Heat supply data prediction method, system and device for self-adaptive hysteresis relation modeling

Abstract

The invention relates to the technical field of multivariable time series data prediction, in particular to a heat supply data prediction method, system and device for adaptive hysteresis relation modeling. The method comprises the following specific steps: carrying out data preprocessing on an industrial data set to obtain a normalized time sequence, and segmenting the normalized time sequence to obtain a true value of time sequence data; lagging relation calculation is carried out on the true value of the time series data to obtain lagging relation representation; performing auxiliary prediction on the lagging relationship representation to obtain a predicted value of the time series data; according to the true value of the time series data and the predicted value of the time series data, calculating hysteresis relation loss and a mean square error, and according to the hysteresis relation loss and the mean square error, performing joint training to obtain a heat supply data prediction model; and inputting the to-be-predicted data into the heat supply data prediction model to obtain a heat supply data prediction result. According to the method, the problems of instability and low robustness of heat supply system data prediction are solved, and the adaptability and robustness of real-time heat supply system data prediction are improved.

Description

Technical Field

[0001] The present invention relates to the technical field of multivariate time series data prediction, and in particular to a method, system and device for predicting heating data based on adaptive lag relationship modeling. Background Art

[0002] Heating system data prediction faces many challenges. On the one hand, the operating data of the heating system often contains a lot of noise, such as sensor failure, data acquisition errors, and data transmission anomalies. These noises may mask the actual operating mode, interfere with the model's learning of data features, and ultimately affect the accuracy and robustness of the prediction. On the other hand, there are complex dynamic relationships between multiple variables in the heating system (such as outdoor temperature, boiler outlet water temperature, return water temperature, boiler load, etc.), especially the hysteresis effect. For example, changes in outdoor temperature usually have a certain time lag effect on the operation of the heating system, and this lag time is affected by the dynamic changes of weather, system load and regional characteristics.

[0003] In recent years, multivariate time series data prediction has gradually become a hot topic in the research of heating system data prediction. Traditional methods such as the ARIMA model (Autoregressive Integrated Moving Average Model) make predictions by capturing the overall trend and periodic characteristics of the time series, but they are limited in their ability to handle complex relationships and nonlinear characteristics of multiple variables. With the rapid development of deep learning technology, recurrent neural networks (RNN), convolutional neural networks (CNN) and transformer-based models have been widely used in time series data prediction. These methods have made certain breakthroughs in prediction accuracy and generalization ability by capturing the global dependencies of time series. However, most existing methods fail to fully explore the dynamic relationships between multiple variables, especially in capturing dynamic lag effects. As a common nonlinear relationship, lag relationships are widely present in multivariate time series data and play a vital role. Lag relationships refer to the impact of the past value of a variable on the current value. Recent studies have begun to focus on capturing and modeling these lag relationships through various methods in order to improve the accuracy of predictions.

[0004] However, current processing methods are often based on statistical methods to estimate lag relationships, which usually require additional time to use the estimated lag amount to correct the model prediction. Although correcting the prediction results by statistical methods can improve the prediction effect to a certain extent, there are several major problems with this approach. First, the process of estimating the lag relationship based on statistical methods will bring additional time overhead, which limits the real-time application capability of the model. Secondly, this method does not really give the model the ability to automatically extract the lag relationship, which means that the process of estimating the lag amount must be repeated every time the data is predicted, which reduces the adaptability and flexibility of the model. The dynamic characteristics of the lag relationship also make the judgment method of fixed rules appear inflexible. For example, the intensity of the lag effect may change over time, and the traditional fixed rule method is difficult to cope with such dynamic changes, which further limits the accuracy of the prediction results.

[0005] In summary, the heating system data prediction method based on adaptive hysteresis relationship modeling represents an innovative research problem with important research significance and application value. Summary of the invention

[0006] The present invention aims to solve at least one of the technical problems existing in the related art. To this end, the present invention provides a heating data prediction method, system and device based on adaptive hysteresis relationship modeling to achieve better completion of the heating system data prediction task.

[0007] The present invention provides a heating data prediction method based on adaptive hysteresis relationship modeling, comprising: S1: performing data preprocessing on the industrial data set to obtain a normalized time series, and segmenting the normalized time series to obtain a segmented normalized time series; S2: performing lag relationship calculation on the segmented normalized time series to obtain a final vector representation; performing auxiliary prediction on the final vector representation to obtain a predicted value of the time series data; S3: deriving a true value of the time series data from the normalized time series, and calculating a lag relationship loss according to the true value of the time series data and the predicted value of the time series data; S4: Calculate the mean square error between the predicted value of the time series data and the true value of the time series data; S5: jointly training a time series predictor according to the lag relationship loss and mean square error to obtain a heating data prediction model; S6: Inputting the normalized time series to be predicted into the heating data prediction model to obtain the predicted value of the time series.

[0008] According to a method for predicting heating data based on adaptive hysteresis relationship modeling provided by the present invention, step S1 comprises: S11: Arrange the industrial data set to obtain a time data series in, is the mth data vector of the time data series, is the data ordinal of the time series data, , is the length of the time data series, is the number of variables, is the set of real numbers; S12: performing data cleaning on the time data sequence to obtain a cleaned time data sequence; S13: Normalize the data vector of the cleaned time series to obtain a normalized time series vector in, is the variable mean of the data vector of the cleaned time data series, The variable standard deviation of the data vector of the time data series after cleaning, , ; The normalized time series vector constitutes the normalized time series ; S14: Divide the normalized time series into several time periods according to the sampling size and sampling interval, and the length of the time period is , the normalized time series after segmentation is expressed as , in, For the current moment, The starting time of the normalized time series after segmentation is , the end time is , .

[0009] According to a method for predicting heating data based on adaptive hysteresis relationship modeling provided by the present invention, step S12 comprises: S121: deleting data vectors whose time steps are non-integer in the time data sequence; S122: Linearly fill the time data sequence with discontinuous time steps; S123: Convert the time step into real time.

[0010] According to a method for predicting heating data based on adaptive hysteresis relationship modeling provided by the present invention, step S2 comprises: S21: Encoding the segmented normalized time series to obtain a multivariate representation : in, is the embedding layer function, implemented by a multi-layer perceptron. is the number of layers of the multilayer perceptron; , is the vector dimension of multivariate representation; S22: Repeat the multivariate characterization Then, concatenate them in the zeroth dimension to get the vector representation of the source variable : in, is a repeated function, ; S23: Multivariate Characterization Repeat for each line separately Get the vector representation of the target variable : in, Each row representing the multivariate representation is repeated separately Second-rate, ; S24: concatenate the vector representation of the source variable and the vector representation of the target variable in the last dimension to obtain a concatenated vector representation : in, Represents a splicing operation; The concatenated vector representation is input into a multi-layer perceptron to obtain a lag relationship representation. : in, represents a multi-layer perceptron; , Dimensions for characterizing lagged relationships; S25: Expand the hysteresis relationship representation dimension to Get the hysteresis relationship representation after expansion , ; S26: Concatenate the multivariate representation and the expanded lag relationship representation to obtain a final vector representation : in, ; The final vector represents Input into the multi-layer perceptron to get the predicted value of the time series data: in, Representation variables from arrive The predicted value at time, variable Including the first variable and the second variable , is the prediction step length.

[0011] According to a method for predicting heating data based on adaptive hysteresis relationship modeling provided by the present invention, step S3 comprises: S31: Filter from arrive The normalized time series corresponding to the moment, to obtain the true value of the time series data ; S32: the true value of the time series data Normalize to get the true value of the normalized time series data ; S33: Prediction of the time series data Normalize to get the predicted value of the normalized time series data ; S34: Calculate the first variable Move Back time steps and the second variable The similarity is: in, is the similarity set, The first variable Displacement time steps and the second variable The similarity of , is the Fourier transform, is the inverse Fourier transform, is the conjugate Fourier transform, For bitwise multiplication, Second variable The predicted value of the normalized time series data, First variable The predicted value of the normalized time series data; S35: Select the similarity with the largest absolute value in the similarity set as the first variable With the second variable The hysteresis : in, Indicates that The largest The value of Indicates taking the absolute value; S36: The first variable of the predicted value of the normalized time series data Move Back time steps, the formula is: in, Represented as the first variable Move Back The predicted value of the normalized time series data at time steps, The first variable from arrive The true value of the normalized time series data at the moment, The first variable from arrive The predicted value of the normalized time series data at the moment, The first variable from arrive Normalized time series data at the moment; S37: Calculate the lagged relationship loss.

[0012] According to a method for predicting heating data based on adaptive hysteresis relationship modeling provided by the present invention, step S37 includes: S371: Calculate the first variable and the second variable The lag relationship loss : ; S372: Calculate all first variables and the second variable Take the average of the lag relationship losses between: get the overall lag relationship loss : .

[0013] According to a method for predicting heating data based on adaptive hysteresis relationship modeling provided by the present invention, step S4 comprises: calculating the mean square error between the predicted value of the time series data and the true value of the time series data : in, For variables In the backward step length The predicted value in For variables In the backward step length The true value in .

[0014] According to a method for predicting heating data based on adaptive hysteresis relationship modeling provided by the present invention, step S5 comprises: S51: Obtaining a total loss function based on the lag relationship loss and the mean square error : in, is a hyperparameter; S52: According to the total loss function ,The optimizer is used to update the time series predictor to obtain the heating data prediction model.

[0015] The present invention also provides a heating data prediction system based on adaptive hysteresis relationship modeling, comprising: Preprocessing module: performing data preprocessing on the industrial data set to obtain a normalized time series, and segmenting the normalized time series to obtain a segmented normalized time series; Model training module: performing lag relationship calculation on the segmented normalized time series to obtain a final vector representation; performing auxiliary prediction on the final vector representation to obtain a predicted value of the time series data; obtaining the true value of the time series data from the normalized time series, and calculating the lag relationship loss according to the true value of the time series data and the predicted value of the time series data; calculating the mean square error between the predicted value of the time series data and the true value of the time series data; jointly training the time series predictor according to the lag relationship loss and the mean square error to obtain a heating data prediction model; Data prediction module: inputs the normalized time series to be predicted into the heating data prediction model to obtain the predicted value of the time series.

[0016] The present invention also provides an electronic device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the computer program, the steps of a heating data prediction method based on adaptive lag relationship modeling as described in any one of the above are implemented.

[0017] The above one or more technical solutions in the embodiments of the present invention have at least one of the following technical effects: The heating data prediction method, system and device based on adaptive hysteresis relationship modeling provided by the present invention solve the problems of instability and low robustness of predicted heating system data by introducing hysteresis relationship representation and hysteresis relationship loss, thereby improving the adaptability and robustness of real-time heating system data prediction.

[0018] Additional aspects and advantages of the present invention will be given in part in the following description and in part will be obvious from the following description, or will be learned through practice of the present invention. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] In order to more clearly illustrate the technical solutions in the present invention or the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0020] Figure 1 It is a flow chart of the heating data prediction method based on adaptive hysteresis relationship modeling provided by the present invention.

[0021] Figure 2 It is a structural block diagram of a heating data prediction device based on adaptive hysteresis relationship modeling provided by the present invention.

[0022] Figure 3 It is a structural schematic diagram of the electronic device provided by the present invention.

[0023] Reference numerals: 101. Preprocessing module; 102. Model training module; 103. Data prediction module; 810. Processor; 820. Communication interface; 830. Memory; 840. Communication bus. DETAILED DESCRIPTION

[0024] In order to make the purpose, technical scheme and advantages of the present invention clearer, the technical scheme in the present invention will be clearly and completely described below. Obviously, the described embodiments are part of the embodiments of the present invention, rather than all the embodiments. Based on the embodiments in the present invention, all other embodiments obtained by ordinary technicians in the field without creative work are within the scope of protection of the present invention. The following embodiments are used to illustrate the present invention, but cannot be used to limit the scope of the present invention.

[0025] In the description of this specification, the description with reference to the terms "one embodiment", "some embodiments", "example", "specific example", or "some examples" etc. means that the specific features, structures, materials or characteristics described in conjunction with the embodiment or example are included in at least one embodiment or example of the embodiment of the present invention. In this specification, the schematic representations of the above terms do not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials or characteristics described may be combined in any one or more embodiments or examples in a suitable manner. In addition, those skilled in the art may combine and combine the different embodiments or examples described in this specification and the features of the different embodiments or examples, without contradiction.

[0026] Combine the following Figures 1 to 3 Description of the Invention Example like Figure 1 As shown, the present invention provides a heating data prediction method based on adaptive hysteresis relationship modeling: S1: performing data preprocessing on the industrial data set to obtain a normalized time series, and segmenting the normalized time series to obtain a segmented normalized time series; S2: performing lag relationship calculation on the segmented normalized time series to obtain a final vector representation; performing auxiliary prediction on the final vector representation to obtain a predicted value of the time series data; S3: deriving a true value of the time series data from the normalized time series, and calculating a lag relationship loss according to the true value of the time series data and the predicted value of the time series data; S4: Calculate the mean square error between the predicted value of the time series data and the true value of the time series data; S5: jointly training a time series predictor according to the lag relationship loss and mean square error to obtain a heating data prediction model; S6: Inputting the normalized time series to be predicted into the heating data prediction model to obtain the predicted value of the time series.

[0027] The predicted values ​​of the time series include variables such as outdoor temperature, boiler outlet water temperature, return water temperature, boiler load, etc.

[0028] Specifically, step S1 includes: S11: Arrange the industrial data set to obtain a time data series in, is the mth data vector of the time data series, is the data ordinal of the time series data, , is the length of the time data series, is the number of variables, is the set of real numbers.

[0029] S12: Cleaning the time data sequence to obtain a cleaned time data sequence, the steps comprising: S121: Delete data vectors whose time steps are non-integer in the time data sequence.

[0030] S122: Linearly fill in the time data sequence with discontinuous time steps.

[0031] S123: Convert the time step into real time.

[0032] Since the data in industrial data sets are all unprocessed data, there are problems such as non-integer time steps, discontinuous time steps, and time steps that cannot correspond to real time, which seriously affects the accuracy of the model. Therefore, it is necessary to clean the time data series.

[0033] S13: Normalize the data vector of the cleaned time series to obtain a normalized time series vector in, is the variable mean of the data vector of the cleaned time data series, The variable standard deviation of the data vector of the time data series after cleaning, , ; The normalized time series vector constitutes the normalized time series .

[0034] S14: Divide the normalized time series into several time periods according to the sampling size and sampling interval, and the length of the time period is , the normalized time series after segmentation is expressed as , in, For the current moment, The starting time of the normalized time series after segmentation is , the end time is , .

[0035] Since there are problems such as missing data and no definite timestamp in the time step of the data in the industrial data set, it is necessary to process the input heating data set. In the embodiment of the present invention, the training set, validation set and test set are divided into 8:1:1 ratio, and the data vectors of all cleaned time data sequences are normalized.

[0036] Specifically, step S2 includes: S21: Encoding the segmented normalized time series to obtain a multivariate representation : in, is the embedding layer function, implemented by a multi-layer perceptron. is the number of layers of the multilayer perceptron; , is the vector dimension of multivariate representation; S22: Repeat the multivariate characterization Then, concatenate them in the zeroth dimension to get the vector representation of the source variable : in, is a repeated function, ; S23: Multivariate Characterization Repeat for each line separately Get the vector representation of the target variable : in, Each row representing the multivariate representation is repeated separately Second-rate, ; S24: concatenate the vector representation of the source variable and the vector representation of the target variable in the last dimension to obtain a concatenated vector representation : in, Represents a splicing operation; The concatenated vector representation is input into a multi-layer perceptron to obtain a lag relationship representation. : in, represents a multi-layer perceptron; , Dimensions for characterizing lagged relationships; S25: Expand the hysteresis relationship representation dimension to Get the hysteresis relationship representation after expansion , ; S26: Concatenate the multivariate representation and the expanded lag relationship representation to obtain a final vector representation : in, ; The final vector represents Input into the multi-layer perceptron to get the predicted value of the time series data: in, Representation variables from arrive The predicted value at time, variable Including the first variable and the second variable , is the prediction step length.

[0037] Specifically, step S3 includes: S31: Filter from arrive The normalized time series corresponding to the moment, to obtain the true value of the time series data ; S32: the true value of the time series data Normalize to get the true value of the normalized time series data ; S33: Prediction value of the time series data Normalize to get the predicted value of the normalized time series data ; S34: Calculate the first variable Move Back time steps and the second variable The similarity is: in, is the similarity set, The first variable Displacement time steps and the second variable The similarity of , is the Fourier transform, is the inverse Fourier transform, is the conjugate Fourier transform, For bitwise multiplication, Second variable The predicted value of the normalized time series data, First variable The predicted value of the normalized time series data; S35: Select the similarity with the largest absolute value in the similarity set as the first variable With the second variable The hysteresis : in, Indicates that The largest The value of Indicates taking the absolute value; S36: The first variable of the predicted value of the normalized time series data Move Back time steps, the formula is: in, Represented as the first variable Move Back The predicted value of the normalized time series data at time steps, The first variable from arrive The true value of the normalized time series data at the moment, The first variable from arrive The predicted value of the normalized time series data at the moment, The first variable from arrive Normalized time series data at the moment, Represents a splicing operation; S37: Calculate the lag relationship loss. Calculate the first variable and the second variable The lag relationship loss : Calculate all first variables and the second variable Take the average of the lag relationship losses between: get the overall lag relationship loss : .

[0038] Specifically, step S4 includes: calculating the mean square error between the predicted value of the time series data and the true value of the time series data : in, For variables In the backward step length The predicted value in For variables In the backward step length The true value in .

[0039] Specifically, step S5 includes: S51: Obtaining a total loss function based on the lag relationship loss and the mean square error : in, is a hyperparameter; S52: According to the total loss function ,The optimizer is used to update the time series predictor to obtain the heating data prediction model.

[0040] After adjustment, in the embodiment of the present invention In the embodiment of the present invention, the Adam optimizer is used to update the model so that the total loss function If there is no decrease for three consecutive times, it is regarded as the total loss function Convergence, get the heating data prediction model .

[0041] S6: Inputting the normalized time series to be predicted into the heating data prediction model to obtain the predicted value of the time series.

[0042] The heating data forecast results are: in, is the normalized time series to be predicted, including outdoor temperature, boiler outlet water temperature, return water temperature, boiler load, etc. variables; is the predicted value of the time series, and is the predicted value of the variable of the corresponding normalized time series.

[0043] After multiple rounds of joint training, a trained model is obtained. The trained model can correctly extract the lag relationship between variables and accurately predict time series data based on the lag relationship. The heating system data containing multiple variables to be predicted is input into the model, and the model outputs the prediction result with the specified prediction step.

[0044] The effectiveness of the multivariate time series data prediction method of the present invention with adaptive hysteresis relationship is verified as follows: The present invention conducted an experiment and used the heating system data set provided by Xinao Gas Group. As shown in Table 1, the experiment used four commonly used evaluation indicators: MSE (Mean Squared Error), MAE (Mean Absolute Error), MAPE (Mean Absolute Percent Error) and SMAPE (Symmetric Average Absolute Percentage Error).

[0045] Table 1 Comparison results with other methods

[0046] Table 1 presents the results of the method of the present invention and other advanced methods in the field of multivariable time series prediction. Among them, Informer, Autoformer, iTransformer are other prediction methods, and w / o relation loss means that the lag relation loss optimization model is not applicable. The experimental results show that the method proposed in the present invention achieves the best performance compared with the previous methods. The above comparison results fully prove that the method proposed in the present invention has achieved excellent results in multivariable heating system data prediction and adaptive lag relation modeling.

[0047] like Figure 2 As shown, the present invention also provides a heating data prediction system based on adaptive hysteresis relationship modeling, comprising: Preprocessing module 101: performing data preprocessing on the industrial data set to obtain a normalized time series, and segmenting the normalized time series to obtain segmented normalized time series; Model training module 102: Calculate the lag relationship of the segmented normalized time series to obtain a lag relationship representation; perform auxiliary prediction on the lag relationship representation to obtain a predicted value of the time series data; derive the true value of the time series data from the normalized time series, and calculate the lag relationship loss according to the true value of the time series data and the predicted value of the time series data; calculate the mean square error between the predicted value of the time series data and the true value of the time series data; jointly train the time series predictor according to the lag relationship loss and the mean square error to obtain a heating data prediction model; Data prediction module 103: inputs the data to be predicted into the heating data prediction model to obtain the heating data prediction result.

[0048] Figure 3 An example of a physical structure diagram of an electronic device is shown in FIG. Figure 3As shown, the electronic device may include: a processor 810, a communication interface 820, a memory 830 and a communication bus 840, wherein the processor 810, the communication interface 820, and the memory 830 communicate with each other through the communication bus 840. The processor 810 may call the logic instructions in the memory 830 to execute the heating data prediction method based on the adaptive hysteresis relationship modeling, and the method includes: S1: performing data preprocessing on the industrial data set to obtain a normalized time series, and segmenting the normalized time series to obtain a segmented normalized time series; S2: performing lag relationship calculation on the segmented normalized time series to obtain a vector representation; performing auxiliary prediction on the vector representation to obtain a predicted value of the time series data; S3: deriving a true value of the time series data from the normalized time series, and calculating a lag relationship loss according to the true value of the time series data and the predicted value of the time series data; S4: Calculate the mean square error between the predicted value of the time series data and the true value of the time series data; S5: jointly training a time series predictor according to the lag relationship loss and mean square error to obtain a heating data prediction model; S6: Inputting the normalized time series to be predicted into the heating data prediction model to obtain the predicted value of the time series.

[0049] In addition, the logic instructions in the above-mentioned memory 830 can be implemented in the form of a software functional unit and can be stored in a computer-readable storage medium when it is sold or used as an independent product. Based on this understanding, the technical solution of the present invention is essentially or the part that contributes to the prior art or the part of the technical solution can be embodied in the form of a software product, and the computer software product is stored in a storage medium, including a number of instructions to enable a computer device (which can be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the method described in each embodiment of the present invention. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), disk or optical disk and other media that can store program codes.

[0050] The device embodiments described above are merely illustrative, wherein the units described as separate components may or may not be physically separated, and the components displayed as units may or may not be physical units, that is, they may be located in one place, or they may be distributed on multiple network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the scheme of this embodiment. Ordinary technicians in this field can understand and implement it without paying creative labor.

[0051] Through the description of the above implementation methods, those skilled in the art can clearly understand that each implementation method can be implemented by means of software plus a necessary general hardware platform, and of course, can also be implemented by hardware. Based on this understanding, the above technical solution is essentially or the part that contributes to the prior art can be embodied in the form of a software product, and the computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, a disk, an optical disk, etc., including a number of instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to execute the methods described in each embodiment or some parts of the embodiments.

[0052] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present invention.

[0053] It should be noted that the embodiments of the present disclosure may be implemented by hardware, software, or a combination of software and hardware. The hardware portion may be implemented using dedicated logic: the software portion may be stored in a memory and executed by an appropriate instruction execution system such as a microprocessor or dedicated design hardware. Those skilled in the art will appreciate that the above-described apparatus and methods may be implemented using computer executable instructions and / or contained in a processor control code, such as a programmable memory or a data carrier such as an optical or electronic signal carrier providing such code.

[0054] In addition, although the operation of the method of the present disclosure is described in a particular order in the accompanying drawings, this does not require or imply that these operations must be performed in this particular order, or that all the operations shown must be performed to achieve the desired result. On the contrary, the steps depicted in the flow chart can change the order of execution. Additionally or alternatively, some steps can be omitted, multiple steps can be combined into one step for execution, and / or one step can be decomposed into multiple steps for execution. It should also be noted that the features and functions of two or more devices according to the present disclosure can be embodied in one device. Conversely, the features and functions of a device described above can be further divided into being embodied by multiple devices.

[0055] Although the present disclosure has been described with reference to several specific embodiments, it should be understood that the present disclosure is not limited to the specific embodiments disclosed. The present disclosure is intended to cover various modifications and equivalent arrangements included within the spirit and scope of the appended claims.

Claims

1. A heating data prediction method based on adaptive hysteresis relationship modeling, characterized in that: include: S1: performing data preprocessing on the industrial data set to obtain a normalized time series, and segmenting the normalized time series to obtain a segmented normalized time series; S2: performing lag relationship calculation on the segmented normalized time series to obtain a final vector representation; performing auxiliary prediction on the final vector representation to obtain a predicted value of the time series data; S3: deriving a true value of the time series data from the normalized time series, and calculating a lag relationship loss according to the true value of the time series data and the predicted value of the time series data; S4: Calculate the mean square error between the predicted value of the time series data and the true value of the time series data; S5: jointly training a time series predictor according to the lag relationship loss and mean square error to obtain a heating data prediction model; S6: Inputting the normalized time series to be predicted into the heating data prediction model to obtain the predicted value of the time series.

2. The method for predicting heating data based on adaptive hysteresis relationship modeling according to claim 1, characterized in that: Step S1 includes: S11: Arrange the industrial data set to obtain a time data series in, The time series data data vectors, is the data ordinal of the time series data, , is the length of the time data series, is the number of variables, is the set of real numbers; S12: performing data cleaning on the time data sequence to obtain a cleaned time data sequence; S13: Normalize the data vector of the cleaned time series to obtain a normalized time series vector in, is the variable mean of the data vector of the cleaned time data series, The variable standard deviation of the data vector of the time data series after cleaning, , ; The normalized time series vector constitutes the normalized time series ; S14: Divide the normalized time series into several time periods according to the sampling size and sampling interval, and the length of the time period is , the normalized time series after segmentation is expressed as , in, For the current moment, The starting time of the normalized time series after segmentation is , the end time is , .

3. The method for predicting heating data based on adaptive hysteresis relationship modeling according to claim 2, characterized in that: Step S12 includes: S121: deleting data vectors whose time steps are non-integer in the time data sequence; S122: Linearly fill the time data sequence with discontinuous time steps; S123: Convert the time step into real time.

4. The method for predicting heating data based on adaptive hysteresis relationship modeling according to claim 2, characterized in that: Step S2 includes: S21: Encoding the segmented normalized time series to obtain a multivariate representation : in, is the embedding layer function, implemented by a multi-layer perceptron. is the number of layers of the multilayer perceptron; , is the vector dimension of multivariate representation; S22: Repeat the multivariate characterization Then, concatenate them in the zeroth dimension to get the vector representation of the source variable : in, is a repeated function, ; S23: Multivariate Characterization Repeat for each line separately Get the vector representation of the target variable : in, Each row representing the multivariate representation is repeated separately Second-rate, ; S24: concatenate the vector representation of the source variable and the vector representation of the target variable in the last dimension to obtain a concatenated vector representation : in, Represents a splicing operation; The concatenated vector representation is input into a multi-layer perceptron to obtain a lag relationship representation. : in, represents a multi-layer perceptron; , Dimensions for characterizing lagged relationships; S25: Expand the hysteresis relationship representation dimension to , and the hysteresis relationship representation after expansion is obtained , ; S26: Concatenate the multivariate representation and the expanded lag relationship representation to obtain a final vector representation : in, ; The final vector represents Input into the multi-layer perceptron to get the predicted value of the time series data: in, Representation variables from arrive The predicted value at time, variable Including the first variable and the second variable , is the prediction step length.

5. The method for predicting heating data based on adaptive hysteresis relationship modeling according to claim 4, characterized in that: Step S3 includes: S31: Filter from arrive The normalized time series corresponding to the moment, to obtain the true value of the time series data ; S32: the true value of the time series data Normalize to get the true value of the normalized time series data ; S33: Prediction value of the time series data Normalize to get the predicted value of the normalized time series data ; S34: Calculate the first variable Move Back time steps and the second variable The similarity is: in, is the similarity set, The first variable Move Back time steps and the second variable The similarity of , is the Fourier transform, is the inverse Fourier transform, is the conjugate Fourier transform, For bitwise multiplication, Second variable The true value of the normalized time series data, First variable The predicted value of the normalized time series data; S35: Select the similarity with the largest absolute value in the similarity set as the first variable With the second variable The hysteresis : in, Indicates that The largest The value of Indicates taking the absolute value; S36: The first variable of the predicted value of the normalized time series data Move Back time steps, the formula is: in, Represented as the first variable Move Back The predicted value of the normalized time series data at time steps, The first variable from arrive The true value of the normalized time series data at the moment, The first variable from arrive The predicted value of the normalized time series data at the moment, The first variable from arrive Normalized time series data at the moment; S37: Calculate the lagged relationship loss.

6. The method for predicting heating data based on adaptive hysteresis relationship modeling according to claim 5, characterized in that: Step S37 includes: S371: Calculate the first variable and the second variable The lag relationship loss : in, Second variable The predicted value of the normalized time series data; S372: Calculate all first variables and the second variable Take the average of the lag relationship losses between: get the overall lag relationship loss : 。 7. The method for predicting heating data based on adaptive hysteresis relationship modeling according to claim 6, characterized in that: Step S4 includes: calculating the mean square error between the predicted value of the time series data and the true value of the time series data in, For variables In the backward step length The predicted value in For variables In the backward step length The true value in .

8. The method for predicting heating data based on adaptive hysteresis relationship modeling according to claim 7, characterized in that: Step S5 includes: S51: Obtaining a total loss function based on the lag relationship loss and the mean square error : in, is a hyperparameter; S52: According to the total loss function ,The optimizer is used to update the time series predictor to obtain the heating data prediction model.

9. A heating data prediction system based on adaptive hysteresis relationship modeling, used to execute a heating data prediction method based on adaptive hysteresis relationship modeling as claimed in any one of claims 1 to 8, characterized in that: include: Preprocessing module: performing data preprocessing on the industrial data set to obtain a normalized time series, and segmenting the normalized time series to obtain a segmented normalized time series; Model training module: performing lag relationship calculation on the segmented normalized time series to obtain a final vector representation; performing auxiliary prediction on the final vector representation to obtain a predicted value of the time series data; Determining a true value of the time series data from the normalized time series, and calculating a lag relationship loss according to the true value of the time series data and the predicted value of the time series data; Calculate the mean square error between the predicted value of the time series data and the true value of the time series data; A time series predictor is trained jointly according to the lag relationship loss and mean square error to obtain a heating data prediction model; Data prediction module: inputs the normalized time series to be predicted into the heating data prediction model to obtain the predicted value of the time series.

10. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein: When the processor executes the computer program, the steps of the heating data prediction method based on adaptive hysteresis relationship modeling as described in any one of claims 1 to 8 are implemented.

Images