old control algorithms and/or models were measured by comparing the energy-saving effects of the new and old control algorithms and/or models after they were actually run for a period of time, thereby evaluating the energy-saving effects of the new and old control algorithms and/or models.

Because the new/old control algorithms and/or models are run in different time periods, the operating conditions during each running time cannot be kept consistent, and thus the fairness of the comparison cannot be guaranteed. Moreover, the operating conditions covered by the comparison time period are limited, and the comparison is not comprehensive. Therefore, this method of evaluating the energy-saving effect based on the actual operating results of the new/old control algorithms and/or models does not produce accurate evaluation results.

Summary of the invention

The embodiments of the present application provide an air conditioning control algorithm evaluation method, device, equipment and storage medium, which can solve the problem of inaccurate evaluation results of existing air conditioning control algorithms and improve the accuracy of air conditioning control algorithm evaluation results.

In a first aspect, an embodiment of the present application provides an air conditioning control algorithm evaluation method, comprising:

At the end of each time step, updating the model parameters of the system model according to the first data, wherein the first data includes at least one of the actual air-conditioning operation parameters, weather information, indoor environment status, indoor number of people information, air-conditioning energy consumption value and air-conditioning cooling output value of the current time step;

According to a plurality of preset alternative control algorithms, the first data and the simulation result of the previous time step are simulated through a system model to obtain first simulation results corresponding to each alternative control algorithm, wherein the first simulation results include a first simulated indoor environment state, a first simulated air-conditioning system energy consumption, and a first simulated air-conditioning cooling output;

At the end of each cycle, a performance indicator of the current cycle is calculated based on the accumulated first data and/or the first simulation result, and a plurality of time steps constitute a cycle;

According to the performance indicators accumulated in the current cycle, the evaluation results of each candidate control algorithm in the current cycle are obtained.

Further, the plurality of candidate control algorithms include a plurality of reference control algorithms;

The method of simulating the first data and the simulation result of the previous time step through a system model according to the preset multiple alternative control algorithms to obtain the first simulation result corresponding to each alternative control algorithm includes:

Input the indoor environment state, weather information, indoor number of people information in the first data of the current time step, and the simulated air-conditioning operation parameters, simulated air-conditioning energy consumption value and simulated air-conditioning cooling output value obtained by simulation of the previous time step into a plurality of preset benchmark control algorithms, and output a plurality of simulated air-conditioning operation parameters of the current time step;

The simulated air-conditioning operating parameters of the multiple current time steps and the corresponding weather information, indoor number information in the first data, and the simulated indoor environmental state obtained by simulation in the previous time step are input into the system model for simulation processing, and the simulation results of the multiple current time steps are output. The simulation results include the simulated indoor environmental state, the simulated air-conditioning system energy consumption value and the simulated air-conditioning cooling output value.

Furthermore, the plurality of candidate control algorithms also include an actual operation algorithm;

The actual air-conditioning operation parameters, indoor environmental status, weather information, and indoor number of people information in the first data of the current time step are input into the system model for simulation processing, and the verification results are output. The verification results include the verification of indoor environmental status, the verification of air-conditioning system energy consumption value, and the verification of air-conditioning cooling output value.

Furthermore, at the end of each cycle, calculating the performance index of the current cycle according to the accumulated first data and/or the first simulation result includes:

At the end of each cycle, based on the simulation results accumulated in the current cycle, the first performance index corresponding to each benchmark control algorithm in the current cycle is calculated, wherein the first performance index is a function of the simulated indoor environment state, the simulated air conditioning system energy consumption value, and the simulated air conditioning cooling output value;

At the end of each cycle, based on the calibration results accumulated in the current cycle, a second performance indicator corresponding to the actual operation algorithm of the current cycle is calculated, where the second performance indicator is a function of the calibration indoor environment state, the calibration air conditioning system energy consumption value, and the calibration air conditioning cooling output value;

At the end of each cycle, the third performance indicator corresponding to the actual operation algorithm of the current cycle is calculated based on the first data accumulated in the current cycle. The third performance indicator is a function of the indoor environment state, air conditioning system energy consumption value and air conditioning cooling output value obtained in the actual operation.

Furthermore, the evaluation results of the candidate control algorithms of the current period are obtained according to the performance indicators accumulated in the current period, including:

The first performance indicator is compared with the second performance indicator and the third performance indicator to obtain evaluation results of each alternative control algorithm in the current cycle.

Furthermore, the system model includes an air conditioning system model and a space heat transfer model;

The input of the air conditioning system includes air conditioning operation parameters, weather information and indoor environmental status of the previous time step, wherein the weather information includes outdoor temperature, outdoor humidity and solar radiation intensity, and the indoor environmental status includes indoor temperature and indoor humidity;

The output of the air conditioning system includes the air conditioning system energy consumption value and the air conditioning cooling output value;

The input of the spatial heat transfer model includes weather information, indoor environmental status of the previous time step, indoor occupancy information and air conditioning cooling output;

The output of the spatial heat transfer model includes indoor environmental conditions.

Furthermore, the air conditioning system model is expressed as:

(P(t),Q(t))=μ _θ (u(t),x _r (t-1), _xa (t)), wherein μ _θ represents a neural network, θ represents a model parameter set of the neural network, P(t) represents the energy consumption value of the air-conditioning system at the t-th time step, Q(t) represents the air-conditioning cooling output value at the t-th time step, u(t) represents the air-conditioning operating parameters at the t-th time step, x _r (t-1) represents the indoor environment state at the t-1-th time step, and _xa (t) represents the weather information at the t-th time step.

Furthermore, the spatial heat transfer model is expressed as:

_xr (t)=H(t)·k(t), where _xr (t) represents a column vector of the indoor environmental states of each indoor area at the t-th time step, H(t) represents a matrix containing weather information, air conditioning cooling output value, and indoor occupancy information, k(t) represents the model parameters of the spatial heat transfer model, and H(t) is expressed as:

represents the transpose of x _r (t), and N _p (t) represents the number of people in each indoor area.

Further, at the end of each time step, updating the model parameters of the system model according to the first data includes:

At the end of each time step, updating the model parameters of the spatial heat transfer model according to the first data corresponding to the time step to determine the first parameter;

At the end of a preset cycle, a model parameter update process is performed on the air conditioning system model according to the first data of all time steps in the cycle to determine a second parameter, wherein a plurality of time steps constitute a cycle;

An updated system model is determined according to the first parameter and/or the second parameter.

In a second aspect, an embodiment of the present application provides an air conditioning system control device, comprising:

a parameter updating unit, configured to update the model parameters of the system model according to first data at the end of each time step, wherein the first data includes at least one of actual air-conditioning operation parameters, weather information, indoor environment status, indoor number of people information, air-conditioning energy consumption value and air-conditioning cooling output value of the current time step;

A simulation unit, configured to simulate the first data and the simulation result of the previous time step through a system model according to a plurality of preset alternative control algorithms, and obtain first simulation results corresponding to each alternative control algorithm, wherein the first simulation results include a first simulated indoor environment state, a first simulated air-conditioning system energy consumption, and a first simulated air-conditioning cooling output;

a performance index calculation unit, configured to calculate the performance index of the current period based on the accumulated first data and/or the first simulation result at the end of each period, wherein a plurality of time steps constitute one period;

The evaluation unit is used to obtain the evaluation results of each candidate control algorithm in the current cycle based on the performance indicators accumulated in the current cycle.

The simulation unit is further used to input the indoor environment state, weather information, indoor number of people information in the first data of the current time step and the simulated air-conditioning operation parameters, simulated air-conditioning energy consumption value and simulated air-conditioning cooling output value obtained by the simulation of the previous time step into a plurality of preset benchmark control algorithms, and output a plurality of simulated air-conditioning operation parameters of the current time step;

The simulation unit is also used to input the actual air-conditioning operation parameters, environmental status, weather information, and indoor number of people information in the first data of the current time step into the system model for simulation processing, and output verification results, which include verification of indoor environmental status, verification of air-conditioning system energy consumption value, and verification of air-conditioning cooling output value.

Furthermore, the performance index calculation unit is also used to calculate, at the end of each cycle, the first performance index corresponding to each benchmark control algorithm of the current cycle according to the simulation results accumulated in the current cycle, wherein the first performance index is a function of the simulated indoor environment state, the simulated air-conditioning system energy consumption value and the simulated air-conditioning cooling output value;

Furthermore, the evaluation unit is also used to compare the first performance indicator with the second performance indicator and the third performance indicator to obtain evaluation results of each alternative control algorithm in the current cycle.

The input of the space heat transfer model includes weather information, indoor environmental status of the previous time step, indoor number of people information and air conditioning cooling output value;

The output of the spatial heat transfer model includes the indoor environmental state at the next time step.

Furthermore, the air conditioning system model is expressed as:

Furthermore, the spatial heat transfer model is expressed as:

Furthermore, the parameter updating unit is further used to perform model parameter updating processing on the spatial heat transfer model according to the first data corresponding to the time step at the end of each time step to determine the first parameter;

In a third aspect, an embodiment of the present application provides an air conditioning system control device, including:

memory and one or more processors;

The memory is used to store one or more programs;

When the one or more programs are executed by the one or more processors, the one or more processors implement the air conditioning control algorithm evaluation method as described in the first aspect.

In a fourth aspect, an embodiment of the present application provides a storage medium storing computer executable instructions, which, when executed by a computer processor, are used to execute the air conditioning control algorithm evaluation method as described in the first aspect.

The embodiment of the present application updates the model parameters of the system model according to the first data at the end of each time step, simulates the first data and the simulation results of the previous time step through the system model according to the preset multiple alternative control algorithms, and obtains the first simulation results corresponding to each alternative control algorithm. At the end of each cycle, the performance index of the current cycle is calculated according to the accumulated first data and/or the first simulation result, and multiple time steps constitute a cycle; according to the performance index accumulated in the current cycle, the evaluation results of each alternative control algorithm in the current cycle are obtained. By adopting the above technical means, the performance index corresponding to each alternative control algorithm in the current cycle can be calculated by simulating the preset alternative control algorithm, and the evaluation result can be obtained by evaluating each alternative control algorithm according to the performance index, so that when evaluating, each alternative control algorithm is carried out under the same operating time and operating conditions, and the evaluation is more fair and comprehensive, thereby improving the accuracy of the evaluation result of the air conditioning control algorithm. In addition, the evaluation of the air conditioning control algorithm through simulation processing avoids the waste of resources caused by the actual operation evaluation of the alternative control algorithm to be evaluated, thereby improving the energy saving effect of the air conditioning system.

BRIEF DESCRIPTION OF THE DRAWINGS

FIG1 is a flow chart of an air conditioning control algorithm evaluation method provided by an embodiment of the present application;

FIG2 is a schematic diagram of a system model structure provided in an embodiment of the present application;

FIG3 is a schematic diagram of the relationship between a performance indicator and an alternative control algorithm provided in an embodiment of the present application;

FIG4 is a schematic structural diagram of an air conditioning system control device provided in an embodiment of the present application;

FIG. 5 is a schematic diagram of the structure of an air conditioning system control device provided in an embodiment of the present application.

Detailed ways

In order to make the purpose, technical scheme and advantages of the present application clearer, the specific embodiments of the present application are further described in detail below in conjunction with the accompanying drawings. It is understood that the specific embodiments described herein are only used to explain the present application, rather than to limit the present application. It should also be noted that, for the convenience of description, only the part related to the present application but not all the contents are shown in the accompanying drawings. Before discussing the exemplary embodiments in more detail, it should be mentioned that some exemplary embodiments are described as processes or methods depicted as flow charts. Although the flow chart describes each operation (or step) as a sequential process, many of the operations therein can be implemented in parallel, concurrently or simultaneously. In addition, the order of each operation can be rearranged. The process can be terminated when its operation is completed, but it can also have additional steps not included in the accompanying drawings. The process can correspond to a method, a function, a procedure, a subroutine, a subprogram, etc.

The air conditioning control algorithm evaluation method, device, equipment and storage medium provided in the present application are intended to evaluate the air conditioning control algorithm by simulating the preset alternative control algorithm, calculating the performance index corresponding to each alternative control algorithm in the current cycle, and evaluating each alternative control algorithm according to the performance index to obtain the evaluation result, so that when evaluating, each alternative control algorithm is performed under the same running time and operating conditions, and the evaluation is more fair and comprehensive, thereby improving the accuracy of the evaluation result of the air conditioning control algorithm. In addition, the evaluation of the air conditioning control algorithm by simulation processing avoids the waste of resources caused by the evaluation of the alternative control algorithm to be evaluated after the actual operation, thereby improving the energy saving effect of the air conditioning system. Compared with the traditional air conditioning control algorithm evaluation method, it usually runs each control algorithm to be evaluated for a period of time before performing performance evaluation. In this way, the running time of each control algorithm to be evaluated is different, and the working conditions during the running time are also different. Therefore, the evaluation index obtained lacks fairness and comprehensiveness, the evaluation result is not true and reliable, and the evaluation result has low accuracy. Based on this, the air conditioning control algorithm evaluation method of the embodiment of the present application is provided to solve the problem of low accuracy of the existing air conditioning control algorithm evaluation results.

FIG1 shows a flow chart of an air conditioning control algorithm evaluation method provided in an embodiment of the present application. The air conditioning control algorithm evaluation method provided in this embodiment can be executed by an air conditioning system control device, which can be implemented by software and/or hardware. The air conditioning system control device can be composed of two or more physical entities, or can be composed of one physical entity. Generally speaking, the air conditioning system control device can be a host computer of the air conditioning system, such as a computer device, etc.

The following description is made by taking a computer device as an example of a subject for executing the air conditioning control algorithm evaluation method. Referring to FIG1 , the air conditioning control algorithm evaluation method specifically includes:

S101. At the end of each time step, the model parameters of the system model are updated according to the first data, wherein the first data includes at least one of the actual air-conditioning operation parameters of the current time step, weather information, indoor environmental status, indoor number of people information, air-conditioning energy consumption value and air-conditioning cooling output value.

In the process of evaluating the air conditioning control algorithm, it is necessary to obtain the corresponding air conditioning operation data and environmental data, and evaluate the corresponding air conditioning control algorithm based on the reference data. When acquiring data, it can be acquired according to the time step and/or cycle so as to perform periodic statistics and processing on the data. Therefore, before the evaluation, the time step can be preset, and multiple time steps constitute a cycle. The specific time step setting can be set according to the actual situation. Generally, the time step is set to be consistent with the time step of the control algorithm.

The first data is obtained according to the time step, and the first data includes air conditioning operation parameters, weather information, indoor environmental status, indoor number of people information, air conditioning energy consumption value and air conditioning cooling output value. Among them, the specific content of the air conditioning operation parameters depends on the type and configuration of the air conditioning system. For example, taking the water-cooled central air conditioner as an example, the air conditioning operation parameters include the chiller outlet water temperature, the chilled water flow rate and the air volume of each fan. The air conditioning operation parameters, air conditioning cooling output value and air conditioning energy consumption value in the corresponding previous time step can be obtained through the communication of each device corresponding to the air conditioning system. Weather information includes outdoor temperature, outdoor humidity and solar radiation intensity, etc., and the weather information in the current time step can be obtained through outdoor sensors and/or Internet communication. The indoor environmental status includes indoor temperature and indoor humidity, etc., and the indoor environmental status such as indoor temperature and indoor humidity in the corresponding time step can be obtained through the indoor temperature and humidity sensor. The indoor number of people information includes the number of people in the room, and the indoor number of people information in the current time step can be obtained through the corresponding access control, gate and/or monitoring system.

By acquiring the first data according to the time step, the first data can be differentiated into stages, which helps to perform corresponding processing on the first data as reference data according to the time step, and can improve the orderliness and efficiency of data processing.

Considering that the air conditioning system is easily affected by weather, seasonal changes, and indoor population density, the system model parameters need to be updated before each evaluation so that the system model can be adaptively adjusted according to the actual environmental conditions and indoor population density. Therefore, at the end of each time step, the model parameters of the system model are updated according to the first data, and subsequent evaluations are performed only after the update is completed. By updating the model parameters at the end of each time step and performing simulation evaluation based on the updated system model, the simulation results are more accurate, thereby improving the accuracy of the final evaluation results.

FIG2 is a schematic diagram of a system model structure provided by an embodiment of the present application. Referring to FIG2 , the system model includes an air conditioning system model 10 and a space heat transfer model 20. The input of the air conditioning system model 10 includes air conditioning operation parameters, weather information, and the indoor environmental state of the previous time step, wherein the weather information includes outdoor temperature, outdoor humidity, and solar radiation intensity, and the indoor environmental state includes indoor temperature and indoor humidity. The output of the air conditioning system model 10 includes the energy consumption value of the air conditioning system and the air conditioning cooling output value. The input of the space heat transfer model 20 includes weather information, the indoor environmental state of the previous time step, indoor number of people information, and the air conditioning cooling output value. The output of the space heat transfer model 20 includes the indoor environmental state.

In one embodiment, the spatial system model is represented as

(P(t),Q(t))＝μ _θ (u(t),x _r (t-1), _xa (t)), where μ _θ represents a neural network, and its network structure includes the number of layers, the number of nodes per layer, and the type of activation function, etc. These network structures can be preset manually. θ represents the model parameter set of the neural network, P(t) represents the energy consumption value of the air-conditioning system at the t-th time step, Q(t) represents the air-conditioning cooling output value at the t-th time step, and the air-conditioning cooling output value may include the cooling capacity of the chiller, the sensible cooling output and the latent cooling output of each air-conditioning outlet according to the control requirements. u(t) represents the air-conditioning operating parameters at the t-th time step, and the specific content of the air-conditioning operating parameters depends on the type and configuration of the air-conditioning system. Taking the water-cooled central air-conditioning system as an example, u(t) may include the chiller outlet water temperature, the chilled water flow rate, and the air volume of each fan, etc. _{x r} (t-1) represents the indoor environmental state at the t-1th time step, and the indoor environmental state may include the indoor temperature and indoor humidity of each indoor area according to the control requirements. x _a (t) represents the weather information at the tth time step, which may include outdoor temperature, outdoor humidity, and solar radiation intensity.

It should be noted that the initial model parameters θ in the spatial system model can be obtained by training using simulation data and/or historical data. The simulation data can be obtained by establishing a simulation model of the air-conditioning system, running the simulation under random working conditions, and recording the input and output of the air-conditioning system at each time step. The generation of historical data requires recording the input and output of the same or similar air-conditioning system during actual operation. After determining the initial model parameters θ, the air-conditioning system model is trained on a part of the simulation data set and/or the historical data set, and θ is adjusted cyclically by a certain artificial neural network parameter optimization method (such as the stochastic gradient descent method) until the air-conditioning system model can appropriately fit the simulation data set and/or the historical data set, that is, given any air-conditioning system input at the tth time step, the output of the air-conditioning system model has a small deviation from the output of the air-conditioning system at the tth time step in the simulation data set and/or the historical data set. In addition, during the operation of the air-conditioning system, when each preset cycle interval is reached, the air-conditioning system model parameter θ is updated according to the first data in the corresponding cycle.

The update of the model parameters of the air conditioning system model requires retraining with multiple sets of data. Therefore, at the end of the preset period, the model parameters of the air conditioning system model are updated according to the first data set of all time steps in the corresponding period to determine the second parameter, which is the model parameter of the air conditioning system model after the update of the period. Multiple time steps constitute a period, and the number of time steps corresponding to a period can be set according to actual conditions.

In one embodiment, the spatial heat transfer model adopts a first-order linear kinetic model. The spatial heat transfer model is expressed as x _r (t) = H (t) · k (t), where x _r (t) represents a column vector formed by sequentially arranging the indoor environmental states of each indoor area at the t-th time step. Assuming that the number of indoor areas is m _d and the number of indoor environmental state contents (such as indoor temperature and indoor humidity) is m _r , the length of x _r (t) is m _d m _r . H (t) represents a matrix containing weather information, air conditioning cooling output value and indoor number of people information, that is, H (t) represents an m _d m _r row and m _d m _r (m _r +m _a +m _q +m _p +1) column matrix, where m _a , m _q , and m _p represent the number of contents contained in weather information, air conditioning cooling output value, and indoor number of people information, respectively. H (t) is expressed as:

represents the transpose of x _a (t), N _p (t) represents the number of people in each indoor area, which can be directly the number of people in each indoor area, or the number of people entering and leaving in the past period of time, etc., which indirectly reflects the number of people. Let (m _r +m _a +m _q +m _p +1) be m _h , then H(t) has the nth row and the ((n-1)m _h +1) to ((n-1)m _h +m _h ) columns,

for

The remaining elements are all 0. k(t) represents the model parameters of the spatial heat transfer model, which is a column vector with m _r m _h elements. There are two ways to set the initial values of the model parameters of the spatial heat transfer system model. One of them is to use the least squares method to determine the initial values when the corresponding historical operation data (such as subway stations or similar subway stations) can be obtained, so that the spatial heat transfer model fits the operation data to obtain the corresponding model parameters. The other way is to set the first value of k(t) when the historical operation data cannot be obtained.

The first element is 1, and the other elements are 0 as initial values to obtain the corresponding model parameters.

When updating the model parameters of the above-mentioned space heat transfer model, at the end of each time step, the model parameter updating process of the space heat transfer model is performed according to the first data set corresponding to the time step to determine the first parameter, which is the space heat transfer model parameter after the current step is updated.

It should be noted that at the end of each time step, the model parameters of the corresponding air conditioning system model in the system model remain unchanged, and the model parameters of the corresponding space heat transfer model are updated, so the updated model parameters are determined according to the first parameter to obtain the updated system model. At the end of each preset period, the model coefficients of the corresponding air conditioning system model and space heat transfer model in the system model are updated, and the updated system model is determined according to the first parameter and the second parameter.

At the end of each time step, the model parameters of the preset model are updated according to the first data of the current time step to obtain new model parameters. The model data is updated at each time step to improve the accuracy of the simulation results obtained by the system model after the model parameters are updated, thereby improving the accuracy of the evaluation results evaluated according to the simulation results.

In one embodiment, the system model updates the model parameters according to the first data in the corresponding period every time a preset period interval is reached. The preset period interval can be the last time step of each day or every certain number of time steps. For the model parameter update of the air-conditioning system model, the x _r (t), _xa (t), u (t), P (t), Q (t) obtained in step S101 are stored in a temporary data set at each time step t. The temporary data set is provided with a maximum storage capacity. If the maximum storage capacity has been reached at this time, the earliest row of data is discarded to accommodate the latest data. The model is updated every time a preset period interval is reached. Subsequently, the air-conditioning system model is trained on a part of the temporary data set, and a certain artificial neural network parameter optimization method (such as stochastic gradient descent method) is used to cyclically adjust the model parameters θ until the model can appropriately fit the temporary data set, that is, given the air-conditioning system input at any time step t, the deviation between the output of the model and the output of the air-conditioning system at the tth time step in the temporary data set is small. For the model parameter update of the space heat transfer model, at each time step, the model parameters of the space heat transfer model are updated using the Kalman filter algorithm according to the first data obtained in step S101 and the existing information in the temporary data set. When the space heat transfer model described above is used, the Kalman filter algorithm can use the following system state transfer equation and observation equation:

X(t)＝AX(t-1)+w(t-1), Z(t)＝H(t)X(t)+v(t), where: X(t)＝h(t), Z(t)＝ _xr (t), A is a unit matrix with the same number of rows as X(t); w(t-1) is random process noise, and its mean and variance are preset; v(t) is the random measurement error of indoor temperature and humidity, and its mean and variance are preset according to the performance of the temperature and humidity sensor.

S102. According to a plurality of preset alternative control algorithms, the first data and the simulation results of the previous time step are simulated through a system model to obtain first simulation results corresponding to each alternative control algorithm, wherein the first simulation results include a first simulated indoor environmental state, a first simulated air-conditioning system energy consumption and a first simulated air-conditioning cooling output.

During the evaluation, in order to achieve fairness and comprehensiveness of the evaluation process, it is necessary to make each alternative control algorithm to be evaluated run at the same operating time and operating conditions. Therefore, the corresponding evaluation data can be obtained by simulation. A plurality of alternative control algorithms to be evaluated are preset. According to the preset plurality of alternative control algorithms, the simulation results of the first data obtained above and the previous time step are input into the system model after the model coefficient is updated in the above step S101 for simulation processing, and the first simulation results corresponding to each alternative control algorithm are obtained. The first simulation results include a first simulated indoor environment state, a first simulated air-conditioning system energy consumption value, and a first simulated air-conditioning cooling output value.

The multiple candidate control algorithms may be multiple benchmark control algorithms, but generally, in order to compare with the existing actual operation algorithm, in order to screen out the control algorithm with better performance according to the evaluation results and put it into practical use, the actual operation algorithm may be added to the multiple candidate control algorithms to increase the practicality of the evaluation results. Therefore, the preset multiple candidate control algorithms include multiple benchmark control algorithms and actual operation algorithms.

In one embodiment, a method for simulating according to a benchmark control algorithm is provided. Since the benchmark control algorithm has not been actually run, there are no ready-made air conditioning operation parameters. Therefore, simulation processing can be performed through each benchmark control algorithm. After obtaining the corresponding simulated air conditioning operation parameters, the corresponding simulation results are obtained by simulating the air conditioning operation parameters obtained by simulation and other parameters in the first data through system model simulation processing. Specifically, the indoor environment state, weather information, indoor number of people information in the first data obtained at the current time step, and the simulated air conditioning operation parameters, simulated air conditioning energy consumption value and simulated air conditioning cooling output value obtained by simulation at the previous time step are input into a plurality of preset benchmark control algorithms, and a plurality of simulated air conditioning operation parameters of the current time step are output. It should be noted that each benchmark control algorithm outputs the corresponding simulated air conditioning operation parameters. The simulated air conditioning operation parameters of the multiple current time steps and the corresponding weather information, indoor number of people information in the first data and the simulated indoor environment state obtained by simulation at the previous time step are input into the system model after the model parameters are updated in the above step S101 for simulation processing, and a plurality of simulation results of the current time step are output, and the simulation results simulate the indoor environment state, simulated air conditioning system energy consumption value and simulated air conditioning cooling output value. It should be noted that each benchmark control algorithm corresponds to a simulation result, and each simulation result simulates the indoor environmental state, simulates the energy consumption value of the air-conditioning system, and simulates the cooling output value of the air-conditioning.

In one embodiment, if the current time step is t=0, part or all of the indoor environment state, weather information, indoor population information, actual air conditioning operation parameters, air conditioning energy consumption value of the previous time step, and air conditioning cooling output value of the previous time step obtained in the above step S101 are input into each benchmark control algorithm to obtain the simulated air conditioning operation parameters of each benchmark control algorithm at this time step. If the current time step is t≠0, part or all of the indoor environment state, weather information, indoor population information within the current time step obtained in step S101, and the simulated air conditioning operation parameters of the previous time step generated by the system model simulation of the corresponding benchmark control algorithm at the previous time step, the simulated indoor environment state of the previous time step, the simulated air conditioning system energy consumption of the previous time step, and the simulated air conditioning cooling output of the previous time step are input into each benchmark control algorithm to obtain the simulated air conditioning operation parameters of each benchmark control algorithm at this time step.

It should be noted that, combined with the system model described above, the input of the benchmark control algorithm is the indoor environment state _xr (t), weather information _xa (t), indoor occupancy information _Np (t), and part or all of the air-conditioning operation parameters u(t), air-conditioning energy consumption value P(t), and air-conditioning cooling output value Q(t) at the current time step, and the output is the simulated air-conditioning operation parameters of this time step.

In one embodiment, if the current time step is t=0, the indoor environment state, weather information, indoor number of people information of the current time step obtained in the above step S101, and the simulated air conditioning operation parameters obtained by simulating the above reference control algorithm are input into the system model to obtain the simulated indoor environment state, simulated air conditioning system energy consumption and simulated air conditioning cooling output corresponding to each reference control algorithm. If the current time step t≠0, the simulated indoor environment state obtained by simulating the previous time step through the system model, the weather information, indoor number of people information of the current time step obtained through the above step S101, and the simulated air conditioning operation parameters obtained by simulating the above reference control algorithm are input into the system model to obtain the simulated indoor environment state, simulated air conditioning system energy consumption and simulated air conditioning cooling output corresponding to each reference control algorithm.

In one embodiment, a method for simulating according to an actual operation algorithm is provided, and the actual air-conditioning operation parameters, indoor environmental status, weather information, and indoor number of people information of the current time step obtained in the above step S101 are input into the current system model for simulation processing according to the actual operation algorithm, and the verification result is output, and the verification result includes the verification of the indoor environmental status, the verification of the air-conditioning system energy consumption value, and the verification of the air-conditioning cooling output value. The corresponding verification result is obtained by performing simulation processing according to the actual operation algorithm, and the indoor environmental data, the air-conditioning system energy consumption value, and the air-conditioning cooling output value obtained by the verification result and the corresponding actual operation are compared to determine whether the simulation operation information is credible. Only when the simulation operation information is credible can the next step of evaluation be performed. If the simulation operation information is not credible, re-simulation processing is performed after the simulation operation information needs to be changed. As for the simulation operation information, it is judged that the credible information is based on the verification result before the next step of evaluation is performed.

In one embodiment, a method for judging whether simulation operation information is credible according to the verification result is provided, and the indoor environment state, air conditioning system energy consumption value and air conditioning cooling output value corresponding to the first data corresponding to the past several time steps are obtained, and the verification results corresponding to these time steps are obtained, and the verification results include the verification of the indoor environment state, the verification of the air conditioning system energy consumption value and the verification of the air conditioning cooling output value. Compare the first data corresponding to a certain time step with the verification result obtained by the simulation of the same time step. For example, the first data of the t-th time step and the verification result obtained by the simulation of the t-th time step. Compare the two sets of data to determine whether the similarity requirements are met. If they are met, the simulation operation information is determined to be credible, otherwise the simulation operation information is determined to be unreliable. If the simulation operation information is unreliable, it is necessary to adjust the simulation operation information and re-perform the steps described above for re-simulation processing. In the specific judgment, it can be that the first data of the corresponding time step is compared and judged after each time step; it can also be that after accumulating the first data of multiple time steps, the multiple first data are compared and judged accordingly. The specific number of time steps for performing a simulation information credible judgment can be set according to the actual situation.

It should be noted that an example of similarity requirement is: within the past several time steps, the absolute values of the differences between the indoor environment state, air conditioning system energy consumption value and air conditioning cooling output stored at any j-th time step and the verified indoor environment state, verified air conditioning system energy consumption value and verified air conditioning cooling output value stored at the corresponding j-th time step are all less than a certain threshold value. The specific threshold value can be set according to the actual situation and is not limited in this embodiment.

It should be noted that the alternative control algorithms include a benchmark control algorithm and an actual operation control algorithm. As described above, simulation results are obtained by simulating according to the benchmark control algorithm, and verification results are obtained by simulating according to the actual operation algorithm. Therefore, the first simulation result obtained by simulating according to multiple alternative control algorithms includes simulation results and verification results.

S103. At the end of each cycle, calculate the performance index of the current cycle according to the accumulated first data and/or the first simulation result, and multiple time steps constitute a cycle.

The air conditioning control algorithm is evaluated every other preset period to select a better air conditioning control algorithm. Therefore, the first data accumulated in the current period and/or the first simulation result can be used to calculate the performance index corresponding to each candidate control algorithm in the current period. As described above, the candidate control algorithm includes a benchmark control algorithm and an actual operation algorithm, and the first energy index corresponding to the benchmark control algorithm and the second performance index corresponding to the actual operation algorithm can be obtained respectively. For the benchmark control algorithm, at the end of each period, the first performance index corresponding to each benchmark control algorithm in the current period is calculated according to the simulation results accumulated in the current period, and the first performance index is a function of the simulated indoor environment state, the simulated air conditioning system energy consumption value and the simulated air conditioning cooling output value. The specific function form can be set according to the actual situation, for example, the function is set to EPI1=KP, EPI1 represents the first performance index, P is the simulated air conditioning energy consumption value accumulated in the current period, K is the function coefficient, which is a fixed value, so according to the set function, it can be known that the smaller the simulated air conditioning energy consumption value accumulated in the current period, the smaller the first performance index value, and the smaller the first performance index value, the better the energy saving effect. For the actual operation algorithm, because the actual operation algorithm has actual data such as the indoor environmental state, air conditioning system energy consumption value, and air conditioning cooling output value obtained by actual operation, the corresponding performance indicators can be calculated for the verification results obtained by simulation and the first data obtained by actual operation. For the verification results, at the end of each cycle, the second performance indicator corresponding to the actual operation algorithm of the current cycle is calculated based on the verification results accumulated in the current cycle. The second performance indicator is a function of the verification of the indoor environmental state, the verification of the air conditioning system energy consumption value, and the verification of the air conditioning cooling output value. The specific function form can be set according to the actual situation and is not limited in this embodiment. For the actual operation results, at the end of each cycle, the third performance indicator corresponding to the actual operation algorithm of the current cycle is calculated based on the first data accumulated in the current cycle. The third performance indicator is a function of the indoor environmental state, the air conditioning system energy consumption value, and the air conditioning cooling output value obtained by actual operation.

By calculating the current cycle cumulative simulation results, verification results and first data, the corresponding first performance index, second performance index and third performance index are obtained to achieve performance quantification, so as to form corresponding evaluation results according to the quantified values of the performance index. Through performance quantification, the credibility of the evaluation results is improved, and the accuracy of the evaluation is improved.

S104. Obtain evaluation results of each candidate control algorithm in the current period according to the accumulated performance indicators in the current period.

As described above, it can be seen that the first performance index of each benchmark algorithm in the current cycle, as well as the second performance index and the third performance index of the actual operation algorithm in the current cycle can be obtained. The first performance index is compared with the second performance index and the third performance index to obtain the evaluation results of each candidate control algorithm in the current cycle. By comparing the first performance index with the second performance index, or the first performance index with the third performance index, it can be obtained which control algorithm has better performance compared with the actual operation algorithm. Based on the fact that the first performance index and the second performance index are calculated by simulation results and verification results, the comparison of the two processes reflects the fairness of the comparison. Based on the fact that the third performance index is calculated by actual operation data, the first performance index is compared with the third performance index, which reflects the authenticity of the comparison. Therefore, by comparing the first performance index with the second performance index and the third performance index respectively, the comparison results between each benchmark control algorithm and the actual operation algorithm are obtained, thereby obtaining the evaluation result of whether the performance of the benchmark control algorithm is better than that of the actual operation algorithm, and improving the accuracy of the evaluation result of the air conditioning control algorithm.

In one embodiment, the third performance index corresponding to the actual operation result of the actual operation algorithm is calculated according to the actual indoor environment state, air conditioning system energy consumption value, air conditioning cooling output value, indoor number of people information, etc. obtained by step S101 in the past t time steps, such as the past 2 to t time steps. The first performance index of each benchmark control algorithm is calculated according to the simulated indoor environment state, simulated air conditioning system energy consumption value, simulated air conditioning cooling output value, indoor number of people information, etc. obtained by simulation in the past t+1 time step. The second performance index of the actual operation algorithm is calculated according to the indoor environment state for verification, energy consumption value of the air conditioning system for verification, cooling output value of the air conditioning for verification, indoor number of people information, etc. obtained by simulation in the past t+1 time step. One example of the performance index is the cumulative energy consumption value of the air conditioning system; another example is the cumulative time steps when the indoor temperature exceeds a certain set temperature. When selecting the target performance index, one way is to compare and evaluate the third performance index and the first performance index to obtain the performance comparison results of each benchmark control algorithm and the actual operation algorithm. This comparison method emphasizes the authenticity of the data. Another way is to compare and evaluate the first performance indicator and the second performance indicator to obtain the performance comparison results of each benchmark control algorithm and the actual operation algorithm. This comparison method emphasizes the fairness of the comparison.

After obtaining the evaluation results, the target performance indicator with the best corresponding performance indicator can be screened out according to the evaluation results, wherein the optimal performance indicator can be determined according to the corresponding function setting to determine the performance indicator with the smallest or largest performance indicator as the optimal performance indicator. For example, as described above, assuming that the function is set to EPI1=KP, the target performance indicator with the smallest performance indicator is the optimal performance indicator.

In one embodiment, Figure 3 is a schematic diagram of the relationship between a performance indicator and an alternative control algorithm provided in an embodiment of the present application. Referring to Figure 3, assuming that preset benchmark control algorithms A, B and C, and actual operation algorithm D constitute multiple alternative control algorithms, at the current time step, simulation processing will be performed through the system model according to the benchmark control algorithms A, B, C and the actual operation algorithm D to obtain a simulation result A1 corresponding to the benchmark control algorithm A, a simulation result B1 corresponding to the benchmark control algorithm B, a simulation result C1 corresponding to the benchmark control algorithm C, and a simulation result D1 corresponding to the actual operation algorithm D. Assuming that a cycle includes 5 time steps, the performance index a of the current cycle cumulative simulation result A corresponding to the benchmark control algorithm A is obtained, wherein the simulation result A includes the simulation results A1, A2, A3, A4 and A5 corresponding to the 5 time steps, the performance index b of the current cycle cumulative simulation result B corresponding to the benchmark control algorithm B, wherein the simulation result B includes the simulation results B1, B2, B3, B4 and B5 corresponding to the 5 time steps, the performance index c of the current cycle cumulative simulation result C corresponding to the benchmark control algorithm C, wherein the simulation result C includes the simulation results C1, C2, C3, C4 and C5 corresponding to the 5 time steps, and the performance index d of the current cycle cumulative simulation result D corresponding to the actual operation algorithm D, wherein the simulation result D includes the simulation results D1, D2, D3, D4 and D5 corresponding to the 5 time steps. The performance indicators a, b and c accumulated in the current cycle are all called the first performance indicators, and the performance indicator d accumulated in the current cycle is called the second performance indicator. Compare the performance indicators a, b, c and d accumulated in the current period, evaluate which performance indicator is the best, and select the target performance indicator with the best performance. Assume that the selected target performance indicator is performance indicator a.

The corresponding target control algorithm is determined according to the selected target performance index, and the target control algorithm is one of the candidate control algorithms. If the target control algorithm is one of the benchmark control algorithms, it proves that there is a benchmark control algorithm with better performance than the actual operation algorithm, and the benchmark control algorithm can be applied to the operation control of the actual air-conditioning system to improve the actual performance of the air-conditioning system and improve the saving effect. If the target control algorithm is the actual operation algorithm, it proves that the performance of the preset benchmark control algorithm is not better than the actual operation algorithm, and the actual operation algorithm can continue to be used to control the air-conditioning system.

The performance index is a reference index for judging the quality of the control algorithm. Therefore, the target control algorithm with the best performance among multiple alternative control algorithms can be screened out according to the evaluation results, wherein the target control algorithm can be a benchmark control algorithm or an actual control algorithm, depending on which control algorithm has the best performance index shown in the evaluation results. Since the performance indicators are all obtained by the corresponding benchmark control algorithm or the actual operation algorithm, the target control algorithm is one of the alternative control algorithms. According to the above, combined with Figure 3, according to the screened target performance index a, the corresponding target control algorithm can be obtained as the benchmark control algorithm A. By simulating and screening out the target performance index with the highest performance index in the current cycle, it means that the target control algorithm corresponding to the target performance index has the best performance compared with other alternative control algorithms. The air-conditioning control algorithm is screened by simulation, which avoids the waste of resources caused by the actual operation of the control algorithm and saves resources. In addition, the target control algorithm can be simulated and screened, and the corresponding cycle can be replaced with a higher-performance control algorithm in time to control the air-conditioning system, which improves the speed of control algorithm replacement, thereby improving the energy-saving effect of the air-conditioning system.

In order to continuously update the air conditioning control algorithm and continuously improve the energy-saving effect of the air conditioning system, the above-mentioned target control algorithm screening process can be performed once at a preset periodic interval. According to the above implementation process, the target control algorithm with the best energy-saving effect corresponding to the current cycle is screened out, and the target control algorithm is replaced with the actual operation algorithm, and the control system is operated and controlled by the target control algorithm in the next cycle. By setting up multiple alternative control algorithms and allowing multiple alternative control algorithms to be compared and evaluated at the same time, time efficiency is improved, thereby shortening the time for the air conditioning system to be replaced with a more energy-saving target control algorithm for operation control, thereby improving the overall time energy-saving effect of the air conditioning system.

It should be noted that if the target control algorithm obtained by screening is the actual operation algorithm, there is no need to perform a replacement action, and the actual operation algorithm is directly continued to be used to control the operation of the air-conditioning system.

This embodiment performs simulation evaluation of other benchmark control algorithms while running the existing actual operation algorithm. The simulation operation conditions are consistent with the actual operation conditions, and the simulated system model is updated and calibrated according to the actual operation data, thereby ensuring the fidelity of the simulation data and the fairness of the comparison. The comparative evaluation is performed during all operation times, covering a wider range of operating conditions and making the comparison more comprehensive. The new control algorithm can be directly adopted during all operation times, so there is no reduction or delay in the benefits of starting the new algorithm, thereby improving the energy-saving effect of the air-conditioning system.

In the above, at the end of each time step, the model parameters of the system model are updated according to the first data, and the simulation results of the first data and the previous time step are simulated through the system model according to the preset multiple alternative control algorithms to obtain the first simulation results corresponding to each alternative control algorithm. At the end of each cycle, the performance index of the current cycle is calculated according to the accumulated first data and/or the first simulation result, and multiple time steps constitute a cycle; according to the performance index accumulated in the current cycle, the evaluation results of each alternative control algorithm in the current cycle are obtained. By adopting the above technical means, the performance index corresponding to each alternative control algorithm in the current cycle can be calculated by simulating the preset alternative control algorithm, and the evaluation results can be obtained by evaluating each alternative control algorithm according to the performance index, so that when evaluating, each alternative control algorithm is performed under the same operating time and operating conditions, and the evaluation is more fair and comprehensive, thereby improving the accuracy of the evaluation results of the air conditioning control algorithm. In addition, the evaluation of the air conditioning control algorithm through simulation processing avoids the waste of resources caused by the evaluation of the alternative control algorithm to be evaluated after actual operation, thereby improving the energy saving effect of the air conditioning system.

Based on the above embodiments, FIG4 is a schematic diagram of the structure of an air conditioning system control device provided in the embodiment of the present application. Referring to FIG4 , the air conditioning system control device provided in the embodiment specifically includes: a parameter updating unit 21, a simulation unit 22, a performance index calculation unit 23 and an evaluation unit 24.

The parameter updating unit 21 is used to update the model parameters of the system model according to the first data at the end of each time step, wherein the first data includes at least one of the actual air-conditioning operation parameters, weather information, indoor environment status, indoor number of people information, air-conditioning energy consumption value and air-conditioning cooling output value of the current time step;

The simulation unit 22 is used to simulate the first data and the simulation result of the previous time step through the system model according to the preset multiple alternative control algorithms to obtain the first simulation result corresponding to each alternative control algorithm, wherein the first simulation result includes a first simulated indoor environment state, a first simulated air-conditioning system energy consumption and a first simulated air-conditioning cooling output;

A performance index calculation unit 23, used to calculate the performance index of the current period according to the accumulated first data and/or the first simulation result at the end of each period, where a plurality of time steps constitute a period;

The evaluation unit 24 is used to obtain the evaluation results of each candidate control algorithm in the current period according to the performance indicators accumulated in the current period.

The simulation unit 22 is further used to input the indoor environment state, weather information, indoor number information in the first data of the current time step, and the simulated air-conditioning operation parameters, simulated air-conditioning energy consumption values, and simulated air-conditioning cooling output values obtained by the simulation of the previous time step into a plurality of preset benchmark control algorithms, and output a plurality of simulated air-conditioning operation parameters of the current time step;

The simulation unit 22 is also used to input the actual air-conditioning operation parameters, indoor environmental status, weather information, and indoor number of people information in the first data of the current time step into the system model for simulation processing, and output verification results, which include verification of indoor environmental status, verification of air-conditioning system energy consumption value, and verification of air-conditioning cooling output value.

Furthermore, the performance index calculation unit 23 is also used to calculate, at the end of each cycle, the first performance index corresponding to each benchmark control algorithm of the current cycle according to the simulation results accumulated in the current cycle, wherein the first performance index is a function of the simulated indoor environment state, the simulated air conditioning system energy consumption value, and the simulated air conditioning cooling output value;

Furthermore, the evaluation unit 24 is further configured to compare the first performance indicator with the second performance indicator and the third performance indicator to obtain evaluation results of each candidate control algorithm in the current cycle.

Furthermore, the air conditioning system model is expressed as:

Furthermore, the spatial heat transfer model is expressed as:

Furthermore, the parameter updating unit 21 is further used to perform model parameter updating processing on the spatial heat transfer model according to the first data corresponding to the time step at the end of each time step to determine the first parameter;

In the above, at the end of each time step, the model parameters of the system model are updated according to the first data, and the simulation results of the first data and the previous time step are simulated through the system model according to the preset multiple alternative control algorithms to obtain the first simulation results corresponding to each alternative control algorithm. At the end of each cycle, the performance index of the current cycle is calculated according to the accumulated first data and/or the first simulation result, and multiple time steps constitute a cycle; according to the performance index accumulated in the current cycle, the evaluation results of each alternative control algorithm in the current cycle are obtained. By adopting the above technical means, the performance index corresponding to each alternative control algorithm in the current cycle can be calculated by simulating the preset alternative control algorithm, and the evaluation results can be obtained by evaluating each alternative control algorithm according to the performance index, so that when evaluating, each alternative control algorithm is performed under the same operating time and operating conditions, and the evaluation is more fair and comprehensive, thereby improving the accuracy of the evaluation results of the air conditioning control algorithm. In addition, the evaluation of the air conditioning control algorithm through simulation processing avoids the waste of resources caused by the actual operation evaluation of the alternative control algorithm to be evaluated, thereby improving the energy saving effect of the air conditioning system.

The air conditioning system control device provided in the embodiment of the present application can be used to execute the air conditioning control algorithm evaluation method provided in the above embodiment, and has corresponding functions and beneficial effects.

The embodiment of the present application provides an air conditioning system control device. Referring to FIG. 5 , the air conditioning system control device includes: a processor 31, a memory 32, a communication module 33, an input device 34, and an output device 35. The number of processors in the air conditioning system control device may be one or more, and the number of memories in the air conditioning system control device may be one or more. The processor, memory, communication module, input device, and output device of the air conditioning system control device may be connected via a bus or other means.

The memory 32, as a computer-readable storage medium, can be used to store software programs, computer executable programs and modules, such as the program instructions/modules corresponding to the air conditioning control algorithm evaluation method described in any embodiment of the present application (for example, the parameter update unit, simulation unit, performance index calculation unit and evaluation unit in the air conditioning system control device). The memory may mainly include a storage program area and a storage data area, wherein the storage program area may store an operating system, an application required for at least one function; the storage data area may store data created according to the use of the device, etc. In addition, the memory may include a high-speed random access memory, and may also include a non-volatile memory, such as at least one disk storage device, a flash memory device, or other non-volatile solid-state storage device. In some instances, the memory may further include a memory remotely arranged relative to the processor, and these remote memories may be connected to the device via a network. Examples of the above-mentioned network include, but are not limited to, the Internet, an intranet, a local area network, a mobile communication network and a combination thereof.

The communication module 33 is used for data transmission.

The processor 31 executes various functional applications and data processing of the device by running the software programs, instructions and modules stored in the memory, that is, realizes the above-mentioned air conditioning control algorithm evaluation method.

The input device 34 may be used to receive input digital or character information and generate key signal input related to user settings and function control of the device. The output device 35 may include a display device such as a display screen.

The air conditioning system control device provided above can be used to execute the air conditioning control algorithm evaluation method provided in the above embodiment, and has corresponding functions and beneficial effects.

The embodiment of the present application also provides a storage medium storing computer executable instructions, which are used to execute an air conditioning control algorithm evaluation method when executed by a computer processor. The air conditioning control algorithm evaluation method includes: at the end of each time step, updating the model parameters of the system model according to the first data, the first data including the actual air conditioning operation parameters of the current time step, weather information, indoor environmental status, indoor number of people information, air conditioning energy consumption value and air conditioning cooling output value at least one; according to a plurality of preset alternative control algorithms, the first data and the simulation results of the previous time step are simulated through the system model to obtain the first simulation results corresponding to each alternative control algorithm, the first simulation results including the first simulated indoor environmental status, the first simulated air conditioning system energy consumption and the first simulated air conditioning cooling output; at the end of each cycle, calculating the performance index of the current cycle according to the accumulated first data and/or the first simulation result, and a plurality of time steps constitute a cycle; according to the accumulated performance index of the current cycle, obtaining the evaluation results of each alternative control algorithm of the current cycle.

Storage medium - any of various types of memory devices or storage devices. The term "storage medium" is intended to include: installation media, such as CD-ROM, floppy disk or tape device; computer system memory or random access memory, such as DRAM, R RAM, SRAM, EDO RAM, Rambus RAM, etc.; non-volatile memory, such as flash memory, magnetic media (such as hard disk or optical storage); registers or other similar types of memory elements, etc. Storage media may also include other types of memory or combinations thereof. In addition, the storage medium may be located in the first computer system in which the program is executed, or may be located in a different second computer system, which is connected to the first computer system via a network (such as the Internet). The second computer system may provide program instructions to the first computer for execution. The term "storage medium" may include two or more storage media residing in different locations (for example, in different computer systems connected by a network). The storage medium may store program instructions (for example, embodied as a computer program) that can be executed by one or more processors.

Of course, the storage medium storing computer executable instructions provided in the embodiment of the present application, whose computer executable instructions are not limited to the air conditioning control algorithm evaluation method described above, can also execute related operations in the air conditioning control algorithm evaluation method provided in any embodiment of the present application.

The air-conditioning system control device, storage medium and air-conditioning system control equipment provided in the above embodiments can execute the air-conditioning control algorithm evaluation method provided in any embodiment of the present application. For technical details not described in detail in the above embodiments, please refer to the air-conditioning control algorithm evaluation method provided in any embodiment of the present application.

The above are only preferred embodiments of the present application and the technical principles used. The present application is not limited to the specific embodiments described herein, and various obvious changes, readjustments and substitutions that can be made by those skilled in the art will not deviate from the scope of protection of the present application. Therefore, although the present application is described in more detail through the above embodiments, the present application is not limited to the above embodiments, and may include more other equivalent embodiments without departing from the concept of the present application, and the scope of the present application is determined by the scope of the claims.

Claims

An air conditioning control algorithm evaluation method, characterized by comprising:

At the end of each time step, updating the model parameters of the system model according to the first data, wherein the first data includes at least one of the actual air-conditioning operation parameters, weather information, indoor environment status, indoor number of people information, air-conditioning energy consumption value and air-conditioning cooling output value of the current time step;

According to a plurality of preset alternative control algorithms, the first data and the simulation result of the previous time step are simulated through a system model to obtain first simulation results corresponding to each alternative control algorithm, wherein the first simulation results include a first simulated indoor environment state, a first simulated air-conditioning system energy consumption, and a first simulated air-conditioning cooling output;

At the end of each cycle, a performance indicator of the current cycle is calculated based on the accumulated first data and/or the first simulation result, and a plurality of time steps constitute a cycle;

According to the performance indicators accumulated in the current cycle, the evaluation results of each candidate control algorithm in the current cycle are obtained.
The method according to claim 1, characterized in that the plurality of candidate control algorithms include a plurality of baseline control algorithms;

The method of simulating the first data and the simulation result of the previous time step through a system model according to the preset multiple alternative control algorithms to obtain the first simulation result corresponding to each alternative control algorithm includes:

Input the indoor environment state, weather information, indoor number of people information in the first data of the current time step, and the simulated air-conditioning operation parameters, simulated air-conditioning energy consumption value and simulated air-conditioning cooling output value obtained by simulation of the previous time step into a plurality of preset benchmark control algorithms, and output a plurality of simulated air-conditioning operation parameters of the current time step;

The simulated air-conditioning operating parameters of the multiple current time steps and the corresponding weather information, indoor number information in the first data, and the simulated indoor environmental state obtained by simulation in the previous time step are input into the system model for simulation processing, and the simulation results of the multiple current time steps are output. The simulation results include the simulated indoor environmental state, the simulated air-conditioning system energy consumption value and the simulated air-conditioning cooling output value.
The method according to claim 2, characterized in that the plurality of alternative control algorithms further comprises an actual operation algorithm;

The method of simulating the first data and the simulation result of the previous time step through a system model according to the preset multiple alternative control algorithms to obtain the first simulation result corresponding to each alternative control algorithm includes:

The actual air-conditioning operation parameters, indoor environmental status, weather information, and indoor number of people information in the first data of the current time step are input into the system model for simulation processing, and the verification results are output. The verification results include the verification of indoor environmental status, the verification of air-conditioning system energy consumption value, and the verification of air-conditioning cooling output value.
The method according to claim 1, characterized in that, at the end of each cycle, calculating the performance index of the current cycle based on the accumulated first data and/or the first simulation result comprises:

At the end of each cycle, according to the simulation results accumulated in the current cycle, the first performance index corresponding to each benchmark control algorithm in the current cycle is calculated, wherein the first performance index is a function of the simulated indoor environment state, the simulated air conditioning system energy consumption value, and the simulated air conditioning cooling output value;

At the end of each cycle, based on the calibration results accumulated in the current cycle, a second performance indicator corresponding to the actual operation algorithm of the current cycle is calculated, where the second performance indicator is a function of the calibration indoor environment state, the calibration air conditioning system energy consumption value, and the calibration air conditioning cooling output value;

At the end of each cycle, the third performance indicator corresponding to the actual operation algorithm of the current cycle is calculated based on the first data accumulated in the current cycle. The third performance indicator is a function of the indoor environment state, air conditioning system energy consumption value and air conditioning cooling output value obtained in the actual operation.
The method according to claim 4 is characterized in that obtaining the evaluation results of each candidate control algorithm in the current period based on the performance indicators accumulated in the current period includes:

The first performance indicator is compared with the second performance indicator and the third performance indicator to obtain evaluation results of each candidate control algorithm in the current cycle.
The method according to claim 1, characterized in that the system model includes an air conditioning system model and a space heat transfer model;

The input of the air conditioning system includes air conditioning operation parameters, weather information and indoor environmental status of the previous time step, wherein the weather information includes outdoor temperature, outdoor humidity and solar radiation intensity, and the indoor environmental status includes indoor temperature and indoor humidity;

The output of the air conditioning system includes the air conditioning system energy consumption value and the air conditioning cooling output value;

The input of the space heat transfer model includes weather information, indoor environmental status of the previous time step, indoor number of people information and air conditioning cooling output value;

The output of the spatial heat transfer model includes indoor environmental conditions.
The method according to claim 6, characterized in that the air conditioning system model is expressed as:

(P(t),Q(t))=μ θ (u(t),x r (t-1), xa (t)), wherein μ θ represents a neural network, θ represents a model parameter set of the neural network, P(t) represents the energy consumption value of the air-conditioning system at the t-th time step, Q(t) represents the air-conditioning cooling output value at the t-th time step, u(t) represents the air-conditioning operating parameters at the t-th time step, x r (t-1) represents the indoor environment state at the t-1-th time step, and xa (t) represents the weather information at the t-th time step.
The method according to claim 6, characterized in that the spatial heat transfer model is expressed as:

xr (t)=H(t)·k(t), where xr (t) represents a column vector of the indoor environmental states of each indoor area at the t-th time step, H(t) represents a matrix containing weather information, air conditioning cooling output value, and indoor occupancy information, k(t) represents the model parameters of the spatial heat transfer model, and H(t) is expressed as:

represents the transpose of x r (t), and N p (t) represents the number of people in each indoor area.
The method according to claim 6, characterized in that the updating of the model parameters of the system model according to the first data at the end of each time step comprises:

At the end of each time step, updating the model parameters of the spatial heat transfer model according to the first data corresponding to the time step to determine the first parameter;

At the end of a preset cycle, a model parameter update process is performed on the air conditioning system model according to the first data of all time steps in the cycle to determine a second parameter, wherein a plurality of time steps constitute a cycle;

An updated system model is determined according to the first parameter and/or the second parameter.
An air conditioning system control device, characterized by comprising:

a parameter updating unit, configured to update the model parameters of the system model according to the first data at the end of each time step, wherein the first data includes at least one of the actual air-conditioning operation parameters, weather information, indoor environment status, indoor number of people information, air-conditioning energy consumption value and air-conditioning cooling output value of the current time step;

A simulation unit, configured to simulate the first data and the simulation result of the previous time step through a system model according to a plurality of preset alternative control algorithms, and obtain first simulation results corresponding to each alternative control algorithm, wherein the first simulation results include a first simulated indoor environment state, a first simulated air-conditioning system energy consumption, and a first simulated air-conditioning cooling output;

a performance index calculation unit, configured to calculate the performance index of the current period based on the accumulated first data and/or the first simulation result at the end of each period, wherein a plurality of time steps constitute one period;

The evaluation unit is used to obtain the evaluation results of each candidate control algorithm in the current period according to the performance indicators accumulated in the current period.
An air conditioning system control device, characterized by comprising:

memory and one or more processors;

The memory is used to store one or more programs;

When the one or more programs are executed by the one or more processors, the one or more processors implement the method according to any one of claims 1 to 9.
A storage medium storing computer executable instructions, characterized in that the computer executable instructions are used to execute any method according to claims 1-9 when executed by a processor.