183 results about "Air conditioning load" patented technology

The invention relates to a multivariable multi-step air conditioner load prediction model based on time sequence convolution and a double attention mechanism, and belongs to the technical field of air conditioner load prediction. The model adopts a parallel encoding structure, in an encoder, a time sequence convolution module is responsible for modeling local dependence and long-term trend in a time dimension, and a double attention mechanism module is used for modeling a dynamic dependence structure among multiple variables and a coupling relation between the variables and a target load from a variable dimension. The two structures are respectively subjected to feature extraction from a time dimension and a variable dimension, and are complementary to each other. In a decoder, a decoding module with a memory ability and a teacher mandatory strategy is designed, and continuous prediction from a historical multivariable sequence to a future target load is realized.

The invention discloses a data center dynamic air-conditioning load rapid calculation method, which comprises the following steps of: 1, summarizing data center cold load influence factors as an input condition of a calculation model; 2, constructing a standardized data center reference model based on building energy consumption simulation software; 3, based on the building geometrical characteristics and thermal performance of the target data center, correcting the enclosing structure load of the reference model by adopting an area correction method, and constructing an enclosing structure load rapid calculation model; 4, automatically resetting an internal heat source load calculation result, and constructing an internal heat source load rapid calculation model; 5, correcting the fresh air load of the reference model by adopting a volume correction method based on the geometric characteristics of the target data center building, and constructing a fresh air load rapid calculation model; and step 6, forming a target data center dynamic air-conditioning load rapid calculation method, and obtaining a hourly cooling load prediction result required for covering the whole year or a design stage.

The invention relates to the technical field of air conditioner load prediction, and provides an air conditioner load prediction method and system, and the method comprises the steps: extracting intra-day meteorological data features and intra-day air conditioner load data features, and forming multi-dimensional data features; performing dimension reduction processing on the multi-dimensional data features to form a comprehensive feature curve; clustering the comprehensive characteristic curve, dividing the air conditioner load data in the historical day into a data set according to a clustering result, and dividing the data set into a training set and a test set; a plurality of prediction models corresponding to different meteorological scenes are constructed, the prediction models are trained through the corresponding training sets, hyper-parameter tuning is conducted on the prediction models through an improved sodat swarm optimization algorithm, and a plurality of air conditioner load prediction models are formed; and inputting the test set corresponding to various meteorological scenes into the corresponding air conditioner load prediction model, and outputting an air conditioner load prediction result. According to the invention, air conditioner load curves with obvious boundaries in different meteorological scenes can be effectively separated, and air conditioner power load prediction errors in extreme weather are reduced.

The invention relates to an off-grid photovoltaic air conditioner self-adaptive control system and method based on rolling self-learning, and aims to solve the problems of low power supply reliability, large battery loss, inflexible temperature control, weak adaptability and the like caused by poor matching between an existing off-grid photovoltaic system and an air conditioner load. The method mainly comprises the steps that parameters are collected in real time, loads are calculated, and air conditioning unit operation frequency bands of the off-grid system are dynamically divided based on the temperature difference; constructing a multi-target model for guaranteeing power supply, prolonging the service life of the battery and optimizing the thermal comfort of the off-grid system, and adaptively adjusting the target by utilizing a dynamic weight coefficient; the optimal sequence is solved through rolling time window optimization, and the off-grid regulation and control strategy is adjusted in real time in combination with temperature gradient feedback. According to the method, the energy dynamic matching efficiency in the off-grid supply and demand system can be improved, power supply is guaranteed, fine adjustment of temperature control is achieved, harmful charging and discharging are limited based on state optimization, loss is reduced, the service life is prolonged, and compared with a traditional regulation and control strategy, the method has remarkable advantages in reliability, comfort and economical efficiency.

The present application relates to an electric power regulation and control method and apparatus for an air conditioning load, a computer device, a storage medium, and a computer program product. The method comprises: acquiring temperature parameters of each target space; inputting the temperature parameters into a pre-constructed air conditioner power prediction model to obtain target power curves corresponding to the target spaces, wherein the air conditioner power prediction model is used for representing an association relationship between the temperature parameters and the air conditioner power, and the target power curves are used for indicating the changes of the air conditioner power in a target time period; and determining an electric power regulation and control strategy on the basis of the target power curves corresponding to the plurality of target spaces.

The invention relates to the technical field of air conditioner load regulation and control, in particular to a quantitative evaluation method for air conditioner load regulation and control potential in the field of modern industry, which comprises the following steps: S1, collecting and preprocessing environment data and crop information of an air conditioner system, and generating a high-dimensional feature vector containing derived quantities such as VPD and DLI; s2, predicting the transpiration rate and the disease risk based on the high-dimensional features and the crop growth stage, and achieving the dynamic representation of crop response; and S3, the prediction result and the environment, physiology and energy consumption parameters are fused, a multi-dimensional regulation potential index system is constructed, the adjustable space and depth of the air conditioner are evaluated, and regulation potential indexes and grades are output. According to the method, the multi-dimensional evaluation index system is constructed by fusing the multi-source environmental parameters and the crop response characteristics, and accurate quantification and intelligent evaluation of the air conditioner regulation and control potential are achieved.

The invention discloses a building photovoltaic integration and air conditioning system fusion system comprising a photovoltaic power generation subsystem used for converting solar energy into electric energy; the energy storage subsystem is used for storing and releasing electric energy; the air conditioning subsystem is used for adjusting the indoor environment of the building; the intelligent management and control subsystem is in communication connection with the photovoltaic power generation subsystem, the energy storage subsystem and the air conditioner subsystem and is used for collecting operation data of the subsystems, predicting photovoltaic output and air conditioner loads and executing scene-divided energy scheduling based on a model prediction control algorithm; therefore, the cooperative operation of preferentially supplying photovoltaic electric energy to the air-conditioning subsystem, storing or grid-connecting surplus electric energy, and supplementing energy by the stored energy or the power grid when the energy is insufficient is realized. According to the system, efficient production, storage and intelligent utilization of building energy can be achieved, a novel building energy system with self-generation and self-use, residual electricity storage and coordinated regulation is constructed, and the core problems that an existing building photovoltaic and air conditioning system is poor in collaboration and low in intelligent level are solved.

The invention discloses an electrical load aggregation scheduling method considering thermal dynamic characteristics of a building. The method comprises the following steps: constructing an equivalent thermal resistance-thermal capacity physical model of each single building in an office park; data are collected in real time, model parameters are updated through an online recursion identification algorithm, and a digital twin heat dynamic model is obtained; aggregating the buildings into a virtual aggregation group based on thermal dynamic response characteristics, and constructing an aggregation load response characteristic curve; double-layer optimization scheduling is executed, the upper layer solves a total air conditioner load power reference curve of each virtual aggregation group with the purpose of minimizing the total power utilization cost of the park, and the lower layer decomposes the reference curve to each single building and generates a control instruction; and calibrating the model parameters according to the fed back actual indoor temperature. According to the method, the capacity of the scheduling system for dealing with uncertainty is remarkably improved, and the physical feasibility and the actual execution effect of the scheduling instruction are guaranteed.

The invention discloses an air conditioner adjustable potential prediction method based on a cloud-side collaborative architecture, electronic equipment, a storage medium and a product, comprising a cloud end deployed on a server and a side end deployed on user side computing equipment, the cloud end shares side end data, and the cloud end performs prediction on the adjustable potential of an air conditioner based on historical air conditioner load time sequence data. Obtaining a bidirectional LSTM prediction model capable of predicting an air conditioner load value at a specific time in the future, and carrying out iterative optimization on the bidirectional LSTM prediction model based on air conditioner load time sequence data collected by an edge end in real time; and the side end carries out lightweight processing on the bidirectional LSTM prediction model after iterative optimization, deploys the bidirectional LSTM prediction model on user side computing equipment, and carries out air conditioner load value prediction based on air conditioner load time sequence data collected in real time to further obtain an air conditioner adjustable potential prediction result. According to the method, for the problems that an existing prediction model is insufficient in generalization ability and difficult in edge deployment, the cross-scene adaptability is improved through MMD transfer learning, ONNX quantization and Docker containerization are combined, and edge lightweight and standardized efficient deployment is achieved.

The invention discloses a DLC hierarchical optimization scheduling method considering air conditioner load state difference, and relates to the field of power system optimization scheduling, and the method comprises the steps: building a double-layer interaction model of a power supply company, a load aggregator and a user; establishing a load aggregator output model, and determining the controllable capacity of the load aggregator output model; establishing a clustering grouping method considering air conditioner load state differences; establishing an air conditioner load double-layer scheduling optimization model; and solving based on a multi-strategy improved whale optimization algorithm to obtain the start-stop state of each group of air conditioner load. The method mainly solves the problems of low regulation and control precision and non-ideal response effect caused by uncertainty of the initial state of the air conditioner load and parameter difference, improves the response speed and execution reliability through central coordination of the load aggregator, also considers the comfort demand of a user, and has important value for enhancing the dispatching efficiency of a power grid and stable operation.

The invention relates to an air conditioner load prediction method based on a self-adaptive double-flow graph attention network, and belongs to the technical field of building energy conservation and intelligent control. The method comprises the following steps: collecting historical power and environmental data of an air conditioner, and after preprocessing and normalization, constructing an input sequence through a sliding window and dividing a data set according to time; a prediction model is constructed, and a causal graph learning module, a multi-scale graph structure learning module, a self-adaptive space-time attention module, an uncertainty quantization module and a self-adaptive sampling module are integrated; a training set and a joint loss function training model are adopted, and a load prediction result and uncertainty estimation are output through Monte Carlo Dropout during testing. According to the method, the dynamic causal relationship between variables and multi-scale space-time dependence can be adaptively learned, reliable uncertainty quantification is provided while the prediction precision is improved, and the method is suitable for intelligent regulation and control and energy efficiency optimization of the air conditioning system.

The invention provides an airport terminal building heating, ventilation and air conditioning load prediction system and method based on a micro-service architecture. The method comprises the following steps: collecting multi-source heterogeneous original data and extracting a key feature set; mining space-time relevance between people flow and equipment operation, and outputting a space-time characteristic matrix; and constructing a multi-algorithm fusion framework, carrying out cross-terminal distributed cooperative training and dynamic weight aggregation through federal learning, introducing a lion group optimization algorithm to optimize global parameters, and outputting an airport terminal heating ventilation air conditioning load prediction result. Fusion data are standardized, key features are extracted in combination with automatic feature engineering, and the data utilization rate and feature relevance are improved; then capturing a space-time association mode by using a space-time diagram convolutional network, constructing a multi-algorithm fusion framework and dynamically allocating weights; then realizing distributed cooperative training by adopting federal learning, and optimizing parameters in combination with a lion group optimization algorithm to enhance the generalization ability of the model; and finally, through incremental learning dynamic iterative optimization, the prediction precision is continuously improved.

The invention discloses a commercial air conditioner load dynamic influence factor prediction method and device. The method comprises the following steps: firstly, collecting historical data according to a power load four-level classification system, and constructing a historical impact factor time sequence; then decomposing the sequence into four components including a reference value, a medium and short term value, a short term value and a disturbance value through multi-scale decomposition; on the basis of the reference value sequence and the prediction day features, similar days are screened through grey correlation analysis, and a training data set is constructed; aiming at the characteristics of different time scale components, a plurality of machine learning models are adopted to train prediction sub-models respectively; and finally, fusing prediction results of the components by optimizing weights to generate a dynamic impact factor time sequence. According to the method, through a multi-scale decomposition and model cooperation mechanism, prediction precision and adaptability are effectively improved, and a reliable basis is provided for fine scheduling of a power system.

The application provides an air conditioner load prediction method and device, electronic equipment, an air conditioner and a storage medium, and relates to the field of air conditioners. The method comprises the following steps: acquiring actual operation environment data of an air conditioner; wherein the actual operation environment data comprises a space occupancy parameter and a door and window operation parameter, the space occupancy parameter is used for indicating information that an object exists in a space where the air conditioner is located, and the door and window operation parameter is used for indicating information that a door and window in the space is opened or closed; encoding the actual operation environment data to obtain a first feature representation; and performing load prediction according to the first feature representation to obtain a predicted load of the air conditioner. Therefore, by capturing the main disturbance factors in the real operation scene of the air conditioner, including the space occupancy parameter reflecting the heat source distribution of objects such as personnel, equipment and pets in the space where the air conditioner is located, and the door and window operation parameter representing the opening and closing behavior or the ventilation intensity of the door and window, and performing load prediction after encoding the above parameters, the accuracy of air conditioner load prediction can be significantly improved.

The application discloses a building cold load cycle prediction method based on boundary feature protection and an HOA-LightGBM model, and belongs to the field of building intelligence, and comprises the following steps: S1, data acquisition and cleaning; S2, GCMWSG filtering for smoothing processing; S3, selecting evaluation indexes; and S4, constructing an HOA-LightGBM hybrid model and comprehensive evaluation. The application adopts the building cold load cycle prediction method based on boundary feature protection and the HOA-LightGBM model, and by introducing a periodic boundary protection mechanism, effectively solves the boundary distortion problem of traditional MWSG filtering when processing periodic air conditioning load data. Experimental results show that the data processed by GCMWSG retains the load characteristics of start and stop moments, and significantly improves the performance of the prediction model; and the HOA-LightGBM hybrid model provided by the application solves the problems of low parameter optimization efficiency and insufficient prediction accuracy of a shallow model by deeply fusing meta-heuristic algorithms and gradient boosting frameworks, and provides an efficient and reliable technical scheme for real-time prediction of building cold load.

To provide an air conditioning system for a vehicle capable of suppressing intrusion of rainwater into an outside air introduction port even when a weather condition and inclination of a vehicle body are changed while achieving high air conditioning efficiency.SOLUTION: The vehicular air-conditioning system 14 includes the inside-outside air switching door 25 configured to adjust the outside air mixed amount E from the outside air introduction port 19 and the inside air mixed amount G from the inside air introduction port 20, and the air-conditioning control unit 50 configured to control the inside-outside air switching door 25. The air-conditioning control unit 50 calculates the first inside air mixed amount G1 based on the air-conditioning load, and calculates the second inside air mixed amount G2 based on the vehicle body angle and the weather condition. The inside air mixing amount G by the inside / outside air switching door 25 is set based on a larger value of the first inside air mixing amount G1 and the second inside air mixing amount G2.SELECTED DRAWING: Figure 2

The application relates to the technical field of power systems, and discloses an air conditioner load analysis method and system based on data fusion and carbon emission reduction accounting. The method comprises the following steps: acquiring air conditioner related data in different dimensions, and performing fusion processing on all the air conditioner related data to obtain air conditioner load fusion knowledge; a plurality of benchmark scenes are set based on first power consumption behavior characteristics, and the first air conditioner carbon emission of each benchmark scene is determined according to the air conditioner load fusion knowledge; a project scene is set based on second power consumption behavior characteristics, and the second air conditioner carbon emission of the project scene is determined according to the air conditioner load fusion knowledge; the air conditioner carbon emission reduction is obtained based on each first air conditioner carbon emission and second air conditioner carbon emission, and the air conditioner carbon emission reduction and the air conditioner power consumption and charging data obtained by analyzing the air conditioner load fusion knowledge are integrated to form an air conditioner load analysis result. The application deeply couples the air conditioner load data and the carbon emission reduction accounting, and helps the energy system to upgrade to intelligentization.

To improve the comfort of a target space conditioned by an air conditioning system. A controller (60) of an air conditioning system (10) performs a first control action when a load index is lower than a reference value, and performs a second control action when the load index is higher than the reference value. The load index correlates with the air conditioning load of a target space (70). The first control action is an action of maintaining the blowout flow rate at a first flow rate and adjusting the blowout temperature so that the temperature of the target space (70) becomes a set temperature. The second control action is an action of maintaining the blowout temperature at the first temperature and adjusting the blowout flow rate so that the temperature of the target space (70) becomes a set temperature.

The invention discloses a resident air conditioner load multi-mode configuration method for various auxiliary service requirements. The resident air conditioner load multi-mode configuration method comprises the steps that a matching mapping relation between differential control modes and auxiliary service types is constructed; establishing a resident user participation willingness response model based on the incentive price; building a physical aggregation power calculation model of the resident air conditioning load in multiple modes; the theoretical aggregation power interval is corrected, and an air conditioner load reliable aggregation power interval is generated; and based on the matching mapping relation, the resident user participation willingness response model and the air conditioner load reliable aggregation power interval, constructing and solving a collaborative configuration optimization model considering both reliability and economy, and outputting an optimal resource configuration scheme. According to different requirements of secondary frequency modulation, peak regulation and spinning reserve, the configuration combination of the three modes of the intelligent control switch, the intelligent temperature controller and non-intrusive reminding can be differentially optimized, and the operation cost of the system is minimized on the premise that the aggregation reliability is guaranteed.

The invention relates to the technical field of building thermal environment regulation and control, and discloses a building ventilation thermal load regulation and control method and system based on thermal plume prediction, and the method comprises the steps: constructing a curtain wall surface three-dimensional coordinate system and a thermal state temperature map, and obtaining a current thermal plume trajectory; similar trajectory matching is carried out in combination with a historical thermal plume trajectory database, and a thermal plume development path is predicted; based on a prediction result, establishing a compensation relationship between a floor load grading sequence and an adjacent floor feed-forward temperature, implementing time sequence dislocation dimming through spiral progressive control, and after time sequence dislocation dimming, performing thermal-optical coupling feedback correction; hot air can be effectively prevented from flowing backwards, the high-rise room temperature and the air conditioner load are reduced, indoor natural lighting is guaranteed, the building thermal environment is optimized, and building ventilation thermal load efficient regulation and control are achieved.

The invention relates to a control method and device of an air conditioning unit, the air conditioning unit, equipment and a storage medium. The method comprises the steps that expected water temperature, pipeline parameters and environment temperature corresponding to a preset air conditioning unit can be obtained; according to the expected water temperature and the pipeline parameters, the air conditioner load demand quantity is determined; according to the air conditioner load demand quantity, the module basic number is determined; furthermore, according to the expected water temperature and the environment temperature, module gain data are determined; and according to the module basic number and the module gain data, the module operation number in the air conditioning unit is adjusted. Due to the fact that environmental factors can generate an inhibiting effect or a promoting effect on the actual energy efficiency of the air conditioning unit, the basic number of the modules is determined firstly, then the correction value for the basic number of the modules is determined according to the environment temperature and the expected water temperature, the correction value is used for adjusting the influence of the environmental factors on the actual energy efficiency of the air conditioning unit, and the actual energy efficiency of the air conditioning unit is adjusted. And the air conditioning unit meets the air conditioning load demand quantity required by a user.

The invention discloses an air conditioner load demand response potential evaluation method based on the Internet of Things, relates to the field of air conditioner load response analysis, and aims to avoid the uncomfortable problem of completely identifying all physical parameters in a traditional first-order equivalent thermal parameter model. The operation state of the air conditioner is decomposed into a steady-state interval and a dynamic interval, and key macroscopic characteristic parameters, namely a steady-state thermal coefficient, a power consumption boundary function and a dynamic time constant, capable of describing system boundaries and inertia are directly identified from data. According to the method, coupling identification of a plurality of internal states and parameters such as indoor temperature, equivalent thermal resistance and equivalent thermal capacity is ingeniously avoided, finally, in a non-intrusive scene only depending on user aggregation power utilization and outdoor temperature data, the macroscopic parameters capable of being accurately identified are utilized to evaluate response potential, and the response potential is evaluated. The accuracy and robustness of air conditioner load demand response potential evaluation are realized.

This invention discloses a collaborative control method for multiple types of air conditioners based on user-equipment dual regulation costs. The method includes acquiring the operating status information of each type of air conditioner within the controlled area and the user's comfort preference settings; constructing a user-equipment dual regulation cost model for each air conditioner, the model including user satisfaction cost and equipment regulation loss cost; establishing a control optimization model with the objective of minimizing the dual regulation costs of all participating air conditioners, constrained by grid power balance and the operating characteristics of each type of air conditioner; and using a distributed optimization algorithm to solve the control optimization model, obtaining optimized control commands for each type of air conditioner and issuing them for execution. This invention achieves collaborative optimization control of multiple types of air conditioners, effectively balancing user experience and equipment health status while meeting grid regulation requirements, and improving the overall benefits of air conditioner load participation in grid interaction.

This device suppresses vibrations that can occur in a moving vehicle due to the operation of the compressor that compresses the refrigerant used for air conditioning. [Solution] The control device comprises an engine and a compressor that compresses a refrigerant for air conditioning driven by the engine. In a vehicle in motion, it performs air conditioning load torque correction control that reflects the change in air conditioning load torque due to a change in the compressor's driving state in the requested torque based on the driver's operation. The air conditioning load torque correction control includes a first torque correction control that increases the intake air volume to increase the actual air volume torque when the compressor transitions from a de-drive state to a drive state, and decreases the intake air volume to decrease the actual air volume torque when the compressor transitions from a drive state to a de-drive state, and a second torque correction control that retards the ignition timing to decrease the actual air volume torque when the compressor transitions from a de-drive state to a drive state and when it transitions from a drive state to a de-drive state.

This invention relates to the field of air conditioning load modeling and parameter identification technology, specifically a method for air conditioning parameter identification and modeling based on a multi-mechanism fusion gray wolf optimization algorithm. The method first collects air conditioning operation data and performs preprocessing and normalization. Then, based on the traditional gray wolf algorithm, it integrates a composite chaotic initialization strategy, a cross-optimization strategy, and a Cauchy inverse cumulative distribution variation strategy to construct a multi-mechanism fusion optimization algorithm. Next, a first-order equivalent thermal parameter model is established, and the algorithm is used to identify the equivalent thermal resistance R and equivalent heat capacity C parameters, with the goal of minimizing the mean square error between the predicted and measured temperatures. Finally, based on the identified parameters, air conditioning load aggregation modeling is performed, and the impact of the diversity of set temperature and initial indoor temperature on aggregated power fluctuations is analyzed. Through algorithm improvement and comprehensive aggregation characteristic analysis, theoretical and technical support is provided for the flexible scheduling of air conditioning loads in modern power systems.

A vehicle air conditioning device is provided which makes it possible to extend the service life of a compressor by limiting the number of on / off cycles during on / off control.A vehicle air conditioning device 1 comprises a refrigerant circuit R with a compressor 2, a heat dissipation unit 4 that heats air, a pressure reducing unit 6, a heat absorption unit 9, a heat transfer circuit 61 which has a heat generating device and is thermally connectable to at least one of the heat dissipation unit 4 and the heat absorption unit 9, and a control device 32, wherein the control device 32 switches to an air conditioning power consumption mode during heating or cooling operation in the event that a speed of the compressor 2 is a minimum speed and there is excess air conditioning power at the refrigerant circuit R with respect to a required air conditioning load, in which a part of the refrigerant circuit R and the heat transfer circuit 61 are thermally connected.

The invention provides an air conditioner load prediction method for a commercial building partition and a related device, and belongs to the technical field of building air conditioner load prediction. The method comprises the following steps: constructing a multi-scale feature based on a plurality of obtained related features of all prediction target partitions; screening out a front set number of features from the multi-scale features to obtain the screened front set number of features; respectively inputting a set number of screened features into an XGBoost model and a BiLSTM model to obtain a target partition load prediction result of the XGBoost model and a target partition load prediction result of the BiLSTM model; and based on the dynamic weight, fusing a target partition load prediction result of the XGBoost model and a target partition load prediction result of the BiLSTM model to obtain a final air conditioner load prediction result of all target partitions of the commercial building. According to the method, the problem that the accuracy of air conditioner load prediction of commercial building partitions is not high is solved.

This invention relates to the field of energy consumption prediction technology for subway station air conditioning systems, and discloses a method and system for constructing an artificial intelligence-based subway station air conditioning load prediction model. The method includes: acquiring multi-source heterogeneous data and merging it into an original dataset; identifying outliers, imputing missing values ​​based on non-missing dimensions, and normalizing the original dataset to obtain a preprocessed dataset; using a feature extraction algorithm to calculate the relevance scores of each data subset and filtering them in descending order to obtain key feature parameters; establishing a probabilistic surrogate model to predict unknown point function values ​​with the goal of minimizing prediction error, and using the acquisition function trade-off to determine the next evaluation point; determining the optimal hyperparameters of the prediction model through Bayesian iterative optimization; and finally training the model based on the optimal parameters and outputting the load prediction results. This invention solves the problems of large simulation errors and low efficiency in hyperparameter optimization in traditional methods, improves the robustness of data processing, and provides accurate decision support for energy-saving operation of air conditioning systems.

The invention relates to an air conditioner load prediction model generation method and system based on hierarchical contrast distillation and an air conditioner load prediction method.The model generation method comprises the steps that training data including industry category data are collected, a time sequence input tensor is generated, and a teacher model and a student model are trained; the teacher model takes the total loss minimization of all users as a training target, and provides global time sequence structure knowledge and prediction trend knowledge; and the student model takes joint loss minimization of prediction supervision loss, hierarchical comparison learning loss and unified prediction distribution distillation loss as a training target, and the trained student model is used as an air conditioner load prediction model. According to the model, air conditioner load prediction tasks of industrial, commercial, non-industrial and other heterogeneous users can be simultaneously processed in a single model framework, the training number and the parameter maintenance scale of the model are reduced, and the real-time performance and engineering deployment requirements of a large-scale user side energy efficiency management system are met.

The invention relates to the field of power grid load regulation, in particular to a GARCH-Copula-based load joint probability distribution prediction method, system and device for an electric vehicle and an air conditioner and a medium. The method comprises the steps that a GARCH model is adopted to analyze load related data of the electric vehicle to obtain edge distribution of the load of the electric vehicle; meanwhile, a GARCH model is adopted to analyze the load related data of the air conditioner to obtain the edge distribution of the air conditioner load; a pre-constructed Copula joint model is adopted to solve the edge distribution of the electric vehicle load and the edge distribution of the air conditioner load to obtain a probability density curved surface reflecting the relation between the electric vehicle load and the air conditioner load; based on a GARCH-Copula coupling framework, the dynamic change characteristics of tail correlation of EV and AC loads in different time periods and different environments can be effectively captured, and the joint risk of synchronous surge of the two types of loads in an extreme scene is accurately described.