Energy management system optimization method based on dynamic demand algorithm

Abstract

The invention discloses an energy management system optimization method based on a dynamic demand algorithm, and the method comprises the steps: firstly obtaining historical energy management data, carrying out the data preprocessing, obtaining target historical energy management data, carrying out the data cleaning of the target historical energy management data, carrying out the arrangement according to time information, and generating time series data; carrying out standardization processing on the time sequence data, and carrying out ADF stability test and processing on the standardized time sequence data to obtain stable time sequence data; training stationary time sequence data after ADF stationary test and processing, and establishing an energy management demand threshold prediction model; and finally, determining a demand threshold value at the current moment by using the energy management demand threshold value prediction model so as to carry out power control based on the demand threshold value at the current moment. The problem that the demand threshold value in an existing EMS system cannot adapt to the real-time energy demand is solved, the influence of disordered data is eliminated through multiple preprocessing, and the reliability of a set result is guaranteed.

Description

Technical Field

[0001] The present invention belongs to the technical field of machine learning and energy management system (EMS), and in particular relates to an energy management system optimization method based on a dynamic demand algorithm. Background Art

[0002] At present, the demand control method commonly used in energy management systems (EMS) is to manually set fixed demand thresholds. Although it can achieve basic power demand management to a certain extent, its inherent limitations are becoming increasingly prominent.

[0003] In practical applications, the power demand of power consumption sites often presents a high degree of dynamics and uncertainty. For example, in the manufacturing industry, enterprises often need to flexibly adjust production plans according to market changes. When faced with a surge in orders, enterprises may need to quickly increase production lines to meet the urgent needs of customers, and the power demand will increase significantly. However, if the EMS system still uses the previously set fixed demand threshold, it may not be able to respond to such changes in time, resulting in insufficient power supply, which in turn affects production progress and product quality; on the contrary, when the order volume decreases and the company decides to reduce production lines to reduce operating costs, the power demand will also decrease accordingly. If the EMS system continues to be configured according to the original high demand threshold, it will not only cause a waste of power resources, but also increase unnecessary electricity bills and reduce the overall operating efficiency of the enterprise. However, the existing dynamic demand calculation is generally a simple calculation, which often makes it affected by factors such as data clutter and data loss, resulting in unreliable results of the demand threshold.

[0004] Therefore, there is an urgent need to provide an energy management system optimization method based on a dynamic demand algorithm that can automatically and dynamically set the demand threshold without the need for manual control and with reliable demand threshold setting results. Summary of the invention

[0005] The purpose of the present invention is to provide an energy management system optimization method based on a dynamic demand algorithm to solve the above problems existing in the prior art.

[0006] In order to achieve the above object, the present invention adopts the following technical solutions:

[0007] In a first aspect, the present invention provides an energy management system optimization method based on a dynamic demand algorithm, comprising:

[0008] Acquire historical energy management data, and perform data preprocessing on the historical energy management data to obtain target historical energy management data;

[0009] Performing data cleaning on the target historical energy management data, and arranging the cleaned target historical energy management data according to time information to generate historical energy management time series data;

[0010] Standardizing the historical energy management time series data to obtain standardized historical energy management time series data, and performing ADF stationarity test and processing on the standardized historical energy management time series data to obtain stationary time series data;

[0011] Using the stable time series data, constructing an energy management demand threshold prediction model;

[0012] Using the energy management demand threshold prediction model, the real-time energy management data is calculated to obtain a real-time demand threshold prediction value;

[0013] Based on the real-time demand threshold prediction value, the demand threshold at the current moment is determined, so as to perform power control based on the demand threshold at the current moment.

[0014] In a possible design, historical energy management data is obtained and data preprocessing is performed on the historical energy management data, including:

[0015] According to the preset time granularity, the historical energy data of each electrical device is obtained, and the historical energy data of each electrical device obtained is used as the historical energy management data;

[0016] Deduplication processing is performed on the device identification in the historical energy management data, wherein the device identification is used to characterize the type of electrical equipment;

[0017] Based on the device identification, the deduplicated historical energy management data is summed according to the data type, and the summed result is used as the target historical energy management data, wherein the data type includes the number of devices and / or the device power consumption and device electricity consumption during peak, valley and normal periods.

[0018] In one possible design, the target historical energy management data includes a plurality of energy data sets, and each energy data set corresponds to a data type;

[0019] The target historical energy management data is cleaned, including:

[0020] According to the data type corresponding to each energy data set, the designated cleaning action corresponding to each energy data set is selected respectively;

[0021] For the target historical energy management data in each energy data set, the corresponding designated cleaning actions are respectively performed to obtain the target historical energy management data after data cleaning.

[0022] In one possible design, the historical energy management time series data includes a plurality of energy time series data sets, and each energy time series data set corresponds to a data type;

[0023] The process of standardizing the historical energy management time series data includes:

[0024] Calculate the data mean and data standard deviation of each energy time series data set;

[0025] According to the data mean and data standard deviation of each energy time series data set, each energy time series data set is standardized to obtain the standardized historical energy management time series data.

[0026] In a possible design, the ADF stationarity test is performed on the standardized historical energy management time series data, including:

[0027] Get the ADF test model;

[0028] Determine an autoregressive model of the standardized historical energy management time series data, and calculate an ADF test statistic and a critical value of the standardized historical energy management time series data based on the autoregressive model and the ADF test model;

[0029] Determine whether the ADF test statistic of the standardized historical energy management time series data is less than or equal to the critical value;

[0030] If so, it is determined that the standardized historical energy management time series data is a non-stationary series;

[0031] Performing differential processing on the standardized historical energy management time series data to obtain differentially processed historical energy management time series data;

[0032] The standardized historical energy management time series data is updated to the historical energy management time series data after difference processing, and the autoregressive model of the standardized historical energy management time series data is re-determined until the ADF test statistic of the standardized historical energy management time series data is greater than the critical value, and the stable time series data is obtained.

[0033] In a possible design, the stationary time series data is used to construct an energy management demand threshold prediction model, including:

[0034] The stationary time series data that has passed the ADF stationarity test and processing is divided into a training data set, a test data set, and a validation data set according to a preset ratio;

[0035] The machine learning model is trained by taking the stationary time series data in the training data set as input and the energy management demand threshold as output. During the training process, the output of the machine learning model is used to adjust the parameters of the machine learning model so as to obtain a pre-energy management demand threshold prediction model after the training is completed.

[0036] Using the test data set, test the pre-energy management demand threshold prediction model, obtain the test results, and determine whether the test results meet the preset conditions;

[0037] If not, the pre-energy management demand threshold prediction model is optimized until the test result meets the preset conditions, and the test energy management demand threshold prediction model is obtained;

[0038] The test energy management demand threshold prediction model is verified using the verification data set to obtain a verification result, and when the verification result is verification passed, the energy management demand threshold prediction model is obtained.

[0039] In a possible design, based on the real-time demand threshold prediction value, the demand threshold at the current moment is determined, including:

[0040] Get the demand threshold value at the previous moment;

[0041] Compare the demand threshold at the previous moment with the predicted value of the real-time demand threshold;

[0042] If the demand threshold at the previous moment is greater than or equal to the real-time demand threshold prediction value, the real-time demand threshold prediction value is discarded, and the demand threshold at the previous moment is used as the demand threshold at the current moment;

[0043] If the demand threshold at the previous moment is less than the real-time demand threshold prediction value, the real-time demand threshold prediction value is used as the demand threshold at the current moment.

[0044] In a second aspect, the present invention provides an energy management system based on a dynamic demand algorithm, comprising:

[0045] A data acquisition module acquires historical energy management data and performs data preprocessing on the historical energy management data to obtain target historical energy management data;

[0046] A data processing module, used for performing data cleaning on the target historical energy management data, and arranging the cleaned target historical energy management data according to time information to generate historical energy management time series data; and also used for performing standardization processing on the historical energy management time series data to obtain standardized historical energy management time series data, and performing ADF stationarity test and processing on the standardized historical energy management time series data to obtain stationary time series data;

[0047] A model training module, used to construct an energy management demand threshold prediction model using the stable time series data;

[0048] The demand threshold update module is used to calculate the real-time energy management data using the energy management demand threshold prediction model to obtain the real-time demand threshold prediction value; it is also used to determine the demand threshold at the current moment based on the real-time demand threshold prediction value, so as to perform power control based on the demand threshold at the current moment.

[0049] In a third aspect, the present invention provides an electronic device comprising a memory, a processor and a transceiver which are communicatively connected in sequence, wherein the memory is used to store a computer program, the transceiver is used to send and receive messages, and the processor is used to read the computer program to execute an energy management system optimization method based on a dynamic demand algorithm as described in the first aspect or any possible design of the first aspect.

[0050] In a fourth aspect, the present invention provides a computer program product comprising instructions, which, when executed on a computer, causes the computer to execute an energy management system optimization method based on a dynamic demand algorithm as described in the first aspect or any possible design of the first aspect.

[0051] Beneficial effect: The present invention provides an energy management system optimization method based on a dynamic demand algorithm, firstly, the historical energy management data is obtained, and the data is preprocessed to obtain the target historical energy management data, and then the target historical energy management data is cleaned and arranged according to the time information to generate the historical energy management time series data; then the historical energy management time series data is standardized, and the standardized historical energy management time series data is tested and processed by ADF stability to obtain stable time series data; then the stable time series data after the ADF stability test and processing is trained to establish an energy management demand threshold prediction model; finally, the energy management demand threshold prediction model is used to calculate the real-time energy management data, and finally determine the demand threshold at the current moment, so as to perform power control based on the demand threshold at the current moment. The problem that the demand threshold in the current EMS system cannot adapt to the real-time energy demand due to the fact that most of the demand thresholds are manually set in the current EMS system is solved, and because the data is preprocessed multiple times, the influence of messy data such as missing values ​​on the results is eliminated, and the reliability of the demand threshold setting results is guaranteed. BRIEF DESCRIPTION OF THE DRAWINGS

[0052] Figure 1 It is a flow chart of the energy management system optimization method based on the dynamic demand algorithm in the present invention. DETAILED DESCRIPTION

[0054] In order to more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the present invention will be briefly introduced below in combination with the drawings and the description of the embodiments or the prior art. Obviously, the following description of the structure of the drawings is only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work. It should be noted that the description of these embodiments is used to help understand the present invention, but does not constitute a limitation of the present invention.

[0055] It should be understood that although the terms first, second, etc. may be used herein to describe various units, these units should not be limited by these terms. These terms are only used to distinguish one unit from another unit. For example, a first unit can be referred to as a second unit, and similarly, a second unit can be referred to as a first unit without departing from the scope of the exemplary embodiments of the present invention.

[0056] It should be understood that the term "and / or" that may appear in this article is merely a description of the association relationship between associated objects, indicating that three relationships may exist. For example, A and / or B can represent three situations: A exists alone, B exists alone, and A and B exist at the same time. The term " / and" that may appear in this article describes another type of association object relationship, indicating that two relationships may exist. For example, A / and B can represent two situations: A exists alone, and A and B exist alone. In addition, the character " / " that may appear in this article generally indicates that the previous and next associated objects are in an "or" relationship.

[0057] Example:

[0058] like Figure 1 As shown, this embodiment provides an energy management system optimization method based on a dynamic demand algorithm, which includes:

[0059] S100. Acquire historical energy management data, and perform data preprocessing on the historical energy management data to obtain target historical energy management data;

[0060] Specifically, in a possible design, in step S100, historical energy management data is obtained and data preprocessing is performed on the historical energy management data, including:

[0061] S1001. According to a preset time granularity, the historical energy data of each electrical device is obtained, and the historical energy data of each electrical device obtained is used as the historical energy management data;

[0062] S1002. De-duplication of equipment identifiers in historical energy management data, where the equipment identifiers are used to characterize the types of electrical equipment;

[0063] S1003. Based on the device identification, the deduplicated historical energy management data is summed according to the data type, and the summed result is used as the target historical energy management data, wherein the data type includes the number of devices and / or the device power consumption and device electricity consumption during peak, valley and normal periods.

[0064] Among them, the data types include the number of devices and / or the power and electricity consumption of devices during peak, valley and normal times. The number of devices refers to the number of devices of the same type (with the same device identification). The power and electricity consumption of devices during peak, valley and normal times refer to the power and electricity consumption data of devices during peak, valley, normal and peak periods in the peak, valley and normal calendars. The data types may also include the power and electricity consumption data of devices during holidays. These data types may be selected and set according to actual needs to ensure their rationality. In addition, the data types may also be selected according to needs, such as the working voltage and working current of the equipment.

[0065] The historical energy management data initially obtained all carry time information. For example, the power consumption and electricity consumption of equipment during peak and valley periods are selected and classified based on their time information.

[0066] S200. Clean the target historical energy management data, and arrange the cleaned target historical energy management data according to time information to generate historical energy management time series data;

[0067] Specifically, in a possible design, in step S200, the target historical energy management data includes a plurality of energy data sets, and each energy data set corresponds to a data type;

[0068] Among them, data cleaning of target historical energy management data includes:

[0069] S2001. According to the data type corresponding to each energy data set, select the designated cleaning action corresponding to each energy data set;

[0070] S2002. Execute corresponding designated cleaning actions on the target historical energy management data in each energy data set to obtain the target historical energy management data after data cleaning.

[0071] Among them, according to the data type corresponding to each energy data set, the designated cleaning actions corresponding to each energy data set are selected respectively. The designated cleaning actions here correspond to the data type corresponding to each energy data set. Different cleaning actions are selected according to the data type to ensure that the target historical energy management data after cleaning can be directly subjected to subsequent standardization processing without affecting key parameters such as mean and standard deviation. For example, for the missing values ​​of the equipment power consumption and equipment power consumption data during peak and valley periods, the first designated cleaning action (the first designated cleaning action is preset as: mean filling) is selected for cleaning; for the equipment working current, equipment working voltage and other data, the first designated cleaning action (the second designated cleaning action is preset as: performing outlier box plot analysis and performing outlier mean correction) is selected for cleaning.

[0072] S300. Standardize the historical energy management time series data to obtain standardized historical energy management time series data, and perform ADF stationarity test and processing on the standardized historical energy management time series data to obtain stable time series data;

[0073] Specifically, in a possible design, in step S300, the historical energy management time series data includes a plurality of energy time series data sets, and each energy time series data set corresponds to a data type;

[0074] Among them, the historical energy management time series data is standardized, including:

[0075] S3001. Calculate the data mean and data standard deviation of each energy time series data set;

[0076] S3002. Perform standardization on each energy time series data set according to the data mean and data standard deviation of each energy time series data set to obtain standardized historical energy management time series data.

[0077] It should be noted that, for each energy time series data set, preferably, the mean value calculation formula is:

[0078] 1ni=1nxi; (1)

[0079] The formula for calculating standard deviation is:

[0080] 1ni=1nxi-μ2; (2)

[0081] xi is the data in any energy time series data set, i is the guide number, μ is the data mean of any energy time series data set, and is the data standard deviation of any energy time series data set. Then the data of each energy time series data set is standardized, and the standardized calculation formula is:

[0082] x-μσ; (3)

[0083] xi is an element in x, Z is a standard data set after standardization of the data set x of any energy time series data set, and the mean of the elements in Z is 0 and the standard deviation is 1 (zero mean).

[0084] Through standardization processing, the standardized historical energy management time series data is distributed more evenly in each dimension.

[0085] By using the above formula (1) and formula (2), the data of each energy time series data set are calculated to obtain the data mean and data standard deviation of each energy time series data set.

[0086] Furthermore, each energy time series data set corresponds to a data type, and the data type here is consistent with the data type corresponding to each of the above energy data sets.

[0087] Accordingly, the standardized historical energy management time series data is subjected to an ADF stationarity test, including:

[0088] S3003. Obtain ADF test model;

[0089] S3004. Determine the autoregressive model of the standardized historical energy management time series data, and calculate the ADF test statistic and critical value of the standardized historical energy management time series data based on the autoregressive model and the ADF test model;

[0090] S3005. Determine whether the ADF test statistic of the standardized historical energy management time series data is less than or equal to the critical value;

[0091] S3006. If yes, determine that the standardized historical energy management time series data is a non-stationary sequence;

[0092] S3007. Perform differential processing on the standardized historical energy management time series data to obtain the differentially processed historical energy management time series data;

[0093] S3008. Update the standardized historical energy management time series data to the historical energy management time series data after difference processing, and re-determine the autoregressive model of the standardized historical energy management time series data until the ADF test statistic of the standardized historical energy management time series data is greater than the critical value, and obtain stable time series data.

[0094] Among them, the ADF (Augmented Dickey-Fuller) stationarity test is a time series analysis method, which is mainly used to determine whether a time series is stationary. The stationarity here refers to the time series mean and variance remain unchanged at different time points. Therefore, the ADF stationarity test is helpful for the subsequent selection of a suitable model for forecasting analysis. Specifically, it is assumed that the time series is non-stationary, and this hypothesis is used as the null hypothesis (H0), that is, the time series has a unit root; and the alternative hypothesis (H1) believes that the time series is stationary, that is, the time series does not have a unit root.

[0095] If the absolute value of the ADF test statistic is less than or equal to the critical value, the time series has a unit root and the null hypothesis (H0) is accepted; if the absolute value of the ADF test statistic is greater than the critical value, the time series does not have a unit root, the null hypothesis (H0) is rejected, and the alternative hypothesis (H1) is selected, indicating that the time series is stationary.

[0096] For the case of the null hypothesis (H0), it is necessary to perform difference processing (first-order difference) on the time series (standardized historical energy management time series data), and perform the ADF stationarity test again after the difference processing. If the absolute value of the ADF test statistic is greater than the critical value, proceed to the next step, otherwise perform difference processing (higher-order difference) again, and repeat the test until the absolute value of the ADF test statistic is greater than the critical value to reject the null hypothesis (H0), select the alternative hypothesis (H1), and proceed to the next step.

[0097] S400. Using the stable time series data, construct an energy management demand threshold prediction model;

[0098] Specifically, in a possible design, in step S400, the steady time series data is used to construct an energy management demand threshold prediction model, including:

[0099] S4001. The stationary time series data after ADF stationarity test and processing is divided into a training data set, a test data set and a validation data set according to a preset ratio;

[0100] S4002. Using each stationary time series data in the training data set as input and the energy management demand threshold as output to train the machine learning model, and during the training process, using the output of the machine learning model to adjust the parameters of the machine learning model, so as to obtain a pre-energy management demand threshold prediction model after the training is completed;

[0101] S4003. Using the test data set, test the pre-energy management demand threshold prediction model, obtain the test results, and determine whether the test results meet the preset conditions;

[0102] S4004. If not, the pre-energy management demand threshold prediction model is optimized until the test result meets the preset conditions, and the test energy management demand threshold prediction model is obtained;

[0103] S4005. Use the verification data set to verify the test energy management demand threshold prediction model to obtain a verification result, and when the verification result is verification passed, obtain the energy management demand threshold prediction model.

[0104] Among them, the stationary time series data is divided into a training data set, a test data set and a verification data set according to a preset ratio. The preset ratio can be set according to actual needs, and the verification data set can be 0. In general, preferably, the preset ratio is 7:2:1 to randomly split the data; and when training the machine learning model, a suitable linear kernel can be selected, and a corresponding penalty parameter can be selected for modeling to establish an SVR model (support vector regression model) to predict the demand threshold. Since the power consumption data of the factory is strongly correlated with the power consumption time, the number of power-consuming equipment, etc., this SVR model can provide better support, and its real-time performance is good, which is very suitable for the energy management demand threshold prediction model of this embodiment. The verification data set verifies the test energy management demand threshold prediction model to obtain a verification result. When the verification result fails, it is necessary to return to S4001 for readjustment until the verification result is verified to obtain the energy management demand threshold prediction model.

[0105] S500. Using the energy management demand threshold prediction model, calculate the real-time energy management data to obtain the real-time demand threshold prediction value.

[0106] S600. Based on the real-time demand threshold prediction value, determine the current demand threshold, and perform power control based on the current demand threshold

[0107] Specifically, in a possible design, in step S600, based on the real-time demand threshold prediction value, the demand threshold at the current moment is determined, including:

[0108] S6001. Get the demand threshold of the previous moment;

[0109] S6002. Compare the demand threshold at the previous moment with the real-time demand threshold prediction value;

[0110] S6003. If the demand threshold at the previous moment is greater than or equal to the real-time demand threshold prediction value, the real-time demand threshold prediction value is discarded, and the demand threshold at the previous moment is used as the demand threshold at the current moment;

[0111] S6004. If the demand threshold at the previous moment is less than the predicted value of the real-time demand threshold, the predicted value of the real-time demand threshold is used as the demand threshold at the current moment.

[0112] Finally, it should be noted that the above are only preferred embodiments of the present invention and are not intended to limit the scope of protection of the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the scope of protection of the present invention.

Claims

1. An energy management system optimization method based on a dynamic demand algorithm, characterized in that: include: Acquire historical energy management data, and perform data preprocessing on the historical energy management data to obtain target historical energy management data; Performing data cleaning on the target historical energy management data, and arranging the cleaned target historical energy management data according to time information to generate historical energy management time series data; Standardizing the historical energy management time series data to obtain standardized historical energy management time series data, and performing ADF stationarity test and processing on the standardized historical energy management time series data to obtain stationary time series data; Using the stable time series data, constructing an energy management demand threshold prediction model; Using the energy management demand threshold prediction model, the real-time energy management data is calculated to obtain a real-time demand threshold prediction value; Based on the real-time demand threshold prediction value, the demand threshold at the current moment is determined, so as to perform power control based on the demand threshold at the current moment.

2. The energy management system optimization method based on dynamic demand algorithm according to claim 1 is characterized in that: Obtain historical energy management data and perform data preprocessing on the historical energy management data, including: According to the preset time granularity, the historical energy data of each electrical device is obtained, and the historical energy data of each electrical device obtained is used as the historical energy management data; Deduplication processing is performed on the device identification in the historical energy management data, wherein the device identification is used to characterize the type of electrical equipment; Based on the device identification, the deduplicated historical energy management data is summed according to the data type, and the summed result is used as the target historical energy management data, wherein the data type includes the number of devices and / or the device power consumption and device electricity consumption during peak, valley and normal periods.

3. The energy management system optimization method based on dynamic demand algorithm according to claim 1 is characterized in that: The target historical energy management data includes a plurality of energy data sets, and each energy data set corresponds to a data type; The target historical energy management data is cleaned, including: According to the data type corresponding to each energy data set, the designated cleaning action corresponding to each energy data set is selected respectively; For the target historical energy management data in each energy data set, the corresponding designated cleaning actions are respectively performed to obtain the target historical energy management data after data cleaning.

4. The energy management system optimization method based on dynamic demand algorithm according to claim 1 is characterized in that: The historical energy management time series data includes a plurality of energy time series data sets, and each energy time series data set corresponds to a data type; The process of standardizing the historical energy management time series data includes: Calculate the data mean and data standard deviation of each energy time series data set; According to the data mean and data standard deviation of each energy time series data set, each energy time series data set is standardized to obtain the standardized historical energy management time series data.

5. The energy management system optimization method based on dynamic demand algorithm according to claim 4 is characterized in that: Perform ADF stationarity test on the standardized historical energy management time series data, including: Get the ADF test model; Determine an autoregressive model of the standardized historical energy management time series data, and calculate an ADF test statistic and a critical value of the standardized historical energy management time series data based on the autoregressive model and the ADF test model; Determine whether the ADF test statistic of the standardized historical energy management time series data is less than or equal to the critical value; If so, it is determined that the standardized historical energy management time series data is a non-stationary series; Performing differential processing on the standardized historical energy management time series data to obtain differentially processed historical energy management time series data; The standardized historical energy management time series data is updated to the historical energy management time series data after difference processing, and the autoregressive model of the standardized historical energy management time series data is re-determined until the ADF test statistic of the standardized historical energy management time series data is greater than the critical value, and the stable time series data is obtained.

6. The energy management system optimization method based on dynamic demand algorithm according to claim 1 is characterized in that: Using the stable time series data, an energy management demand threshold prediction model is constructed, including: The stationary time series data that has passed the ADF stationarity test and processing is divided into a training data set, a test data set, and a validation data set according to a preset ratio; The machine learning model is trained by taking the stationary time series data in the training data set as input and the energy management demand threshold as output. During the training process, the output of the machine learning model is used to adjust the parameters of the machine learning model so as to obtain a pre-energy management demand threshold prediction model after the training is completed. Use the test data set to test the pre-energy management demand threshold prediction model, obtain the test results, and determine whether the test results meet the preset conditions; If not, the pre-energy management demand threshold prediction model is optimized until the test result meets the preset conditions, and the test energy management demand threshold prediction model is obtained; The test energy management demand threshold prediction model is verified using the verification data set to obtain a verification result, and when the verification result is verification passed, the energy management demand threshold prediction model is obtained.

7. The energy management system optimization method based on dynamic demand algorithm according to claim 1 is characterized in that: Based on the real-time demand threshold prediction value, the demand threshold at the current moment is determined, including: Get the demand threshold value at the previous moment; Compare the demand threshold at the previous moment with the predicted value of the real-time demand threshold; If the demand threshold at the previous moment is greater than or equal to the real-time demand threshold prediction value, the real-time demand threshold prediction value is discarded, and the demand threshold at the previous moment is used as the demand threshold at the current moment; If the demand threshold at the previous moment is less than the real-time demand threshold prediction value, the real-time demand threshold prediction value is used as the demand threshold at the current moment.

8. An energy management system based on a dynamic demand algorithm, characterized in that: include: A data acquisition module acquires historical energy management data and performs data preprocessing on the historical energy management data to obtain target historical energy management data; A data processing module, used for performing data cleaning on the target historical energy management data, and arranging the cleaned target historical energy management data according to time information to generate historical energy management time series data; and also used for performing standardization processing on the historical energy management time series data to obtain standardized historical energy management time series data, and performing ADF stationarity test and processing on the standardized historical energy management time series data to obtain stationary time series data; A model training module, used to construct an energy management demand threshold prediction model using the stable time series data; A demand threshold updating module, used to calculate the real-time energy management data using the energy management demand threshold prediction model to obtain a real-time demand threshold prediction value; It is also used to determine the demand threshold at the current moment based on the real-time demand threshold prediction value, so as to perform power control based on the demand threshold at the current moment.

9. An electronic device, characterized in that: It includes a memory, a processor and a transceiver which are communicatively connected in sequence, wherein the memory is used to store a computer program, the transceiver is used to send and receive messages, and the processor is used to read the computer program to execute the energy management system optimization method based on the dynamic demand algorithm as described in any one of claims 1 to 7.

10. A computer program product comprising a computer program or instructions, characterized in that When the computer program or the instruction is executed by a computer, the energy management system optimization method based on a dynamic demand algorithm as described in any one of claims 1 to 7 is implemented.

Images