CN115574437A - Air conditioner energy-saving control method, device, equipment and storage medium - Google Patents

Air conditioner energy-saving control method, device, equipment and storage medium Download PDF

Info

Publication number
CN115574437A
CN115574437A CN202211379920.XA CN202211379920A CN115574437A CN 115574437 A CN115574437 A CN 115574437A CN 202211379920 A CN202211379920 A CN 202211379920A CN 115574437 A CN115574437 A CN 115574437A
Authority
CN
China
Prior art keywords
air conditioner
operation parameter
conditioner operation
humidity
time step
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Pending
Application number
CN202211379920.XA
Other languages
Chinese (zh)
Inventor
陈峥
韩星
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
PCI Technology Group Co Ltd
PCI Technology and Service Co Ltd
Original Assignee
PCI Technology Group Co Ltd
PCI Technology and Service Co Ltd
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by PCI Technology Group Co Ltd, PCI Technology and Service Co Ltd filed Critical PCI Technology Group Co Ltd
Priority to PCT/CN2022/138833 priority Critical patent/WO2024073944A1/en
Publication of CN115574437A publication Critical patent/CN115574437A/en
Pending legal-status Critical Current

Links

Images

Classifications

    • F—MECHANICAL ENGINEERING; LIGHTING; HEATING; WEAPONS; BLASTING
    • F24—HEATING; RANGES; VENTILATING
    • F24F—AIR-CONDITIONING; AIR-HUMIDIFICATION; VENTILATION; USE OF AIR CURRENTS FOR SCREENING
    • F24F11/00—Control or safety arrangements
    • F24F11/62—Control or safety arrangements characterised by the type of control or by internal processing, e.g. using fuzzy logic, adaptive control or estimation of values
    • F24F11/63—Electronic processing
    • F24F11/64—Electronic processing using pre-stored data
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/08—Learning methods
    • F—MECHANICAL ENGINEERING; LIGHTING; HEATING; WEAPONS; BLASTING
    • F24—HEATING; RANGES; VENTILATING
    • F24F—AIR-CONDITIONING; AIR-HUMIDIFICATION; VENTILATION; USE OF AIR CURRENTS FOR SCREENING
    • F24F2110/00—Control inputs relating to air properties
    • F24F2110/10—Temperature
    • F—MECHANICAL ENGINEERING; LIGHTING; HEATING; WEAPONS; BLASTING
    • F24—HEATING; RANGES; VENTILATING
    • F24F—AIR-CONDITIONING; AIR-HUMIDIFICATION; VENTILATION; USE OF AIR CURRENTS FOR SCREENING
    • F24F2110/00—Control inputs relating to air properties
    • F24F2110/10—Temperature
    • F24F2110/12—Temperature of the outside air
    • F—MECHANICAL ENGINEERING; LIGHTING; HEATING; WEAPONS; BLASTING
    • F24—HEATING; RANGES; VENTILATING
    • F24F—AIR-CONDITIONING; AIR-HUMIDIFICATION; VENTILATION; USE OF AIR CURRENTS FOR SCREENING
    • F24F2110/00—Control inputs relating to air properties
    • F24F2110/20—Humidity
    • F—MECHANICAL ENGINEERING; LIGHTING; HEATING; WEAPONS; BLASTING
    • F24—HEATING; RANGES; VENTILATING
    • F24F—AIR-CONDITIONING; AIR-HUMIDIFICATION; VENTILATION; USE OF AIR CURRENTS FOR SCREENING
    • F24F2110/00—Control inputs relating to air properties
    • F24F2110/20—Humidity
    • F24F2110/22—Humidity of the outside air

Abstract

The embodiment of the application discloses an air conditioner energy-saving control method, device, equipment and storage medium, wherein when each time step is finished, a preset model is updated according to a second data set to obtain a first model; inputting a first data set and a plurality of preset alternative air conditioner operation parameter sequences into a first model, and outputting corresponding prediction results; optimizing according to the prediction result to obtain a first air conditioner operation parameter sequence, if the optimization termination condition is not met, adjusting the alternative air conditioner operation parameter sequence, inputting the adjusted alternative air conditioner operation parameter sequence into the first model again to output a corresponding new prediction result, and optimizing again according to the new prediction result until the first air conditioner operation parameter sequence is obtained; the first air conditioner operation parameter in the first air conditioner operation parameter sequence is used as the second air conditioner operation parameter and is output, the operation of the air conditioning system at the next time step length is controlled according to the second air conditioner operation parameter, the adaptivity of air conditioner control can be improved, and the energy-saving effect is improved.

Description

Air conditioner energy-saving control method, device, equipment and storage medium
Technical Field
The embodiment of the application relates to the technical field of air conditioner control, in particular to an air conditioner energy-saving control method, device, equipment and storage medium.
Background
With the development of social economy, the energy consumption of buildings is increased year by year, and the energy consumption accounts for about four times of the energy demand of the world. In China, the energy consumption of buildings accounts for more than three times of the energy consumption of the whole society, meanwhile, the air-conditioning and heating systems account for about half of the total energy consumption of the buildings, the proportion of the air-conditioning and heating systems is increased continuously in recent years, and the energy-saving standard-reaching rate of public buildings is not the first time, so that the energy conservation of the air-conditioning system is the primary task of building energy conservation.
The traditional air conditioning system has two control modes, namely manual control and automatic parameter setting control. The two control modes are preset based on human experience, and the air conditioning system has great subjective randomness and limitation in energy-saving control and environment body feeling comfort level perception in the actual operation process. The two control modes lack the overall consideration of the air conditioning system, so that the energy-saving effect of the air conditioning control process is poor.
Disclosure of Invention
The embodiment of the application provides an air conditioner energy-saving control method, device, equipment and storage medium, which can solve the problem of poor air conditioner energy-saving effect, improve the adaptivity of air conditioner control and improve the energy-saving effect of an air conditioner.
In a first aspect, an embodiment of the present application provides an energy saving control method for an air conditioner, including:
when each time step is finished, updating model parameters of a preset model according to a second data set to obtain a first model, wherein the second data set comprises indoor temperature and humidity, outdoor temperature and humidity, people number information, actual air conditioner operation parameters, an air conditioner system energy consumption value and an air conditioner cold output quantity of the current time step;
inputting a first data set and a plurality of preset alternative air conditioner operation parameter sequences into the first model, and outputting corresponding prediction results, wherein the prediction results comprise indoor temperature and humidity prediction values and air conditioner system energy consumption prediction values, the first data set comprises indoor temperature and humidity, outdoor temperature and humidity and people number information of the current time step, and outdoor temperature and humidity prediction values and people number information prediction values of a plurality of time steps in the future;
performing optimization processing according to the prediction result to obtain a first air conditioner operation parameter sequence, if the optimization processing does not meet the optimization termination condition, adjusting the alternative air conditioner operation parameter sequence, inputting the adjusted alternative air conditioner operation parameter sequence into the first model again to output a corresponding new prediction result, and performing optimization processing again according to the new prediction result until the first air conditioner operation parameter sequence is obtained, wherein the first air conditioner operation parameter sequence is one of a plurality of alternative air conditioner operation parameter sequences corresponding to the first air conditioner operation parameter sequence when the optimization termination condition is met;
and taking the first air conditioner operation parameter in the first air conditioner operation parameter sequence as a second air conditioner operation parameter and outputting the second air conditioner operation parameter, and controlling the operation of the air conditioning system at the next time step according to the second air conditioner operation parameter.
Further, the inputting the first data set and the preset candidate air conditioner operation parameter sequences into the first model and outputting the corresponding prediction results includes:
inputting the indoor temperature and humidity, the outdoor temperature and humidity, the number of people information, the outdoor temperature and humidity predicted values of a plurality of time steps in the future, the number of people information predicted values of a plurality of time steps in the future and a plurality of preset alternative air conditioner operation parameter sequences of the current time step into a first model for prediction processing;
and according to the prediction processing, obtaining an air conditioning system energy consumption prediction value, an air conditioning cold output prediction value and an indoor temperature and humidity prediction value of a plurality of time step lengths in the future corresponding to each preset alternative air conditioning operation parameter sequence.
Further, the first model comprises an air conditioning system model and a space heat transfer model;
the step of inputting a first data set and a plurality of preset alternative air conditioner operation parameter sequences into the first model and outputting corresponding prediction results comprises the following steps:
inputting the indoor temperature and humidity, the outdoor temperature and humidity and a plurality of preset alternative air conditioner operation parameter sequences of the current time step into an air conditioner system model, and performing first prediction processing to obtain a corresponding first air conditioner system energy consumption prediction value and a first air conditioner cold output prediction value;
inputting the first air conditioner cold output predicted value and the corresponding outdoor temperature and humidity, indoor temperature and humidity and the number information of people into a space heat transfer model, and performing second prediction processing to obtain a corresponding first indoor temperature and humidity predicted value;
inputting the first indoor temperature and humidity predicted value, the outdoor temperature and humidity predicted value and the number information predicted value of the next time step and the plurality of preset candidate air conditioner operation parameter sequences into a first model, and performing first prediction processing and second prediction processing to obtain a corresponding second air conditioner system energy consumption predicted value, a second air conditioner cold output predicted value and a second indoor temperature and humidity predicted value;
and repeating the first prediction processing and the second prediction processing until the energy consumption prediction value of the air conditioning system, the cold output prediction value of the air conditioner and the indoor temperature and humidity prediction value at a plurality of time steps in the future are obtained.
Further, if the optimization termination condition is not satisfied in the optimization process, adjusting the candidate air conditioner operation parameter sequence, inputting the adjusted candidate air conditioner operation parameter sequence into the first model again to output a corresponding new prediction result, and performing the optimization process again according to the new prediction result until the first air conditioner operation parameter sequence is obtained, including:
if the optimization termination condition is not met, calculating a loss value corresponding to each alternative air conditioner operation parameter sequence according to a preset loss function and the prediction result;
calculating the gradient of each alternative air conditioner operation parameter sequence according to the loss function, and adjusting the alternative air conditioner operation parameter sequences according to the gradient to obtain a plurality of new alternative air conditioner operation parameter sequences;
inputting the new alternative air conditioner operation parameter sequence and the first data set into a first model, and outputting a new prediction result;
and carrying out optimization processing according to the new prediction result until the first air conditioner operation parameter sequence is obtained.
Further, the optimizing according to the prediction result to obtain a first air conditioner operation parameter sequence includes:
when the optimization termination condition is met, calculating a loss value corresponding to each alternative air conditioner operation parameter sequence through a preset loss function according to indoor temperature and humidity predicted values corresponding to a plurality of time steps in the future and an energy consumption predicted value of an air conditioning system by combining a plurality of corresponding alternative air conditioner operation parameter sequences;
comparing the numerical value of the loss value corresponding to each alternative air conditioner operation sequence, and screening out the minimum loss value with the minimum numerical value;
and determining an alternative air conditioner operation parameter sequence corresponding to the minimum loss value as the first air conditioner operation parameter sequence.
Further, the air conditioning system model is expressed as:
(P(t+1),Q(t+1),Q lat (t+1))＝μ θ (u(t+1),T r (t),W r (t),T a (t),W a (t))，
wherein, mu θ Represents an artificial neural network, theta representsThe model parameter set of the artificial neural network, P (t + 1) represents the predicted value of the energy consumption of the air conditioning system at the t +1 th time step, the predicted value of the air conditioning cold output comprises the predicted value of the air conditioning sensible heat and cold output and the predicted value of the air conditioning latent heat and cold output, Q (t + 1) represents the predicted value of the air conditioning sensible heat and cold output at the t +1 th time step, and Q lat (T + 1) represents the predicted value of the output of the latent heat and the cold of the air conditioner at the T +1 time step, u (T + 1) is the preset air conditioner operation parameter at the T +1 time step, and T r (t) represents the room temperature of the t-th time step, W r (T) represents the indoor humidity at the tth time step, T a (t) outdoor temperature, W, for the t-th time step a (t) represents the outdoor humidity for the t time step;
inputting the indoor temperature and humidity, the outdoor temperature and humidity and a plurality of preset alternative air conditioner operation parameter sequences of the current time step into an air conditioner system model, and performing first prediction processing to obtain a corresponding first air conditioner system energy consumption prediction value and a first air conditioner cooling output prediction value, wherein the method comprises the following steps:
the outdoor temperature T in the T time step a (t) outdoor humidity W a (T) indoor temperature T r (t) indoor humidity W r (t) and alternative air conditioner operating parameter sequence U n Inputting the energy consumption predicted value P (t + 1), the sensible heat and cold output predicted value Q (t + 1) and the latent heat and cold output predicted value Q of the air conditioner into the air conditioner system model to obtain the t +1 time step length of the energy consumption predicted value P (t + 1), the sensible heat and cold output predicted value Q of the air conditioner lat (t+1)。
Further, the candidate air conditioner operation parameter sequence includes candidate air conditioner operation parameters of a plurality of time steps in the future, and the candidate air conditioner operation parameter sequence U n Expressed as:
Figure BDA0003927919060000031
wherein u is n (t) the alternative air conditioner operation parameter for the t time step in the nth alternative air conditioner operation parameter sequence, u n (t + 1) represents t +1 time step in nth candidate air conditioner operation parameter sequenceAlternative air conditioner operating parameter, u n (t+K max ) Represents t + K in the nth candidate air conditioner operation parameter sequence max Alternative air-conditioning operating parameters of individual time steps, K max Is a positive integer.
Further, the spatial heat transfer model is expressed as
Figure BDA0003927919060000041
Figure BDA0003927919060000042
Wherein, T r (t + 1) represents the predicted indoor temperature value of the t +1 th time step, W r (t + 1) represents the predicted value of the indoor humidity at the t +1 th time step, N in (t) information on the number of persons entering in the t-th time step, N out (t) the leaving people number information of the t-th time step, when t is the current time step, N in (t)、N out (t) taking the actual value, otherwise taking the predicted value, k 0 ～k 4 、b、j 0 ～j 4 Model parameters representing the spatial heat transfer model;
inputting the first air conditioner cold output prediction value and the corresponding outdoor temperature and humidity, the indoor temperature and humidity and the number information into a space heat transfer model, and performing second prediction processing to obtain a corresponding first indoor temperature and humidity prediction value, wherein the method comprises the following steps:
outputting a predicted value Q (t + 1) of the sensible heat and cold output of the air conditioner and a latent heat and cold output of the air conditioner at the t-th time step lat (T + 1), outdoor temperature T a (t) outdoor humidity W a (T) indoor temperature T r (t) indoor humidity W r (t) information of the number of persons who entered N in (t) and leaving population information N out (T) inputting the indoor temperature into the space heat transfer model to obtain an indoor temperature predicted value T of the T +1 time step r (t + 1) and predicted indoor humidity value W r (t+1)。
Further, the preset loss function is represented by loss = obj + pen, where obj represents an objective function and pen represents a penalty function, where the objective function is represented by:
Figure BDA0003927919060000043
wherein, K max Representing the predicted number of steps, T r (T + k) represents the predicted indoor temperature value of the T + k time step, T tg Representing the preset target indoor temperature, P (t + k) representing the predicted value of the energy consumption of the air conditioning system at the t + k time step, w 1 Representing a preset comfort weight coefficient, w 2 Representing a preset energy consumption weight coefficient of the air conditioning system;
the penalty function is expressed as:
Figure BDA0003927919060000044
wherein n is p Number of representing constraints, c n，k Representing a constraint function, relu a linear rectification function, p n,k Representing a constraint function c n,k The penalty factor of (2).
Further, when each time step is finished, updating the model parameters of the preset model according to the second data set to obtain the first model, including:
when each time step is finished, carrying out model parameter updating processing on the space heat transfer model according to a second data set corresponding to the time step to determine a first parameter;
when a preset period is finished, carrying out model parameter updating processing on the air conditioning system model according to a second data set of all time step lengths in the period to determine a second parameter, wherein a plurality of time step lengths form a period;
and determining the model parameters of the first model according to the first parameters and/or the second parameters to obtain the first model.
In a second aspect, an embodiment of the present application provides an energy saving control device for an air conditioner, including:
the model parameter updating unit is used for updating model parameters of a preset model according to a second data set when each time step is finished to obtain a first model, wherein the second data set comprises indoor temperature and humidity, outdoor temperature and humidity, people number information, actual air conditioner operation parameters, an air conditioner system energy consumption value and an air conditioner cooling capacity output quantity of the current time step;
the prediction unit is used for inputting a first data set and a plurality of preset alternative air conditioner operation parameter sequences into the first model and outputting corresponding prediction results, wherein the prediction results comprise indoor temperature and humidity prediction values and air conditioner system energy consumption prediction values, the first data set comprises indoor temperature and humidity, outdoor temperature and humidity and people number information of the current time step, and outdoor temperature and humidity prediction values and people number information prediction values of a plurality of time steps in the future;
the optimization unit is used for carrying out optimization processing according to the prediction result to obtain a first air conditioner operation parameter sequence, adjusting the alternative air conditioner operation parameter sequence and inputting the adjusted alternative air conditioner operation parameter sequence into the first model again to output a corresponding new prediction result if the optimization processing does not meet the optimization termination condition, and carrying out optimization processing again according to the new prediction result until the first air conditioner operation parameter sequence is obtained, wherein the first air conditioner operation parameter sequence is one alternative air conditioner operation parameter sequence in a plurality of alternative air conditioner operation parameter sequences corresponding to the first air conditioner operation parameter sequence when the optimization termination condition is met;
and the control unit is used for taking the first air conditioner operation parameter in the first air conditioner operation parameter sequence as a second air conditioner operation parameter and outputting the second air conditioner operation parameter, and controlling the operation of the air conditioning system at the next time step length according to the second air conditioner operation parameter.
Further, the prediction unit is further configured to input the indoor temperature and humidity of the current time step, the outdoor temperature and humidity, the number of people information, the predicted outdoor temperature and humidity values at a plurality of time steps in the future, the predicted number of people information at a plurality of time steps in the future, and a plurality of preset candidate air conditioner operation parameter sequences into the first model for prediction processing;
and according to the prediction processing, obtaining an air conditioning system energy consumption prediction value, an air conditioning cold output prediction value and an indoor temperature and humidity prediction value of a plurality of time step lengths in the future corresponding to each preset alternative air conditioning operation parameter sequence.
Further, the first model comprises an air conditioning system model and a space heat transfer model;
the prediction unit is further used for inputting the indoor temperature and humidity, the outdoor temperature and humidity and the plurality of preset alternative air conditioner operation parameter sequences of the current time step into an air conditioner system model, and performing first prediction processing to obtain a corresponding first air conditioner system energy consumption prediction value and a first air conditioner cooling output prediction value;
inputting the first air conditioner cold output predicted value and the corresponding outdoor temperature and humidity, the indoor temperature and humidity and the number information into a space heat transfer model, and performing second prediction processing to obtain a corresponding first indoor temperature and humidity predicted value;
inputting the first indoor temperature and humidity predicted value, the outdoor temperature and humidity predicted value and the number information predicted value of the next time step and the plurality of preset candidate air conditioner operation parameter sequences into the first model, and performing the first prediction processing and the second prediction processing to obtain a corresponding second air conditioner system energy consumption predicted value, a second air conditioner cold output predicted value and a second indoor temperature and humidity predicted value;
and repeating the first prediction processing and the second prediction processing until obtaining the energy consumption prediction value of the air conditioning system, the air conditioning cold output prediction value and the indoor temperature and humidity prediction value at a plurality of time steps in the future.
Further, the optimization unit is further configured to calculate a loss value corresponding to each candidate air conditioner operation parameter sequence according to a preset loss function and the prediction result if the optimization termination condition is not satisfied;
calculating the gradient of each alternative air conditioner operation parameter sequence according to the loss function, and adjusting the alternative air conditioner operation parameter sequences according to the gradient to obtain a plurality of new alternative air conditioner operation parameter sequences;
inputting the new alternative air conditioner operation parameter sequence and the first data set into a first model, and outputting a new prediction result;
and carrying out optimization processing according to the new prediction result until the first air conditioner operation parameter sequence is obtained.
Further, the prediction unit is further configured to calculate, when the optimization termination condition is met, a loss value corresponding to each alternative air conditioner operation parameter sequence through a preset loss function according to a corresponding predicted value of indoor temperature and humidity and a predicted value of energy consumption of the air conditioning system in combination with the corresponding multiple alternative air conditioner operation parameter sequences;
comparing the numerical value of the loss value corresponding to each alternative air conditioner operation sequence, and screening out the minimum loss value with the minimum numerical value;
and determining the alternative air conditioner operation parameter sequence corresponding to the minimum loss value as the first air conditioner operation parameter sequence.
Further, the air conditioning system model is expressed as:
(P(t+1),Q(t+1),Q lat (t+1))＝μ θ (u(t+1),T r (t),W r (t),T a (t),W a (t))，
wherein, mu θ Representing an artificial neural network, theta represents a model parameter set of the artificial neural network, P (t + 1) represents an air conditioning system energy consumption predicted value of t +1 time step, the air conditioning cold output predicted value comprises an air conditioning sensible heat and cold output predicted value and an air conditioning latent heat and cold output predicted value, Q (t + 1) represents an air conditioning sensible heat and cold output predicted value of t +1 time step, and Q (t + 1) represents an air conditioning sensible heat and cold output predicted value of t +1 time step lat (T + 1) represents the predicted value of the output of the latent heat and the cold of the air conditioner at the T +1 time step, u (T + 1) is the preset air conditioner operation parameter at the T +1 time step, and T r (t) represents the indoor temperature, W, of the t-th time step r (T) represents the indoor humidity at the tth time step, T a (t) outdoor temperature, W, for the t-th time step a (t) represents the outdoor humidity for the t time step;
the prediction unit is also used for converting the outdoor temperature T in the T time step a (t) outdoor humidity W a (T) indoor temperature T r (t) indoor humidity W r (t) and alternative air conditioner operating parameter sequence U n Inputting the time step into the model of the air conditioning system to obtain the t +1 time stepAn air conditioning system energy consumption predicted value P (t + 1), an air conditioning sensible heat and cold output predicted value Q (t + 1) and an air conditioning latent heat and cold output predicted value Q lat (t+1)。
Further, the candidate air conditioner operation parameter sequence includes candidate air conditioner operation parameters of a plurality of time steps in the future, and the candidate air conditioner operation parameter sequence U n Expressed as:
Figure BDA0003927919060000071
wherein u is n (t) the alternative air conditioner operation parameter for the t time step in the nth alternative air conditioner operation parameter sequence, u n (t + 1) represents the alternative air conditioner operation parameter of the t +1 time step in the nth alternative air conditioner operation parameter sequence, u n (t+K max ) Represents t + K in the nth candidate air conditioner operation parameter sequence max Alternative air-conditioning operating parameters of individual time steps, K max Is a positive integer.
Further, the spatial heat transfer model is represented as:
Figure BDA0003927919060000072
Figure BDA0003927919060000073
wherein, T r (t + 1) represents the predicted indoor temperature value of the t +1 th time step, W r (t + 1) represents the predicted value of the indoor humidity at the t +1 th time step, N in (t) information on the number of persons entering the t-th time step, N out (t) leaving people information representing the t-th time step, k 0 ～k 4 、b、j 0 ～j 4 Model parameters representing the spatial heat transfer model;
the prediction unit is also used for outputting the air conditioner sensible heat and cold quantity predicted value Q (t + 1) and the air conditioner latent heat and cold quantity output value of the tth time stepQ lat (T + 1), outdoor temperature T a (t) outdoor humidity W a (T) indoor temperature T r (t) indoor humidity W r (t) number of persons entering information N in (t) and leaving population information N out (T) inputting the indoor temperature into the space heat transfer model to obtain an indoor temperature predicted value T of the T +1 time step r (t + 1) and predicted indoor humidity value W r (t+1)。
Further, the preset loss function is represented by loss = obj + pen, where obj represents an objective function and pen represents a penalty function, where the objective function is represented by:
Figure BDA0003927919060000074
wherein, K max Representing the predicted step number, T r (T + k) represents the predicted indoor temperature value of the T + k time step, T tg Representing the preset target indoor temperature, P (t + k) represents the predicted value of the energy consumption of the air conditioning system with t + k time steps, w 1 Represents a preset comfort weight coefficient, w 2 Representing a preset energy consumption weight coefficient of the air conditioning system;
the penalty function is expressed as:
Figure BDA0003927919060000075
wherein n is p Number of representing constraints, c n，k Representing a constraint function, relu a linear rectification function, p n,k Representing a constraint function c n,k The penalty factor of (2).
Further, the model parameter updating unit is further configured to, when each time step ends, perform model parameter updating processing on the spatial heat transfer model according to the second data set corresponding to the time step, and determine a first parameter;
when the preset period is finished, carrying out model parameter updating processing on the air conditioning system model according to the second data set of all time step lengths in the period, determining a second parameter, and forming a period by a plurality of time step lengths;
and determining the model parameters of the first model according to the first parameters and/or the second parameters to obtain the first model.
In a third aspect, an embodiment of the present application provides an energy saving control apparatus for an air conditioner, including:
a memory and one or more processors;
the memory 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 are caused to implement the energy-saving control method of the air conditioner according to the first aspect.
In a fourth aspect, embodiments of the present application provide a storage medium storing computer-executable instructions for performing the air conditioner energy saving control method according to the first aspect when executed by a computer processor.
According to the method and the device, when each time step is finished, model parameters of a preset model are updated according to a second data set to obtain a first model, the first data set and a plurality of preset alternative air conditioner operation parameter sequences are input into the first model, corresponding prediction results are output, optimization processing is carried out according to the prediction results to obtain a first air conditioner operation parameter sequence, a first air conditioner operation parameter in the first air conditioner operation parameter sequence is used as a second air conditioner operation parameter and is output, and the operation of the next time step of an air conditioning system is controlled according to the second air conditioner operation parameter. By adopting the technical means, the first air conditioner operation parameter sequence with the optimal energy-saving effect is screened out by predicting and optimizing the plurality of preset alternative air conditioner operation parameters, the first air conditioner operation parameter in the first air conditioner operation parameter sequence is used as the second air conditioner operation parameter of the air conditioning system corresponding to the next time step length actual control, the optimization of the air conditioner operation parameters can be realized, and the energy-saving effect of the air conditioning system can be improved by controlling the air conditioning system through the optimized air conditioner operation parameters. In addition, model parameters are updated through the second data set, the adaptivity of the model can be improved, the model is more suitable for actual operation conditions, errors between air conditioner energy consumption predicted values corresponding to predicted second air conditioner operation parameters and actual air conditioner energy consumption values corresponding to actual control through the second air conditioner operation parameters can be reduced, the accuracy of prediction results is improved, and the energy-saving effect of an air conditioning system is further improved.
Drawings
Fig. 1 is a flowchart of an energy-saving control method for an air conditioner according to an embodiment of the present disclosure;
FIG. 2 is a schematic structural diagram of a default model according to an embodiment of the present disclosure;
fig. 3 is a schematic structural diagram of an energy-saving control device of an air conditioner according to an embodiment of the present application;
fig. 4 is a schematic structural diagram of an air conditioner energy-saving control device provided in an embodiment of the present application.
Detailed Description
In order to make the objects, technical solutions and advantages of the present application more apparent, specific embodiments of the present application will be described in detail with reference to the accompanying drawings. It is to be understood that the specific embodiments described herein are merely illustrative of the application and are not limiting of the application. It should be further noted that, for the convenience of description, only some but not all of the relevant portions of the present application are shown in the drawings. Before discussing exemplary embodiments in more detail, it should be noted that some exemplary embodiments are described as processes or methods depicted as flowcharts. Although a flowchart may describe the operations (or steps) as a sequential process, many of the operations can be performed in parallel, concurrently, or simultaneously. In addition, the order of the operations may be re-arranged. The process may be terminated when its operations are completed, but may have additional steps not included in the figure. The processes may correspond to methods, functions, procedures, subroutines, and the like.
The application provides an air conditioner energy-saving control method and device, and aims to screen out a first air conditioner operation parameter sequence with the optimal energy-saving effect by predicting and optimizing a plurality of preset alternative air conditioner operation parameters when an air conditioner system is controlled, and the first air conditioner operation parameter in the first air conditioner operation parameter sequence is used as a second air conditioner operation parameter of the air conditioner system corresponding to the next time step length actual control, so that the optimization of the air conditioner operation parameters is realized, and the air conditioner system is controlled through the optimized air conditioner operation parameters, so that the energy-saving effect of the air conditioner system is improved. And updating the model parameters through the second data set so as to improve the adaptivity of the model, reduce the error between the predicted air conditioner energy consumption value corresponding to the predicted second air conditioner operation parameter and the actual air conditioner energy consumption value corresponding to the actual control through the second air conditioner operation parameter, and improve the accuracy of the prediction result. Compared with the traditional control mode of the air conditioning system, the control mode usually presets corresponding air conditioning operation parameters based on manual experience, and has great subjectivity and limitation in the aspects of energy-saving control of the air conditioning system and perception of environment somatosensory comfort in the actual operation process, so that the energy-saving effect in the air conditioning control process is poor. Therefore, the air conditioner energy-saving control method provided by the embodiment of the application is provided to solve the problem that the existing air conditioner control process is poor in energy-saving effect.
Fig. 1 is a flowchart of an air conditioner energy saving control method provided in an embodiment of the present application, where the air conditioner energy saving control method provided in this embodiment may be executed by an air conditioner energy saving control device, the air conditioner energy saving control device may be implemented by software and/or hardware, and the air conditioner energy saving control device may be formed by two or more physical entities or may be formed by one physical entity. Generally, the air-conditioning energy-saving control device can be an upper computer of an air-conditioning system, such as a computer device.
The following description will be given taking a computer device as an example of a subject that executes the air-conditioning energy-saving control method. Referring to fig. 1, the air conditioner energy-saving control method specifically includes:
s101, when each time step is finished, updating model parameters of a preset model according to a second data set to obtain a first model, wherein the second data set comprises indoor temperature and humidity, outdoor temperature and humidity, people number information, actual air conditioner operation parameters, an air conditioner system energy consumption value and an air conditioner cooling output quantity of the current time step.
The number of people in large-scale activity places such as stations or shopping malls is particularly important for determining the operation parameters in the air conditioner control process. The more the number of people, the more the cooling output of the corresponding control air conditioner, but the energy consumption of the air conditioner is increased at the same time. Therefore, how to integrate the information of the number of people and the indoor and outdoor temperature and humidity to consider the determination of the air conditioner operation parameters is a main problem to be solved by the embodiment, so that the energy consumption of the air conditioner is reduced as much as possible while the comfort of people is ensured. Therefore, the alternative air conditioner operation parameter with a good energy saving effect needs to be acquired according to the existing passenger flow volume and environmental data and is used as the air conditioner operation parameter of the next control period, so that the air conditioner system is controlled through the air conditioner operation parameter in the next control period, and the good energy saving effect is achieved.
The number information indicates the intensity of the passenger flow at a station, a subway station, or the like, and can be represented by the number of persons entering the passenger flow and the number of persons leaving the passenger flow within a single time step. The information of the number of people in a large building such as a shopping mall or an office building represents the number of people in a room in the building, and can be represented by information of the cumulative number of people entering the building on the day and information of the cumulative number of people leaving the building on the day.
Before the acquisition of the air conditioner operation parameters of the next control period is carried out, the specific duration of the control period needs to be determined. A time step may be preset, one time step being one control period. One time step belongs to one time period, and the specific numerical value of the duration corresponding to one time step can be set according to the actual situation. And obtaining alternative air conditioner operation parameters with better energy-saving effect according to the passenger flow and the environmental condition of the current time step as the air conditioner operation parameters of the next time step so as to improve the energy-saving effect of the air conditioning system of the next time step. Therefore, it is necessary to acquire the corresponding passenger flow volume data, environmental data, air conditioner energy consumption data, and the like at each time step.
And acquiring a second data set according to the time step length, wherein the second data set comprises the indoor temperature and humidity, the outdoor temperature and humidity, the information of the number of people, the actual air conditioner operation parameters, the energy consumption value of the air conditioning system and the air conditioning cold output quantity of the current time step length. The first data set comprises indoor temperature and humidity, outdoor temperature and humidity and people number information of the current time step, and outdoor temperature and humidity predicted values and people number information predicted values of a plurality of time steps in the future. The outdoor temperature and humidity predicted values of a plurality of time steps in the future can be obtained by predicting corresponding weather information through the Internet, and the corresponding people number information prediction can take the people number information corresponding to the same time step in the previous day as the corresponding people number information predicted value or predict the people number information corresponding to the time step by using other passenger flow/people number prediction modes. And forming a period based on a plurality of step lengths, and obtaining a plurality of second data sets and a first data set according to the step lengths in the period, wherein each second data set comprises actual air conditioner operation parameters, indoor temperature and humidity values, outdoor temperature and humidity values, people number information, air conditioner energy consumption values and air conditioner cold output values of corresponding time step lengths. And an auxiliary data basis is provided for the prediction process of the subsequent alternative air conditioner operation parameters by acquiring the corresponding first data set according to the time step.
In an embodiment, when acquiring data, the indoor temperature and humidity value and the outdoor temperature and humidity value may be acquired by the temperature and humidity sensor. And in each time step, acquiring data of the current indoor temperature and the current outdoor temperature through temperature and humidity sensors arranged indoors and outdoors. When a plurality of temperature and humidity sensors are arranged indoors, the average value of the indoor temperature and humidity values acquired by all the temperature and humidity sensors is calculated, and the obtained average temperature and humidity are used as the current indoor temperature and humidity value T r (t),W r (t) of (d). Similarly, when a plurality of outdoor temperature and humidity sensors exist outdoors, the average value of the outdoor temperature and humidity acquired by all the outdoor temperature and humidity sensors is calculated, and the obtained average temperature and humidity value is used as the current outdoor temperature and humidity value T a (t),W a (t) of (d). Obtaining an outdoor temperature and humidity predicted value, namely T, of each time step within a predicted step range through an Internet weather application interface a (t+k),W a (t+k),
Figure BDA0003927919060000101
The predicted value of (2). Wherein t represents the current time step, t + k represents the kth time step in the future, the outdoor temperature and humidity predicted value of the next time step (t + 1) and the outdoor temperature and humidity predicted value of the next time step (t + 2) can be obtained through Internet weather applicationAnd measuring the values until obtaining the predicted outdoor temperature and humidity values of k time steps (t + k) in the future. Corresponding entering people number information and leaving people number information are obtained through the passenger flow volume statistical equipment in the corresponding place, taking a station as an example, the people number information represents the number of people in and out of the station, and the number of people in and out of the station with the current time step length, namely N, can be obtained through communication with an AFC system corresponding to the station in (t),N out (t) obtaining the predicted value of the passenger flow volume in and out at each time step within the predicted step range by the existing passenger flow prediction mode, namely
Figure BDA0003927919060000111
The predicted value of (2). The predicted value of the number of people in the next time step (t + 1) and the predicted value of the number of people in the next time step (t + 2) can be obtained through the existing method for predicting the number of people in and out until the predicted value of the number of people in the next k time steps (t + k) is obtained. The energy consumption value of the air conditioning system, the air conditioning cooling output quantity and the actual operation air conditioning operation parameters, namely P (t), Q (t) and Q (t) of the current time step length are obtained through communication with all equipment of the air conditioning system lat (t),u(t)。
Before the air conditioning system control of the next time step is carried out, the prediction and optimization processing of the air conditioner energy consumption value and the indoor temperature and humidity value can be carried out according to the actual indoor temperature and humidity, the outdoor temperature and humidity and the number information of people of the current time step and the preset optional air conditioner operation parameter data, and then the air conditioner operation parameter with better energy saving effect is obtained and used as the air conditioner operation parameter of the next time step. And predicting the corresponding air conditioner energy consumption value and indoor temperature and humidity value through a preset model. Before prediction, in order to guarantee accuracy of a prediction result, parameters of a preset model can be updated, and prediction is performed after a new model is obtained. And when updating the model parameters, updating the model parameters of the preset model according to the second data set of the current time step when each time step is finished to obtain new model parameters, and determining the first model with updated model coefficients according to the new model parameters. The model data is updated at each time step to improve the accuracy of the subsequent prediction result output through the first model, so that the accuracy of the prediction of the corresponding energy-saving effect of the predicted air-conditioning operation parameter is improved, and the energy-saving effect of the air-conditioning system at the next time step is further ensured.
When the model is preset, training can be carried out through the sample data set to obtain corresponding initial model parameters. The sample data set is composed of a second data set of a plurality of time steps acquired historically. And performing data storage on a sample data set consisting of sample data such as actual air conditioner operation parameters, indoor temperature and humidity, outdoor temperature and humidity, people number information, energy consumption values of an air conditioning system, air conditioner cold output quantity and the like corresponding to a plurality of second data sets which are acquired in history, and training a preset model according to the stored sample data set to obtain corresponding initial model parameters and an initial preset model. And when each time step is finished, updating the model parameters of the preset model according to the second data set obtained from the current time step to obtain new model parameters, and determining the updated first model according to the new model parameters. By updating the model parameters at the end of each time step, the updated first model of the model parameters is closer to the actual operation condition, so that the accuracy of model prediction is improved, the adaptability of the model is improved, and the accuracy of the model prediction result is improved.
In one embodiment, the sample data set may be simulation data or historical data, or both simulation data and historical data trained in combination. The simulation data can be obtained by establishing a simulation model of the air conditioning system, running simulation under random working conditions and recording input and output values corresponding to a preset model in each time step. The generation of the historical data needs to record input and output values of the same or similar air conditioning system in the actual operation process.
Fig. 2 is a schematic structural diagram of a preset model according to an embodiment of the present application, and referring to fig. 2, the preset model is composed of two models, namely, an air conditioning system model 10 and a space heat transfer model 20. It should be noted that the first model is a model obtained by updating the model parameters of the preset model, and therefore the representation form of the first model is the same as that of the preset model, except that the corresponding model parameters are different, that is, the first model also includes the air conditioning system model 10 and the space heat transfer model 20.
Wherein the preset air conditioning system model 10 is represented as:
(P(t+1),Q(t+1),Q lat (t+1))＝μ θ (u(t+1),T r (t),W r (t),T a (t),W a (t))
wherein, mu θ Representing an artificial neural network, theta represents a model parameter set of the artificial neural network, P (t + 1) represents an air conditioning system energy consumption predicted value of t +1 time step, the air conditioning cold output predicted value comprises an air conditioning sensible heat and cold output predicted value and an air conditioning latent heat and cold output predicted value, Q (t + 1) represents an air conditioning sensible heat and cold output predicted value of t +1 time step, and Q (t + 1) represents an air conditioning sensible heat and cold output predicted value of t +1 time step lat (T + 1) represents the predicted value of the output of the latent heat and the cold of the air conditioner at the T +1 time step, u (T + 1) is the preset air conditioner operation parameter at the T +1 time step, and T r (t) represents the indoor temperature, W, of the t-th time step r (T) represents the indoor humidity at the tth time step, T a (t) outdoor temperature, W, for the t-th time step a (t) represents the outdoor humidity for the t time step.
It should be noted that, when T is the current time step, T is r (t),W r (t),T a (t),W a (t) taking the actual value, otherwise taking the predicted value.
Illustratively, the specific content of the air-conditioning operation parameters depends on the type and configuration of the air-conditioning system, and taking a water-cooled central air-conditioning system commonly found in a subway station as an example, the air-conditioning operation parameters u (t) may include the outlet water temperature of a chiller, the flow rate of chilled water, the air volume of each fan, and the like. And u (t) represents the actual air conditioner operation parameter of the t-th time step when the air conditioner operation parameter u (t) is trained or the model parameter is updated, and represents the air conditioner operation parameter in the alternative air conditioner operation parameter sequence of the t-th time step when the prediction is carried out.
Based on the preset air conditioning system model, the preset air conditioning system model can be trained through the sample data set to obtainCorresponding initial model parameter mu θ . It should be noted that, when the air conditioning system model is trained through the sample parameters, the sample data set includes multiple sets of data, and theoretically, the more data of the sample data set, the more accurate the training result is.
Updating of the model coefficients for the air conditioning system model requires retraining of multiple sets of data to obtain. Therefore, when the preset period is finished, the model parameter updating processing is carried out on the air conditioning system model according to the second data set of all time step lengths in the period, and a second parameter is determined, wherein the second parameter is the updated model parameter of the air conditioning system model in the period. A plurality of time steps form a period, and the number of the time steps divided corresponding to the period can be set according to actual conditions.
The preset spatial heat transfer model 20 in the preset model is represented as:
Figure BDA0003927919060000131
Figure BDA0003927919060000132
wherein, T r (t + 1) represents the predicted indoor temperature value of the t +1 th time step, W r (t + 1) represents the predicted indoor humidity value of the t +1 time step, N in (t) information on the number of persons entering the t-th time step, N out (t) leaving people information representing the t-th time step, k 0 ～k 4 、b、j 0 ～j 4 Model parameters representing the spatial heat transfer model.
The number of persons entering/leaving in the t-th time step is the cumulative number of persons entering/leaving up to the t-th time step for a general building, and the number of persons entering/leaving in the t-th time step for a building with high mobility of people, such as a subway station.
Through the space heat transfer model, a plurality of calendars corresponding to the sample data set can be collected through the sample dataTraining the space heat transfer model to obtain corresponding initial model parameters k according to the number information acquired by history and the air conditioner cold output value 0 ～k 4 、b、j 0 ～j 4 . The number of people and the air conditioner cold output value can be obtained through a plurality of histories corresponding to the sample data set, and k is determined by adopting a least square method 0 ～k 4 、b、j 0 ～j 4 Fitting a spatial heat transfer model to the operating data. Or, if the historical operation data cannot be obtained, adopting the following initial values:
Figure BDA0003927919060000133
it should be noted that, when the model is trained through the sample data set, the sample data set includes multiple groups of data, and theoretically, the more data of the sample data set, the more accurate the training result is.
When the model parameters of the space heat transfer model are updated and each time step is finished, the model parameters of the space heat transfer model are updated according to the second data set corresponding to the time step, and a first parameter is determined, wherein the first parameter is the updated space heat transfer model parameter of the current step.
It should be noted that, when each time step is finished, the model parameters of the corresponding air conditioning system model in the obtained first model are not changed, and the model parameters of the corresponding space heat transfer model are updated. And at the end of each preset period, updating the model coefficients of the corresponding air conditioning system model and the space heat transfer model in the obtained first model.
By updating the model parameters when each time step is finished, the prediction result of the first model after the model parameters are updated is more accurate, and the accuracy of the energy-saving effect prediction corresponding to the subsequently obtained second air conditioner operation parameters can be improved.
In one embodiment, an implementation of model parameter updates to a pre-defined spatial heat transfer model is provided. At the end of each time step the time is,and acquiring data of the second data set according to the data acquisition mode, and storing the data in the temporary data set. Spatial heat transfer model parameter k by using Kalman filtering algorithm 0 ～k 4 、b、j 0 ～j 4 And (6) updating. In one setting mode of the kalman filter algorithm, a state transition equation and an observation equation in the kalman filter algorithm can be respectively set as follows:
x (t) = Ax (t-1) + w (t-1), Z (t) = H (t) x (t) + v (t), wherein a = I 11 I.e., a is an identity matrix of order 11. w (t-1) represents random process noise and can be preset. v (t) represents the random measurement error of the indoor temperature and humidity, and can be preset according to the performance of the temperature and humidity sensor.
Figure BDA0003927919060000141
Figure BDA0003927919060000142
In one embodiment, an implementation of model parameter updating for a predetermined air conditioning system model is provided. A preset period is formed by a plurality of time steps, and when each preset period is finished, a second data set corresponding to the plurality of time steps in the period is obtained according to the data obtaining mode, wherein the second data set comprises the indoor temperature and humidity value T of the current time step r (t),W r (T), outdoor temperature and humidity value T of current time step a (t),W a (t), entering number of people information N of current time step in (t) and leaving population information N out (t), the energy consumption value P (t) of the air conditioning system of the current time step, the air conditioning cooling output quantity Q (t) of the current time step, Q lat (t), and the actual air conditioner operation parameter u (t) for the current time step. The data acquired at each time step is stored in a temporary data set. Updating the model parameters at the end of the last time step of a predetermined periodic interval, e.g. one day, the last time step of each dayAnd updating the model parameters after the time step is finished. The specific preset period interval can be set according to actual conditions, and can be set to one day or specific hours. When updating the model parameters of the air conditioning system model, training the current air conditioning system model on a part of temporary data sets, and circularly adjusting the model parameter set theta by using an artificial neural network parameter optimization method, such as a random gradient descent method and the like, until the air conditioning system model can be properly fitted to the temporary data sets, namely the air conditioning system model input of the temporary data set corresponding to any t-th time step is given, and the deviation of the output of the air conditioning system model and the output value of the air conditioning system corresponding to the t-th time step in the temporary data sets meets a preset difference range. Typically, this preset difference range is set relatively small. The air conditioning system model is periodically updated, so that the adaptivity of the air conditioning system model is improved, and the accuracy of subsequent energy saving prediction is improved.
S102, inputting a first data set and a plurality of preset alternative air conditioner operation parameter sequences into the first model, and outputting corresponding prediction results, wherein the prediction results comprise indoor temperature and humidity prediction values and air conditioner system energy consumption prediction values, the first data set comprises indoor temperature and humidity, outdoor temperature and humidity and people number information of the current time step, and outdoor temperature and humidity prediction values and people number information prediction values of a plurality of time steps in the future.
On the basis of the air conditioner operation parameters of actual operation, in order to reduce the energy consumption of the air conditioner as much as possible and improve the energy-saving effect on the basis of meeting the comfort requirement under the current number information, the air conditioner operation parameters can be further optimized according to the current number information condition and the environmental condition. The actual operation parameters can be used as reference, a plurality of candidate air conditioner operation parameter sequences are preset, and optimization processing is carried out according to the preset candidate air conditioner operation parameter sequences to obtain corresponding target air conditioner operation parameters. Wherein the alternative air conditioner operation parameter sequence represents alternative air conditioner operation parameters of a plurality of time steps in the future. And randomly generating a plurality of alternative air conditioner operation parameter sequences by referring to the air conditioner operation parameters in actual operation. The nth alternative air conditioner operation parameter sequence can be expressed as follows:
Figure BDA0003927919060000151
wherein u is n (t + 1) represents the alternative air conditioner operation parameter of t +1 time step in the nth alternative air conditioner operation parameter sequence, u n (t + 2) represents the alternative air conditioner operation parameter of the t +2 time step in the nth alternative air conditioner operation parameter sequence, u n (t+K max ) Represents t + K in the nth candidate air conditioner operation parameter sequence max Alternative air-conditioning operating parameters of individual time steps, K max Is a positive integer and is a non-zero integer,
Figure BDA0003927919060000152
when predicting, inputting the indoor temperature and humidity of the current time step, the outdoor temperature and humidity, the information of people number, the predicted value of the outdoor temperature and humidity of a plurality of time steps in the future, the predicted value of the information of the people number of the plurality of time steps in the future and a plurality of preset candidate air conditioner operation parameter sequences into the first model for prediction processing. And according to the prediction processing of the first model, obtaining an air conditioning system energy consumption predicted value, an air conditioning cold output predicted value and an indoor temperature and humidity predicted value of a plurality of time step lengths in the future corresponding to each preset alternative air conditioning operation parameter sequence.
When prediction is carried out, the indoor temperature and humidity, the outdoor temperature and humidity and a plurality of preset alternative air conditioner operation parameter sequences of the current time step are input into an air conditioner system model, first prediction processing is carried out, and a corresponding first air conditioner system energy consumption prediction value and a first air conditioner cooling output prediction value are obtained. And inputting the first air conditioner cold output value predicted value, the corresponding outdoor temperature and humidity, indoor temperature and humidity, the number information of people and a plurality of preset optional air conditioner operation parameter sequences into a space heat transfer model, and performing second prediction processing to obtain the corresponding first indoor temperature and humidity predicted value. Wherein the predicted value of the first air conditioner energy consumption system, the predicted value of the first air conditioner cold output and the predicted value of the first indoor temperature and humidity are predicted values of the next time step, namely the predicted value of the t +1 time step, therefore, a plurality of alternative prediction processing is utilized in the first prediction processingThe alternative air conditioner operation parameter u corresponding to the next time step length in the air conditioner operation parameter sequence n (t + 1). On the basis, when prediction is carried out once, the predicted value of the next time step, namely the predicted value of the t +2 time step, is predicted. Outputting the first air-conditioning cold quantity, the outdoor temperature and humidity predicted value and the number information predicted value of the next time step and a plurality of preset alternative air-conditioning operation parameter sequences, inputting the first air-conditioning cold quantity, the outdoor temperature and humidity predicted value and the number information predicted value of the next time step and the plurality of preset alternative air-conditioning operation parameter sequences into the first model, respectively performing first prediction processing on the corresponding space system model and second prediction processing on the space heat transfer model to obtain the corresponding second air-conditioning system energy consumption predicted value, the second air-conditioning cold quantity output predicted value and the second indoor temperature and humidity predicted value, and utilizing the corresponding alternative air-conditioning operation parameter u corresponding to the next time step (t + 2) in the plurality of alternative air-conditioning operation parameter sequences in the first prediction processing in the prediction process n (t + 2). And repeating the processes of the first prediction processing and the second prediction processing until the energy consumption prediction value, the air-conditioning cooling output prediction value and the indoor temperature and humidity prediction value of the air-conditioning system corresponding to a plurality of time steps in the future are obtained, namely the prediction values corresponding to k time steps in the future. Based on that the alternative air conditioner operation parameter sequence represents alternative air conditioner operation parameters of a plurality of time steps in the future, an air conditioner energy consumption predicted value and an indoor temperature and humidity predicted value corresponding to the alternative air conditioner operation parameter corresponding to each future time step in each alternative air conditioner operation parameter sequence can be obtained by inputting the indoor temperature and humidity, the outdoor temperature and humidity, the number information of people and a plurality of preset alternative air conditioner operation parameter sequences of the current time step into the first model.
Specifically, a specific prediction process of one air conditioner operation parameter sequence is taken as an example for explanation, the air conditioner operation parameter sequence includes alternative air conditioner operation parameters of a plurality of time steps in the future, the indoor temperature and the outdoor temperature of the current time step and the alternative air conditioner operation parameters of the next time step in the alternative air conditioner operation parameter sequence are input into the air conditioner system model, and a first air conditioner system energy consumption prediction value and a first air conditioner cooling output prediction value of the next time step in the future, which correspond to the alternative air conditioner operation parameters of the next time step in the future, are output. And inputting the first air conditioning system energy consumption predicted value and the first air conditioning cold output predicted value corresponding to the alternative air conditioning operation parameter of the next time step in the future and the number of people information of the current time step into the space heat transfer model to obtain a corresponding first indoor temperature and humidity predicted value. Therefore, the indoor temperature and humidity of the current time step, the outdoor temperature and humidity, the number information of people and the next time step of the alternative air conditioner operation parameters in the alternative air conditioner operation parameter sequence in the future are input into the first model, and the first air conditioner system energy consumption predicted value and the first indoor temperature and humidity predicted value corresponding to the next time step of the alternative air conditioner operation parameters in the alternative air conditioner operation parameter sequence can be output. And repeating the process, and inputting the indoor temperature and humidity, the outdoor temperature and humidity, the number information, the number prediction information of a plurality of future time steps, the outdoor temperature and humidity prediction information of a plurality of future time steps and a plurality of preset alternative air conditioner operation parameter sequences corresponding to the current time step into the first model to obtain the air conditioner system energy consumption prediction value and the indoor temperature and humidity prediction value corresponding to the air conditioner alternative operation parameter of each future time step corresponding to the corresponding preset alternative air conditioner operation parameter sequence.
For example, the specific prediction process of the first alternative air conditioner operation parameter sequence is taken as an example for explanation, and the first alternative air conditioner operation parameter sequence is expressed as
Figure BDA0003927919060000171
First alternative air conditioner operation parameter sequence U 1 Alternative air conditioner operating parameters u including multiple time steps in the future 1 (t+1)、u 1 (t+2)....u 1 (t+K max ). The indoor temperature and humidity (T) of the T time step r (t),W r (T)), outdoor temperature and humidity (T) a (t),W a (t)) and the alternative air conditioner operating parameter sequence U 1 The alternative air conditioner operation parameter u of the future first time step (t + 1) 1 (t + 1) inputting the data into an air conditioning system model, and outputting an alternative air conditioning operation parameter u 1 (t + 1) corresponds toPredicted value P of energy consumption of air conditioning system 1 (t + 1) and air conditioner cooling output predicted value
Figure BDA0003927919060000172
The obtained alternative air conditioner operation parameter u 1 (t + 1) corresponding air conditioner cold output prediction value
Figure BDA0003927919060000173
And the number of people in the t time step (N) in (t),N out (t)) inputting the parameters into a space heat transfer model to obtain an alternative air conditioner operation parameter u 1 Indoor temperature and humidity predicted value (T) corresponding to (T + 1) r1 (t+1),W r1 (t + 1)), completing the alternative air conditioner operation parameter u 1 (t + 1) to obtain an alternative air conditioner operation parameter u 1 (t + 1) corresponding air-conditioning system energy consumption predicted value P 1 (t + 1) air conditioner cold output prediction value
Figure BDA0003927919060000174
And predicted value of indoor temperature and humidity (T) r1 (t+1),W r1 (t + 1)). Then, the alternative air conditioner operation parameter u is carried out 1 (t + 2) prediction processing of the alternative air conditioner operation parameter u obtained as described above 1 (T + 1) corresponds to the predicted value of indoor temperature and humidity (T) r1 (t+1),W r1 (T + 1)), and outdoor temperature and humidity predicted value (T) of T +1 th time step a (t+1),W a (t + 1)) and an alternative air conditioner operating parameter u 1 (t + 2) inputting the air-conditioning system model, and outputting the alternative air-conditioning operation parameter u 1 (t + 2) corresponding predicted value P of energy consumption of air conditioning system 1 (t + 2) and air conditioner cooling output prediction value
Figure BDA0003927919060000175
The obtained alternative air conditioner operation parameter u 1 (t + 2) corresponding air conditioner cold output prediction value
Figure BDA0003927919060000176
And the predicted value (N) of the information of the number of people at the t +1 time step in (t+1),N out (t + 1)) is input into a space heat transfer model to obtain an alternativeAir conditioner operation parameter u 1 Indoor temperature and humidity predicted value (T) corresponding to (T + 2) r1 (t+2),W r1 (t + 2)), completing the alternative air conditioner operation parameter u 1 (t + 2) to obtain an alternative air conditioner operation parameter u 1 (t + 2) corresponding predicted value P of energy consumption of air conditioning system 1 (t + 2) air conditioner cooling output predicted value
Figure BDA0003927919060000177
And predicted value of indoor temperature and humidity (T) r1 (t+2),W r1 (t + 2)). Then, the alternative air conditioner operation parameter u is carried out 1 (t+3)......u 1 (t+K max ) Obtaining each corresponding candidate air conditioner operation parameter u through the prediction processing 1 (t+1)、u 1 (t+2)....u 1 (t+K max ) Corresponding predicted values of energy consumption, cold output and indoor temperature and humidity of air conditioning system, such as candidate air conditioner operation parameter u 1 (t+K max ) Corresponding predicted value P of energy consumption of air conditioning system 1 (t+K max ) Air conditioner cooling output predicted value
Figure BDA0003927919060000178
And predicted value of indoor temperature and humidity (T) r1 (t+K max ),W r1 (t+K max )). Note that K is max The specific value of (b) can be set according to actual conditions, and is not limited in this embodiment.
Through the embodiment, the indoor temperature and humidity T according to the tth time step is realized r (T) outdoor temperature and humidity T a (t) information on the number of persons (N) in (t),N out (T)), outdoor temperature and humidity predicted values (T) for a plurality of time steps in the future a (t+K max ),W a (t+K max ) Predicted value (N) of information on the number of people at a plurality of time steps in the future in (t+K max ),N out (t+K max ) And a first alternative air conditioner operating parameter sequence U 1 All alternative air conditioner operation parameters u in 1 (t+1)、u 1 (t+2)....u 1 (t+K max ) Inputting the first model to obtain a first alternative air conditioner operation parameter sequence U 1 InAir conditioning system energy consumption predicted value P corresponding to each alternative air conditioning operation parameter 1 (t)、P 1 (t+1)...P 1 (T + K) and predicted indoor temperature and humidity value (T) r1 (t+1),W r1 (t+1))、(T r1 (t+2),W r1 (t+2))...(T r1 (t+K),W r1 (t + K)). Furthermore, a second alternative air conditioner operation parameter sequence U can be obtained in the same way 2 The predicted value of energy consumption and indoor temperature and humidity of the air conditioning system corresponding to each alternative air conditioning operation parameter in the sequence U of the third alternative air conditioning operation parameter 3 The energy consumption predicted value and the indoor temperature and humidity predicted value of the air conditioning system corresponding to each candidate air conditioning operation parameter, and the nth candidate air conditioning operation parameter sequence U n And the air-conditioning system energy consumption predicted value and the indoor temperature and humidity predicted value corresponding to each alternative air-conditioning operation parameter are obtained. The specific n value setting may be limited according to actual situations, and is not limited in this embodiment.
S103, carrying out optimization processing according to the prediction result to obtain a first air conditioner operation parameter sequence, if the optimization processing summary does not meet the optimization termination condition, adjusting the alternative air conditioner operation parameter sequence, inputting the adjusted alternative air conditioner operation parameter sequence into the first model again to output a corresponding new prediction result, and carrying out optimization processing again according to the new prediction result until the first air conditioner operation parameter sequence is obtained, wherein the first air conditioner operation parameter sequence is one alternative air conditioner operation parameter sequence in a plurality of alternative air conditioner operation parameter sequences corresponding to the first air conditioner operation parameter sequence when the optimization termination condition is met.
If the predicted value of the energy consumption of the air conditioning energy system in the prediction result obtained based on the preset alternative air conditioning operation parameter sequence is randomly generated and the predicted value is not ideal, the alternative air conditioning operation parameters in the alternative air conditioning operation parameter sequence need to be adjusted to output the prediction result again until the obtained predicted value of the energy consumption of the air conditioning system conforms to the ideal air conditioning operation parameter sequence, and the process is called optimization processing. The first air conditioner operation parameter sequence obtained through optimization is an air conditioner operation parameter sequence which accords with the air conditioner energy consumption predicted value and accords with ideal requirements, optimization of the air conditioner operation parameters can be achieved, air conditioner operation parameters in the optimized air conditioner operation parameter sequence are used for air conditioner system operation control, the energy consumption value of the air conditioner system can be reduced, and the energy saving effect is improved.
After an air conditioner energy consumption predicted value and an indoor temperature and humidity predicted value corresponding to each alternative air conditioner operation parameter in each alternative air conditioner operation parameter sequence are obtained, in optimization processing, calculation processing is carried out according to a preset loss function and an indoor temperature predicted value and an air conditioner energy consumption predicted value corresponding to a prediction result, and a loss value corresponding to each alternative air conditioner operation parameter sequence is obtained. And the loss value of each alternative air conditioner operation parameter sequence is the sum of the sub-loss values corresponding to each air conditioner operation parameter in each alternative air conditioner operation parameter sequence. Calculating the gradient of each alternative air conditioner operation parameter sequence according to the loss function, adjusting the multiple alternative air conditioner operation parameter sequences according to the gradient to obtain multiple new alternative air conditioner operation parameter sequences, inputting the new alternative air conditioner operation parameter sequences and first data into the first model, predicting again, outputting new prediction results, carrying out optimization processing again according to the new prediction results, calculating the loss value and the gradient of each corresponding alternative air conditioner operation parameter sequence, judging whether optimization termination conditions are met according to the loss value and the gradient of each alternative air conditioner operation parameter sequence, and if the optimization termination conditions are met, defining the multiple corresponding alternative air conditioner operation parameter sequences in the last optimization processing as target alternative air conditioner operation parameter sequences. And comparing the values of the loss values corresponding to the target alternative air conditioner operation parameter sequences, and determining that one alternative air conditioner operation parameter sequence with the minimum loss value in the target alternative air conditioner operation parameter sequences is the first air conditioner operation parameter sequence. If the optimization termination condition is not met, the alternative air conditioner operation parameter sequences are readjusted according to a gradient optimization algorithm (such as an Adam algorithm or a quasi-Newton method), then input into the first model again for re-prediction to output a new prediction result, when the optimization termination condition is met, a target alternative air conditioner operation parameter sequence is obtained according to the satisfaction of the optimization termination condition, and an alternative air conditioner operation parameter sequence with the minimum value is obtained from the target alternative air conditioner operation parameter sequence according to the corresponding loss value and serves as the first air conditioner operation parameter sequence.
It should be noted that the optimization termination condition may be set such that the calculated gradient value meets a preset threshold, or the number of times the optimization process is executed reaches a preset threshold, and the like. The specific termination condition may be set according to actual conditions, and is not limited in this embodiment.
The preset loss function is represented as loss = obj + pen, wherein obj represents an objective function, pen represents a penalty function, and the loss value loss can be obtained through calculation according to the preset loss function.
The objective function can be any continuous derivative function related to the temperature and humidity in the room in the predicted step number, the information of the number of people, the predicted value of the energy consumption of the air conditioner and the like. One of the representations of the objective function is:
Figure BDA0003927919060000191
wherein, K max Representing the predicted number of steps, T r (T + k) represents the predicted indoor temperature value in the T + k time step, T tg The representative preset target indoor temperature can be set manually or according to the preset requirement of human comfort, and if the representative preset target indoor temperature is set to be 26 ℃, the human temperature sensitivity is most comfortable. P (t + k) represents the predicted value of the energy consumption of the air conditioning system in the t + k time step, w 1 Representing a preset comfort weight coefficient, w 2 Representing a preset energy consumption weight coefficient of the air conditioning system. By a preset target indoor temperature T tg And the predicted value P (T + k) of the energy consumption of the air conditioning system output by the air conditioning system model and the predicted value T of the indoor temperature output by the space heat transfer model r And (t + k) calculating in the optimization model according to the objective function to obtain a corresponding objective function value.
The penalty function is expressed as:
Figure BDA0003927919060000192
wherein n is p Representing the number of constraints, c n，k Representing a constraint function, relu a linear rectification function, p n,k Representative constraint function c n,k The penalty factor of (2). The linear rectification function is expressed as:
Figure BDA0003927919060000193
the constraint condition means that the value of any available function such as the temperature and humidity in the room, the passenger flow person and the air conditioner energy consumption predicted value in the predicted step number is less than 0. Specific constraint conditions can be set according to actual conditions, and some constraint conditions are listed for reference. For example, setting 3 constraints, where the 1 st constraint is that the indoor temperature must not exceed the target temperature by +2 ℃ in any kth time step within the prediction step number range, the corresponding constraint function is obtained as:
Figure BDA0003927919060000201
the 2 nd constraint condition is that within any kth time step within the prediction step number range, the air conditioner operation parameter is not higher than a preset maximum threshold value, and then the corresponding constraint function is obtained as follows:
Figure BDA0003927919060000202
wherein u is max Is a preset maximum threshold. The 3 rd constraint condition is that the air conditioner operation parameter is not lower than a preset minimum threshold value in any kth time step within the prediction step number range, and then a corresponding constraint function is obtained as follows:
Figure BDA0003927919060000203
wherein u is min Is a preset minimum threshold.
C above 1,k 、c 2,k And c 3,k A constraint function called a constraint.
And performing constraint function calculation on the indoor temperature predicted value output by the space heat transfer model and other parameters in the first data set according to constraint conditions to obtain a constraint function value pen. And combining the calculated objective function value obj, and obtaining a corresponding loss value loss by using a preset loss function loss = obj + pen.
According to a preset target indoor temperature T tg And outputting an air conditioning system energy consumption predicted value P (T + k) and a corresponding indoor temperature predicted value T output by the space heat transfer model by the air conditioning system model r And (t + k) performing optimization processing, wherein in the optimization processing, when an optimization termination condition is met, according to a plurality of candidate air conditioner operation parameter sequences which meet the optimization termination condition and are subjected to sub-optimization processing, values of loss values corresponding to the candidate air conditioner operation parameter sequences are compared, a minimum loss value with the minimum value is screened out, and the candidate air conditioner operation parameter sequence corresponding to the minimum loss value is obtained and used as a first air conditioner operation parameter sequence. And the screened first air conditioner operation parameter sequence is a group of air conditioner operation parameter sequences which are obtained by optimization processing and have the smallest energy consumption prediction value of the air conditioner system and meet the target indoor temperature requirement.
For example, suppose that 3 candidate air conditioner operation parameter sequences are preset, which are:
Figure BDA0003927919060000204
according to the above, the alternative air conditioner operation parameter sequence U can be calculated 1 、U 2 、U 3 The corresponding prediction result can obtain the alternative air conditioner operation parameter sequence U according to the preset loss function and the prediction result 1 、U 2 、U 3 The corresponding loss values loss1, loss2 and loss3 and the loss function formula are used for calculating the partial derivatives, so that the alternative air conditioner operation parameter sequence U can be obtained 1 、U 2 、U 3 Grads gra1, gra2, and gra3. It is assumed that the optimization termination condition is set such that the gradient reaches a preset threshold GRA. Judging the alternative air conditioner operation parameter sequence U obtained in the current optimization process 1 、U 2 、U 3 Whether there are gradients GRA1, GRA2, and GRA3 that reach the preset threshold GRA, if there is no gradient GRA that reaches the preset threshold GRA,adjusting the values of the alternative air conditioner operation parameters in the alternative air conditioner operation parameter sequence through a gradient optimization algorithm (such as Adam algorithm, quasi-newton method, etc.), so as to form 3 new alternative air conditioner operation parameter sequences, which are respectively:
Figure BDA0003927919060000211
according to the above, the alternative air conditioner operation parameter sequence U can be calculated 4 、U 5 、U 6 The corresponding prediction result can obtain the alternative air conditioner operation parameter sequence U according to the preset loss function and the prediction result 4 、U 5 、U 6 The corresponding loss values loss4, loss5 and loss6 and the loss function formula are used for calculating the partial derivatives, so that the alternative air conditioner operation parameter sequence U can be obtained 4 、U 5 、U 6 Grads gra4, gra5, and gra6. It is assumed that the optimization termination condition is set such that the gradient reaches the preset threshold GRA. Judging the alternative air conditioner operation parameter sequence U obtained in the current optimization process 4 、U 5 、U 6 If the gradients GRA4, GRA5 and GRA6 reach the preset threshold GRA, continuously readjusting the values of the alternative air conditioner operation parameters in the alternative operation parameter sequence according to the gradient optimization algorithm, and repeating the process until the obtained gradients reach the preset threshold GRA, so that the optimization termination condition is met. If the gradient GRA4, GRA5 and GRA6 reach the preset threshold GRA, the optimization termination condition is met, the target alternative air conditioner operation parameter sequence meeting the optimization termination condition is screened out, and if the gradient GRA5 and GRA6 both reach the preset threshold GRA, the corresponding alternative air conditioner operation parameter sequence U is selected 5 And U 6 And selecting an air conditioner operation parameter sequence for the target. To target alternative operating parameter sequence U 5 And U 6 Comparing the corresponding loss values loss5 and loss6, screening out the minimum loss value with the minimum value, and determining the alternative air conditioner operation parameter sequence U corresponding to the minimum loss value loss5 if the loss value loss5 is the minimum loss value 5 Is a first air conditioner operating parameter sequence.
It should be noted that, the preset 3 candidate air conditioner operation parameter sequences in the foregoing embodiment are only examples, and when in actual application, the number of the preset candidate air conditioner operation parameter sequences subjected to the optimization processing may be set according to an actual situation.
It should be noted that, if the optimization termination condition is that the number of times of the optimization process reaches a preset threshold, the multiple candidate air conditioner operation parameter sequences corresponding to the last optimization process are used as target candidate air conditioner operation parameter sequences, and one candidate air conditioner operation parameter sequence with the smallest loss value is selected from the target candidate air conditioner operation parameter sequences as the first air conditioner operation parameter sequence.
The alternative air conditioner operation parameter sequences containing a plurality of future time steps are used as the minimum unit of optimization, so that the plurality of future time steps can be wholly predicted, the alternative air conditioner operation parameter sequences with the minimum overall loss value of the alternative air conditioner operation parameters corresponding to the plurality of future time steps are screened out and used as the optimal group of alternative air conditioner operation parameter sequences, optimization from the perspective of long-distance benefits is realized, and the long-distance energy-saving effect of actual operation control of a subsequent air conditioning system is improved.
And S104, taking the first air conditioner operation parameter in the first air conditioner operation parameter sequence as a second air conditioner operation parameter and outputting the second air conditioner operation parameter, and controlling the operation of the air conditioning system at the next time step according to the second air conditioner operation parameter.
As can be known from the foregoing, the candidate air conditioner operation parameter sequence is a candidate air conditioner operation parameter for a plurality of time steps in the future, so the first air conditioner operation parameter sequence obtained after the optimization processing is also a candidate air conditioner operation parameter for a plurality of time steps in the future. For example, the first air conditioner operation parameter sequence obtained after the optimization processing in step S103 is:
Figure BDA0003927919060000221
then u is 5 (t + 1) is the alternative air conditioner operation parameter for the next time step,u 5 (t + 2) is the predicted alternative air conditioner operating parameter of the next two time steps, u 5 (t+K max ) Is the following K max . The first air conditioner operation parameter sequence obtained through the optimization processing can be regarded as a group of air conditioner operation parameter sequences with the optimal energy-saving effect in the current optimization process. And outputting the first alternative air conditioner operation parameter in the first air conditioner operation parameter sequence as a second air conditioner operation parameter. And transmitting the second air conditioner operation parameter to an air conditioning system control end so that the air conditioning system control end can operate the air conditioning system according to the second air conditioner operation parameter. The method comprises the steps of predicting an air conditioner energy consumption value and an indoor temperature and humidity value through a first model to obtain a prediction result, carrying out optimization processing according to the prediction result, screening a first air conditioner operation parameter sequence with an optimal energy-saving effect, carrying out operation control on an air conditioning system with a next time step by using a second air conditioner operation parameter which is a first candidate air conditioner operation parameter representing the next time step in the first air conditioner operation parameter sequence with the optimal energy-saving effect, achieving optimization adjustment of the air conditioner operation parameters, reducing an energy consumption prediction value of the air conditioning system on the premise of meeting target indoor temperature and humidity requirements, and enabling the energy consumption value of the air conditioning system to be reduced in actual operation when the air conditioning system is subsequently controlled according to the second air conditioner operation parameter obtained through optimization processing, so that comfort of passengers is guaranteed, energy consumption of the air conditioning system is reduced as much as possible, and an energy-saving effect is improved.
In an embodiment, the second air conditioner operation parameter output in the implementation process is issued to each air conditioner or the bottom controller through the command issuing module, so that each air conditioner or the bottom controller controls the air conditioning system to control according to the second air conditioner operation parameter. For example, the cold machine water outlet temperature parameter in the second air conditioner operation parameter is sent to the cold machine, and the chilled water flow parameter is sent to the chilled water pump frequency regulator. And repeating the steps at the next time step to update and obtain the corresponding second air conditioner operation parameter as the optimal air conditioner operation parameter, so that the adaptive updating of the air conditioner system operation parameter is realized, and the energy-saving effect of the air conditioner system is improved.
The optimal control effect can not be realized by the combination of the operation parameters of the air-conditioning equipment through the existing rule-based control method, but the technical scheme provided by the embodiment adopts an operation parameter optimization mode based on a preset model, and the control effect under different air-conditioning operation parameters can be predicted by utilizing an air-conditioning system model and a space heat transfer model, so that the optimal air-conditioning equipment operation parameter combination is determined.
The air conditioner operation parameters determined by the existing rule-based control method only consider the current operation state and requirements, and the technical scheme provided by the embodiment optimizes the air conditioner operation parameters of a plurality of future cycles. Through the technical scheme of this embodiment can reduce the total energy consumption of air conditioning system overall time, and then play energy-conserving technological effect.
According to the existing model prediction control method, a fixed parameter model is usually used, the calibration cannot be automatically updated along with the actual condition of the air conditioner operation, and the control precision is easily influenced by the misalignment of the model. The air conditioning system model, the spatial heat transfer model and the optimization model in the embodiment adopt a data-driven mode, namely, the model is continuously calibrated and updated by utilizing the actual operation data of the air conditioner, machine learning and a model identification algorithm, so that the adaptivity, the control precision and the rapid deployment capability of the air conditioning energy control are improved.
Most of the existing air-conditioning models are in the form of mechanism models or simulation models, and are discontinuous and non-conductive, so that when the air-conditioning models are used for optimization solution, an optimization method based on gradients cannot be utilized, and the solution efficiency is low. In the embodiment, the artificial neural network is used as the air conditioning system model and can be regarded as a continuous conductible function, so that a gradient-based optimization method can be used for the model, the solving efficiency is higher, the working efficiency of outputting the optimal air conditioning operation parameters is improved, and the working efficiency of energy-saving control of the air conditioning system is improved.
When each time step is finished, updating model parameters of a preset model according to a second data set to obtain a first model, inputting the first data and a plurality of preset alternative air conditioner operation parameter sequences into the first model, outputting corresponding prediction results, performing optimization processing according to the prediction results to obtain a first air conditioner operation parameter sequence, taking a first air conditioner operation parameter in the first air conditioner operation parameter sequence as a second air conditioner operation parameter and outputting the second air conditioner operation parameter, and controlling the operation of the next time step of the air conditioning system according to the second air conditioner operation parameter. By adopting the technical means, the first air conditioner operation parameter sequence with the optimal energy-saving effect is screened out by predicting and optimizing the plurality of preset alternative air conditioner operation parameters, the first air conditioner operation parameter in the first air conditioner operation parameter sequence is used as the second air conditioner operation parameter of the air conditioning system corresponding to the next time step length actual control, the optimization of the air conditioner operation parameters can be realized, and the energy-saving effect of the air conditioning system can be improved by controlling the air conditioning system through the optimized air conditioner operation parameters. In addition, model parameters are updated through the second data set, the adaptivity of the model can be improved, the model is more suitable for actual operation conditions, errors between air conditioner energy consumption predicted values corresponding to predicted second air conditioner operation parameters and actual air conditioner energy consumption values corresponding to actual control through the second air conditioner operation parameters can be reduced, the accuracy of prediction results is improved, and the energy-saving effect of an air conditioning system is further improved.
On the basis of the foregoing embodiments, fig. 3 is a schematic structural diagram of an energy-saving control device for an air conditioner according to an embodiment of the present application. Referring to fig. 3, the air conditioner energy saving control device provided in this embodiment specifically includes: a model parameter updating unit 21, a prediction unit 22, an optimization unit 23, and a control unit 24.
The model parameter updating unit 21 is configured to update the model parameters of the preset model according to a second data set when each time step is ended, so as to obtain a first model, where the second data set includes indoor temperature and humidity, outdoor temperature and humidity, information of people number, actual air conditioner operation parameters, an energy consumption value of an air conditioning system, and an air conditioning cooling output quantity of the current time step;
the prediction unit 22 is configured to input a first data set and a plurality of preset candidate air conditioner operation parameter sequences into the first model, and output a corresponding prediction result, where the prediction result includes an indoor temperature and humidity prediction value and an air conditioner system energy consumption prediction value, the first data set includes an indoor temperature and humidity, an outdoor temperature and humidity and people number information of a current time step, and an outdoor temperature and humidity prediction value and people number information prediction value of a plurality of time steps in the future;
an optimizing unit 23, configured to perform optimization processing according to the prediction result to obtain a first air conditioner operation parameter sequence, adjust the candidate air conditioner operation parameter sequence if the optimization termination condition is not satisfied in the optimization processing, input the adjusted candidate air conditioner operation parameter sequence into the first model again to output a corresponding new prediction result, and perform optimization processing again according to the new prediction result until the first air conditioner operation parameter sequence is obtained, where the first air conditioner operation parameter sequence is one of the multiple candidate air conditioner operation parameter sequences corresponding to the first air conditioner operation parameter sequence when the optimization termination condition is satisfied, and the first air conditioner operation parameter sequence is one of the multiple candidate air conditioner operation parameter sequences corresponding to the first air conditioner operation parameter sequence when the optimization termination condition is satisfied
And the control unit 24 is configured to take the first air conditioner operation parameter in the first air conditioner operation parameter sequence as a second air conditioner operation parameter and output the second air conditioner operation parameter, and control the operation of the air conditioning system at the next time step according to the second air conditioner operation parameter.
Further, the prediction unit 22 is further configured to input the indoor temperature and humidity of the current time step, the outdoor temperature and humidity, the number of people information, the predicted values of the outdoor temperature and humidity at a plurality of time steps in the future, the predicted values of the number of people information at a plurality of time steps in the future, and a plurality of preset candidate air conditioner operation parameter sequences into the first model for prediction processing;
and according to the prediction processing, obtaining an air conditioning system energy consumption prediction value, an air conditioning cold output prediction value and an indoor temperature and humidity prediction value of a plurality of time step lengths in the future corresponding to each preset alternative air conditioning operation parameter sequence.
Further, the first model comprises an air conditioning system model and a space heat transfer model;
the prediction unit 22 is further configured to input the indoor temperature and humidity of the current time step, the outdoor temperature and humidity, and a plurality of preset candidate air conditioner operation parameter sequences into an air conditioner system model, and perform a first prediction process to obtain a corresponding first air conditioner system energy consumption prediction value and a first air conditioner cooling output prediction value;
inputting the first air conditioner cold output predicted value and the corresponding outdoor temperature and humidity, the indoor temperature and humidity and the number information into a space heat transfer model, and performing second prediction processing to obtain a corresponding first indoor temperature and humidity predicted value;
inputting the first indoor temperature and humidity predicted value, the outdoor temperature and humidity predicted value and the number information predicted value of the next time step and the plurality of preset candidate air conditioner operation parameter sequences into the first model, and performing the first prediction processing and the second prediction processing to obtain a corresponding second air conditioner system energy consumption predicted value, a second air conditioner cold output predicted value and a second indoor temperature and humidity predicted value;
and repeating the first prediction processing and the second prediction processing until obtaining the energy consumption prediction value of the air conditioning system, the air conditioning cold output prediction value and the indoor temperature and humidity prediction value at a plurality of time steps in the future.
Further, the optimizing unit 23 is further configured to calculate a loss value corresponding to each candidate air conditioner operation parameter sequence according to a preset loss function and the prediction result if the optimization termination condition is not satisfied;
calculating the gradient of each alternative air conditioner operation parameter sequence according to the loss function, and adjusting the alternative air conditioner operation parameter sequences according to the gradient to obtain a plurality of new alternative air conditioner operation parameter sequences;
inputting the new candidate air conditioner operation parameter sequence and the first data set into a first model, and outputting a new prediction result;
and carrying out optimization processing according to the new prediction result until the first air conditioner operation parameter sequence is obtained.
Further, the prediction unit 22 is further configured to calculate, when the optimization termination condition is met, a loss value corresponding to each alternative air conditioner operation parameter sequence through a preset loss function according to a corresponding predicted value of indoor temperature and humidity and a predicted value of energy consumption of the air conditioning system, in combination with the corresponding multiple alternative air conditioner operation parameter sequences;
comparing the numerical value of the loss value corresponding to each alternative air conditioner operation sequence, and screening out the minimum loss value with the minimum numerical value;
and determining an alternative air conditioner operation parameter sequence corresponding to the minimum loss value as the first air conditioner operation parameter sequence.
Further, the air conditioning system model is expressed as:
(P(t+1),Q(t+1),Q lat (t+1))＝μ θ (u(t+1),T r (t),W r (t),T a (t),W a (t))，
wherein, mu θ Representing an artificial neural network, theta represents a model parameter set of the artificial neural network, P (t + 1) represents an air conditioning system energy consumption predicted value of t +1 time step, the air conditioning cold output predicted value comprises an air conditioning sensible heat and cold output predicted value and an air conditioning latent heat and cold output predicted value, Q (t + 1) represents an air conditioning sensible heat and cold output predicted value of t +1 time step, and Q (t + 1) represents an air conditioning sensible heat and cold output predicted value of t +1 time step lat (T + 1) represents the predicted value of the output of the latent heat and the cold of the air conditioner at the T +1 time step, u (T + 1) is the preset air conditioner operation parameter at the T +1 time step, and T r (t) represents the room temperature of the t-th time step, W r (T) represents the indoor humidity at the tth time step, T a (t) outdoor temperature, W, for the t-th time step a (t) represents the outdoor humidity for the t time step;
the prediction unit 22 is further configured to predict the outdoor temperature T in the T-th time step a (t) outdoor humidity W a (T) indoor temperature T r (t) indoor humidity W r (t) and alternative air conditioner operating parameter sequence U n Inputting the energy consumption predicted value P (t + 1), the sensible heat and cold output predicted value Q (t + 1) and the latent heat and cold output predicted value Q of the air conditioner into the air conditioner system model to obtain the t +1 time step length of the energy consumption predicted value P (t + 1), the sensible heat and cold output predicted value Q of the air conditioner lat (t+1)。
Further, the candidate air conditioner operation parameter sequence includes candidate air conditioner operation parameters of a plurality of time steps in the future, and the candidate air conditioner operation parameter sequence includes candidate air conditioner operation parameters of a plurality of time steps in the futureAir conditioner operation parameter sequence U n Expressed as:
Figure BDA0003927919060000251
wherein u is n (t) the alternative air conditioner operation parameter for the t time step in the nth alternative air conditioner operation parameter sequence, u n (t + 1) represents the alternative air conditioner operation parameter of t +1 time step in the nth alternative air conditioner operation parameter sequence, u n (t+K max ) Represents t + K in the nth candidate air conditioner operation parameter sequence max Alternative air conditioner operating parameters of individual time step, K max Is a positive integer.
Further, the spatial heat transfer model is represented as:
Figure BDA0003927919060000261
Figure BDA0003927919060000262
wherein, T r (t + 1) represents the predicted indoor temperature value of the t +1 th time step, W r (t + 1) represents the predicted indoor humidity value of the t +1 time step, N in (t) information on the number of persons entering in the t-th time step, N out (t) leaving people information representing the t-th time step, k 0 ～k 4 、b、j 0 ～j 4 Model parameters representing the spatial heat transfer model;
the prediction unit 22 is further configured to output a prediction value Q (t + 1) of sensible heat and cooling capacity of the air conditioner and an output value Q of latent heat and cooling capacity of the air conditioner at the tth time step lat (T + 1), outdoor temperature T a (t) outdoor humidity W a (T) indoor temperature T r (t) indoor humidity W r (t) number of persons entering information N in (t) and leaving population information N out (t) inputting into the space heat transfer model to obtain the t +1 time stepIndoor temperature predicted value T r (t + 1) and predicted indoor humidity value W r (t+1)。
Further, the preset loss function is represented by loss = obj + pen, where obj represents an objective function and pen represents a penalty function, where the objective function is represented by:
Figure BDA0003927919060000263
wherein, K max Representing the predicted number of steps, T r (T + k) represents the predicted indoor temperature value of the T + k time step, T tg Representing the preset target indoor temperature, P (t + k) representing the predicted value of the energy consumption of the air conditioning system at the t + k time step, W 1 Represents a preset comfort weight coefficient, w 2 Representing a preset energy consumption weight coefficient of the air conditioning system;
the penalty function is expressed as:
Figure BDA0003927919060000264
wherein n is p Representing the number of constraints, c n，k Representing a constraint function, relu a linear rectification function, p n,k Representative constraint function c n,k The penalty factor of (2).
Further, the model parameter updating unit 21 is further configured to, when each time step is finished, perform model parameter updating processing on the spatial heat transfer model according to the second data set corresponding to the time step, and determine a first parameter;
when the preset period is finished, carrying out model parameter updating processing on the air conditioning system model according to the second data set of all time step lengths in the period, determining a second parameter, and forming a period by a plurality of time step lengths;
and determining the model parameters of the first model according to the first parameters and/or the second parameters to obtain the first model.
The air conditioner energy-saving control device provided by the embodiment of the application can be used for executing the air conditioner energy-saving control method provided by the embodiment, and has corresponding functions and beneficial effects.
An embodiment of the present application provides an energy-saving control device for an air conditioner, and referring to fig. 4, the energy-saving control device for an air conditioner 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 energy-saving control device can be one or more, and the number of memories in the air-conditioning energy-saving control device can be one or more. The processor, the memory, the communication module, the input device and the output device of the air conditioner energy-saving control equipment can be connected through a bus or other modes.
The memory 32 is a computer-readable storage medium, and can be used for storing software programs, computer-executable programs, and modules, such as program instructions/modules corresponding to the air-conditioning energy saving control method according to any embodiment of the present application (for example, a model parameter updating unit, a prediction unit, an optimization unit, and a control unit in the air-conditioning energy saving control device). The memory can mainly comprise a program storage area and a data storage area, wherein the program storage area can store an operating system and an application program required by at least one function; the storage data area may store data created according to use of the device, and the like. Further, the memory may include high speed random access memory, and may also include non-volatile memory, such as at least one magnetic disk storage device, flash memory device, or other non-volatile solid state storage device. In some examples, the memory may further include memory remotely located from the processor, which may be connected to the device through a network. Examples of such networks include, but are not limited to, the internet, intranets, local area networks, mobile communication networks, and combinations 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 software programs, instructions and modules stored in the memory, so as to implement the air conditioner energy-saving control method.
The input device 34 may be used to receive entered numeric or character information and to generate key signal inputs relating to user settings and function controls of the apparatus. The output device 35 may include a display device such as a display screen.
The air conditioner energy-saving control device provided by the embodiment can be used for executing the air conditioner energy-saving control method provided by the embodiment, and has corresponding functions and beneficial effects.
Embodiments of the present application also provide a storage medium storing computer-executable instructions, which when executed by a computer processor, are configured to perform an air conditioner energy saving control method, including: when each time step is finished, updating model parameters of a preset model according to a second data set to obtain a first model, wherein the second data set comprises indoor temperature and humidity, outdoor temperature and humidity, information of people, actual air conditioner operation parameters, an energy consumption value of an air conditioning system and an air conditioning cold output quantity of the current time step; inputting a first data set and a plurality of preset alternative air conditioner operation parameter sequences into the first model, and outputting corresponding prediction results, wherein the prediction results comprise indoor temperature and humidity prediction values and air conditioner system energy consumption prediction values, the first data set comprises indoor temperature and humidity, outdoor temperature and humidity and people number information of the current time step, and outdoor temperature and humidity prediction values and people number information prediction values of a plurality of time steps in the future; performing optimization processing according to the prediction result to obtain a first air conditioner operation parameter sequence, if the optimization processing does not meet the optimization termination condition, adjusting the alternative air conditioner operation parameter sequence, inputting the adjusted alternative air conditioner operation parameter sequence into the first model again to output a corresponding new prediction result, and performing optimization processing again according to the new prediction result until the first air conditioner operation parameter sequence is obtained, wherein the first air conditioner operation parameter sequence is one of a plurality of alternative air conditioner operation parameter sequences corresponding to the first air conditioner operation parameter sequence when the optimization termination condition is met; and taking the first air conditioner operation parameter in the first air conditioner operation parameter sequence as a second air conditioner operation parameter and outputting the second air conditioner operation parameter, and controlling the operation of the air conditioning system at the next time step according to the second air conditioner operation parameter.
Storage medium-any of various types of memory devices or storage devices. The term "storage medium" is intended to include: mounting media such as CD-ROM, floppy disk, or tape devices; computer system memory or random access memory such as DRAM, DDR RAM, SRAM, EDO RAM, lanbas (Rambus) RAM, etc.; non-volatile memory, such as flash memory, magnetic media (e.g., hard disk or optical storage); registers or other similar types of memory elements, etc. The storage medium may also include other types of memory or combinations thereof. In addition, the storage medium may be located in a first computer system in which the program is executed, or may be located in a different second computer system connected to the first computer system through 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, e.g., in different computer systems connected by a network. The storage medium may store program instructions (e.g., embodied as a computer program) that are executable by one or more processors.
Of course, the storage medium storing the computer-executable instructions provided in the embodiments of the present application is not limited to the air-conditioning energy saving control method described above, and may also perform related operations in the air-conditioning energy saving control method provided in any embodiment of the present application.
The air-conditioning energy-saving control device, the storage medium and the air-conditioning energy-saving control device provided in the above embodiments may execute the air-conditioning energy-saving control method provided in any embodiment of the present application, and refer to the air-conditioning energy-saving control method provided in any embodiment of the present application without detailed technical details described in the above embodiments.
The foregoing is considered as illustrative of the preferred embodiments of the invention and the technical principles employed. The present application is not limited to the particular embodiments described herein, but is capable of various obvious changes, rearrangements and substitutions as will now become apparent to those skilled in the art without departing from the scope of the invention. Therefore, although the present application has been described in more detail with reference to the above embodiments, the present application is not limited to the above embodiments, and may include other equivalent embodiments without departing from the spirit of the present application, and the scope of the present application is determined by the scope of the claims.

Claims (13)

1. An energy-saving control method for an air conditioner is characterized by comprising the following steps:
when each time step is finished, updating model parameters of a preset model according to a second data set to obtain a first model, wherein the second data set comprises indoor temperature and humidity, outdoor temperature and humidity, information of people, actual air conditioner operation parameters, an energy consumption value of an air conditioning system and an air conditioning cold output quantity of the current time step;
inputting a first data set and a plurality of preset alternative air conditioner operation parameter sequences into the first model, and outputting corresponding prediction results, wherein the prediction results comprise indoor temperature and humidity prediction values and air conditioner system energy consumption prediction values, the first data set comprises indoor temperature and humidity, outdoor temperature and humidity and people number information of the current time step, and outdoor temperature and humidity prediction values and people number information prediction values of a plurality of time steps in the future;
carrying out optimization processing according to the prediction result to obtain a first air conditioner operation parameter sequence, if the optimization processing does not meet the optimization termination condition, adjusting the alternative air conditioner operation parameter sequence, inputting the adjusted alternative air conditioner operation parameter sequence into the first model again to output a corresponding new prediction result, and carrying out optimization processing again according to the new prediction result until the first air conditioner operation parameter sequence is obtained, wherein the first air conditioner operation parameter sequence is one of a plurality of corresponding alternative air conditioner operation parameter sequences when the optimization termination condition is met;
and taking the first air conditioner operation parameter in the first air conditioner operation parameter sequence as a second air conditioner operation parameter and outputting the second air conditioner operation parameter, and controlling the operation of the air conditioning system at the next time step according to the second air conditioner operation parameter.
2. The method of claim 1, wherein inputting the first data set and the plurality of preset candidate air conditioner operation parameter sequences into the first model and outputting corresponding prediction results comprises:
inputting the indoor temperature and humidity, the outdoor temperature and humidity, the number of people information, the outdoor temperature and humidity predicted values of a plurality of time steps in the future, the number of people information predicted values of a plurality of time steps in the future and a plurality of preset alternative air conditioner operation parameter sequences of the current time step into a first model for prediction processing;
and according to the prediction processing, obtaining an air conditioning system energy consumption prediction value, an air conditioning cold output prediction value and an indoor temperature and humidity prediction value of a plurality of time step lengths in the future corresponding to each preset alternative air conditioning operation parameter sequence.
3. The method of claim 2, wherein the first model comprises an air conditioning system model and a spatial heat transfer model;
the step of inputting a first data set and a plurality of preset alternative air conditioner operation parameter sequences into the first model and outputting corresponding prediction results comprises the following steps:
inputting the indoor temperature and humidity, the outdoor temperature and humidity and a plurality of preset alternative air conditioner operation parameter sequences of the current time step into an air conditioner system model, and performing first prediction processing to obtain a corresponding first air conditioner system energy consumption prediction value and a first air conditioner cold output prediction value;
inputting the first air conditioner cold output predicted value and the corresponding outdoor temperature and humidity, the indoor temperature and humidity and the number information into a space heat transfer model, and performing second prediction processing to obtain a corresponding first indoor temperature and humidity predicted value;
inputting the first indoor temperature and humidity predicted value, the outdoor temperature and humidity predicted value and the number information predicted value of the next time step and the plurality of preset candidate air conditioner operation parameter sequences into the first model, and performing the first prediction processing and the second prediction processing to obtain a corresponding second air conditioner system energy consumption predicted value, a second air conditioner cold output predicted value and a second indoor temperature and humidity predicted value;
and repeating the first prediction processing and the second prediction processing until the predicted values of the energy consumption of the air conditioning system, the output predicted value of the cooling capacity of the air conditioner and the predicted values of the indoor temperature and the indoor humidity of a plurality of time steps in the future are obtained.
4. The method of claim 1, wherein if the optimization termination condition is not satisfied during the optimization process, the candidate air conditioner operation parameter sequence is adjusted and then input into the first model again to output a corresponding new prediction result, and the optimization process is performed again according to the new prediction result until the first air conditioner operation parameter sequence is obtained, including:
if the optimization termination condition is not met, calculating a loss value corresponding to each alternative air conditioner operation parameter sequence according to a preset loss function and the prediction result;
calculating the gradient of each alternative air conditioner operation parameter sequence according to the loss function, and adjusting the alternative air conditioner operation parameter sequences according to the gradient to obtain a plurality of new alternative air conditioner operation parameter sequences;
inputting the new alternative air conditioner operation parameter sequence and the first data set into a first model, and outputting a new prediction result;
and carrying out optimization processing according to the new prediction result until the first air conditioner operation parameter sequence is obtained.
5. The method of claim 4, wherein the performing optimization processing according to the prediction result to obtain a first air conditioner operation parameter sequence comprises:
when the optimization termination condition is met, calculating a loss value corresponding to each alternative air conditioner operation parameter sequence through a preset loss function according to the corresponding indoor temperature and humidity predicted value and the air conditioner system energy consumption predicted value and combining the corresponding multiple alternative air conditioner operation parameter sequences;
comparing the numerical value of the loss value corresponding to each alternative air conditioner operation sequence, and screening out the minimum loss value with the minimum numerical value;
and determining an alternative air conditioner operation parameter sequence corresponding to the minimum loss value as the first air conditioner operation parameter sequence.
6. The method of claim 3, wherein the air conditioning system model is represented as:
(P(t+1)，Q(t+1)，Q lat (t+1))＝μ θ (u(t+1)，T r (t)，W r (t)，T a (t)，W a (t))，
wherein, mu θ Representing an artificial neural network, theta represents a model parameter set of the artificial neural network, P (t + 1) represents an air conditioning system energy consumption predicted value of t +1 time step, the air conditioning cold output predicted value comprises an air conditioning sensible heat and cold output predicted value and an air conditioning latent heat and cold output predicted value, Q (t + 1) represents an air conditioning sensible heat and cold output predicted value of t +1 time step, and Q (t + 1) represents an air conditioning sensible heat and cold output predicted value of t +1 time step lat (T + 1) represents the predicted value of the output of the latent heat and the cold of the air conditioner at the T +1 time step, u (T + 1) is the preset air conditioner operation parameter at the T +1 time step, and T r (t) represents the indoor temperature, W, of the t-th time step r (T) represents the indoor humidity at the tth time step, T a (t) outdoor temperature, W, for the t-th time step a (t) represents the outdoor humidity for the t time step;
inputting the indoor temperature and humidity, the outdoor temperature and humidity and a plurality of preset alternative air conditioner operation parameter sequences of the current time step into an air conditioner system model, and performing first prediction processing to obtain a corresponding first air conditioner system energy consumption prediction value and a first air conditioner cooling output prediction value, wherein the method comprises the following steps:
the outdoor temperature T in the T time step a (t) outdoor humidity W a (T) indoor temperature T r (t) indoor humidity W r (t) and alternative air conditioner operating parameter sequence U n Inputting the predicted values into the air-conditioning system model to obtain a predicted value P (t + 1) of energy consumption of the air-conditioning system, a predicted value Q (t + 1) of sensible heat and cold output of the air conditioner and a predicted value Q of latent heat and cold output of the air conditioner at t +1 time step lat (t+1)。
7. The method of claim 6, wherein the sequence of alternative air conditioner operating parameters comprises alternative air conditioner operating parameters for a plurality of time steps in the future, and wherein the sequence of alternative air conditioner operating parameters U n Expressed as:
Figure FDA0003927919050000031
wherein u is n (t + 1) represents the alternative air conditioner operation parameter of t +1 time step in the nth alternative air conditioner operation parameter sequence, u n (t + 2) represents the alternative air conditioner operation parameter of t +2 time step in the nth alternative air conditioner operation parameter sequence, u n (t+K max ) Represents t + K in the nth candidate air conditioner operation parameter sequence max Alternative air conditioner operating parameters of individual time step, K max Is a positive integer.
8. The method of claim 6, wherein the spatial heat transfer model is represented as:
Figure FDA0003927919050000032
Figure FDA0003927919050000041
wherein, T r (t + 1) represents the predicted indoor temperature value of the t +1 th time step, W r (t + 1) represents the predicted value of the indoor humidity at the t +1 th time step, N in (t) information on the number of persons entering the t-th time step, N out (t) leaving people information representing the t-th time step, k 0 ～k 4 、b、j 0 ～j 4 Model parameters representing the spatial heat transfer model;
inputting the first air conditioner cold output prediction value and the corresponding outdoor temperature and humidity, the indoor temperature and humidity and the number information into a space heat transfer model, and performing second prediction processing to obtain a corresponding first indoor temperature and humidity prediction value, wherein the method comprises the following steps:
outputting a predicted value Q (t + 1) of sensible heat and cold output of the air conditioner and a predicted value Q of latent heat and cold output of the air conditioner in the tth time step lat (T + 1), outdoor temperature T a (t) outdoor humidity W a (T) indoor temperature T r (t) indoor humidity W r (t) number of persons entering information N in (t) and leaving population information N out (T) inputting the indoor temperature into the space heat transfer model to obtain an indoor temperature predicted value T of the T +1 time step r (t + 1) and predicted indoor humidity value W r (t+1)。
9. The method of claim 8, wherein the predetermined loss function is represented as loss = obj + pen, where obj represents an objective function and pen represents a penalty function, wherein the objective function is represented as:
Figure FDA0003927919050000042
wherein, K max Representing the predicted step number, T r (T + k) represents the predicted indoor temperature value of T + k time steps, T tg Representing the preset target indoor temperature, P (t + k) representing the predicted value of the energy consumption of the air conditioning system at the t + k time step, w 1 Representing a preset comfort weight coefficient, w 2 Representing a preset energy consumption weight coefficient of the air conditioning system;
the penalty function is expressed as:
Figure FDA0003927919050000043
wherein n is p Number of representing constraints, c n，k Representing a constraint function, relu a linear rectification function, p n，k Representing a constraint function c n，k The penalty factor of (2).
10. The method of claim 3, wherein updating the model parameters of the predetermined model according to the second data set at the end of each time step to obtain the first model comprises:
when each time step is finished, carrying out model parameter updating processing on the space heat transfer model according to the second data set corresponding to the time step to determine a first parameter;
when a preset period is finished, carrying out model parameter updating processing on the air conditioning system model according to a second data set of all time step lengths in the period to determine a second parameter, wherein a plurality of time step lengths form a period;
and determining the model parameters of the first model according to the first parameters and/or the second parameters to obtain the first model.
11. An energy-saving control device for an air conditioner is characterized by comprising:
the model parameter updating unit is used for updating model parameters of a preset model according to a second data set when each time step is finished to obtain a first model, wherein the second data set comprises indoor temperature and humidity, outdoor temperature and humidity, people number information, actual air conditioner operation parameters, an air conditioner system energy consumption value and an air conditioner cooling capacity output quantity of the current time step;
the prediction unit is used for inputting a first data set and a plurality of preset candidate air conditioner operation parameter sequences into the first model and outputting corresponding prediction results, wherein the prediction results comprise indoor temperature and humidity prediction values and air conditioner system energy consumption prediction values, the first data set comprises indoor temperature and humidity, outdoor temperature and humidity and people number information of the current time step, and outdoor temperature and humidity prediction values and people number information prediction values of a plurality of time steps in the future;
the optimization unit is used for carrying out optimization processing according to the prediction result to obtain a first air conditioner operation parameter sequence, adjusting the alternative air conditioner operation parameter sequence and inputting the adjusted alternative air conditioner operation parameter sequence into the first model again to output a corresponding new prediction result if the optimization processing does not meet the optimization termination condition, and carrying out optimization processing again according to the new prediction result until the first air conditioner operation parameter sequence is obtained, wherein the first air conditioner operation parameter sequence is one alternative air conditioner operation parameter sequence in a plurality of alternative air conditioner operation parameter sequences corresponding to the first air conditioner operation parameter sequence when the optimization termination condition is met;
and the control unit is used for taking the first air conditioner operation parameter in the first air conditioner operation parameter sequence as a second air conditioner operation parameter and outputting the second air conditioner operation parameter, and controlling the operation of the air conditioning system at the next time step length according to the second air conditioner operation parameter.
12. An energy-saving control device for an air conditioner, comprising:
a memory and one or more processors;
the memory to store one or more programs;
the one or more programs, when executed by the one or more processors, cause the one or more processors to perform the method recited in any of claims 1-10.
13. A storage medium storing computer-executable instructions, which when executed by a processor are configured to perform the method of any one of claims 1-10.
CN202211379920.XA 2022-10-08 2022-11-04 Air conditioner energy-saving control method, device, equipment and storage medium Pending CN115574437A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
PCT/CN2022/138833 WO2024073944A1 (en) 2022-10-08 2022-12-13 Air conditioner energy-saving control method, apparatus and device, and storage medium

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN2022112245502 2022-10-08
CN202211224550 2022-10-08

Publications (1)

Publication Number Publication Date
CN115574437A true CN115574437A (en) 2023-01-06

Family

ID=84589249

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202211379920.XA Pending CN115574437A (en) 2022-10-08 2022-11-04 Air conditioner energy-saving control method, device, equipment and storage medium

Country Status (2)

Country Link
CN (1) CN115574437A (en)
WO (1) WO2024073944A1 (en)

Family Cites Families (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN107062548B (en) * 2017-04-25 2019-10-15 天津大学 A kind of central air-conditioning varying load rate adjustment control method based on argument sequence
KR102212663B1 (en) * 2018-05-22 2021-02-05 주식회사 석영시스템즈 An apparatus for hvac system input power control based on target temperature and method thereof
WO2020246631A1 (en) * 2019-06-04 2020-12-10 엘지전자 주식회사 Temperature prediction model generation device and simulation environment provision method
CN111735178A (en) * 2020-07-15 2020-10-02 西安木牛能源技术服务有限公司 Air conditioner energy-saving system based on elite meaning and SVR regression algorithm and optimization method
CN114398939A (en) * 2021-12-06 2022-04-26 深圳市英维克信息技术有限公司 Method, device and equipment for detecting system fault and storage medium

Also Published As

Publication number Publication date
WO2024073944A1 (en) 2024-04-11

Similar Documents

Publication Publication Date Title
CN107120782B (en) A kind of HVAC system control method based on multi-user's hot comfort data
CN111684370A (en) System and method for controlling operation
CN110298487A (en) It is a kind of for meeting the room temperature prediction technique of users ' individualized requirement
CN111795484A (en) Intelligent air conditioner control method and system
CN102844773B (en) Based on building energy-saving unit and the system of interactive learning
CN113112077B (en) HVAC control system based on multi-step prediction deep reinforcement learning algorithm
CN114370698B (en) Indoor thermal environment learning efficiency improvement optimization control method based on reinforcement learning
CN112963946B (en) Heating, ventilating and air conditioning system control method and device for shared office area
CN111580382A (en) Unit-level heat supply adjusting method and system based on artificial intelligence
CN111609534A (en) Temperature control method and device and central temperature control system
CN114923265B (en) Central air conditioning energy-saving control system based on Internet of things
CN111561734A (en) Heating household valve adjusting method, system and equipment based on artificial intelligence
JP2020060331A (en) Air conditioner, data transmission method and air conditioning system
CN115220351B (en) Intelligent energy-saving optimization control method for building air conditioning system based on cloud side end
CN110377084A (en) A kind of Building Indoor Environment regulation method based on wisdom control strategy
CN111788435A (en) Composite control device and method in air conditioning system
CN107576015A (en) A kind of building air conditioning model predictive control method and device for realizing Demand Side Response
CN113028610B (en) Method and device for global optimization and energy-saving control of dynamic load of central air conditioner
Zhou et al. A Comprehensive Review of the Applications of Machine Learning for HVAC
CN116907036A (en) Deep reinforcement learning water chilling unit control method based on cold load prediction
CN115574437A (en) Air conditioner energy-saving control method, device, equipment and storage medium
CN114200839B (en) Intelligent office building energy consumption control model for dynamic monitoring of coupling environment behaviors
CN114565167B (en) Dynamic load prediction and regulation method for thermal inlet
CN116307074A (en) Method for acquiring real thermal data, system and method for constructing neural network model of thermal data, and method for predicting thermal load
CN113825955A (en) Method, apparatus, system, storage medium, and processor for determining a temperature setting

Legal Events

Date Code Title Description
PB01 Publication
PB01 Publication
SE01 Entry into force of request for substantive examination
SE01 Entry into force of request for substantive examination