Electrical equipment and control system thereof

Abstract

The invention relates to the technical field of electric appliance control, in particular to electric appliance equipment and a control system thereof, and the system comprises a feature extraction module, a distribution alignment module, a parameter trend modeling module, a sensitivity analysis module and an operation strategy optimization module. According to the method, through accurate screening of environment parameters and sparse basis vector calculation of characteristic parameters, data redundancy is remarkably reduced, the analysis efficiency and pertinence are improved, data consistency is optimized by combining statistical adjustment of differentiation time interval distribution with equipment performance parameters, accurate adaptive capacity is provided for complex working conditions, and the method is suitable for large-scale popularization and application. According to the dynamic characteristic modeling based on the time sequence trend, future change prediction of environmental parameters is realized, the pre-judgment capability and the regulation and control efficiency of the system are effectively enhanced, the influence of each characteristic parameter is clear through sensitivity weight distribution, cooperative control among equipment is optimized, balance between performance and energy consumption is achieved, and dynamic optimization of multifunctional linkage is realized.

Description

Technical Field

[0001] The present invention relates to the field of electrical appliance control technology, and in particular to an electrical appliance and a control system thereof. Background Art

[0002] The field of electrical control technology includes control and management methods and system design related content of electrical equipment. The core content of this technical field is to realize the intelligent, efficient and precise control of electrical equipment, involving multiple links such as the detection of electrical equipment operating parameters, the generation and transmission of control signals, and the operation of actuators. The overall technical field systematically covers the design and optimization of hardware control devices, the development and implementation of control logic, system communication and coordination, and linkage control between various electrical equipment.

[0003] Among them, the electrical control system refers to the technical means of combining hardware and software for realizing centralized or decentralized management of electrical equipment. The subject of this patent is aimed at the automated control and intelligent operation of electrical equipment, covering the design and collaboration of signal acquisition and processing modules, circuit control modules, and logic execution units. By establishing an efficient data transmission path and an accurate logic judgment mechanism, it realizes the switch control of electrical equipment, working state adjustment, and multi-functional linkage task allocation.

[0004] Existing technologies rely on static analysis for parameter screening and fail to effectively eliminate non-critical parameters, resulting in redundant data affecting control accuracy. The parameter distribution alignment method lacks dynamic adaptation capabilities and is difficult to handle nonlinear data deviations under complex working conditions, affecting system consistency. Time series analysis is limited to simple modeling of historical data and lacks the ability to accurately predict dynamic changes in the environment. The sensitivity analysis between characteristic parameters has not formed a systematic strategy, making it impossible to achieve precise parameter control and optimal resource allocation, and it is difficult to meet the needs of intelligent control and multi-functional linkage. Summary of the invention

[0005] The purpose of the present invention is to solve the shortcomings in the prior art and to propose an electrical device and a control system thereof.

[0006] In order to achieve the above object, the present invention adopts the following technical solution: an electrical control system comprises:

[0007] The feature extraction module calculates the sparse basis vectors of the environmental parameters in the feature space based on the set of environmental parameters, eliminates parameters based on relevance, compares the temperature control impact with the operating data of the compressor and refrigerant flow, selects the features directly related to the temperature control performance fluctuation, and generates a key feature parameter set;

[0008] The distribution alignment module calculates the statistical distribution form of the parameters of the differentiated time periods based on the key characteristic parameter set, adjusts the parameters by comparing the deviation values ​​in the distribution, establishes mapping rules in combination with the expansion valve and condenser performance data, adjusts the characteristic distribution values ​​to the unified reference distribution, calculates the consistency of the adjusted data, and generates the aligned characteristic distribution values;

[0009] The parameter trend modeling module analyzes the time series feature change trend based on the aligned feature distribution value, establishes a feature evolution model in combination with the evaporator and compressor operation data, fits the future distribution form through the time series feature change, infers the dynamic change of the environmental parameters, and generates a feature time series prediction value;

[0010] The sensitivity analysis module calculates the sensitivity changes of multiple features to the temperature control performance based on the feature time series prediction value, analyzes the effect trend of the refrigerant flow change on the performance index, extracts the gradient relationship between sensitivity and performance, adjusts the feature set weight distribution, and generates the sensitivity adjustment weight value;

[0011] The operation strategy optimization module adjusts the weight value based on the sensitivity, adjusts the equipment operation parameters, updates the control strategy in combination with the operation status of the compressor and the expansion valve, gradually optimizes the operation parameters of multiple key components, completes the optimization through multiple parameter calculations, and generates a dynamic control optimization parameter set.

[0012] As a further solution of the present invention, the feature extraction module includes:

[0013] The sparse basis vector extraction submodule calculates the sparse basis vectors of the environmental parameters in the feature space based on the environmental parameter set, normalizes the multi-parameter value range in the environmental parameter set, decomposes the feature space using matrix operations, extracts the sparse basis vectors of the parameters in the spatial dimension based on the decomposition results, and generates a sparse basis vector set;

[0014] The correlation elimination submodule analyzes the correlation between environmental parameters based on the sparse basis vector set, eliminates parameters with correlation lower than a specified threshold based on the calculation results, and verifies the screened parameter set using the correlation calculation model to obtain a correlation screening parameter set;

[0015] The key feature screening submodule is based on the correlation screening parameter set, combined with the operating data of the compressor and the refrigerant flow, and uses a correlation calculation model to compare the correlation between the temperature control performance fluctuation and the parameters, screens the characteristic parameters with the largest contribution to the temperature control performance fluctuation, and generates a key feature parameter set.

[0016] As a further solution of the present invention, the feature space is decomposed by matrix operation, using the formula:

[0017]

[0018] Calculate the matrix decomposition weight M ij , extract the sparse basis vectors of the parameters in the spatial dimension according to the decomposition results, and generate a sparse basis vector set;

[0019] Among them, M ij represents the weighted coefficient matrix generated by the environmental parameters i and j in the feature space decomposition process, E i and E j Respectively represent the standardized values ​​of the i-th and j-th items in the environmental parameter set, P ik and P jk They represent the projection values ​​of the parameters of the i-th item and the j-th item in the environmental parameter set on the k-th dimension of the feature space, n represents the number of dimensions of the feature space, S il and S jl They respectively represent the absolute value weights of the i-th item and the j-th item in the environmental parameter set in the l-th subspace after the feature space is decomposed, and m represents the number of subspaces after the feature space is decomposed.

[0020] As a further solution of the present invention, the distribution alignment module includes:

[0021] The distribution morphology calculation submodule extracts parameter values ​​within differentiated time periods based on the key feature parameter set, statistically classifies the parameter set, respectively calculates the central value, distribution range and skewness of multiple parameters in differentiated time periods, performs data morphology feature analysis on the statistical results, and obtains a time period distribution feature set;

[0022] The parameter deviation adjustment submodule calculates the difference between the distribution values ​​based on the time period distribution feature set using a difference calculation model based on pairwise comparison, determines the deviation direction through data comparison, adjusts the parameter value range based on the difference, and corrects the adjustment result in combination with the expansion valve and condenser performance data to generate a deviation adjustment parameter set;

[0023] The unified distribution mapping submodule establishes mapping rules between parameters and unified reference distribution based on the deviation adjustment parameter set, adjusts the distribution data range by calculating the mapping positions of multi-parameter values ​​in the unified distribution, recalculates the data consistency between the adjusted parameters, and generates alignment feature distribution values.

[0024] As a further solution of the present invention, the parameter trend modeling module includes:

[0025] The time series feature analysis submodule extracts the parameter change sequence in the time dimension based on the alignment feature distribution value, analyzes the parameter change rate in multiple time periods on the time axis, summarizes the periodicity and mutation points of the time distribution features by calculating the growth rate, decline amplitude and fluctuation frequency of the parameters in multiple time periods, and generates a time series distribution feature set;

[0026] The feature evolution modeling submodule establishes a correlation matrix for the change pattern of the feature parameters on the time axis based on the time series distribution feature set and the operation data of the evaporator and the compressor, calculates and derives the dynamic change trend of the parameters according to the correlation between time periods, and establishes a parameter evolution model;

[0027] The dynamic change prediction submodule uses a time series fitting method to infer the distribution of characteristic values ​​in the future based on the parameter evolution model, and obtains characteristic time series prediction values ​​by calculating the parameter change trend and generating distribution results at future time points.

[0028] As a further solution of the present invention, the parameter change trend is calculated and the distribution result at a future time point is generated by using the formula:

[0029]

[0030] Calculate the characteristic change trend parameter T t , obtain the feature time series prediction value;

[0031] Among them, T t The parameter value representing the characteristic change trend at time t, w i Represents the weight factor of parameter i, which is determined based on the contribution ratio of historical time series data, P i represents the observed value of parameter i in the historical time series, and P represents the weighted average of all parameters in the time series. The calculation formula is: D j represents the absolute change of the jth distribution parameter, n represents the number of historical time series parameters involved in the calculation, and m represents the total number of distribution parameters in the current time period.

[0032] As a further solution of the present invention, the sensitivity analysis module includes:

[0033] The sensitivity change calculation submodule extracts the effect data of the characteristic value on the temperature control performance in the differentiated time period based on the characteristic time series prediction value, calculates the fluctuation range and offset direction of the characteristic sensitivity value by analyzing the corresponding relationship between the change amplitude of the characteristic value and the performance fluctuation amount, summarizes the sensitivity characteristic change trend in the differentiated time period, and generates a characteristic sensitivity data set;

[0034] The gradient relationship extraction submodule analyzes the dynamic relationship between the sensitivity value and the performance index based on the characteristic sensitivity data set and the data of the refrigerant flow change, calculates the gradient ratio between the sensitivity value and the performance parameter increment, summarizes the nonlinear gradient mode of the sensitivity value changing with the performance, and generates a sensitivity gradient relationship set;

[0035] The weight allocation adjustment submodule calculates the weight adjustment ratio of multiple features based on the sensitivity gradient relationship set, redistributes the weight ratio of the feature set according to the change distribution of the sensitivity value, adjusts the matching relationship between the feature weight and the performance sensitivity, and generates a sensitivity adjustment weight value.

[0036] As a further solution of the present invention, the weight ratio of the feature set is redistributed according to the change distribution of the sensitivity value, using the formula:

[0037]

[0038] Calculate the feature weight adjustment value W i , generate sensitivity adjustment weight value;

[0039] Among them, W i represents the final weight ratio of feature i after weight adjustment, S i represents the sensitivity value of feature i, which is obtained by monitoring the sensitivity gradient relationship set. Represents the mean of all feature sensitivity values, and the calculation formula is R i It represents the original weight ratio of feature i, which is obtained through the initial allocation of features, and n represents the total number of features in the feature set.

[0040] As a further solution of the present invention, the operation strategy optimization module includes:

[0041] The operating parameter adjustment submodule adjusts the weight value based on the sensitivity, extracts the key influencing items in the equipment operating parameters, analyzes the distribution law of the weight value in multiple operating parameters, adjusts the initial configuration and matching range of the equipment operating parameters by calculating the offset of the weight effect on the parameter, gradually corrects the influence of the characteristic parameters on the operating state, and generates an adjusted operating parameter set;

[0042] The control strategy update submodule is based on the adjusted operating parameter set, combined with the real-time operating status data of the compressor and the expansion valve, by comparing the current operating status with the matching value after parameter adjustment, analyzing the deviation relationship between the equipment operating parameters and the control logic, updating the dynamic control strategy of the equipment and calculating the adjusted difference, and generating an updated equipment control strategy;

[0043] The dynamic optimization calculation submodule is based on the updated equipment control strategy, iteratively calculates the operating parameters of key equipment, extracts the dynamic change rules of equipment parameters in multiple calculation results, optimizes the adaptation relationship between control strategy and parameters, reallocates the operating adjustment rules between equipment, and generates a dynamic control optimization parameter set.

[0044] An electrical appliance comprises a memory and a processor, wherein the memory stores a computer program, and the processor implements the electrical appliance control system when executing the computer program.

[0045] Compared with the prior art, the advantages and positive effects of the present invention are:

[0046] In the present invention, data redundancy is significantly reduced and the efficiency and pertinence of analysis are improved through precise screening of environmental parameters and calculation of sparse basis vectors of characteristic parameters. Statistical adjustment of differentiated time period distribution combined with equipment performance parameters optimizes data consistency and provides accurate adaptability to complex working conditions. Dynamic feature modeling based on time series trends enables prediction of future changes in environmental parameters, effectively enhancing the system's predictive ability and regulation efficiency. Sensitivity weight allocation clarifies the impact of each characteristic parameter, optimizes collaborative control between devices, achieves a balance between performance and energy consumption, and realizes dynamic optimization of multi-functional linkage. BRIEF DESCRIPTION OF THE DRAWINGS

[0047] Figure 1 is a system flow chart of the present invention;

[0048] Figure 2 is a system block diagram of the present invention;

[0049] Figure 3 This is a flow chart of the feature extraction module of the present invention;

[0050] Figure 4 This is a flow chart of the distribution alignment module of the present invention;

[0051] Figure 5 This is a flow chart of the parameter trend modeling module of the present invention;

[0052] Figure 6 This is a flow chart of the sensitivity analysis module of the present invention;

[0053] Figure 7 This is a flow chart of the operation strategy optimization module of the present invention. DETAILED DESCRIPTION

[0054] In order to make the purpose, technical solution and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention.

[0055] In the description of the present invention, it should be understood that the terms "length", "width", "up", "down", "front", "back", "left", "right", "vertical", "horizontal", "top", "bottom", "inside", "outside" and the like indicate positions or positional relationships based on the positions or positional relationships shown in the drawings, and are only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and therefore cannot be understood as limiting the present invention. In addition, in the description of the present invention, "multiple" means two or more, unless otherwise clearly and specifically defined.

[0056] Embodiment 1

[0057] See also Figure 1 The present invention provides a technical solution: an electrical control system comprising:

[0058] The feature extraction module calculates the sparse basis vectors of the environmental parameters in the feature space based on the set of environmental parameters, eliminates parameters based on relevance, compares the temperature control impact with the operating data of the compressor and refrigerant flow, selects the features directly related to the temperature control performance fluctuation, and generates a key feature parameter set;

[0059] The distribution alignment module calculates the statistical distribution of parameters in differentiated time periods based on the key feature parameter set, adjusts the parameters by comparing the deviation values ​​in the distribution, establishes mapping rules based on the expansion valve and condenser performance data, adjusts the feature distribution values ​​to the unified reference distribution, calculates the consistency of the adjusted data, and generates the aligned feature distribution values;

[0060] The parameter trend modeling module analyzes the time series feature change trend based on the alignment feature distribution value, establishes a feature evolution model based on the evaporator and compressor operation data, fits the future distribution form through the time series feature change, infers the dynamic change of environmental parameters, and generates feature time series prediction values;

[0061] The sensitivity analysis module calculates the sensitivity changes of multiple features to temperature control performance based on the feature time series prediction value, analyzes the effect trend of refrigerant flow changes on performance indicators, extracts the gradient relationship between sensitivity and performance, adjusts the feature set weight distribution, and generates sensitivity adjustment weight values;

[0062] The operation strategy optimization module adjusts the weight value based on sensitivity, adjusts the equipment operating parameters, updates the control strategy based on the operating status of the compressor and expansion valve, gradually optimizes the operating parameters of multiple key components, completes the optimization through multiple parameter calculations, and generates a dynamic control optimization parameter set.

[0063] The key feature parameter set includes sparse basis vectors, eliminated unrelated parameters, and features directly related to temperature control performance fluctuations. The aligned feature distribution values ​​include adjusted parameter distribution values, unified reference distribution, and data consistency. The feature evolution model specifically includes feature change trends, evaporator and compressor operating characteristics, and future distribution forms. The feature time series prediction values ​​include time series feature changes, dynamic environmental parameters, and future feature distribution forms. The sensitivity adjustment weight values ​​include multi-feature sensitivity gradient relationships, performance indicator action trends, and feature set weight allocations. The dynamic control optimization parameter set includes equipment operating parameters, control strategy update values, and multi-key component optimization parameters.

[0064] See also Figures 2 to 7 The feature extraction module includes a sparse basis vector extraction submodule, a correlation elimination submodule, and a key feature screening submodule. The distribution alignment module includes a distribution morphology calculation submodule, a parameter deviation adjustment submodule, and a uniform distribution mapping submodule. The parameter trend modeling module includes a time series feature analysis submodule, a feature evolution modeling submodule, and a dynamic change prediction submodule. The sensitivity analysis module includes a sensitivity change calculation submodule, a gradient relationship extraction submodule, and a weight distribution adjustment submodule. The operation strategy optimization module includes an operation parameter adjustment submodule, a control strategy update submodule, and a dynamic optimization calculation submodule.

[0065] See also Figure 3 , the feature extraction module includes:

[0066] The sparse basis vector extraction submodule calculates the sparse basis vectors of the environmental parameters in the feature space based on the environmental parameter set, normalizes the multi-parameter value range in the environmental parameter set, decomposes the feature space using matrix operations, extracts the sparse basis vectors of the parameters in the spatial dimension based on the decomposition results, and generates a sparse basis vector set;

[0067] The sparse basis vector extraction submodule is based on the environmental parameter set. It adjusts the multi-parameter value range in the environmental parameter set through the normalization method. First, the environmental parameter set needs to be analyzed, decomposed and extracted. This step can be performed in the form of matrix operations. For a specific feature space, the environmental parameters are normalized according to the formula Complete standardization, where x is the original parameter value, μ is the mean, and σ is the standard deviation. After standardization, these parameters are input into the sparse basis vector extraction algorithm, and the feature space is decomposed by a sparse matrix decomposition algorithm (such as the K-SVD algorithm). The result of the decomposition is a set of sparse basis vectors. A specific sparse vector screening model is further used to calculate the validity of the basis vectors. The contribution threshold of the basis vectors in the feature space is used for comparison and the valid vector set is extracted. A stable sparse basis vector set is obtained through multiple decomposition iterations.

[0068] The correlation elimination submodule analyzes the correlation between environmental parameters based on the sparse basis vector set, eliminates parameters with correlation lower than the specified threshold based on the calculation results, and verifies the screened parameter set using the correlation calculation model to obtain the correlation screening parameter set;

[0069] The correlation elimination submodule is based on a sparse basis vector set and uses a correlation calculation model to analyze the correlation between environmental parameters. First, a correlation model needs to be constructed to calculate the correlation index between each basis vector in the sparse basis vector set. The Pearson correlation coefficient can be used. where x i ,y i Represent the data points in the vector, is the mean, and the calculated correlation coefficient can be used to analyze the correlation between basis vectors. According to the calculation results, the parameters with correlation coefficients lower than the threshold (for example, 0.3) are eliminated, and the remaining parameters after elimination form a new parameter set. The selected parameters are further verified using the correlation verification model. The verification method can adopt association network analysis, build a parameter association graph through a graph theory model, analyze the connectivity indicators of parameter nodes (such as degree centrality and betweenness centrality), and after obtaining the final association screening parameter set, record the parameter list of the elimination and screening process.

[0070] The key feature screening submodule is based on the correlation screening parameter set. Combined with the operating data of the compressor and refrigerant flow, the correlation calculation model is used to compare the correlation between the temperature control performance fluctuation and the parameters, and the characteristic parameters with the largest contribution to the temperature control performance fluctuation are screened to generate the key feature parameter set.

[0071] The key feature screening submodule is based on the correlation screening parameter set, combined with the operating data of the compressor and refrigerant flow, and compares the correlation between the temperature control performance fluctuation and the parameters by building a correlation calculation model. First, the operating data needs to be collected and cleaned. The operating data of the compressor includes real-time power, speed, temperature, etc. The operating data of the refrigerant flow includes pressure, flow, etc. After cleaning the data, the data is normalized, and then the objective function of the temperature control performance fluctuation is constructed. The correlation calculation formula is used Where Cov(X,Y) is the covariance, σ X ,σ Y are the standard deviations of X and Y respectively. The characteristic parameters with larger contribution values ​​are screened through regression analysis of multidimensional parameters and temperature control performance. The specific calculation method can adopt stepwise regression method to sort the parameters from large to small according to the contribution degree, and screen out the key parameters according to the set contribution threshold to generate the final key characteristic parameter set.

[0072] The feature space is decomposed by matrix operation, using the formula:

[0073]

[0074] Calculate the matrix decomposition weight M ij , extract the sparse basis vectors of the parameters in the spatial dimension according to the decomposition results, and generate a sparse basis vector set;

[0075] Among them, M ij represents the weighted coefficient matrix generated by the environmental parameters i and j in the feature space decomposition process, E i and E j Respectively represent the standardized values ​​of the i-th and j-th items in the environmental parameter set, P ik and P jk They represent the projection values ​​of the parameters of the i-th item and the j-th item in the environmental parameter set on the k-th dimension of the feature space, n represents the number of dimensions of the feature space, S il and S jl They respectively represent the absolute value weights of the i-th item and the j-th item in the environmental parameter set in the l-th subspace after the feature space is decomposed, and m represents the number of subspaces after the feature space is decomposed.

[0076] formula:

[0077]

[0078] Detailed explanation of the formula and the process of formula calculation and derivation:

[0079] In this formula, M ij Represents the weighted coefficient matrix generated by environmental parameters i and j during the feature space decomposition process.

[0080] First, E i and E j They represent the standardized values ​​of the i-th and j-th items in the environmental parameter set respectively.

[0081] The standardization process is usually achieved through the following steps:

[0082] Data collection: The raw data of environmental parameters are obtained through sensors or other measuring devices. For example, the value of the i-th environmental parameter collected is e i =75, the value of the jth environmental parameter is e j =60.

[0083] Calculate the mean and standard deviation: Calculate the mean μ and standard deviation σ of the environmental parameter based on the collected data. Assume that the mean μ = 65 and the standard deviation σ = 10.

[0084] Standardization: Using the formula Normalize the data.

[0085] For item i:

[0086] For the jth item:

[0087] Next, P ik and P jk They respectively represent the projection values ​​of the parameters of the i-th item and the j-th item in the environmental parameter set on the k-th dimension of the feature space.

[0088] These projection values ​​are obtained as follows:

[0089] Feature extraction: Dimensionality reduction techniques such as principal component analysis (PCA) are applied to project high-dimensional environmental parameter data into a low-dimensional feature space.

[0090] Calculate the projection value: According to the feature vector, calculate the projection value of each parameter in each dimension of the feature space. Assume that in the kth dimension, the projection value of the i-th item is P ik =0.8, the projection value of the jth item is P jk =0.5.

[0091] Then, S il and S jl They respectively represent the absolute value weights of the i-th item and the j-th item in the environmental parameter set in the l-th subspace after the feature space is decomposed.

[0092] These weights are determined as follows:

[0093] Matrix decomposition: Decompose the matrix of the feature space (such as singular value decomposition) to obtain the basis vectors and corresponding singular values ​​of each subspace.

[0094] Calculate weights: Determine the weight of each parameter in each subspace based on the singular value. Assume that in the lth subspace, the weight of the i-th item is S il =0.6, the weight of the jth item is S jl =0.4.

[0095] Finally, substitute the above values ​​into the formula for calculation:

[0096] Calculate the numerator: |E i -E j |=|1.0-(-0.5)|=1.5;

[0097] Compute the sum of squared differences:

[0098] Compute the square root of the sum of squared differences:

[0099] Calculate the square root sum of the absolute values ​​of the products of the weights:

[0100] ≈0.49;

[0101] Calculate the denominator: 0.3×0.49≈0.147;

[0102] Compute the weighting coefficient matrix elements:

[0103] The result shows that the weighted coefficient matrix element generated by environmental parameters i and j in the feature space decomposition process is 10.2. This value reflects the similarity between parameters i and j in the feature space, and the larger the value, the greater the difference. In the subsequent content, the sparse basis vectors of the parameters in the spatial dimension can be further extracted according to the elements of the matrix M to generate a sparse basis vector set.

[0104] See also Figure 4 , the distribution alignment module includes:

[0105] The distribution morphology calculation submodule extracts parameter values ​​within differentiated time periods based on the key feature parameter set, statistically classifies the parameter set, calculates the central value, distribution range and skewness of multiple parameters in differentiated time periods, analyzes the data morphology characteristics of the statistical results, and obtains the time period distribution feature set;

[0106] The distribution morphology calculation submodule is based on the key feature parameter set. It extracts the parameter values ​​in the differentiated time periods for analysis. First, the differentiated time periods need to be classified and the parameter values ​​in each time period are treated as independent data sets. The central value, distribution range and skewness value are obtained by statistical analysis of each set. The central value can be calculated using the mean formula. The distribution range can be calculated by statistical extreme value difference Range = max(x)-min(x), and the skewness calculation uses the skewness coefficient formula in is the mean, σ is the standard deviation. The above formula can be used to obtain the statistical indicators of each time period. These indicators are integrated to form a time period distribution feature set. The morphological characteristics are analyzed by statistical chart tools for data visualization. An analysis report is generated and stored as a specific data file for subsequent calculation and correction.

[0107] The parameter deviation adjustment submodule is based on the time period distribution feature set, and uses a difference calculation model based on pairwise comparison to calculate the difference between distribution values. The deviation direction is determined by data comparison, and the parameter value range is adjusted based on the difference. The adjustment result is corrected in combination with the expansion valve and condenser performance data to generate a deviation adjustment parameter set.

[0108] The parameter deviation adjustment submodule is based on the time period distribution feature set. It uses a pairwise comparison difference calculation model to calculate the difference between distribution values. First, it is necessary to sort out the central value, distribution range and skewness of the parameter distribution in each time period, and calculate the difference by pairwise comparison. The difference calculation can be used using the following formula Among them, x1 and x2 represent the statistical characteristic values ​​of the parameters in the two time periods respectively. The calculated difference matrix can be displayed through the heat map to determine the distribution pairs with higher difference values. The deviation direction is determined by combining the parameter data comparison results. The performance data of the expansion valve and the condenser are introduced for correction. During the correction process, the collection results of the performance data need to be curve fitted and analyzed. The fitting method can use the quadratic polynomial regression formula y=ax 2 +bx+c, where a, b, and c are fitting coefficients, and the corrected parameter values ​​are adjusted by updating the deviation parameter set for subsequent use.

[0109] The uniform distribution mapping submodule establishes mapping rules between parameters and the uniform reference distribution based on the deviation adjustment parameter set, adjusts the distribution data range by calculating the mapping positions of multi-parameter values ​​in the uniform distribution, recalculates the data consistency between the adjusted parameters, and generates the alignment feature distribution value.

[0110] The uniform distribution mapping submodule establishes the mapping rules between the parameters and the uniform reference distribution based on the deviation adjustment parameter set. By comparing the position relationship between the deviation-adjusted parameter value and the uniform reference distribution, the uniform reference distribution needs to be set first. The standard normal distribution or uniform distribution can be selected as a reference. By constructing the mapping function Map the deviation-adjusted parameter values ​​to a uniform distribution and calculate the consistency index between the mapped data. The consistency analysis can be performed using the standard deviation formula Calculate, where is the mean, σ represents the consistency of the data, and the adjusted parameter distribution data range is obtained by calculation. The statistical indicators of each parameter are compared and updated, and finally the alignment feature distribution value is generated for subsequent feature analysis or model optimization.

[0111] See also Figure 5 , parameter trend modeling module includes:

[0112] The time series feature analysis submodule extracts the parameter change sequence in the time dimension based on the alignment feature distribution value, analyzes the parameter change rate in multiple time periods on the time axis, and summarizes the periodicity and mutation points of the time distribution characteristics by calculating the growth rate, decline amplitude and fluctuation frequency of the parameters in multiple time periods, thus generating a time series distribution feature set.

[0113] The time series feature analysis submodule is based on aligning feature distribution values ​​and extracting the change sequence of parameters in the time dimension for analysis. First, the time change values ​​of each feature parameter need to be classified and sorted. For the multi-period change rate on the time axis, Among them, Δx represents the increase or decrease value of the parameter in a certain period of time, and Δt represents the time interval. By comparing the change rate in multiple time periods, we can preliminarily analyze the parameter growth rate and decline range. Then, we introduce the fluctuation frequency analysis and use the periodic analysis method based on the Fourier transform formula. Extract the main fluctuation frequency components of each parameter, mark the periodicity and mutation points of the parameters, and combine the cumulative change formula when marking the mutation points Cumulative analysis is performed to obtain the precise location of the mutation point by comparing the time series data, and the above results are organized into a time series distribution feature set.

[0114] The feature evolution modeling submodule is based on the time series distribution feature set, combined with the operating data of the evaporator and the compressor, to establish a correlation matrix for the change pattern of the feature parameters on the time axis, calculate and deduce the dynamic change trend of the parameters based on the correlation between time periods, and establish a parameter evolution model;

[0115] The feature evolution modeling submodule is based on the time series distribution feature set. By combining the operating data of the evaporator and the compressor, a correlation matrix is ​​established for the change pattern of the characteristic parameters on the time axis. First, the key performance data of the evaporator and the compressor operation need to be extracted, including the evaporator's refrigeration efficiency and energy consumption coefficient, the compressor's power curve and load change curve, and these data are normalized. Combined with the time dimension, the correlation between the parameters is calculated. Where Cov(X,Y) is the covariance, σ X ,σ Y are the standard deviations of the two parameters respectively, and the correlation matrix between time periods is calculated. The main dynamic change characteristics are extracted through principal component analysis of the matrix. Further, a time evolution model is established according to the correlation law, and the dynamic change law is modeled in the form of a differential equation. The dynamic change trend of the parameters is determined by the model fitting results.

[0116] The dynamic change prediction submodule is based on the parameter evolution model and uses the time series fitting method to infer the distribution of characteristic values ​​in the future. It obtains the characteristic time series prediction value by calculating the parameter change trend and generating the distribution results at future time points.

[0117] The dynamic change prediction submodule is based on the parameter evolution model and uses the time series fitting method to infer the distribution of feature values ​​in the future. First, the fitting function is extracted through the time change trend in the parameter evolution model. Commonly used fitting methods include the ARIMA model, whose basic form is Y t =c+1Yt-1 +2Y t-2 +…+ p Y t-p + t , where c is a constant term, 1, 2, …, p is the autoregression coefficient, t To avoid random errors, the model parameters are fitted by training data. When predicting future time points, the known data are input into the model for iterative calculation to obtain the characteristic value distribution. Then, the distribution results of multiple time points are used to generate characteristic time series prediction values. The prediction results are cross-validated with historical distribution data to verify their accuracy and credibility.

[0118] By calculating the parameter change trend and generating the distribution results at future time points, the formula is used:

[0119]

[0120] Calculate the characteristic change trend parameter T t , obtain the feature time series prediction value;

[0121] Among them, T t The parameter value representing the characteristic change trend at time t, w i Represents the weight factor of parameter i, which is determined based on the contribution ratio of historical time series data, P i represents the observed value of parameter i in the historical time series, Represents the weighted average of all parameters of the time series, and the calculation formula is D j represents the absolute change of the jth distribution parameter, n represents the number of historical time series parameters involved in the calculation, and m represents the total number of distribution parameters in the current time period.

[0122] formula:

[0123]

[0124] Detailed explanation of the formula and the process of formula calculation and derivation:

[0125] When calculating the characteristic change trend parameters, the required parameters are obtained in the following ways:

[0126] Parameter w i The weight factor of parameter i is determined based on the contribution ratio of historical time series data. Assume that through the time series analysis of historical data, the parameter weight factors are set to w1 = 0.4, w2 = 0.3, w3 = 0.2, and w4 = 0.1.

[0127] Parameter P iIt represents the observed value of parameter i in the historical time series. The data is acquired through sensor monitoring or real-time collection. The observed values ​​are set as P1=10, P2=15, P3=8, and P4=12.

[0128] Parameter P represents the weighted average of all parameters of the time series, and the calculation formula is

[0129] Substitute the parameters into the calculation:

[0130]

[0131] Parameter D j It represents the absolute change of the jth distribution parameter, which is calculated by monitoring the distribution difference at each time point in the time series. The absolute change of the distribution parameters is set to D1=5, D2=4, and D3=6, and the total change is calculated as:

[0132] Substituting the above results into the formula to calculate T t :

[0133] Calculate the molecular part

[0134] Calculate item by item:

[0135]

[0136]

[0137] The sum of the numerators is:

[0138]

[0139] Calculate the denominator

[0140]

[0141] Calculate T t

[0142]

[0143] The result shows that the characteristic change trend parameter value is 0.382, which reflects the change intensity of the parameter in the time series at the current moment. The larger the value, the more significant the change intensity. This trend parameter can be used to generate the characteristic value distribution at future time points and predict the time series value.

[0144] See also Figure 6 , the sensitivity analysis module includes:

[0145] The sensitivity change calculation submodule extracts the data of the effect of the characteristic value on the temperature control performance in the differentiated time period based on the characteristic time series prediction value. By analyzing the corresponding relationship between the change amplitude of the characteristic value and the performance fluctuation amount, the fluctuation range and offset direction of the characteristic sensitivity value are calculated, and the sensitivity characteristic change trend in the differentiated time period is summarized to generate a characteristic sensitivity data set.

[0146] The sensitivity change calculation submodule is based on the characteristic time series prediction value. By extracting the effect of the characteristic value on the temperature control performance in the differentiated time period, the characteristic value change data in the differentiated time period must be collected first, and the distribution value of the time dimension is extracted and compared with the temperature control performance fluctuation item by item. The calculation of the characteristic value change amplitude can be calculated using the formula Δx=x t+1 -x t , where x t+1 is the characteristic value of the current period, x t is the characteristic value of the previous time period, and the corresponding relationship between the change amplitude and the performance fluctuation is further analyzed. The fitting relationship between the characteristic change amount and the temperature control performance fluctuation can be constructed through the linear regression model. For example, the fitting equation is y=β0+β1x+, where β0 is a constant term, β1 is the influence coefficient of the characteristic change amplitude on the temperature control performance fluctuation, and is a random error. The sensitivity value range of the change interval is determined by fitting calculation, and the direction of sensitivity value offset is statistically analyzed. The characteristic sensitivity data set is generated according to the sensitivity change trend of each time period, and the data set will be used for subsequent sub-module processing.

[0147] The gradient relationship extraction submodule analyzes the dynamic relationship between the sensitivity value and the performance index based on the characteristic sensitivity data set and the data of the refrigerant flow change. By calculating the gradient ratio between the sensitivity value and the performance parameter increment, the nonlinear gradient pattern of the sensitivity value changing with the performance is summarized to generate a sensitivity gradient relationship set.

[0148] The gradient relationship extraction submodule is based on the characteristic sensitivity data set. By combining the data of refrigerant flow changes, it analyzes the dynamic relationship between the sensitivity value and the performance index. First, it is necessary to perform a differential calculation on each characteristic value in the sensitivity data set. The differential formula is ΔS = S t+1 -S t , where S t is the sensitivity value of the current period, S t+1 is the sensitivity value for the next period. At the same time, the real-time change value of the refrigerant flow rate needs to be extracted. Combined with the flow change data, the gradient ratio of the sensitivity value and the performance parameter increment is calculated. The gradient ratio calculation formula is: Where ΔP represents the incremental value of the performance parameter. By calculating the nonlinear relationship between sensitivity and performance parameters in segments, the gradient pattern of sensitivity value changing with performance is summarized. Pattern extraction can be combined with nonlinear fitting methods, such as using the polynomial fitting equation y=ax 2+bx+c for fitting verification and finally generate a sensitivity gradient relationship set.

[0149] The weight allocation adjustment submodule calculates the weight adjustment ratio of multiple features based on the sensitivity gradient relationship set, redistributes the weight ratio of the feature set according to the change distribution of the sensitivity value, adjusts the matching relationship between the feature weight and the performance sensitivity, and generates the sensitivity adjustment weight value.

[0150] The weight distribution adjustment submodule optimizes the sensitivity distribution by calculating the weight adjustment ratio of multiple features based on the sensitivity gradient relationship set. First, each feature in the sensitivity gradient relationship set needs to be normalized. The normalization formula is: Where x is the original gradient value, x min ,x max They are the minimum and maximum values ​​respectively. After normalization, the weight distribution in the feature set needs to be updated. By calculating the weight adjustment ratio of each feature, the distribution matching update of the feature weight is completed, and the final sensitivity adjustment weight value is generated. The weight value update result is output in matrix form for system model optimization.

[0151] The weight ratio of the feature set is redistributed according to the change distribution of the sensitivity value, using the formula:

[0152]

[0153] Calculate the feature weight adjustment value W i , generate sensitivity adjustment weight value;

[0154] Among them, W i represents the final weight ratio of feature i after weight adjustment, S i represents the sensitivity value of feature i, which is obtained by monitoring the sensitivity gradient relationship set. Represents the mean of all feature sensitivity values, and the calculation formula is R i It represents the original weight ratio of feature i, which is obtained through the initial allocation of features, and n represents the total number of features in the feature set.

[0155] formula:

[0156]

[0157] Detailed explanation of the formula and the process of formula calculation and derivation:

[0158] The sensitivity adjustment weight value formula is used to redistribute the weight ratio of the feature set. The relevant parameters are obtained and calculated in the following way:

[0159] Sensitivity value S iRepresents the sensitivity value of feature i. The sensitivity value is calculated by collecting the sensitivity gradient relationship in the data set and using the gradient change relationship between the feature performance and the target performance change. The sensitivity monitoring results are as follows:

[0160] S1=0.6, S2=0.8, S3=0.4, S4=0.5.

[0161] Mean sensitivity Represents the mean of all feature sensitivity values, and the calculation formula is:

[0162]

[0163] Substitute the sensitivity value:

[0164]

[0165] Original weight ratio R i It represents the initial weight ratio of feature i, which is determined by scoring the importance of the feature set. The initial weight distribution is as follows:

[0166] R1=0.3, R2=0.4, R3=0.2, R4=0.1.

[0167] The weight normalization coefficient of the denominator in the weight adjustment formula is:

[0168]

[0169] Calculate item by item:

[0170]

[0171] The sum is:

[0172]

[0173] The molecular part is:

[0174]

[0175] Calculate W item by item i :

[0176] For i=1:

[0177]

[0178] For i=2:

[0179]

[0180] For i=3:

[0181]

[0182] For i=4:

[0183]

[0184] The weight adjustment values ​​are W1≈0.143, W2≈0.542, W3≈0.238, and W4≈0.078.

[0185] The results show that the adjusted weight values ​​reflect the degree of influence of each feature sensitivity value. After the feature weight ratio is redistributed, the matching relationship between sensitivity and weight is achieved. The generated sensitivity adjustment weight value is used for further analysis and processing.

[0186] See also Figure 7 , the operation strategy optimization module includes:

[0187] The operating parameter adjustment submodule adjusts the weight value based on sensitivity, extracts the key influencing items in the equipment operating parameters, analyzes the distribution law of the weight value in multiple operating parameters, and adjusts the initial configuration and matching range of the equipment operating parameters by calculating the offset of the weight effect on the parameter, gradually corrects the influence of the characteristic parameters on the operating status, and generates the adjusted operating parameter set;

[0188] The operating parameter adjustment submodule adjusts the weight value based on the sensitivity. It extracts the key influencing items in the equipment operating parameters for analysis. First, it is necessary to collect data on the equipment operating parameters, including real-time power, temperature, flow, etc., and extract the distribution law of the sensitivity adjustment weight value. Where W i is the sensitivity weight value of a single feature, is the sum of weight values. By analyzing the distribution values ​​of each characteristic parameter, the key influencing items are determined, and then the adjustment offset calculation formula ΔP=W is used. i ·S i , where ΔP is the parameter offset, S i The initial configuration range of the parameters is adjusted based on the weight ratio of the characteristic values, and finally the adjusted operating parameter set is formed by gradually correcting the influence of the characteristic parameters.

[0189] The control strategy update submodule is based on the adjusted operating parameter set and the real-time operating status data of the compressor and the expansion valve. By comparing the current operating status with the matching value after parameter adjustment, the deviation relationship between the equipment operating parameters and the control logic is analyzed, the dynamic control strategy of the equipment is updated, and the adjusted difference is calculated to generate the updated equipment control strategy.

[0190] The control strategy update submodule is based on the adjusted operating parameter set. It analyzes the real-time operating status data of the compressor and the expansion valve by combining them. First, the power curve of the compressor and the flow curve data of the expansion valve need to be collected, and the adjusted operating parameter set is compared with these real-time data one by one. During the comparison process, the parameter matching difference value needs to be calculated. The matching difference calculation formula is ΔM=|P adjusted -P real-time |, where P adjusted is the adjusted matching value, P real-time For real-time operation values, the deviation relationship between parameters and control logic is determined through point-by-point difference value analysis, and dynamic correction is performed based on the difference. The updated control strategy is verified through a real-time feedback mechanism to generate the final equipment control strategy.

[0191] The dynamic optimization calculation submodule is based on the updated equipment control strategy. It iteratively calculates the operating parameters of key equipment, extracts the dynamic change rules of equipment parameters in multiple calculation results, optimizes the adaptation relationship between control strategy and parameters, reallocates the operating adjustment rules between equipment, and generates a dynamic control optimization parameter set.

[0192] The dynamic optimization calculation submodule extracts the dynamic change law by iteratively calculating the key equipment operating parameters based on the updated equipment control strategy. First, it is necessary to perform multiple rounds of iterative calculations on the key operating parameters and use dynamic optimization algorithms such as the gradient descent method. The gradient update formula is: where θ (t) is the parameter value of the tth iteration, α is the learning rate, The gradient of the parameter to the objective function. The multiple calculation results of dynamic optimization are analyzed and fitted to extract the law. The fitting equation is y=ax 2 +bx+c, where a, b, and c are fitting coefficients. The final optimization control strategy and parameter adaptation relationship is achieved, the operation rules between devices are redistributed, a dynamic control optimization parameter set is generated, and verification is performed.

[0193] An electrical appliance comprises a memory and a processor. The memory stores a computer program, and the processor implements the electrical appliance control system when executing the computer program.

[0194] The above are only preferred embodiments of the present invention and are not intended to limit the present invention in other forms. Any technician familiar with the profession may use the technical contents disclosed above to change or modify them into equivalent embodiments with equivalent changes and apply them to other fields. However, any simple modification, equivalent change and modification made to the above embodiments based on the technical essence of the present invention without departing from the technical solution of the present invention still falls within the protection scope of the technical solution of the present invention.

Claims

1. An electrical control system, characterized in that: The system comprises: The feature extraction module calculates the sparse basis vectors of the environmental parameters in the feature space based on the set of environmental parameters, eliminates parameters based on relevance, compares the temperature control impact with the operating data of the compressor and refrigerant flow, selects the features directly related to the temperature control performance fluctuation, and generates a key feature parameter set; The distribution alignment module calculates the statistical distribution form of the parameters of the differentiated time periods based on the key characteristic parameter set, adjusts the parameters by comparing the deviation values ​​in the distribution, establishes mapping rules in combination with the expansion valve and condenser performance data, adjusts the characteristic distribution values ​​to the unified reference distribution, calculates the consistency of the adjusted data, and generates the aligned characteristic distribution values; The parameter trend modeling module analyzes the time series feature change trend based on the aligned feature distribution value, establishes a feature evolution model in combination with the evaporator and compressor operation data, fits the future distribution form through the time series feature change, infers the dynamic change of the environmental parameters, and generates a feature time series prediction value; The sensitivity analysis module calculates the sensitivity changes of multiple features to the temperature control performance based on the feature time series prediction value, analyzes the effect trend of the refrigerant flow change on the performance index, extracts the gradient relationship between sensitivity and performance, adjusts the feature set weight distribution, and generates the sensitivity adjustment weight value; The operation strategy optimization module adjusts the weight value based on the sensitivity, adjusts the equipment operation parameters, updates the control strategy in combination with the operation status of the compressor and the expansion valve, gradually optimizes the operation parameters of multiple key components, completes the optimization through multiple parameter calculations, and generates a dynamic control optimization parameter set.

2. The electrical appliance control system according to claim 1, characterized in that: The feature extraction module comprises: The sparse basis vector extraction submodule calculates the sparse basis vectors of the environmental parameters in the feature space based on the environmental parameter set, normalizes the multi-parameter value range in the environmental parameter set, decomposes the feature space using matrix operations, extracts the sparse basis vectors of the parameters in the spatial dimension based on the decomposition results, and generates a sparse basis vector set; The correlation elimination submodule analyzes the correlation between environmental parameters based on the sparse basis vector set, eliminates parameters with correlation lower than a specified threshold based on the calculation results, and verifies the screened parameter set using the correlation calculation model to obtain a correlation screening parameter set; The key feature screening submodule is based on the correlation screening parameter set, combined with the operating data of the compressor and the refrigerant flow, and uses a correlation calculation model to compare the correlation between the temperature control performance fluctuation and the parameters, screens the characteristic parameters with the largest contribution to the temperature control performance fluctuation, and generates a key feature parameter set.

3. The electrical appliance control system according to claim 2, characterized in that: The matrix operation method is used to decompose the feature space, and the formula is: Calculate the matrix decomposition weight M ij , extract the sparse basis vectors of the parameters in the spatial dimension according to the decomposition results, and generate a sparse basis vector set; Among them, M ij represents the weighted coefficient matrix generated by the environmental parameters i and j in the feature space decomposition process, E i and E j Respectively represent the standardized values ​​of the i-th and j-th items in the environmental parameter set, P ik and P jk They represent the projection values ​​of the parameters of the i-th item and the j-th item in the environmental parameter set on the k-th dimension of the feature space, n represents the number of dimensions of the feature space, S il and S jl They respectively represent the absolute value weights of the i-th item and the j-th item in the environmental parameter set in the l-th subspace after the feature space is decomposed, and m represents the number of subspaces after the feature space is decomposed.

4. The electrical appliance control system according to claim 1, characterized in that: The distribution alignment module comprises: The distribution morphology calculation submodule extracts parameter values ​​within differentiated time periods based on the key feature parameter set, statistically classifies the parameter set, respectively calculates the central value, distribution range and skewness of multiple parameters in differentiated time periods, performs data morphology feature analysis on the statistical results, and obtains a time period distribution feature set; The parameter deviation adjustment submodule calculates the difference between the distribution values ​​based on the time period distribution feature set using a difference calculation model based on pairwise comparison, determines the deviation direction through data comparison, adjusts the parameter value range based on the difference, and corrects the adjustment result in combination with the expansion valve and condenser performance data to generate a deviation adjustment parameter set; The unified distribution mapping submodule establishes mapping rules between parameters and unified reference distribution based on the deviation adjustment parameter set, adjusts the distribution data range by calculating the mapping positions of multi-parameter values ​​in the unified distribution, recalculates the data consistency between the adjusted parameters, and generates alignment feature distribution values.

5. The electrical appliance control system according to claim 1, characterized in that: The parameter trend modeling module includes: The time series feature analysis submodule extracts the parameter change sequence in the time dimension based on the alignment feature distribution value, analyzes the parameter change rate in multiple time periods on the time axis, summarizes the periodicity and mutation points of the time distribution features by calculating the growth rate, decline amplitude and fluctuation frequency of the parameters in multiple time periods, and generates a time series distribution feature set; The feature evolution modeling submodule establishes a correlation matrix for the change pattern of the feature parameters on the time axis based on the time series distribution feature set and the operation data of the evaporator and the compressor, calculates and derives the dynamic change trend of the parameters according to the correlation between time periods, and establishes a parameter evolution model; The dynamic change prediction submodule uses a time series fitting method to infer the distribution of characteristic values ​​in the future based on the parameter evolution model, and obtains characteristic time series prediction values ​​by calculating the parameter change trend and generating distribution results at future time points.

6. The electrical appliance control system according to claim 5, characterized in that: The parameter change trend is calculated and the distribution results at future time points are generated using the formula: Calculate the characteristic change trend parameter T t , obtain the feature time series prediction value; Among them, T t The parameter value representing the characteristic change trend at time t, w i Represents the weight factor of parameter i, which is determined based on the contribution ratio of historical time series data, P i represents the observed value of parameter i in the historical time series, Represents the weighted average of all parameters of the time series, and the calculation formula is D j represents the absolute change of the jth distribution parameter, n represents the number of historical time series parameters involved in the calculation, and m represents the total number of distribution parameters in the current time period.

7. The electrical appliance control system according to claim 1, characterized in that: The sensitivity analysis module includes: The sensitivity change calculation submodule extracts the effect data of the characteristic value on the temperature control performance in the differentiated time period based on the characteristic time series prediction value, calculates the fluctuation range and offset direction of the characteristic sensitivity value by analyzing the corresponding relationship between the change amplitude of the characteristic value and the performance fluctuation amount, summarizes the sensitivity characteristic change trend in the differentiated time period, and generates a characteristic sensitivity data set; The gradient relationship extraction submodule analyzes the dynamic relationship between the sensitivity value and the performance index based on the characteristic sensitivity data set and the data of the refrigerant flow change, calculates the gradient ratio between the sensitivity value and the performance parameter increment, summarizes the nonlinear gradient mode of the sensitivity value changing with the performance, and generates a sensitivity gradient relationship set; The weight allocation adjustment submodule calculates the weight adjustment ratio of multiple features based on the sensitivity gradient relationship set, redistributes the weight ratio of the feature set according to the change distribution of the sensitivity value, adjusts the matching relationship between the feature weight and the performance sensitivity, and generates a sensitivity adjustment weight value.

8. The electrical appliance control system according to claim 7, characterized in that: The weight ratio of the feature set is redistributed according to the change distribution of the sensitivity value, using the formula: Calculate the feature weight adjustment value W i , generate sensitivity adjustment weight value; Among them, W i represents the final weight ratio of feature i after weight adjustment, S i represents the sensitivity value of feature i, which is obtained by monitoring the sensitivity gradient relationship set. Represents the mean of all feature sensitivity values, and the calculation formula is R i It represents the original weight ratio of feature i, which is obtained through the initial allocation of features, and n represents the total number of features in the feature set.

9. The electrical appliance control system according to claim 1, characterized in that: The operation strategy optimization module includes: The operating parameter adjustment submodule adjusts the weight value based on the sensitivity, extracts the key influencing items in the equipment operating parameters, analyzes the distribution law of the weight value in multiple operating parameters, adjusts the initial configuration and matching range of the equipment operating parameters by calculating the offset of the weight effect on the parameter, gradually corrects the influence of the characteristic parameters on the operating state, and generates an adjusted operating parameter set; The control strategy update submodule is based on the adjusted operating parameter set, combined with the real-time operating status data of the compressor and the expansion valve, by comparing the current operating status with the matching value after parameter adjustment, analyzing the deviation relationship between the equipment operating parameters and the control logic, updating the dynamic control strategy of the equipment and calculating the adjusted difference, and generating an updated equipment control strategy; The dynamic optimization calculation submodule is based on the updated equipment control strategy, iteratively calculates the operating parameters of key equipment, extracts the dynamic change rules of equipment parameters in multiple calculation results, optimizes the adaptation relationship between control strategy and parameters, reallocates the operating adjustment rules between equipment, and generates a dynamic control optimization parameter set.

10. An electrical device, comprising a memory and a processor, characterized in that: The memory stores a computer program, and the processor implements the electrical appliance control system according to any one of claims 1 to 9 when executing the computer program.

Images