Air conditioning apparatus, control method thereof, and computer readable storage medium

Abstract

The application discloses an air conditioning device and a control method and a computer readable storage medium thereof, and relates to the technical field of air conditioning devices. The control method of the air conditioning device comprises the following steps: obtaining multiple candidate parameter values of a control parameter of the air conditioning device in a next operation cycle, and predicting control deviation values of each control target of the air conditioning device under each candidate parameter value of the control parameter based on a feedforward control strategy and a feedback control strategy; performing weighted fusion processing on the control deviation values of each control target under each candidate parameter value based on target weight values of the control targets, so as to obtain comprehensive deviation values of each candidate parameter value; taking a candidate parameter value with the minimum comprehensive deviation value among the candidate parameter values as a target parameter value, and controlling the air conditioning device to operate according to the target parameter value of the control parameter in the next operation cycle. The application can improve the control effect and the control adaptability of the air conditioning device simultaneously.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of air conditioning equipment, in particular to an air conditioning equipment, a control method thereof and a computer readable storage medium. BACKGROUND

[0002] At present, PID (Proportion Integral Differential) frequency conversion technology is usually used to control air conditioning equipment. However, this control technology has a single control target and is difficult to realize multi-target collaborative control, thereby affecting the control effect. Moreover, this control technology is prone to large response delay in the control process and is difficult to adapt to a dynamic and variable real environment, thereby causing poor control adaptability. SUMMARY

[0003] The main purpose of the present application is to provide an air conditioning equipment, a control method thereof and a computer readable storage medium, which aims to simultaneously improve the control effect and control adaptability of the air conditioning equipment.

[0004] To achieve the above purpose, the present application provides a control method of an air conditioning equipment, which comprises the following steps:

[0005] obtaining a plurality of candidate parameter values of a control parameter of the air conditioning equipment in a next operation cycle, and predicting, based on a feedforward control strategy and a feedback control strategy, a control deviation value of each control target of the air conditioning equipment under each of the candidate parameter values of the control parameter; the control target is used to represent a performance index that the air conditioning equipment needs to achieve, and the control deviation value is used to represent a degree to which the control target meets a set control target requirement value;

[0006] obtaining a target weight value of each control target under a current control requirement of the air conditioning equipment;

[0007] based on the target weight value of each control target, performing weighted fusion processing on the control deviation value of each control target under each of the candidate parameter values to obtain a comprehensive deviation value of each candidate parameter value;

[0008] taking a candidate parameter value with the minimum comprehensive deviation value among the candidate parameter values as a target parameter value, and controlling the air conditioning equipment to operate according to the target parameter value of the control parameter in the next operation cycle.

[0009] In an embodiment, the step of predicting, based on the feedforward control strategy and the feedback control strategy, the control deviation value of each control target of the air conditioning equipment under each of the candidate parameter values of the control parameter comprises:

[0010] A comprehensive prediction model that couples feedforward control strategy and feedback control strategy is obtained, as well as the current state data of the air conditioning equipment is obtained;

[0011] The current state data and the candidate parameter values ​​of the control parameters are input into the comprehensive prediction model to obtain the first predicted value of each control target under the candidate parameter values ​​of the control parameters;

[0012] The difference between the required value of each of the set control targets and the first predicted value of each of the control targets under each of the candidate parameter values ​​is calculated to obtain the control deviation value of each of the control targets under each of the candidate parameter values.

[0013] In one embodiment, the step of obtaining the integrated prediction model that couples the feedforward control strategy and the feedback control strategy includes:

[0014] The difference between the required value and the actual value of each control target in multiple historical operating cycles prior to the current operating cycle is obtained to obtain the correction loss of each control target in multiple historical operating cycles;

[0015] A training sample set is constructed based on the correction loss of each control target over multiple historical operating cycles;

[0016] Based on the training sample set, the pre-trained feedforward prediction model is retrained to generate the comprehensive prediction model.

[0017] In one embodiment, the step of retraining the pre-trained feedforward prediction model based on the training sample set to generate the comprehensive prediction model includes:

[0018] Freeze a predetermined number of consecutive network layers in the feedforward prediction model;

[0019] Based on the training sample set, transfer learning is performed on the unfrozen network layers in the feedforward prediction model to generate the comprehensive prediction model.

[0020] In one embodiment, the step of retraining the pre-trained feedforward prediction model based on the training sample set to generate the comprehensive prediction model further includes:

[0021] Based on the training sample set, some network parameters in the feedforward prediction model are fine-tuned to generate the comprehensive prediction model.

[0022] In one embodiment, for the environmental targets among the control objectives, the integrated prediction model includes a multi-task expert network model;

[0023] The step of inputting the current state data and the candidate parameter values ​​of the control parameters into the comprehensive prediction model to obtain a first predicted value for each control target under the candidate parameter values ​​of the control parameters includes:

[0024] For any of the candidate parameter values ​​of the control parameters, the current state data and the candidate parameter value are input into the multi-task expert network model;

[0025] Based on each expert network in the multi-task expert network model, feature extraction processing is performed on the current state data and the candidate parameter values ​​to obtain the data processed by each expert network; each expert network includes a temperature expert network, a humidity expert network, an air volume expert network, and a public network.

[0026] Based on the gating structure in the multi-task expert network model, a self-balancing training mechanism is used to fuse the data processed by each expert network to obtain fused data.

[0027] The fused data is input into each discriminant network in the multi-task expert network model to obtain temperature prediction, humidity prediction, and air volume prediction, which together serve as the first predicted value of the environmental target under the candidate parameter values.

[0028] In one embodiment, for the energy-saving target among the control objectives, the comprehensive prediction model includes a sensible heat ratio prediction model and an energy efficiency prediction model;

[0029] The step of inputting the current state data and the candidate parameter values ​​of the control parameters into the comprehensive prediction model to obtain a first predicted value for each control target under the candidate parameter values ​​of the control parameters includes:

[0030] For any of the candidate parameter values ​​of the control parameters, the current state data and the candidate parameter value are input into the input layer of the sensible heat ratio prediction model;

[0031] Based on the processing of the feature extraction layer and temperature and humidity energy distribution layer of the sensible heat ratio prediction model, the first sensible heat ratio of the next operating cycle and the second sensible heat ratio of the current operating cycle are obtained.

[0032] In the output layer of the sensible heat ratio prediction model, the first sensible heat ratio and the second sensible heat ratio are constrained to obtain the predicted sensible heat ratio for the next operating cycle.

[0033] Furthermore, the current state data and the candidate parameter values ​​are input into the energy efficiency prediction model;

[0034] Based on the processing of the low-dimensional feature extraction layer and the high-dimensional feature mapping layer of the energy efficiency prediction model, a high-dimensional feature matrix is ​​obtained.

[0035] In the energy efficiency prediction layer of the energy efficiency prediction model, the high-dimensional feature matrix is ​​constrained to obtain the predicted energy efficiency.

[0036] The predicted sensible heat ratio and the predicted energy efficiency are used together as the first predicted value of the energy-saving target under the candidate parameter values.

[0037] In one embodiment, for the comfort objective among the control objectives, the integrated prediction model includes a comfort parameter prediction model;

[0038] The step of inputting the current state data and the candidate parameter values ​​of the control parameters into the comprehensive prediction model to obtain a first predicted value for each control target under the candidate parameter values ​​of the control parameters includes:

[0039] For any of the candidate parameter values ​​of the control parameters, the current state data and the candidate parameter value are input into the comfort parameter prediction model to obtain the predicted human body parameter value and the predicted thermal comfort value; wherein, the comfort parameter prediction model includes a first model trained with the distribution dispersion of human body parameters before and after the switching of the air conditioning equipment's operating state as a constraint, and a second model trained with the difference in the absolute value of the thermal comfort value before and after the switching of the air conditioning equipment's operating state as a constraint.

[0040] The predicted values ​​of the human body parameters and the predicted values ​​of thermal comfort are used together as the first predicted value of the comfort target under the candidate parameter values.

[0041] In one embodiment, the step of predicting the control deviation value of each control target of the air conditioning device under each of the candidate parameter values ​​of the control parameters based on the feedforward control strategy and the feedback control strategy further includes:

[0042] Obtain the current status data of the air conditioning equipment;

[0043] The current state data and the candidate parameter values ​​of the control parameters are input into the feedforward prediction model associated with the feedforward control strategy to obtain the second predicted value of each control objective under each candidate parameter value;

[0044] Calculate the difference between the required value of each of the set control targets and the second predicted value of each of the control targets under each of the candidate parameter values ​​to obtain the first deviation value of each of the control targets under each of the candidate parameter values;

[0045] Obtain the third predicted value of each control objective under the current parameter value of the control parameter, as predicted by the feedback prediction model associated with the feedback control strategy in the previous running cycle.

[0046] The difference between the actual value of each control target in the current operating cycle and the third predicted value of each control target is calculated to obtain the second deviation value of each control target;

[0047] The first deviation value and the second deviation value of each control target under each candidate parameter value are weighted and fused to obtain the control deviation value of each control target under each candidate parameter value.

[0048] In one embodiment, the construction step of the feedback prediction model includes:

[0049] The difference between the demand value and the actual value of each control target in multiple historical operating cycles prior to the current operating cycle is obtained to obtain the correction loss of each control target in multiple historical operating cycles;

[0050] A training sample set is constructed based on the correction loss of each control target over multiple historical operating cycles;

[0051] Based on the training sample set, reinforcement learning is performed on the model to be trained to generate the feedback prediction model.

[0052] In one embodiment, the step of acquiring the current status data of the air conditioning device includes:

[0053] Obtain the current compressor energy of the air conditioning equipment in the current operating cycle, and the historical compressor energy of multiple historical operating cycles prior to the current operating cycle;

[0054] Based on the compressor's historical energy from the multiple historical operating cycles, the inertial energy transferred from each historical operating cycle to the current operating cycle is obtained;

[0055] Based on the inertial energy and the current energy of the compressor, the cumulative energy of the current operating cycle is obtained;

[0056] Based on the accumulated energy, the current status data of the air conditioning equipment is determined.

[0057] In one embodiment, the step of obtaining the inertial energy transferred from each historical operating cycle to the current operating cycle based on the compressor's historical energy from the plurality of historical operating cycles includes:

[0058] Obtain the energy transfer inertia coefficient of the compressor from the historical energy transfer to the current operating cycle for each historical operating cycle;

[0059] Based on the compressor's historical energy in each historical operating cycle and the energy transfer inertia coefficient from the compressor's historical energy in each historical operating cycle to the current operating cycle, the inertial energy transferred from each historical operating cycle to the current operating cycle is calculated.

[0060] In one embodiment, the step of determining the current state data of the air conditioning device based on the accumulated energy includes:

[0061] Based on a preset mapping relationship between compressor energy and status data, the status data corresponding to the accumulated energy is obtained as the current status data.

[0062] In one embodiment, each of the control objectives includes an environmental objective, a comfort objective, and an energy-saving objective. The step of obtaining the target weight values ​​of each of the control objectives under the current control requirements of the air conditioning equipment includes:

[0063] The time interval between the current operating cycle and the moment when the user last set the required values ​​for each of the control targets is calculated, and the current electricity price is obtained.

[0064] Based on the time interval, the initial weight values ​​of the environmental target and the initial weight values ​​of the comfort target are determined; the initial weight values ​​of the environmental target are negatively correlated with the time interval, and the initial weight values ​​of the comfort target are positively correlated with the time interval.

[0065] Based on the electricity price, an initial weight value for the energy-saving target is determined; the initial weight value for the energy-saving target is positively correlated with the electricity price.

[0066] The initial weight values ​​of the environmental target, the comfort target, and the energy-saving target are normalized to obtain the target weight values ​​of the environmental target, the comfort target, and the energy-saving target.

[0067] In one embodiment, the step of obtaining the target weight values ​​of each of the control targets under the current control requirements of the air conditioning equipment further includes:

[0068] Based on the preset mapping relationship between operating functions and the weight values ​​of each control target, the weight values ​​of each control target corresponding to the current operating function of the air conditioning equipment are obtained and used as the target weight values ​​of each control target.

[0069] In one embodiment, the step of obtaining multiple candidate parameter values ​​for the control parameters of the air conditioning device includes:

[0070] Obtain the historical parameter values ​​of the control parameters in the previous operating cycle;

[0071] Based on the historical parameter values, multiple candidate parameter values ​​for the control parameters are determined.

[0072] In addition, to achieve the above objectives, this application also provides an air conditioning device, the air conditioning device comprising: a memory, a processor, and a computer program stored in the memory and executable on the processor, the computer program being configured to implement the steps of the control method for the air conditioning device as described above.

[0073] In addition, to achieve the above objectives, this application also provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the steps of the control method for the air conditioning device as described above.

[0074] In addition, to achieve the above objectives, this application also provides a computer program product, which includes a computer program that, when executed by a processor, implements the steps of the control method for the air conditioning device as described above.

[0075] This application provides a control method for an air conditioning device. The method involves acquiring multiple candidate parameter values ​​for the control parameters of the air conditioning device in the next operating cycle, and predicting the control deviation value of each control objective of the air conditioning device under each candidate parameter value based on a feedforward control strategy and a feedback control strategy. The control objective characterizes the performance indicators that the air conditioning device needs to achieve, and the control deviation value characterizes the degree to which the control objective meets the set control objective requirement. The method obtains the target weight value of each control objective under the current control requirements of the air conditioning device. Based on the target weight value of each control objective, the control deviation values ​​of each control objective under each candidate parameter value are weighted and fused to obtain a comprehensive deviation value for each candidate parameter value. The candidate parameter value with the smallest comprehensive deviation value among all candidate parameter values ​​is taken as the target parameter value, and the air conditioning device is controlled to operate according to the target parameter value of the control parameters in the next operating cycle.

[0076] Therefore, the technical solution provided in this application determines the target parameter values ​​of the air conditioning equipment's control parameters in the next operating cycle by simultaneously combining a feedforward control strategy, a feedback control strategy, and the target weight values ​​of each control objective under the current control requirements of the air conditioning equipment. This ensures that each control objective meets its respective control objective requirements as much as possible, thereby achieving multi-objective coordinated control of the air conditioning equipment with better control performance. Furthermore, because the technical solution provided in this application employs a feedforward control strategy with a priori effect in achieving multi-objective coordinated control, it can adapt to dynamically changing real environments.

[0077] In summary, the technical solution provided in this application can simultaneously improve the control effect and control adaptability of air conditioning equipment. Attached Figure Description

[0078] The accompanying drawings, which are incorporated in and form part of this specification, illustrate embodiments consistent with this application and, together with the description, serve to explain the principles of this application.

[0079] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, for those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0080] Figure 1 A schematic diagram illustrating the overall implementation principle of the control method for the air conditioning equipment provided in this application embodiment;

[0081] Figure 2 A schematic flowchart illustrating the control method for an air conditioning device provided in the first embodiment of this application;

[0082] Figure 3 This is a schematic diagram illustrating the working principle of the multi-task expert network model provided in the third embodiment of this application.

[0083] Figure 4 This is a schematic diagram illustrating the implementation principle of the sensible heat ratio prediction model provided in the fourth embodiment of this application.

[0084] Figure 5 A diagram illustrating the architecture of the comfort parameter prediction model provided in the fifth embodiment of this application;

[0085] Figure 6 A schematic diagram illustrating the principle of accumulating compressor energy across multiple operating cycles, provided in the seventh embodiment of this application;

[0086] Figure 7 This is a schematic diagram of the hardware operating environment involved in the embodiments of this application.

[0087] The purpose, features, and advantages of this application will be further explained in conjunction with the embodiments and with reference to the accompanying drawings. Detailed Implementation

[0088] It should be understood that the specific embodiments described herein are merely illustrative of the technical solutions of this application and are not intended to limit this application.

[0089] To better understand the technical solution of this application, a detailed description will be provided below in conjunction with the accompanying drawings and specific implementation methods.

[0090] Currently, PID variable frequency technology is commonly used to control air conditioning equipment. However, this control technology has a single control objective, making it difficult to achieve multi-objective coordinated control, thus affecting the control effect; furthermore, this control technology is prone to generating significant response delays during the control process, making it difficult to adapt to dynamically changing real environments, resulting in poor control adaptability.

[0091] Furthermore, this control technology is limited by its control granularity and mechanism, making it difficult to handle complex nonlinear systems and prone to significant control errors. This affects control accuracy and leads to phenomena such as overheating, underheating, and temperature oscillations in air conditioning equipment. Moreover, this control technology struggles to achieve good energy-saving and comfort effects.

[0092] Based on this, this application provides a control method for an air conditioning device. The method involves obtaining multiple candidate parameter values ​​for the control parameters of the air conditioning device in the next operating cycle, and predicting the control deviation value of each control objective of the air conditioning device under each candidate parameter value based on a feedforward control strategy and a feedback control strategy. The control objective characterizes the performance indicators that the air conditioning device needs to achieve, and the control deviation value characterizes the degree to which the control objective meets the set control objective requirement value. The method obtains the target weight value of each control objective under the current control requirements of the air conditioning device. Based on the target weight value of each control objective, the control deviation values ​​of each control objective under each candidate parameter value are weighted and fused to obtain a comprehensive deviation value for each candidate parameter value. The candidate parameter value with the smallest comprehensive deviation value among all candidate parameter values ​​is taken as the target parameter value, and the air conditioning device is controlled to operate according to the target parameter value of the control parameters in the next operating cycle.

[0093] Therefore, the technical solution provided in this application determines the target parameter values ​​of the air conditioning equipment's control parameters in the next operating cycle by simultaneously combining a feedforward control strategy, a feedback control strategy, and the target weight values ​​of each control objective under the current control requirements of the air conditioning equipment. This ensures that each control objective meets its respective control objective requirements as much as possible, thereby achieving multi-objective coordinated control of the air conditioning equipment with better control performance. Furthermore, because the technical solution provided in this application employs a feedforward control strategy with a priori effect in achieving multi-objective coordinated control, it can adapt to dynamically changing real environments.

[0094] In summary, the technical solution provided in this application can simultaneously improve the control effect and control adaptability of air conditioning equipment.

[0095] Furthermore, the technical solution provided in this application, in the process of predicting the predicted values ​​of each control target, not only considers the current energy of the compressor in the current operating cycle, but also the historical energy of the compressor in multiple historical operating cycles before the current operating cycle, and the inertial energy transferred to the current operating cycle, so as to effectively solve the impact on the control accuracy of the air conditioning equipment caused by the existence of energy transfer inertia, thereby improving the control accuracy of the air conditioning equipment.

[0096] The overall implementation principle of the control method for the air conditioning equipment mentioned above can be referred to... Figure 1 Specifically, this application constructs a feedforward control strategy by utilizing the predicted and demand values ​​of each control objective (i.e., multi-objective predicted and multi-objective demand values) to determine the prediction loss of each control objective; and by utilizing the demand and actual values ​​of each control objective (i.e., multi-objective demand and multi-objective actual values) to determine the correction loss of each control objective, thereby constructing a feedback control strategy. Therefore, under the control cycle of the air conditioning equipment, this application simultaneously utilizes both feedforward and feedback control strategies to analyze the predicted values ​​of each control objective, thereby rationally determining the target parameter values ​​of the control parameters to simultaneously improve the control effect and control adaptability of the air conditioning equipment. Furthermore, this application also considers the influence of energy transfer inertia on the predicted values ​​of each control objective to solve the control lag problem, thereby improving the control accuracy of the air conditioning equipment.

[0097] Among them, the actual value of the control objective is the value that the control objective actually achieves; the predicted value of the control objective is the predicted value that the control objective will theoretically achieve; and the required value of the control objective is the value that the control objective needs to achieve. Predictive loss is the difference between the required value and the predicted value of the control objective; and corrective loss is the difference between the required value and the actual value of the control objective.

[0098] The execution subject of the control method for the air conditioning equipment in this application can be the air conditioning equipment, or it can be a server, central controller, wired controller, or other equipment that controls other equipment to regulate the air. It can also be a control system or control circuit with corresponding functions. This embodiment does not specifically limit it in this regard.

[0099] The following description uses an air conditioning device as the main implementer to illustrate the various embodiments.

[0100] Based on this, this application proposes a control method for an air conditioning device according to a first embodiment. Please refer to [link / reference]. Figure 2 The control method for air conditioning equipment includes steps S10 to S40:

[0101] Step S10: Obtain multiple candidate parameter values ​​of the control parameters of the air conditioning equipment in the next operating cycle, and based on the feedforward control strategy and the feedback control strategy, predict the control deviation value of each control target of the air conditioning equipment under each candidate parameter value of the control parameters; the control target is used to characterize the performance index that the air conditioning equipment needs to achieve, and the control deviation value is used to characterize the degree to which the control target meets the set control target requirement value;

[0102] It should be noted that control parameters may include, but are not limited to, fan speed, compressor frequency, and / or expansion valve opening, etc., and this embodiment does not specifically limit them. Candidate parameter values ​​refer to the control parameters of the air conditioning equipment, which are parameter values ​​that can be selected to be achieved in the next operating cycle. Multiple control objectives of the air conditioning equipment may include environmental objectives, energy-saving objectives, and comfort objectives, etc., and may also include multiple indicators under environmental objectives (such as temperature, humidity, and air volume), multiple indicators under energy-saving objectives (such as capacity, energy efficiency, and energy consumption), and multiple indicators under comfort objectives (such as PMV (Predicted Mean Vote) and UIC (User's Individual Comfort)), and this embodiment does not specifically limit them.

[0103] Additionally, it should be noted that when obtaining multiple candidate parameter values ​​for the control parameters of the air conditioning equipment in the next operating cycle, the user can determine them based on experience; all set parameter values ​​can be directly used as candidate parameter values; or the control parameters can be determined based on historical parameter values ​​in the previous operating cycle. This embodiment does not impose specific limitations on these methods.

[0104] When determining multiple candidate parameter values ​​based on the historical parameter values ​​of the control parameters in the previous operating cycle, the parameter value range can be determined based on the historical parameter values, the preset maximum parameter change value, and the minimum parameter change value. Then, all or some parameter values ​​within this parameter value range can be used as candidate parameter values.

[0105] Step S20: Obtain the target weight values ​​of each control target under the current control requirements of the air conditioning equipment;

[0106] It should be noted that in this embodiment, the weight value of the control target under the current control requirements of the air conditioning equipment is referred to as the target weight value.

[0107] In one feasible implementation, where the control objectives include environmental objectives, comfort objectives, and energy-saving objectives, step S20 may include steps S21 to S24:

[0108] Step S21: Calculate the time interval between the current operating cycle and the moment when the user last set the required values ​​for each control target, and obtain the current electricity price;

[0109] Step S22: Determine the initial weight values ​​of the environmental target and the comfort target based on the time interval; the initial weight values ​​of the environmental target are negatively correlated with the time interval, and the initial weight values ​​of the comfort target are positively correlated with the time interval.

[0110] It should be noted that the initial weight values ​​of environmental objectives are negatively correlated with the time interval; that is, the longer the time interval, the smaller the initial weight value of the environmental objective, and vice versa. The initial weight values ​​of comfort objectives are positively correlated with the time interval; that is, the longer the time interval, the larger the initial weight value of the comfort objective, and vice versa.

[0111] When determining the initial weight values ​​of environmental targets and comfort targets based on time intervals, the weight values ​​of environmental targets corresponding to the time intervals can be obtained based on the preset mapping relationship between the time intervals and the weight values ​​of environmental targets, and used as the initial weight values ​​of environmental targets. Similarly, the weight values ​​of comfort targets corresponding to the time intervals can be obtained based on the preset mapping relationship between the time intervals and the weight values ​​of comfort targets, and used as the initial weight values ​​of comfort targets.

[0112] The mapping relationship between time intervals and the weight values ​​of environmental targets, as well as the mapping relationship between time intervals and the weight values ​​of comfort targets, can be recorded in the form of relational tables, relational curves, or key-value pairs. This embodiment does not impose specific limitations on this.

[0113] Step S23: Determine the initial weight value of the energy-saving target based on the electricity price; the initial weight value of the energy-saving target is positively correlated with the electricity price.

[0114] It should be noted that the initial weight of the energy-saving target is positively correlated with the electricity price; that is, the higher the electricity price, the greater the weight of the energy-saving target; and the lower the electricity price, the smaller the weight of the energy-saving target.

[0115] When determining the initial weight value of the energy-saving target based on the electricity price, the weight value of the energy-saving target corresponding to the electricity price can be obtained based on the preset mapping relationship between the electricity price and the weight value of the energy-saving target, and used as the initial weight value of the energy-saving target.

[0116] The mapping relationship between electricity price and the weight value of energy saving target can be recorded in the form of relational table, relational curve or key-value pair, and this embodiment does not make specific limitations on this.

[0117] Step S24: Normalize the initial weight values ​​of the environmental target, the comfort target, and the energy-saving target to obtain the target weight values ​​of the environmental target, the comfort target, and the energy-saving target.

[0118] It should be noted that when normalizing the initial weight values ​​of the environmental target, comfort target, and energy-saving target to obtain their target weight values, one can first calculate the sum of their initial weight values ​​to obtain the total weight value. Then, one can calculate the ratio of each initial weight value to the total weight value to obtain their respective target weight values.

[0119] By normalizing the initial weight values ​​of the environmental target, the comfort target, and the energy-saving target, the initial weight values ​​of the three can be normalized to the same data dimension. The sum of the target weight values ​​of the environmental target, the comfort target, and the energy-saving target will be 1.

[0120] It is understandable that, in the actual operation of air conditioning equipment, generally speaking, the longer the air conditioning equipment runs, the higher the user's focus on comfort goals and the lower their focus on environmental goals; furthermore, the higher the electricity price, the higher the user's focus on energy-saving goals. Based on this, this embodiment, by utilizing the time interval between the current operating cycle of the air conditioning equipment and the moment when the user last set the required values ​​for each control goal, as well as the current electricity price, can accurately and reasonably determine the target weight values ​​of environmental goals, comfort goals, and energy-saving goals under the current control requirements of the air conditioning equipment.

[0121] In another feasible implementation, step S20 may include step S25:

[0122] Step S25: Based on the preset mapping relationship between operating functions and the weight values ​​of each control target, obtain the weight values ​​of each control target corresponding to the current operating function of the air conditioning equipment, and use them as the target weight values ​​of each control target.

[0123] It should be noted that the mapping relationship between the operating function and the weight values ​​of each control objective can be recorded using relational tables, key-value pairs, etc., and this embodiment does not impose specific limitations on this. The current operating function is the function that the air conditioning equipment is currently operating. The operating function can be cooling, heating, dehumidifying, etc., and this embodiment does not impose specific limitations on this.

[0124] This embodiment utilizes the current operating functions of the air conditioning equipment to accurately and reasonably determine the target weight values ​​of each control objective under the current control requirements of the air conditioning equipment, in accordance with the actual operating conditions of the air conditioning equipment.

[0125] The above are only two feasible implementation methods of step S20 provided in this embodiment. This embodiment does not specifically limit the specific implementation method of step S20.

[0126] Step S30: Based on the target weight values ​​of each control objective, perform weighted fusion processing on the control deviation values ​​of each control objective under each candidate parameter value to obtain the comprehensive deviation value of each candidate parameter value;

[0127] It should be noted that, when performing weighted fusion processing on the control deviation values ​​of each control objective under each candidate parameter value based on the target weight values ​​of each control objective to obtain the comprehensive deviation value of each candidate parameter value, for any candidate parameter value, the product of the control deviation value of each control objective under that candidate parameter value and the target weight value of each control objective is calculated to obtain the weighted deviation value of each control objective; then, the sum of the weighted deviation values ​​of each control objective is calculated to obtain the comprehensive deviation value of that candidate parameter value. The specific implementation process can be referred to Formula 1 below.

[0128] Formula 1;

[0129] in, This is the overall deviation value of a candidate parameter. ~ The target weight values ​​for each control objective are: ~ Let be the control deviation value of each control objective under a candidate parameter value.

[0130] Step S40: Select the candidate parameter value with the smallest comprehensive deviation value among all candidate parameter values ​​as the target parameter value, and control the air conditioning equipment to operate according to the target parameter value of the control parameters in the next operating cycle.

[0131] It should be noted that the smaller the overall deviation of the candidate parameter values, the higher the degree to which the control parameters of the air conditioning equipment will meet the corresponding demand values ​​when operated according to the candidate parameter values.

[0132] As can be seen from the above, the technical solution provided in this embodiment determines the target parameter values ​​of the air conditioning equipment's control parameters in the next operating cycle by simultaneously combining a feedforward control strategy, a feedback control strategy, and the target weight values ​​of each control objective under the current control requirements of the air conditioning equipment. This ensures that each control objective meets its respective control objective requirements as much as possible, thereby achieving multi-objective coordinated control of the air conditioning equipment with better control performance. Furthermore, because the technical solution provided in this embodiment employs a feedforward control strategy with a priori effect in achieving multi-objective coordinated control, the technical solution provided in this embodiment can adapt to dynamically changing real environments.

[0133] In summary, the technical solution provided in this embodiment can simultaneously improve the control effect and control adaptability of air conditioning equipment.

[0134] Based on the first embodiment described above, a second embodiment of the control method for the air conditioning equipment of this application is proposed. In the second embodiment, step S10 may include steps S11 to S13:

[0135] Step S11: Obtain the integrated prediction model that couples the feedforward control strategy and the feedback control strategy, and obtain the current state data of the air conditioning equipment;

[0136] It should be noted that the current status data refers to the status data of the air conditioning equipment in the current operating cycle. This status data may include, but is not limited to, data reflecting the operating status of the air conditioning equipment, such as temperature, humidity, fan speed and / or compressor frequency. This embodiment does not make specific limitations on this.

[0137] In one feasible implementation, step S11 may include steps S111 to S113:

[0138] Step S111: Obtain the difference between the required value and the actual value of each control target in multiple historical operating cycles before the current operating cycle, and obtain the correction loss of each control target in multiple historical operating cycles;

[0139] Step S112: Construct a training sample set based on the correction loss of each control target in multiple historical operating cycles;

[0140] It should be noted that when constructing a training sample set based on the correction loss of each control target in multiple historical operating cycles, the state data of the air conditioning equipment and the parameter values ​​of the control parameters in the historical operating cycles corresponding to the correction loss can be used as input features, and the calibration loss can be used as the sample label to construct the training sample set.

[0141] Step S113: Based on the training sample set, retrain the pre-trained feedforward prediction model to generate a comprehensive prediction model.

[0142] It should be noted that when retraining the feedforward prediction model based on the training sample set, the training sample set can be divided into a training set and a validation set according to a certain ratio. Thus, the feedforward prediction model can be retrained using the training set, and the model can be verified to converge using the validation set. After verifying that the model has converged, the comprehensive prediction model can be generated.

[0143] In this embodiment, by obtaining the difference between the demand value and the actual value of each control target in multiple historical operating cycles before the current operating cycle, multiple losses of each control target under the feedback mechanism are determined, namely, correction losses. Then, a training sample set is constructed using the correction losses of each control target in multiple historical operating cycles. Afterwards, by using this training sample set to retrain the pre-trained feedforward prediction model, the feedback error information (i.e., correction losses) can be integrated into the feedforward prediction model. The final generated model can both retain the predictive ability of the feedforward prediction model for future states and use historical error experience to correct the prediction results, thereby forming a comprehensive prediction model that couples the feedforward control strategy and the feedback control strategy.

[0144] Furthermore, in a feasible implementation, step S113 may include steps S101~S102:

[0145] Step S101: Freeze a preset number of consecutive network layers in the feedforward prediction model;

[0146] Step S102: Based on the training sample set, perform transfer learning on the unfrozen network layers in the feedforward prediction model to generate a comprehensive prediction model.

[0147] It should be noted that the preset number can be a default value or can be flexibly set by the user according to the actual situation. This embodiment does not impose a specific limitation on this. When freezing a preset number of consecutive network layers in the feedforward prediction model, a preset number of consecutive network layers close to the input layer of the feedforward prediction model can be frozen, or other consecutive network layers in the feedforward prediction model can be frozen. This embodiment does not impose a specific limitation on this.

[0148] In this embodiment, a predetermined number of consecutive network layers in the feedforward prediction model are frozen to preserve the basic patterns learned by the model and prevent the loss of core capabilities due to subsequent model training. Then, using the training sample set, transfer learning is performed on the unfrozen network layers in the feedforward prediction model, enabling the model to learn feedback control strategies in a targeted manner. This allows for rapid learning of feedback control strategies without compromising the core capabilities (i.e., the feedforward control strategy). Thus, a comprehensive prediction model coupling the feedforward control strategy and the feedback control strategy is formed. It is understood that this "freeze + iterative training" mode significantly reduces the number of parameters that need to be updated in the model, lowering training costs. Furthermore, because the generalization basis of the feedforward prediction model is preserved during model training, overfitting of historical error data can be avoided, ensuring the predictive stability and accuracy of the trained comprehensive prediction model.

[0149] In another feasible implementation, step S113 may include step S103:

[0150] Step S103: Based on the training sample set, fine-tune some network parameters in the feedforward prediction model to generate a comprehensive prediction model.

[0151] It should be noted that when fine-tuning some network parameters in the feedforward prediction model based on the training sample set, the parameters (such as scaling factors, offset factors, and feature weights) of key network layers (such as output layers, layers near the output, and normalization layers) in the feedforward prediction model can be fine-tuned based on the training sample set. Specifically, fine-tuning of the parameters of key network layers can be achieved by inserting low-rank matrices into the key network layers of the feedforward prediction model.

[0152] In this embodiment, by utilizing the training sample set, some network parameters in the feedforward prediction model are fine-tuned, thus preserving the original core prediction capabilities of the feedforward prediction model while incorporating new error correction rules. This "local fine-tuning" approach allows the model to learn feedback control strategies in a targeted manner, thereby quickly learning feedback control strategies without compromising the core capabilities (i.e., the feedforward control strategy). This results in a comprehensive prediction model that couples the feedforward control strategy and the feedback control strategy.

[0153] The above are only two feasible implementations of step S113 provided in this embodiment. This embodiment does not specifically limit the specific implementation of step S113.

[0154] Step S12: Input the current state data and the candidate parameter values ​​of the control parameters into the comprehensive prediction model to obtain the first predicted value of each control target under the candidate parameter values ​​of the control parameters;

[0155] It should be noted that in this embodiment, the predicted value of each control objective under each candidate parameter value of the control parameter obtained by the comprehensive prediction model is referred to as the first predicted value.

[0156] Step S13: Calculate the difference between the required value of each set control target and the first predicted value of each control target under each candidate parameter value, and obtain the control deviation value of each control target under each candidate parameter value.

[0157] In this embodiment, a comprehensive prediction model that couples feedforward and feedback control strategies is used. This model combines the current state data of the air conditioning equipment with candidate parameter values ​​for each control parameter to accurately generate the first predicted value for each control objective under different candidate parameter values. This prediction mechanism leverages the feedforward control strategy to anticipate environmental changes and equipment operating trends, while incorporating historical operational deviation correction logic through the feedback control strategy. This effectively avoids the problems of susceptibility to unknown interference in single feedforward prediction or significant lag in simple feedback prediction, significantly improving the reliability of the prediction results. Based on this, by subtracting the first predicted value of each control objective from the first predicted value of each control objective under each candidate parameter value, the control deviation value of each control objective under each candidate parameter value can be accurately determined.

[0158] Based on the second embodiment described above, a third embodiment of the control method for the air conditioning equipment of this application is proposed. In the third embodiment, for the environmental target among the control objectives, the comprehensive prediction model includes a multi-task expert network model; step S12 may include steps S211~S214:

[0159] Step S211: For any candidate parameter value of the control parameter, input the current state data and the candidate parameter value into the multi-task expert network model;

[0160] It should be noted that the Multigate Mixture of Experts model can predict temperature, humidity, and airflow simultaneously.

[0161] Step S212: Based on each expert network in the multi-task expert network model, feature extraction processing is performed on the current state data and candidate parameter values ​​to obtain the data processed by each expert network; each expert network includes a temperature expert network, a humidity expert network, an air volume expert network, and a public network;

[0162] Step S213: Based on the gating structure in the multi-task expert network model, a self-balancing training mechanism is used to fuse the data processed by each expert network to obtain fused data.

[0163] It should be noted that self-balancing training is an adaptive training strategy that dynamically adjusts the weights of various components, tasks, or loss terms during the training of machine learning or deep learning models to achieve multi-objective optimization and alleviate data imbalance or component performance differences. The implementation principle of self-balancing training can be referred to in Equation 2 below.

[0164] Formula 2;

[0165] in, This represents the total loss function (used to measure the deviation between the model's predictions and the actual results). Let be the weight coefficient of the i-th expert network, which reflects the importance of each expert network in the final model decision. Let be the loss function of the i-th expert network (for example, the loss function of the temperature expert network can measure the difference between the temperature prediction result and the actual temperature value).

[0166] It is a conditional probability, representing the probability given a model. In the case where W is the model's parametric geometry and x is the input data, the output is: The probability of; and It is the loss term related to the model parameters W; and It is a hyperparameter, similar to the variance term, used to adjust the weight and influence of different loss terms.

[0167] Step S214: Input the fused data into each discriminant network in the multi-task expert network model to obtain the predicted values ​​of temperature, humidity, and air volume, which together serve as the first predicted value of the environmental target under the candidate parameter values.

[0168] It should be noted that each discrimination network may include a temperature discrimination network, a humidity discrimination network, and an airflow discrimination network. The implementation principle of steps S211-S214 in this embodiment can be referred to... Figure 3 .

[0169] Understandably, the existing coupling mechanisms limit the simultaneous prediction of the three targets of temperature, humidity, and airflow. Specifically, when air conditioning equipment lowers the temperature, the surface temperature of the evaporator decreases, and water vapor in the air condenses on the evaporator surface, resulting in a decrease in humidity. Thus, humidity changes are subject to temperature changes. Changes in airflow speed affect the rate of heat and moisture exchange, thereby causing changes in temperature and humidity. For example, high airflow speeds can cause rapid cooling of the air outlet, which in turn accelerates water vapor condensation and leads to a sharp drop in humidity.

[0170] To address this, this embodiment designs a multi-task expert network model that includes multiple expert networks, a gating structure employing a self-balancing training mechanism, and multiple discriminant networks. By utilizing this multi-task expert network model, combined with current state data and candidate parameter values, accurate predictions of temperature, humidity, and airflow can be achieved simultaneously. This allows for precise determination of the first predicted value of the environmental target under the corresponding candidate parameter values, thereby further improving the control effect and accuracy of air conditioning equipment.

[0171] Based on the second embodiment described above, a fourth embodiment of the control method for the air conditioning equipment of this application is proposed. In the fourth embodiment, for the energy-saving target among the control objectives, the comprehensive prediction model includes a sensible heat ratio prediction model and an energy efficiency prediction model; step S12 may include steps S221~S227:

[0172] Step S221: For any candidate parameter value of the control parameter, input the current state data and the candidate parameter value into the input layer of the sensible heat ratio prediction model;

[0173] Step S222: Based on the processing of the feature extraction layer and temperature and humidity energy distribution layer of the sensible heat ratio prediction model, the first sensible heat ratio of the next operating cycle and the second sensible heat ratio of the current operating cycle are obtained.

[0174] It should be noted that, in this embodiment, the sensible heat ratio of the next operating cycle and the sensible heat ratio of the current operating cycle obtained after processing by the feature extraction layer and the temperature and humidity energy distribution layer based on the sensible heat ratio prediction model are respectively referred to as the first sensible heat ratio and the second sensible heat ratio.

[0175] Step S223: In the output layer of the sensible heat ratio prediction model, constrain the first sensible heat ratio and the second sensible heat ratio to obtain the predicted sensible heat ratio for the next running cycle.

[0176] It should be noted that, in the output layer of the sensible heat ratio prediction model, when constraining the first and second sensible heat ratios, a physical trend loss function can be designed based on the correlation between the sensible heat ratio and temperature, humidity, airflow, and compressor frequency. This function can be used to constrain the direction of change of the sensible heat ratio between the two periods. Furthermore, a sensible heat ratio physical value loss function can be designed based on the energy conservation equation, which can be used to constrain the sensible heat ratio values ​​between the two periods. Based on this, the output layer of the sensible heat ratio prediction model can utilize this physical trend loss function and the sensible heat ratio physical value loss function to constrain the first and second sensible heat ratios.

[0177] Additionally, it should be noted that the sensible heat ratio prediction model can be a two-channel network, and its implementation principle can be found in [reference needed]. Figure 4Specifically, after inputting the current state data and candidate parameter values ​​into the input layer of the sensible heat ratio prediction model, the current state data will serve as the input data for the first channel, and the candidate parameter values ​​will serve as the input data for the second channel. After processing by the feature extraction layer and the temperature and humidity energy allocation layer of the sensible heat ratio prediction model, the first sensible heat ratio (SHR) for the next operating cycle can be obtained. i The second sensible heat ratio SHR in the current operating cycle i-1 Then, in the output layer of the sensible heat ratio prediction model, the first sensible heat ratio (SHR) is... i Second sensible heat ratio SHR i-1 By applying constraints, the predicted sensible heat ratio for the next operating cycle can be obtained.

[0178] In the case of a dual-channel network for the sensible heat ratio prediction model, the process of obtaining the first sensible heat ratio for the next operating cycle and the second sensible heat ratio for the current operating cycle based on the feature extraction layer and the temperature and humidity energy allocation layer of the sensible heat ratio prediction model can include: in the feature extraction layer, performing feature extraction processing on the input data of the first channel and the input data of the second channel to obtain processed first channel data and second channel data; in the temperature and humidity energy allocation layer, determining the first energy consumption caused by temperature change and the second energy consumption caused by humidity change in the next operating cycle based on the processed first channel data and second channel data, and determining the first sensible heat ratio for the next operating cycle and the second sensible heat ratio for the current operating cycle based on the first energy consumption and the second energy consumption, respectively.

[0179] It should be noted that the next operating cycle is adjacent to the current operating cycle, and the energy consumption caused by temperature changes and humidity changes is approximately the same for both. Therefore, by using the energy consumption caused by temperature changes and humidity changes in the next operating cycle, the sensible heat ratio of the next operating cycle and the sensible heat ratio of the current operating cycle can be determined simultaneously.

[0180] Additionally, it should be noted that the Gaussian elimination algorithm can be used to calculate the first energy consumption caused by temperature change and the second energy consumption caused by humidity change in the next operating cycle based on the temperature change, humidity change and energy consumption in the processed first channel data, as well as the temperature change, humidity change and energy consumption in the second channel data.

[0181] Step S224: Input the current state data and candidate parameter values ​​into the energy efficiency prediction model;

[0182] Step S225: Based on the processing of the low-dimensional feature extraction layer and the high-dimensional feature mapping layer of the energy efficiency prediction model, a high-dimensional feature matrix is ​​obtained;

[0183] It should be noted that in the processing of the low-dimensional feature extraction layer and high-dimensional feature mapping layer based on the energy efficiency prediction model to obtain the high-dimensional feature matrix, firstly, in the low-dimensional feature extraction layer, the time-series data in the current state data and candidate parameter values ​​can be transformed into a graph structure. Then, the "node embedding" technique is used to compress the information of each node in the graph structure into a low-dimensional feature vector to obtain each key low-dimensional feature. Next, in the high-dimensional feature mapping layer, each key low-dimensional feature is aggregated, and the mathematical operation of "star operation" is used to map the aggregated features into a high-dimensional nonlinear space to obtain the high-dimensional feature matrix.

[0184] Step S226: In the energy efficiency prediction layer of the energy efficiency prediction model, the high-dimensional feature matrix is ​​constrained to obtain the predicted energy efficiency.

[0185] It should be noted that, in the energy efficiency prediction layer of the energy efficiency prediction model, when constraining the high-dimensional feature matrix, physical equations can be designed based on physical constraints such as the law of conservation of energy (e.g., input electrical energy = cooling capacity + heat loss), and these equations can be embedded into the loss function as constraints. Therefore, the energy efficiency prediction layer of the energy efficiency prediction model can use these constraints embedded into the loss function to constrain the high-dimensional feature matrix.

[0186] Step S227: The predicted sensible heat ratio and the predicted energy efficiency are used together as the first predicted value of the energy-saving target under the candidate parameter values.

[0187] Understandably, there are two main paths to achieving energy-saving goals: sensible heat and energy efficiency. From the perspective of sensible heat, processing sensible heat consumes less energy than processing latent heat, making systems with high sensible heat capacity more energy-efficient. From the perspective of energy efficiency, the energy efficiency of air conditioning equipment refers to the ratio of cooling capacity to the input power of the air conditioning equipment. The higher the energy efficiency, the higher the cooling capacity produced per unit of power, and the lower the power required for the same cooling demand, making the equipment more energy-efficient.

[0188] This embodiment designs a sensible heat ratio prediction model and an energy efficiency prediction model, and combines current state data and candidate parameter values ​​to accurately determine the predicted sensible heat ratio and predicted energy efficiency of the air conditioning equipment under the corresponding candidate parameter values. This allows for the accurate determination of the first predicted value of the energy-saving target under the corresponding candidate parameter values, which in turn helps to further improve the control effect, control adaptability, and control accuracy of the air conditioning equipment.

[0189] Based on the second embodiment described above, a fifth embodiment of the control method for the air conditioning equipment of this application is proposed. In the fifth embodiment, for the comfort target among the control objectives, the comprehensive prediction model includes a comfort parameter prediction model; step S12 may include steps S231~S232:

[0190] Step S231: For any candidate parameter value of the control parameter, input the current state data and the candidate parameter value into the comfort parameter prediction model to obtain the predicted human body parameter value and the predicted thermal comfort value; wherein, the comfort parameter prediction model includes a first model trained with the distribution dispersion of human body parameters before and after the switching of the air conditioning equipment's operating state as a constraint, and a second model trained with the difference in the absolute value of the thermal comfort value before and after the switching of the air conditioning equipment's operating state as a constraint.

[0191] It's important to note that user anatomy parameters can include metabolic rate and clothing thermal resistance (i.e., clothing resistance). Metabolic rate primarily depends on information such as the user's gender, age, weight, and activity level. Among these, activity level is a short-term, variable characteristic with a periodic pattern; for example, the user might exercise between 5 PM and 6 PM and sleep between 10 PM and 8 AM. Clothing resistance, on the other hand, depends on clothing, bedding, and lifestyle. Lifestyle is a key characteristic with a periodic pattern; for example, the user might dress for work between 8 AM and 6 PM, do housework between 6 PM and 7 PM, and sleep under the covers between 10 PM and 8 AM. Thus, due to the periodic patterns of activity level and lifestyle, user behavior patterns exhibit a consistent distribution within the same time period. That is, within the same time period, the dispersion of user anatomy parameters before and after switching air conditioning operation is relatively small.

[0192] Additionally, it should be noted that thermal comfort values ​​are used to quantitatively assess the degree of thermal comfort of the human body in a specific thermal environment. It is understandable that users' air conditioning operation behavior has a monotonic improvement trend; when a user controls the air conditioning equipment to adjust from state A to state B, the user's thermal comfort should be such that they are more comfortable in state B than in state A. The first model can be the PMV model, and its specific construction and workflow can be referenced as follows:

[0193] First, a TCN (Temporal Convolutional Network) time series model is constructed using the status data of the air conditioning equipment and the correlation coefficient between its operating parameters. This TCN time series model is then used to predict the operating parameters. Next, based on this model, user behavior inertia and behavior optimization criteria are introduced to construct a dual-channel neural network. This dual-channel neural network is then used to predict human body parameters.

[0194] In addition, it should be noted that the second model can be a UIC model, which can be built based on user behavior. The specific construction process can be referred to as follows:

[0195] First, based on user adjustment behavior, a comfort triplet (<anchor sample, positive sample, negative sample>) can be constructed. Anchor samples can be understood as basic user behavior-thermal comfort state samples used for comparison; positive samples are samples that are similar to anchor samples in terms of thermal comfort experience and belong to the same type of comfort feeling; negative samples are samples that are significantly different from anchor samples in terms of thermal comfort experience and belong to the opposite type of feeling.

[0196] Then, the comfort triples can be input into DASN (Domain Adaptation Smoothing Network) to aggregate similar triples and isolate dissimilar triples. It can be understood that DASN processes comfort triples by using feature mapping and domain adaptation mechanisms, which allows similar comfort triples to be more tightly aggregated in the feature space, while isolating dissimilar triples from each other, thereby clearly distinguishing the sample distribution of different thermal comfort sensations.

[0197] Then, a triplet loss function is used to expand the metric distance between positive and negative examples in the feature space through metric learning (positive examples are anchor samples and positive samples, and negative examples are anchor samples and negative samples). That is, the features of similar thermal comfort samples are more aggregated, and the features of dissimilar samples are further apart. By continuously optimizing this distance relationship, the model learns to distinguish different thermal comfort states, and thus, a second model can be constructed.

[0198] To help understand the implementation principle of the comfort parameter prediction model, please refer to... Figure 5The diagram shows the architecture of a comfort parameter prediction model, which includes an input layer, a feature extraction layer, a data transformation layer, a PMV parameter layer, and an output layer. The input layer is the first layer of the comfort parameter prediction model. During model training, the input layer receives the model's output parameters, which include historical period data of the air conditioning equipment, state current data of the air conditioning equipment in state A, and state current data of the air conditioning equipment in state B. State A represents the air conditioning operating state before the user-instructed state switch within a first preset time period, and state B represents the air conditioning operating state after the user-instructed state switch within the first preset time period. Each state feature set includes nine features: return air temperature, condenser temperature, evaporator temperature, outdoor temperature, exhaust temperature, return air humidity, indoor fan speed, compressor frequency, and temperature difference between arrival and departure. The feature extraction layer uses an MLP (Multi-Layer Perceptron) network, denoted as the MLPExtracting Layer, to extract features from the air conditioning state data before and after the state switch and output human body parameters, including the metabolic rate under state A. Clothes and metabolic rate in state B Clothes The data transformation layer, represented as the Frozen CTCN Model, is used to determine thermal environment parameters based on historical state data, air conditioner state data in state A, and air conditioner state data in state B. These thermal environment parameters include the air temperature in state A. relative humidity convective heat transfer coefficient and mean radiation temperature and the air temperature in state B. relative humidity convective heat transfer coefficient and mean radiation temperature The PMV parameter layer is used to align human body parameters and thermal environment parameters under state A and state B, respectively, as inputs to the output layer; the output layer is used to calculate the thermal comfort values ​​under state A and state B based on the PMV value calculation formula. and In other words, it can be understood that the first and second models can be trained and applied as a whole, and the overall comfort parameter prediction model has the ability to predict human body parameters and thermal comfort values.

[0199] Step S232: Combine the predicted values ​​of human body parameters and thermal comfort with the predicted values ​​of thermal comfort as the first predicted value of the comfort target under the candidate parameter values.

[0200] This embodiment designs a comfort parameter prediction model, which includes a first model for predicting human body parameters and a second model for predicting thermal comfort. Therefore, by utilizing this comfort parameter prediction model and combining current state data with candidate parameter values, accurate predictions of human body parameters and thermal comfort values ​​can be obtained. Then, the predicted human body parameters reflecting thermal comfort from a physiological perspective, and the predicted thermal comfort values ​​reflecting thermal comfort from a user behavior perspective, are used together as the first predicted value of the comfort target under the candidate parameter values. This allows for accurate analysis of how well the comfort target can be achieved under the corresponding candidate parameter values, enabling diversified and differentiated comfort assessments, thereby further improving the control effect and accuracy of air conditioning equipment.

[0201] Based on the first embodiment described above, a sixth embodiment of the control method for the air conditioning equipment of this application is proposed. In the sixth embodiment, step S10 may include steps S14 to S19:

[0202] Step S14: Obtain the current status data of the air conditioning equipment;

[0203] Step S15: Input the current state data and the candidate parameter values ​​of the control parameters into the feedforward prediction model associated with the feedforward control strategy to obtain the second predicted value of each control objective under each candidate parameter value.

[0204] It should be noted that in this embodiment, the predicted value of each control objective obtained by the feedforward control strategy under each candidate parameter value is referred to as the second predicted value. The feedforward prediction model is a model pre-trained before the air conditioning equipment is deployed; the feedback prediction model is a model trained after the air conditioning equipment is deployed, based on the actual operating conditions of the air conditioning equipment.

[0205] Step S16: Calculate the difference between the required value of each set control target and the second predicted value of each control target under each candidate parameter value, and obtain the first deviation value of each control target under each candidate parameter value;

[0206] Step S17: Obtain the third predicted value of each control objective under the current parameter value of the control parameters, which was predicted by the feedback prediction model associated with the feedback control strategy in the previous running cycle.

[0207] It should be noted that in this embodiment, the predicted value of each control objective predicted by the feedback prediction model under the current parameter value of the control parameter is referred to as the third predicted value.

[0208] In one feasible implementation, the steps for constructing the feedback prediction model may include steps S171-S173:

[0209] Step S171: Obtain the difference between the required value and the actual value of each control target in multiple historical operating cycles before the current operating cycle, and obtain the correction loss of each control target in multiple historical operating cycles;

[0210] Step S172: Construct a training sample set based on the correction loss of each control target in multiple historical operating cycles;

[0211] Step S173: Based on the training sample set, perform reinforcement learning on the model to be trained to generate a feedback prediction model.

[0212] It should be noted that when performing reinforcement learning on the model to be trained based on the training sample set, the training sample set can be divided into a training set and a validation set according to a certain ratio. Thus, the training set is used to perform reinforcement learning on the model to be trained, and the validation set is used to verify whether the model after reinforcement learning has converged. After verifying that the model after reinforcement learning has converged, a feedback prediction model can be generated.

[0213] It should be noted that the model to be trained can be a neural network, a linear regression model, etc., and this embodiment does not make specific limitations on it.

[0214] This implementation method constructs a training sample set by utilizing the difference between the demand value and the actual value of each control target in multiple historical operating cycles before the current operating cycle, i.e. the correction loss of each control target in multiple historical operating cycles. Then, by directly using the training sample set, reinforcement learning is performed on the model to be trained to generate a feedback prediction model.

[0215] Step S18: Calculate the difference between the actual value of each control target in the current operating cycle and the third predicted value of each control target to obtain the second deviation value of each control target;

[0216] Step S19: Perform weighted fusion processing on the first deviation value and the second deviation value of each control target under each candidate parameter value to obtain the control deviation value of each control target under each candidate parameter value.

[0217] It should be noted that during the weighted fusion process of the first and second deviation values ​​of each control objective under each candidate parameter value, the weight values ​​of each control objective under the feedforward control strategy and the feedback control strategy can be a default value or can be flexibly determined according to the actual situation (for example, the weight values ​​of the feedforward control strategy and the feedback control strategy will change accordingly under different operating functions of the air conditioning equipment). This embodiment does not make specific limitations on this.

[0218] In this embodiment, by utilizing both the feedforward prediction model associated with the feedforward control strategy and the feedback prediction model associated with the feedback control strategy, the system can simultaneously predict environmental changes and equipment operating trends using the feedforward control strategy and incorporate historical operational deviation correction logic using the feedback control strategy. This effectively avoids the problems of susceptibility to unknown interference in single feedforward prediction or large lag in simple feedback prediction, significantly improving the reliability of the prediction results. Furthermore, by performing a weighted fusion process on the deviation values ​​determined by the feedforward control strategy and the feedback control strategy, the control deviation value of each control objective under each candidate parameter value can be accurately determined.

[0219] Based on the second, third, fourth, fifth, and / or sixth embodiments described above, a seventh embodiment of the control method for the air conditioning equipment of this application is proposed. In the seventh embodiment, the step of obtaining the current state data of the air conditioning equipment may include steps S01 to S04:

[0220] Step S01: Obtain the current energy of the compressor in the current operating cycle of the air conditioning equipment, and the historical energy of the compressor in multiple historical operating cycles prior to the current operating cycle;

[0221] It should be noted that the compressor's current energy refers to the energy of the compressor in the current operating cycle, while the compressor's historical energy refers to the energy of the compressor in its historical operating cycles. This energy can be related to compressor frequency, cooling capacity, heating capacity, energy consumption, power, etc., but this embodiment does not impose specific limitations on it. The amount of historical compressor energy obtained can be set according to the maximum number of operating cycles that the energy will lag behind.

[0222] Step S02: Based on the compressor's historical energy from multiple historical operating cycles, obtain the inertial energy transferred from each historical operating cycle to the current operating cycle;

[0223] It should be noted that inertial energy refers to the energy transferred from the compressor's historical operating cycle to the current operating cycle due to the influence of energy transfer inertia.

[0224] In one feasible implementation, step S02 may include steps S021-S022:

[0225] Step S021: Obtain the energy transfer inertia coefficient of the compressor in the historical energy transfer to the current operating cycle for each historical operating cycle;

[0226] It should be noted that the energy transfer inertia coefficient is used to quantify the hysteresis effect that exists in the energy transfer process.

[0227] In one feasible implementation, step S021 may include: calculating the time interval between the current operating cycle and each historical operating cycle; and determining the energy transfer inertia coefficient of the compressor's historical energy transfer to the current operating cycle for each historical operating cycle based on the time interval between the current operating cycle and each historical operating cycle.

[0228] This embodiment does not specifically limit the implementation of step S021. For example, in other feasible implementations, the time interval and operating parameters (such as compressor frequency, expansion valve angle, etc.) corresponding to each historical operating cycle can be considered simultaneously to determine the energy transfer inertia coefficient of the compressor's historical energy transfer to the current operating cycle for each historical operating cycle (i.e., a time interval and an operating parameter together correspond to an energy transfer inertia coefficient), so as to further ensure the accuracy of the energy transfer inertia coefficient corresponding to each historical operating cycle, thereby further improving the control accuracy of the air conditioning equipment.

[0229] In determining the energy transfer inertia coefficient of the compressor's historical energy transfer from the current operating cycle to the current operating cycle based on the time intervals between the current operating cycle and each historical operating cycle, the energy transfer inertia coefficient corresponding to each time interval can be found in a preset relationship table using the time intervals between the current operating cycle and each historical operating cycle as an index. This value is then used as the energy transfer inertia coefficient of the compressor's historical energy transfer from each historical operating cycle to the current operating cycle. Alternatively, the time intervals between the current operating cycle and each historical operating cycle can be input into a pre-trained energy transfer inertia model to obtain the energy transfer inertia coefficient of the compressor's historical energy transfer from each historical operating cycle to the current operating cycle. This embodiment does not impose specific limitations on this method.

[0230] It should be noted that the relationship table records the mapping relationship between time intervals and the energy transfer inertia coefficient. This relationship table can be constructed using the enthalpy difference measurement method or based on expert experience. When constructing the relationship table using the enthalpy difference measurement method, parameters such as compressor frequency, fan speed, and expansion valve angle can be fixed in a static enthalpy difference experiment. Then, a load generator is used to bring the indoor environment to a steady state, where energy transfer inertia disappears (i.e., the energy accumulated from historical operating cycles to the current operating cycle ≈ the energy transferred from the current operating cycle to future operating cycles). Based on this experiment, with the operating conditions and load remaining constant, the magnitude and time of the increase in cooling capacity are measured when the compressor frequency parameter is increased. Using the controlled variable method, based on multiple rounds of incremental difference experiments, the energy transfer inertia coefficient can be estimated.

[0231] Additionally, it should be noted that the energy transfer inertia model is used to predict the energy transfer inertia coefficient, and it can be constructed using a numerical modeling method. Specifically, in dynamic environment experimental measurements, the environmental and operational states of the current operating cycle and multiple historical operating cycles are used as independent variables, the energy transferred from the current operating cycle to the next W cycles is used as the dependent variable, and the cooling capacity obtained from the enthalpy difference experiment is used as the calibration to train the energy transfer inertia model so that the error between the cumulative cooling capacity of the current operating cycle and the cooling capacity obtained from the enthalpy difference experiment is minimized (i.e., the constructed loss function converges).

[0232] Step S022: Calculate the inertial energy transferred from each historical operating cycle to the current operating cycle based on the compressor's historical energy for each historical operating cycle and the energy transfer inertia coefficient from the compressor's historical energy for each historical operating cycle to the current operating cycle.

[0233] It should be noted that the inertial energy transferred from each historical operating cycle to the current operating cycle can be obtained by multiplying the compressor's historical energy for each historical operating cycle with its respective energy transfer inertia coefficient.

[0234] Step S03: Based on the inertial energy and the current energy of the compressor, obtain the cumulative energy of the current operating cycle;

[0235] It should be noted that accumulated energy is the energy accumulated in the current operating cycle by the inertial energy and the current energy of the compressor.

[0236] In one feasible implementation, step S03 may include: calculating the sum of each inertial energy and the current energy of the compressor to obtain the cumulative energy of the current operating cycle.

[0237] In another feasible implementation, step S03 may include: obtaining the energy transfer inertia coefficient corresponding to the current operating cycle; calculating the product of the compressor's current energy and the energy transfer inertia coefficient corresponding to the current operating cycle to obtain the target energy of the compressor's current energy transfer in the current operating cycle; and calculating the sum of each inertial energy and the target energy to obtain the cumulative energy of the current operating cycle.

[0238] It is understandable that, considering the inherent lag effect of energy transfer inertia, the current energy of the compressor in the current operating cycle may not be fully utilized due to the influence of energy transfer inertia. Therefore, in this embodiment, when obtaining the cumulative energy of the current operating cycle based on the inertial energies and the current energy of the compressor, it first uses the energy transfer inertia coefficient corresponding to the current operating cycle to determine the target energy of the compressor's current energy transfer in the current operating cycle; then, it integrates the inertial energies and the target energy to determine the cumulative energy of the current operating cycle. Therefore, this embodiment not only considers the impact of the compressor's historical energy in previous operating cycles on the current operating cycle under energy transfer inertia, but also considers the influence of energy transfer inertia on the current energy of the compressor in the current operating cycle, to more effectively address the impact of energy transfer inertia on the control accuracy of the air conditioning equipment, thereby further improving the control accuracy of the air conditioning equipment.

[0239] The above are only two feasible implementation methods of step S03 provided in this embodiment. This embodiment does not specifically limit the specific implementation method of step S03.

[0240] Furthermore, combining the implementation method of step S03 and step S022 of this embodiment, the resulting implementation flow can be expressed as the following formula 3. The implementation principle represented by formula 3 can be specifically referred to Figure 6 .

[0241] Formula 3;

[0242] in, The accumulated energy for the current operating cycle. Let be the energy transfer coefficient from the j-th running cycle to the i-th running cycle (i.e., the current running cycle). Let W be the compressor energy in the j-th operating cycle, and W be the maximum number of operating cycles that the energy will lag behind (the maximum number of operating cycles that the energy will lag behind includes one current operating cycle and multiple historical operating cycles, which can be a default value, such as 10, which can include 1 current operating cycle and 9 historical operating cycles. It can also be flexibly set by the user according to the actual situation, and this embodiment does not make specific limitations on this).

[0243] Step S04: Determine the current status data of the air conditioning equipment based on the accumulated energy.

[0244] In one feasible implementation, step S04 may include step S041:

[0245] Step S041: Based on the preset mapping relationship between compressor energy and status data, obtain the status data corresponding to the accumulated energy as the current status data.

[0246] It should be noted that the mapping relationship between compressor energy and status data can be recorded in the form of relational tables, key-value pairs, etc., and this embodiment does not make specific limitations on this.

[0247] As can be seen from the above, when determining the current state data of the air conditioning equipment, this embodiment not only considers the current energy of the compressor in the current operating cycle, but also the historical energy of the compressor in multiple previous operating cycles, as well as the inertial energy transferred to the current operating cycle. This effectively solves the impact of energy transfer inertia on the control accuracy of the air conditioning equipment, thereby accurately determining the actual energy status of the compressor in the current operating cycle and thus accurately determining the current state data of the air conditioning equipment, thereby improving the control accuracy of the air conditioning equipment.

[0248] This application also provides an air conditioning device, which includes: at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores instructions executable by the at least one processor, which are executed by the at least one processor to enable the at least one processor to perform the control method of the air conditioning device in the above embodiments.

[0249] The following is for reference. Figure 7 It shows a structural schematic diagram of an air conditioning device suitable for implementing the embodiments of this application. Figure 7 The air conditioning device shown is merely an example and should not be construed as limiting the functionality and scope of the embodiments of this application.

[0250] like Figure 7As shown, the air conditioning device may include a processing unit 101 (e.g., a central processing unit, a graphics processing unit, etc.), which can perform various appropriate actions and processes according to a program stored in a read-only memory 102 or a program loaded from a storage device 103 into a random access memory 104. The random access memory 104 also stores various programs and data required for the operation of the air conditioning device. The processing unit 101, the read-only memory 102, and the random access memory 104 are interconnected via a bus 105. An input / output interface 106 is also connected to the bus 105. Typically, the following systems can be connected to the input / output interface 106: input devices 107 including, for example, a touchscreen, touchpad, keyboard, mouse, image sensor, microphone, accelerometer, gyroscope, etc.; output devices 108 including, for example, a liquid crystal display (LCD), speaker, vibrator, etc.; storage devices 103 including, for example, magnetic tape, hard disk, etc.; and communication devices 109. The communication device 109 allows the air conditioning device to communicate wirelessly or wiredly with other devices to exchange data. Although the diagram shows air conditioning equipment with various systems, it should be understood that it is not required to implement or have all of the systems shown. More or fewer systems may be implemented alternatively.

[0251] In particular, according to embodiments of this disclosure, the processes described above with reference to the flowcharts can be implemented as computer software programs. For example, embodiments of this disclosure include a computer program product comprising a computer program carried on a computer-readable medium, the computer program containing program code for performing the methods shown in the flowcharts. In such embodiments, the computer program can be downloaded and installed from a network via a communication device, or installed from storage device 103, or installed from read-only memory 102. When the computer program is executed by processing device 101, it performs the functions defined in the methods of the embodiments of this application.

[0252] The air conditioning device provided in this application embodiment, employing the control method of the air conditioning device in the above embodiments, can simultaneously improve the control effect and control adaptability of the air conditioning device. Compared with the prior art, the beneficial effects of the air conditioning device provided in this application embodiment are the same as the beneficial effects of the control method of the air conditioning device provided in the above embodiments, and other technical features in this air conditioning device are the same as the features disclosed in the methods of the above embodiments, and will not be repeated here.

[0253] It should be understood that various parts of the embodiments of this application can be implemented using hardware, software, firmware, or a combination thereof. In the description of the above embodiments, specific features, structures, materials, or characteristics can be combined in any suitable manner in one or more embodiments or examples.

[0254] The above description is merely a specific implementation of the embodiments of this application, but the protection scope of the embodiments of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the protection scope of the embodiments of this application. Therefore, the protection scope of the embodiments of this application should be determined by the protection scope of the above claims.

[0255] This application also provides a computer-readable storage medium storing a computer program that can run on a processor. The computer program is used to execute the control method of the air conditioning device in the above embodiments.

[0256] The computer-readable storage medium provided in this application embodiment may be, for example, a USB flash drive, but is not limited to, electrical, magnetic, optical, electromagnetic, infrared, or semiconductor systems or devices, or any combination thereof. More specific examples of computer-readable storage media may include, but are not limited to: electrical connections with one or more wires, portable computer disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fibers, portable compact disk read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination thereof. In this embodiment, the computer-readable storage medium may be any tangible medium containing or storing a program that can be used by or in conjunction with an instruction execution system, system, or device. The program code contained on the computer-readable storage medium may be transmitted using any suitable medium, including but not limited to: wires, optical cables, RF (Radio Frequency), etc., or any suitable combination thereof.

[0257] The aforementioned computer-readable storage medium may be included in the air conditioning equipment; or it may exist independently and not be assembled into the air conditioning equipment.

[0258] The aforementioned computer-readable storage medium carries one or more programs. When the air conditioning device executes the aforementioned one or more programs, the air conditioning device causes the air conditioning device to: acquire multiple candidate parameter values ​​of the air conditioning device's control parameters in the next operating cycle, and predict the control deviation value of each control objective of the air conditioning device under each candidate parameter value of the control parameters based on a feedforward control strategy and a feedback control strategy; the control objective is used to characterize the performance index that the air conditioning device needs to achieve, and the control deviation value is used to characterize the degree to which the control objective meets the set control objective requirement value; acquire the target weight value of each control objective under the current control requirements of the air conditioning device; based on the target weight value of each control objective, perform weighted fusion processing on the control deviation values ​​of each control objective under each candidate parameter value to obtain a comprehensive deviation value for each candidate parameter value; and take the candidate parameter value with the smallest comprehensive deviation value among the candidate parameter values ​​as the target parameter value, and control the air conditioning device to operate according to the target parameter value of the control parameters in the next operating cycle.

[0259] Computer program code for performing the operations of this disclosure can be written in one or more programming languages ​​or a combination thereof, including object-oriented programming languages ​​such as Java, Smalltalk, and C++, and conventional procedural programming languages ​​such as the "C" language or similar programming languages. The program code can be executed entirely on the user's computer, partially on the user's computer, as a standalone software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In cases involving remote computers, the remote computer can be connected to the user's computer via any type of network—including a local area network (LAN) or a wide area network (WAN)—or can be connected to an external computer (e.g., via the Internet using an Internet service provider).

[0260] The flowcharts and block diagrams in the accompanying drawings illustrate the architecture, functionality, and operation of possible implementations of systems, methods, and computer program products according to various embodiments of this application. In this regard, each block in a flowchart or block diagram may represent a module, segment, or portion of code containing one or more executable instructions for implementing a specified logical function. It should also be noted that in some alternative implementations, the functions indicated in the blocks may occur in a different order than those indicated in the drawings. For example, two consecutively indicated blocks may actually be executed substantially in parallel, and they may sometimes be executed in reverse order, depending on the functions involved. It should also be noted that each block in the block diagrams and / or flowcharts, and combinations of blocks in the block diagrams and / or flowcharts, can be implemented using a dedicated hardware-based system that performs the specified function or operation, or using a combination of dedicated hardware and computer instructions.

[0261] The modules described in the embodiments of this application can be implemented in software or hardware. The names of the modules do not necessarily limit the functionality of the unit itself.

[0262] The computer-readable storage medium provided in this application embodiment stores computer-readable program instructions for executing the control method of the air conditioning equipment described above, which can simultaneously improve the control effect and control adaptability of the air conditioning equipment. Compared with the prior art, the beneficial effects of the computer-readable storage medium provided in this application embodiment are the same as the beneficial effects of the control method of the air conditioning equipment provided in the above embodiments, and will not be repeated here.

[0263] This application also provides a computer program product, including a computer program that, when executed by a processor, implements the steps of the control method for the air conditioning equipment as described above.

[0264] The computer program product provided in this application can simultaneously improve the control effect and control adaptability of air conditioning equipment. Compared with the prior art, the beneficial effects of the computer program product provided in this application are the same as the beneficial effects of the air conditioning equipment control method provided in the above embodiments, and will not be repeated here.

[0265] The above are merely preferred embodiments of this application and do not limit the patent scope of this application. Any equivalent structural or procedural transformations made using the content of this application's specification and drawings, or direct or indirect applications in other related technical fields, are similarly included within the patent scope of this application.

Claims

1. A control method for an air conditioning device, characterized in that, The method includes: The system acquires multiple candidate parameter values ​​for the control parameters of the air conditioning equipment in the next operating cycle, and predicts the control deviation value of each control target of the air conditioning equipment under each of the candidate parameter values ​​based on the feedforward control strategy and the feedback control strategy. The control target is used to characterize the performance index that the air conditioning equipment needs to achieve, and the control deviation value is used to characterize the degree to which the control target meets the set control target requirement value. Obtain the target weight values ​​of each of the control objectives under the current control requirements of the air conditioning equipment; Based on the target weight values ​​of each control objective, the control deviation values ​​of each control objective under each candidate parameter value are weighted and fused to obtain the comprehensive deviation value of each candidate parameter value. The candidate parameter value with the smallest comprehensive deviation value among all the candidate parameter values ​​is taken as the target parameter value, and the air conditioning equipment is controlled to operate according to the target parameter value of the control parameters in the next operating cycle.

2. The method as described in claim 1, characterized in that, The method of predicting the control deviation value of each control target of the air conditioning equipment under each of the candidate parameter values ​​of the control parameters based on the feedforward control strategy and the feedback control strategy includes: A comprehensive prediction model that couples feedforward control strategy and feedback control strategy is obtained, as well as the current state data of the air conditioning equipment is obtained; The current state data and the candidate parameter values ​​of the control parameters are input into the comprehensive prediction model to obtain the first predicted value of each control target under the candidate parameter values ​​of the control parameters; The difference between the required value of each of the set control targets and the first predicted value of each of the control targets under each of the candidate parameter values ​​is calculated to obtain the control deviation value of each of the control targets under each of the candidate parameter values.

3. The method as described in claim 2, characterized in that, The acquisition of the integrated prediction model that couples the feedforward control strategy and the feedback control strategy includes: The difference between the required value and the actual value of each control target in multiple historical operating cycles prior to the current operating cycle is obtained to obtain the correction loss of each control target in multiple historical operating cycles; A training sample set is constructed based on the correction loss of each control target over multiple historical operating cycles; Based on the training sample set, the pre-trained feedforward prediction model is retrained to generate the comprehensive prediction model.

4. The method as described in claim 3, characterized in that, The step of retraining the pre-trained feedforward prediction model based on the training sample set to generate the comprehensive prediction model includes: Freeze a predetermined number of consecutive network layers in the feedforward prediction model; Based on the training sample set, transfer learning is performed on the unfrozen network layers in the feedforward prediction model to generate the comprehensive prediction model.

5. The method as described in claim 3, characterized in that, The step of retraining the pre-trained feedforward prediction model based on the training sample set to generate the comprehensive prediction model further includes: Based on the training sample set, some network parameters in the feedforward prediction model are fine-tuned to generate the comprehensive prediction model.

6. The method as described in claim 2, characterized in that, For the environmental targets among the control objectives, the integrated prediction model includes a multi-task expert network model; The step of inputting the current state data and each of the candidate parameter values ​​of the control parameters into the comprehensive prediction model to obtain a first predicted value for each control target under each of the candidate parameter values ​​of the control parameters includes: For any of the candidate parameter values ​​of the control parameters, the current state data and the candidate parameter value are input into the multi-task expert network model; Based on each expert network in the multi-task expert network model, feature extraction processing is performed on the current state data and the candidate parameter values ​​to obtain the data processed by each expert network; each expert network includes a temperature expert network, a humidity expert network, an air volume expert network, and a public network. Based on the gating structure in the multi-task expert network model, a self-balancing training mechanism is used to fuse the data processed by each expert network to obtain fused data. The fused data is input into each discriminant network in the multi-task expert network model to obtain temperature prediction, humidity prediction, and air volume prediction, which together serve as the first predicted value of the environmental target under the candidate parameter values.

7. The method as described in claim 2, characterized in that, For the energy-saving target among the control objectives, the comprehensive prediction model includes a sensible heat ratio prediction model and an energy efficiency prediction model; The step of inputting the current state data and each of the candidate parameter values ​​of the control parameters into the comprehensive prediction model to obtain a first predicted value for each control target under each of the candidate parameter values ​​of the control parameters includes: For any of the candidate parameter values ​​of the control parameters, the current state data and the candidate parameter value are input into the input layer of the sensible heat ratio prediction model; Based on the processing of the feature extraction layer and temperature and humidity energy distribution layer of the sensible heat ratio prediction model, the first sensible heat ratio of the next operating cycle and the second sensible heat ratio of the current operating cycle are obtained. In the output layer of the sensible heat ratio prediction model, the first sensible heat ratio and the second sensible heat ratio are constrained to obtain the predicted sensible heat ratio for the next running cycle. Furthermore, the current state data and the candidate parameter values ​​are input into the energy efficiency prediction model; The high-dimensional feature matrix is ​​obtained by processing the low-dimensional feature extraction layer and the high-dimensional feature mapping layer based on the energy efficiency prediction model. In the energy efficiency prediction layer of the energy efficiency prediction model, the high-dimensional feature matrix is ​​constrained to obtain the predicted energy efficiency. The predicted sensible heat ratio and the predicted energy efficiency are used together as the first predicted value of the energy-saving target under the candidate parameter values.

8. The method as described in claim 2, characterized in that, For the comfort objective among the control objectives, the integrated prediction model includes a comfort parameter prediction model; The step of inputting the current state data and each of the candidate parameter values ​​of the control parameters into the comprehensive prediction model to obtain a first predicted value for each control target under each of the candidate parameter values ​​of the control parameters includes: For any of the candidate parameter values ​​of the control parameters, the current state data and the candidate parameter value are input into the comfort parameter prediction model to obtain the predicted human body parameter value and the predicted thermal comfort value; wherein, the comfort parameter prediction model includes a first model trained with the distribution dispersion of human body parameters before and after the switching of the air conditioning equipment's operating state as a constraint, and a second model trained with the difference in the absolute value of the thermal comfort value before and after the switching of the air conditioning equipment's operating state as a constraint. The predicted values ​​of the human body parameters and the predicted values ​​of thermal comfort are used together as the first predicted value of the comfort target under the candidate parameter values.

9. The method as described in claim 1, characterized in that, The method for predicting the control deviation value of each control target of the air conditioning equipment under each of the candidate parameter values ​​of the control parameters, based on the feedforward control strategy and the feedback control strategy, further includes: Obtain the current status data of the air conditioning equipment; The current state data and the candidate parameter values ​​of the control parameters are input into the feedforward prediction model associated with the feedforward control strategy to obtain the second predicted value of each control objective under each candidate parameter value; Calculate the difference between the required value of each of the set control targets and the second predicted value of each of the control targets under each of the candidate parameter values ​​to obtain the first deviation value of each of the control targets under each of the candidate parameter values; Obtain the third predicted value of each control objective under the current parameter value of the control parameter, as predicted by the feedback prediction model associated with the feedback control strategy in the previous running cycle. The difference between the actual value of each control target in the current operating cycle and the third predicted value of each control target is calculated to obtain the second deviation value of each control target; The first deviation value and the second deviation value of each control target under each candidate parameter value are weighted and fused to obtain the control deviation value of each control target under each candidate parameter value.

10. The method as described in claim 9, characterized in that, The steps for constructing the feedback prediction model include: The difference between the demand value and the actual value of each control target in multiple historical operating cycles prior to the current operating cycle is obtained to obtain the correction loss of each control target in multiple historical operating cycles; A training sample set is constructed based on the correction loss of each control target over multiple historical operating cycles; Based on the training sample set, reinforcement learning is performed on the model to be trained to generate the feedback prediction model.

11. The method according to any one of claims 2 to 10, characterized in that, The acquisition of the current status data of the air conditioning equipment includes: Obtain the current compressor energy of the air conditioning equipment during the current operating cycle, as well as the historical compressor energy of multiple historical operating cycles prior to the current operating cycle; Based on the compressor's historical energy from the multiple historical operating cycles, the inertial energy transferred from each historical operating cycle to the current operating cycle is obtained; Based on the inertial energy and the current energy of the compressor, the cumulative energy of the current operating cycle is obtained; Based on the accumulated energy, the current status data of the air conditioning equipment is determined.

12. The method as described in claim 11, characterized in that, The step of obtaining the inertial energy transferred from each historical operating cycle to the current operating cycle based on the compressor's historical energy from the multiple historical operating cycles includes: Obtain the energy transfer inertia coefficient of the compressor from the historical energy transfer to the current operating cycle for each historical operating cycle; Based on the compressor's historical energy in each historical operating cycle and the energy transfer inertia coefficient from the compressor's historical energy in each historical operating cycle to the current operating cycle, the inertial energy transferred from each historical operating cycle to the current operating cycle is calculated.

13. The method as described in claim 11, characterized in that, The process of determining the current status data of the air conditioning equipment based on the accumulated energy includes: Based on a preset mapping relationship between compressor energy and status data, the status data corresponding to the accumulated energy is obtained as the current status data.

14. The method according to any one of claims 1 to 10, characterized in that, The control objectives include environmental objectives, comfort objectives, and energy-saving objectives. Obtaining the target weight values ​​of each control objective under the current control requirements of the air conditioning equipment includes: The time interval between the current operating cycle and the moment when the user last set the required values ​​for each control target is calculated, and the current electricity price is obtained. Based on the time interval, the initial weight values ​​of the environmental target and the initial weight values ​​of the comfort target are determined; the initial weight values ​​of the environmental target are negatively correlated with the time interval, and the initial weight values ​​of the comfort target are positively correlated with the time interval. Based on the electricity price, an initial weight value for the energy-saving target is determined; the initial weight value for the energy-saving target is positively correlated with the electricity price. The initial weight values ​​of the environmental target, the comfort target, and the energy-saving target are normalized to obtain the target weight values ​​of the environmental target, the comfort target, and the energy-saving target.

15. The method according to any one of claims 1 to 10, characterized in that, The step of obtaining the target weight values ​​of each of the control targets under the current control requirements of the air conditioning equipment further includes: Based on the preset mapping relationship between operating functions and the weight values ​​of each control target, the weight values ​​of each control target corresponding to the current operating function of the air conditioning equipment are obtained and used as the target weight values ​​of each control target.

16. The method according to any one of claims 1 to 10, characterized in that, The acquisition of multiple candidate parameter values ​​for the control parameters of the air conditioning equipment in the next operating cycle includes: Obtain the historical parameter values ​​of the control parameters in the previous operating cycle; Based on the historical parameter values, multiple candidate parameter values ​​for the control parameters are determined.

17. An air conditioning device, characterized in that, The air conditioning device includes: a memory, a processor, and a computer program stored in the memory and executable on the processor, the computer program being configured to implement the steps of the control method for the air conditioning device as claimed in any one of claims 1 to 16.

18. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a computer program that, when executed by a processor, implements the steps of the control method for the air conditioning equipment as described in any one of claims 1 to 16.

Images