A virtual power plant control terminal capable of realizing new load monitoring functions

Abstract

The present invention relates to the field of data processing technology, and in particular to a virtual power plant control terminal that can realize a new load monitoring function. The present invention collects power load data to form a load monitoring data sequence; uses an initial load forecasting model to predict to obtain an initial load forecasting data sequence, and obtains the abnormality of the load monitoring data sequence based on the data difference between the initial load forecasting data sequence and the load monitoring data sequence; when the abnormality of the load monitoring data sequence is greater than a set abnormality threshold, obtains the stability of the load monitoring data sequence, optimizes the initial load forecasting model, obtains an optimized load forecasting data sequence, and performs load abnormality monitoring on the power system based on the data difference between the optimized load forecasting data sequence and the load monitoring data sequence. The present invention optimizes the initial load forecasting model according to the stability change characteristics of the load monitoring data, so that the monitoring results of the power system are more accurate.

Description

Technical Field

[0001] The present invention relates to the technical field of data processing, and in particular to a virtual power plant control terminal capable of realizing a novel load monitoring function. Background Art

[0002] A virtual power plant that can realize new load monitoring functions is a system that optimizes power supply and demand by integrating and coordinating multiple distributed energy resources. The virtual power plant control terminal integrates and optimizes multiple energy resources to provide flexible and intelligent power resource scheduling solutions, which not only improves the intelligence level of power management, but also supports the efficient, reliable and economical operation of the energy system.

[0003] The existing virtual power plant control terminal mainly establishes a load forecasting model by analyzing historical load monitoring data, and realizes the dispatch and control of power resources according to the established load forecasting model. However, due to factors such as changes in factory production demand, only the established load forecasting model is used to identify abnormal load data in the power load monitoring process, which has poor adaptability to actual power demand and regular changes. As a result, the established load forecasting model cannot always conform to the actual abnormal load data characteristics and regularities, and it is easy to mistakenly identify the actual power demand adjusted by humans as abnormal load data, resulting in inaccurate monitoring results and affecting the normal operation of factories and other systems.

[0004] Therefore, how to optimize the load forecasting model to make the monitoring results of the power system more accurate has become an urgent problem to be solved. Summary of the invention

[0005] In view of this, an embodiment of the present invention provides a virtual power plant control terminal that can realize a new load monitoring function to solve the problem of how to optimize the load forecasting model to make the monitoring results of the power system more accurate.

[0006] In an embodiment of the present invention, a virtual power plant control terminal capable of realizing a novel load monitoring function is provided, comprising a memory, a processor, and a computer program stored in the memory and running on the processor, wherein the processor implements the following method when executing the computer program:

[0007] For any power system monitored by the virtual power plant, collect the power load data of the power system at each sampling time within a preset period to form a load monitoring data sequence;

[0008] Obtaining an initial load forecasting model, using the initial load forecasting model to forecast the power load data within a preset time period to obtain an initial load forecasting data sequence, and obtaining the abnormality degree of the load monitoring data sequence according to the data difference between the initial load forecasting data sequence and the load monitoring data sequence;

[0009] When the abnormality of the load monitoring data sequence is greater than a set abnormality threshold, the stability of the load monitoring data sequence is obtained according to the power load data fluctuation of the load monitoring data sequence, and the initial load forecasting model is optimized by using the stability of the load monitoring data sequence to obtain an optimized load forecasting model;

[0010] An optimized load forecasting model is used to predict the power load data within a preset time period to obtain an optimized load forecasting data sequence. Based on the data difference between the optimized load forecasting data sequence and the load monitoring data sequence, an abnormality possibility index of each power load data in the load monitoring data sequence is obtained. Based on the abnormality possibility index of each power load data in the load monitoring data sequence, load abnormality monitoring of the power system is performed.

[0011] Preferably, after obtaining the abnormality degree of the load monitoring data sequence, the method further includes:

[0012] When the abnormality degree of the load monitoring data sequence is less than or equal to the set abnormality degree threshold, the abnormality possibility index of each power load data in the load monitoring data sequence is obtained according to the data difference between the initial load forecast data sequence and the load monitoring data sequence, and the power system is subjected to load abnormality monitoring according to the abnormality possibility index of each power load data in the load monitoring data sequence.

[0013] Preferably, obtaining the abnormality degree of the load monitoring data sequence according to the data difference between the initial load forecasting data sequence and the load monitoring data sequence includes:

[0014] For any power load data in the load monitoring data sequence, obtain the initial predicted power load data corresponding to the sampling time of the power load data in the initial load prediction data sequence, calculate the absolute value of the difference between the power load data and the initial predicted power load data, and record it as the prediction difference index of the power load data;

[0015] Obtain a prediction difference index for each electric load data in the load monitoring data sequence, accumulate all the prediction difference indexes to obtain a corresponding accumulated value, normalize the accumulated value, and obtain the abnormality degree of the load monitoring data sequence.

[0016] Preferably, obtaining the stability of the load monitoring data sequence according to the power load data fluctuation of the load monitoring data sequence includes:

[0017] For any power load data in the load monitoring data sequence, a time window of a preset length with the power load data as the starting moment is constructed, and a stability index of the time window corresponding to the power load data is obtained according to the power load data in the time window;

[0018] According to the fluctuation of the power load data in the time window corresponding to the power load data, obtaining a normal possibility index of the power load data;

[0019] Obtain a normal possibility index for each power load data in the load monitoring data sequence and a stability index for the corresponding time window, and obtain the stability degree of the load monitoring data sequence based on the normal possibility index for each power load data in the load monitoring data sequence and the stability index for the corresponding time window.

[0020] Preferably, the step of obtaining the stability index of the time window corresponding to the power load data according to the power load data in the time window includes:

[0021] Obtain the standard deviation of the power load data in the time window corresponding to the power load data, calculate a first addition result of the standard deviation and a constant 1, and obtain the inverse of the first addition result as a stability indicator of the time window corresponding to the power load data.

[0022] Preferably, obtaining the normal possibility index of the power load data according to the fluctuation of the power load data in the time window corresponding to the power load data includes:

[0023] When the stability index of the time window corresponding to the power load data is greater than or equal to a preset stability index threshold, the normal possibility index of the power load data is a preset value;

[0024] When the stability index of the time window corresponding to the power load data is less than a preset stability index threshold, the difference between every two adjacent power load data in the time window corresponding to the power load data is calculated respectively to obtain a corresponding difference sequence, the absolute value of the difference between every two adjacent differences in the difference sequence is obtained respectively to obtain the accumulated value of the absolute value of the difference, the second addition result of the accumulated value of the absolute value of the difference and the constant 1 is calculated, and the inverse of the second addition result is normalized to obtain the normal possibility index of the power load data.

[0025] Preferably, obtaining the stability of the load monitoring data sequence according to the normal possibility index of each power load data in the load monitoring data sequence and the stability index of the corresponding time window includes:

[0026] The addition results between the normal possibility index of each power load data in the load monitoring data sequence and the stability index of the corresponding time window are respectively obtained, all the addition results are accumulated to obtain the cumulative value of the addition results, the cumulative value of the addition results is normalized to obtain the stability of the load monitoring data sequence.

[0027] Preferably, the optimizing the initial load forecasting model by utilizing the stability of the load monitoring data sequence to obtain an optimized load forecasting model comprises:

[0028] The lag coefficient of the initial load forecasting model is obtained as the first lag coefficient, and a real-time forecasting model is constructed according to the power load data in the load monitoring data sequence, and a corresponding second lag coefficient is obtained. The stability of the load monitoring data sequence is used as the weight coefficient of the second lag coefficient, and the difference between a constant 1 and the stability of the load monitoring data sequence is used as the weight coefficient of the first lag coefficient. The first lag coefficient and the second lag coefficient are weightedly summed, and the weighted sum result is used as the optimized lag coefficient. The lag coefficient of the initial load forecasting model is adjusted to the optimized lag coefficient, and an optimized load forecasting model is obtained.

[0029] Preferably, obtaining the abnormal possibility index of each power load data in the load monitoring data sequence according to the data difference between the optimized load forecasting data sequence and the load monitoring data sequence includes:

[0030] For any power load data in the load monitoring data sequence, obtain the optimized predicted power load data corresponding to the sampling time of the power load data in the optimized load prediction data sequence, calculate the absolute value of the difference between the power load data and the optimized predicted power load data, normalize the absolute value of the difference, and use the normalized result as the abnormal possibility indicator of the power load data.

[0031] Preferably, the performing load anomaly monitoring on the power system according to the anomaly possibility index of each power load data in the load monitoring data sequence includes:

[0032] When the abnormality possibility indicators of all the power load data in the load monitoring data sequence are less than the set abnormality possibility indicator threshold, it is determined that the power system is in a normal state within the preset time period; when the abnormality possibility indicator of any one of the power load data in the load monitoring data sequence is greater than or equal to the set abnormality possibility indicator threshold, it is determined that the power system has a load abnormality within the preset time period.

[0033] Compared with the prior art, the embodiments of the present invention have the following beneficial effects:

[0034] The present invention aims at any power system monitored by a virtual power plant, collects power load data of the power system at each sampling moment within a preset time period to form a load monitoring data sequence; obtains an initial load prediction model, uses the initial load prediction model to predict the power load data within the preset time period, obtains the initial load prediction data sequence, and obtains the abnormality of the load monitoring data sequence according to the data difference between the initial load prediction data sequence and the load monitoring data sequence; when the abnormality of the load monitoring data sequence is greater than a set abnormality threshold, obtains the stability of the load monitoring data sequence according to the power load data fluctuation of the load monitoring data sequence, and optimizes the initial load prediction model according to the stability of the load monitoring data sequence to obtain an optimized load prediction model; uses the optimized load prediction model to predict the power load data within the preset time period to obtain an optimized load prediction data sequence, obtains the abnormality possibility index of each power load data in the load monitoring data sequence according to the data difference between the optimized load prediction data sequence and the load monitoring data sequence, and performs load abnormality monitoring on the power system according to the abnormality possibility index of each power load data in the load monitoring data sequence. The present invention analyzes the stability and change characteristics of abnormal load data and artificial scheduling data, distinguishes between power load data changes caused by artificial scheduling data and power load data changes caused by real abnormalities, and optimizes the load forecasting model according to the stability change characteristics of load monitoring data to obtain more accurate forecasting results, thereby improving the adaptability of the forecasting model to actual changes in electricity demand and making the monitoring results of the power system more accurate. BRIEF DESCRIPTION OF THE DRAWINGS

[0035] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, 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 only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative labor.

[0036] Figure 1 It is a method flow chart of a virtual power plant control terminal that can realize a new load monitoring function provided by the first embodiment of the present invention. DETAILED DESCRIPTION

[0037] Embodiments of the present disclosure are described in detail below, and examples of the embodiments are shown in the accompanying drawings. The embodiments described below with reference to the accompanying drawings are exemplary and are intended to be used to explain the present disclosure, but should not be understood as limiting the present disclosure.

[0038] It should be noted that the terms "first", "second", etc. in the specification of the present disclosure and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence. It should be understood that the data used in this way can be interchangeable where appropriate, so that the embodiments of the present disclosure described herein can be implemented in an order other than those illustrated or described herein. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with the present disclosure. Instead, they are merely examples of devices and methods consistent with some aspects of the present disclosure.

[0039] In order to illustrate the technical solution of the present invention, specific embodiments are provided below for illustration.

[0040] See also Figure 1 , is a flowchart of a method for a virtual power plant control terminal capable of realizing a new load monitoring function provided by the first embodiment of the present invention, such as Figure 1 As shown, the virtual power plant control terminal capable of realizing a novel load monitoring function includes a memory, a processor, and a computer program stored in the memory and running on the processor. When the processor executes the computer program, the following method is implemented:

[0041] Step S101, for any power system monitored by the virtual power plant, collect power load data of the power system at each sampling moment within a preset time period to form a load monitoring data sequence.

[0042] Taking one of the power systems monitored by the virtual power plant as an example, the virtual power plant control terminal monitors the power load data every 1 second, and collects the power load data within 2 minutes to form a load monitoring data sequence. There is no restriction here, and it can be set according to the specific implementation scenario. Power demand or power generation data is usually expressed in power, in kilowatts (kW), so the power load data monitored every 1 second is power data.

[0043] Step S102, obtain an initial load forecasting model, use the initial load forecasting model to predict the power load data within a preset time period, obtain an initial load forecasting data sequence, and obtain the abnormality degree of the load monitoring data sequence based on the data difference between the initial load forecasting data sequence and the load monitoring data sequence.

[0044] Under normal conditions, according to the requirements and processes of the power system, the power consumption of the power system should be consistent with the production process, that is, it has a certain change pattern over time. Therefore, the initial load forecasting model of the power system is usually generated based on the power load data of the power system using the ARIMA model. It is used to judge the load anomaly of the power load data in the load monitoring data sequence according to the predicted value.

[0045] In one embodiment, the initial load forecasting model is generated based on the historical power load data before the load monitoring data sequence as the training data of the ARIMA model. Because the ARIMA model is an existing method, a brief description of the steps is given: (1) the historical power load data is plotted into a "time-power" scatter plot in chronological order; (2) the "time-power" scatter plot is fitted using the least squares method to obtain a fitting curve, which can show the change law of the historical power load data of the power system; (3) the obtained fitting curve is subjected to an ADF test to determine whether the historical power load data is stable. If it is not stable, differential processing is performed until stable historical power load data is obtained and the number of differentials is obtained; (4) the parameters of the ARIMA model are obtained based on the autocorrelation coefficient diagram and partial autocorrelation coefficient diagram of the historical power load data; (5) based on the parameters and the number of differentials of the ARIMA model, an initial load forecasting model based on the historical power load data is generated. The ADF test belongs to the prior art and will not be repeated here.

[0046] After obtaining the initial load forecasting model, the initial load forecasting model is used to predict the power load data at each sampling moment in the load monitoring data sequence, and the initial predicted power load data corresponding to the sampling moment of the power load data is obtained to form an initial load forecasting data sequence. Then, the abnormality degree of the load monitoring data sequence is obtained according to the data difference between the initial load forecasting data sequence and the load monitoring data sequence, which is used to judge the load abnormality of the power system in the time period corresponding to the load monitoring data sequence.

[0047] Specifically, for any power load data in the load monitoring data sequence, the initial predicted power load data corresponding to the sampling time of the power load data in the initial load prediction data sequence is obtained, and the absolute value of the difference between the power load data and the initial predicted power load data is calculated, which is recorded as the prediction difference index of the power load data;

[0048] Obtain a prediction difference index for each electric load data in the load monitoring data sequence, accumulate all the prediction difference indexes to obtain a corresponding accumulated value, normalize the accumulated value, and obtain the abnormality degree of the load monitoring data sequence.

[0049] In one embodiment, the i-th power load data of the load monitoring data sequence is Take the initial load forecast data sequence as an example to obtain The initial predicted power load data corresponding to the sampling time , calculate the abnormality degree of the load monitoring data sequence:

[0050]

[0051] Where E is the abnormality degree of the load monitoring data sequence; is the i-th power load data of the load monitoring data sequence; For The initial predicted power load data in the initial load prediction data sequence corresponding to the sampling time of; i is the sequence number of the power load data in the load monitoring data sequence; is the total number of power load data in the load monitoring data sequence; is the normalization function; | | is the absolute value symbol.

[0052] It should be noted that the initial predicted power load data is the predicted value of the initial load forecasting model for the power load data in the load monitoring data sequence. The greater the difference between the initial predicted power load data and the corresponding power load data, the less the power load data conforms to the prediction result of the initial load forecasting model, resulting in a greater prediction difference index of the power load data, that is, The larger the value of is, the higher the abnormality of the corresponding load monitoring data sequence is.

[0053] After obtaining the abnormality degree of the load monitoring data sequence, considering that the initial predicted power load data and the corresponding power load data will not be exactly the same, the abnormality threshold is set to 0.3. There is no restriction here and it can be set according to the specific implementation scenario. When the abnormality degree E of the load monitoring data sequence is less than or equal to 0.3, it is considered that the change law of the load monitoring data sequence conforms to the initial load forecasting model, and the power system can be directly monitored for load abnormalities based on the initial load forecasting data sequence and the load monitoring data sequence.

[0054] Specifically, based on the data difference between the initial load forecast data sequence and the load monitoring data sequence, an abnormality possibility index of each power load data in the load monitoring data sequence is obtained, and load abnormality monitoring is performed on the power system based on the abnormality possibility index of each power load data in the load monitoring data sequence.

[0055] In one implementation, taking the ith power load data in the load monitoring data sequence as an example, the abnormal possibility index of the ith power load data in the load monitoring data sequence is calculated:

[0056]

[0057] in, is the abnormal possibility index of the ith power load data in the load monitoring data sequence; is the initial predicted electric load data corresponding to the sampling time of the i-th electric load data in the initial load prediction data sequence; is the i-th power load data in the load monitoring data sequence; i is the sequence number of the power load data in the load monitoring data sequence; is the normalization function; | | is the absolute value symbol.

[0058] It should be noted that the abnormal possibility index represents the abnormal possibility of the power load data in the load monitoring data sequence. It ensures that the abnormal possibility of the power load data in the load monitoring data sequence can only be positive, and the absolute value of the difference between the i-th power load data in the load monitoring data sequence and the corresponding initial predicted power load data The larger the value is, the higher the abnormal possibility index of the i-th power load data in the load monitoring data sequence is.

[0059] Similarly, according to the method for obtaining the abnormal possibility index of the i-th power load data in the load monitoring data sequence, the abnormal possibility index of each power load data in the load monitoring data sequence is obtained respectively. After that, the load abnormality monitoring of the power system can be performed according to the abnormal possibility index of each power load data in the load monitoring data sequence. The specific process is:

[0060] The abnormal possibility index threshold is set to 0.4. There is no restriction here and it can be set according to the specific implementation scenario. When the abnormal possibility index of all power load data in the load monitoring data sequence is less than 0.4, it is considered that the power system is in a normal state within the preset time period (that is, 2 minutes); when the abnormal possibility index of any power load data in the load monitoring data sequence is greater than or equal to 0.4, it is considered that the power system has a load abnormality within the preset time period (that is, 2 minutes).

[0061] At this point, the load anomaly monitoring of the load monitoring data sequence in the power system whose anomaly degree is less than or equal to the set anomaly degree threshold can be completed.

[0062] When the abnormality degree E of the load monitoring data sequence is greater than the set abnormality degree threshold of 0.3, it is considered that the change pattern of the load monitoring data sequence does not conform to the initial load forecasting model. At this time, the initial load forecasting model needs to be optimized to perform load abnormality monitoring on the load monitoring data sequence in the power system whose abnormality degree is greater than the set abnormality degree threshold.

[0063] Step S103, when the abnormality level of the load monitoring data sequence is greater than the set abnormality level threshold, the stability level of the load monitoring data sequence is obtained according to the power load data fluctuation of the load monitoring data sequence, and the initial load forecasting model is optimized using the stability level of the load monitoring data sequence to obtain an optimized load forecasting model.

[0064] When the changing pattern of the load monitoring data sequence does not conform to the initial load forecasting model, it means that the power load data in the load monitoring data sequence may be abnormal, but these anomalies may be caused by power system failures, or changes in power demand. When the power system is operating normally, the power load data generally fluctuates stably within a small range. If the power system makes artificial adjustments to power consumption, such as adjusting the intensity of the cooling system, the power load data will also change, but this change will often quickly stabilize within a new range; and when the power system is abnormal, the power load data may show a continuous increase or decrease, or at certain moments there may be abnormal sudden changes in the power load data. Therefore, it is necessary to analyze the stability of the load monitoring data sequence to determine whether the abnormality in the power load data in the load monitoring data sequence is caused by power system failures or changes in power demand.

[0065] The method for obtaining the stability of the load monitoring data sequence includes:

[0066] (1) For any power load data in the load monitoring data sequence, a time window of a preset length with the power load data as the starting time is constructed, and a stability index of the time window corresponding to the power load data is obtained based on the power load data in the time window.

[0067] When adjusting electricity demand, the power load data will not change suddenly immediately, but will gradually increase or decrease to the demand range. Therefore, during the adjustment of the power load data, it may also show abnormal performance similar to continuous increase or decrease in power. In order to avoid misidentifying the normal load monitoring data sequence as an abnormal load monitoring data sequence due to such changes, for any power load data in the load monitoring data sequence, a time window with the power load data as the starting moment and a preset duration of 5s is constructed. There is no restriction here and it can be set according to the specific implementation scenario, and then the stability index of the power load data is obtained according to the power load data in the time window.

[0068] Obtain the stability index of the time window corresponding to the power load data in the load monitoring data sequence, including:

[0069] Obtain the standard deviation of the power load data in the time window corresponding to the power load data, calculate a first addition result of the standard deviation and a constant 1, and obtain the inverse of the first addition result as a stability indicator of the time window corresponding to the power load data.

[0070] In one implementation, taking the ith power load data in the load monitoring data sequence as an example, the stability index of the time window corresponding to the ith power load data is calculated:

[0071]

[0072] in, is the stability index of the time window corresponding to the i-th power load data in the load monitoring data sequence; is the standard deviation of the power load data in the time window corresponding to the i-th power load data in the load monitoring data sequence; i is the sequence number of the power load data in the load monitoring data sequence.

[0073] It should be noted that Ensure that the stability index is meaningful; the standard deviation of the power load data in the time window corresponding to the i-th power load data in the load monitoring data sequence The larger the value is, the more unstable the power load data in the time window corresponding to the i-th power load data is. The stability index of the time window corresponding to the i-th power load data in the load monitoring data sequence is The smaller.

[0074] (2) Obtaining a normal possibility index for the power load data based on fluctuations of the power load data in a time window corresponding to the power load data.

[0075] When the stability index of the time window corresponding to the power load data is greater than or equal to a preset stability index threshold, the normal possibility index of the power load data is a preset value;

[0076] When the stability index of the time window corresponding to the power load data is less than a preset stability index threshold, the difference between every two adjacent power load data in the time window corresponding to the power load data is calculated respectively to obtain a corresponding difference sequence, the absolute value of the difference between every two adjacent differences in the difference sequence is obtained respectively to obtain the accumulated value of the absolute value of the difference, the second addition result of the accumulated value of the absolute value of the difference and the constant 1 is calculated, and the inverse of the second addition result is normalized to obtain the normal possibility index of the power load data.

[0077] In one embodiment, 0.5 is a preset stability index threshold, which is not limited here and can be set according to the specific implementation scenario. When the stability index of the time window corresponding to the power load data in the load monitoring data sequence is less than 0.5, it is considered that the stability in the time window corresponding to the power load data in the load monitoring data sequence is poor.

[0078] When the stability index of the time window corresponding to the power load data in the load monitoring data sequence is greater than or equal to 0.5, the normal possibility index of the power load data is considered to be 1. 1 is the preset value of the normal possibility index. There is no restriction here and it can be set according to the specific implementation scenario.

[0079] When the power consumption is adjusted manually, the power load data will first show a gradual increase or decrease, so this abnormal performance is lower than the abnormal degree of large fluctuations or mutations. Therefore, when the stability index of the time window corresponding to the power load data in the load monitoring data sequence is less than 0.5, for the power load data whose stability index of the time window corresponding to the power load data is less than the preset stability index threshold, it is necessary to calculate the normal possibility index of the power load data according to the fluctuation of the power load data in the window corresponding to the power load data. The calculation expression of the normal possibility index of the power load data is:

[0080]

[0081] in, is the normal possibility index of the i-th power load data in the load monitoring data sequence; represents the number of electric load data contained in the time window corresponding to the i-th electric load data in the load monitoring data sequence; i is the sequence number of the electric load data in the load monitoring data sequence; j is the sequence number of the electric load data in the time window corresponding to the i-th electric load data; represents the jth power load data in the time window corresponding to the i-th power load data in the load monitoring data sequence; represents the j+1th power load data in the time window corresponding to the i-th power load data in the load monitoring data sequence; represents the j-1th power load data in the time window corresponding to the i-th power load data in the load monitoring data sequence; is the normalization function; |.| is the absolute value symbol.

[0082] It should be noted that It indicates the fluctuation degree of the power load data in the time window corresponding to the i-th power load data in the load monitoring data sequence. The greater the fluctuation degree, the smaller the normal possibility index.

[0083] (3) Obtaining a normal possibility index for each power load data in the load monitoring data sequence and a stability index for the corresponding time window, and obtaining the stability degree of the load monitoring data sequence based on the normal possibility index for each power load data in the load monitoring data sequence and the stability index for the corresponding time window.

[0084] Specifically, the addition results between the normal possibility index of each power load data in the load monitoring data sequence and the stability index of the corresponding time window are obtained respectively, all the addition results are accumulated to obtain the cumulative value of the addition results, the cumulative value of the addition results is normalized to obtain the stability of the load monitoring data sequence.

[0085] In one embodiment, the stability of the load monitoring data series is calculated:

[0086]

[0087] Where T is the stability of the load monitoring data sequence; is the normalization function; The stability index of the time window corresponding to the i-th power load data in the load monitoring data sequence; is the normal possibility index of the i-th power load data in the load monitoring data sequence; i is the sequence number of the power load data in the load monitoring data sequence; It is the total number of power load data in the load monitoring data sequence.

[0088] It should be noted that if the normal possibility index of any power load data in the load monitoring data sequence or the stability index of the corresponding time window increases, the stability of the load monitoring data sequence will increase.

[0089] Since changes in actual electricity demand and abnormal load in the power system will cause the change pattern of the load monitoring data sequence to be inconsistent with the initial load forecasting model, but the load anomaly caused by changes in actual electricity demand means that the initial load forecasting model is not applicable to the current electricity consumption pattern, the initial load forecasting model needs to be corrected according to the stability of the load monitoring data sequence monitored in real time to improve the robustness of the initial load forecasting model to different electricity demands.

[0090] Specifically, the initial load forecasting model is optimized by using the stability of the load monitoring data sequence to obtain an optimized load forecasting model, including:

[0091] The lag coefficient of the initial load forecasting model is obtained as the first lag coefficient, and a real-time forecasting model is constructed according to the power load data in the load monitoring data sequence, and a corresponding second lag coefficient is obtained. The stability of the load monitoring data sequence is used as the weight coefficient of the second lag coefficient, and the difference between a constant 1 and the stability of the load monitoring data sequence is used as the weight coefficient of the first lag coefficient. The first lag coefficient and the second lag coefficient are weightedly summed, and the weighted sum result is used as the optimized lag coefficient. The lag coefficient of the initial load forecasting model is adjusted to the optimized lag coefficient, and an optimized load forecasting model is obtained.

[0092] In one embodiment, the first lag coefficient is first obtained by the initial load forecasting model, and then the power load data of the load monitoring data sequence is used as the training data of the ARIMA model to obtain the real-time forecasting model, and the second lag coefficient is obtained by the real-time forecasting model, and then the first lag coefficient is optimized according to the stability of the load monitoring data sequence, the first lag coefficient and the second lag coefficient to obtain the optimized lag coefficient. The formula for calculating the optimized lag coefficient is:

[0093]

[0094] in, is to optimize the hysteresis coefficient; T is the stability of the load monitoring data sequence; is the first lag coefficient of the initial load forecasting model; The second lag coefficient of the real-time prediction model constructed for the power load data in the load monitoring data series.

[0095] It should be noted that the stability of the load monitoring data sequence can reflect the possibility that the abnormality of the power load data in the load monitoring data sequence is caused by human adjustment. The greater the stability of the load monitoring data sequence, the greater the proportion of the second lag coefficient of the real-time prediction model corresponding to the load monitoring data sequence in the optimized lag coefficient, which means that the possibility of the abnormality of the power load data in the load monitoring data sequence is caused by human adjustment is greater. The smaller the stability of the load monitoring data sequence, the smaller the proportion of the second lag coefficient of the real-time prediction model corresponding to the load monitoring data sequence in the optimized lag coefficient, which means that the possibility of the abnormality of the power load data in the load monitoring data sequence is caused by human adjustment is smaller, that is, the possibility that the abnormality of the power load data in the load monitoring data sequence is caused by real abnormality is greater.

[0096] After the optimized hysteresis coefficient is obtained, the first hysteresis coefficient of the initial load forecasting model is adjusted to the optimized hysteresis coefficient to obtain the optimized load forecasting model.

[0097] At this point, the optimized load forecasting model is obtained through the load monitoring data sequence whose abnormality level is greater than the set abnormality level threshold.

[0098] Step S104, using the optimized load forecasting model to predict the power load data within a preset time period to obtain an optimized load forecasting data sequence, and based on the data difference between the optimized load forecasting data sequence and the load monitoring data sequence, obtain the abnormality possibility index of each power load data in the load monitoring data sequence, and perform load abnormality monitoring on the power system based on the abnormality possibility index of each power load data in the load monitoring data sequence.

[0099] After obtaining the optimized load forecasting model, the optimized load forecasting model is used to predict the power load data at each sampling moment in the load monitoring data sequence, and the optimized predicted power load data corresponding to the sampling moment of the power load data is obtained to form an optimized load forecasting data sequence. Then, based on the data difference between the optimized load forecasting data sequence and the load monitoring data sequence, the abnormal possibility index of the power load data in the load monitoring data sequence is obtained.

[0100] Specifically, according to the data difference between the optimized load forecasting data sequence and the load monitoring data sequence, the abnormal possibility index of each power load data in the load monitoring data sequence is obtained, including:

[0101] For any power load data in the load monitoring data sequence, obtain the optimized predicted power load data corresponding to the sampling time of the power load data in the optimized load prediction data sequence, calculate the absolute value of the difference between the power load data and the optimized predicted power load data, normalize the absolute value of the difference, and use the normalized result as the abnormal possibility indicator of the power load data.

[0102] In one implementation, taking the ith power load data in the load monitoring data sequence as an example, the abnormal possibility index of the ith power load data in the load monitoring data sequence is calculated:

[0103]

[0104] Among them, V is the abnormal possibility index of the i-th power load data in the load monitoring data sequence; The optimized predicted electric load data corresponding to the sampling time of the i-th electric load data in the optimized load prediction data sequence; is the i-th power load data in the load monitoring data sequence; i is the sequence number of the power load data in the load monitoring data sequence; is the normalization function; | | is the absolute value symbol.

[0105] It should be noted that the abnormal possibility index represents the abnormal possibility of the power load data in the load monitoring data sequence. It ensures that the abnormal possibility of power load data in the load monitoring data sequence can only be positive, and optimizes the absolute value of the difference between the predicted power load data and the corresponding power load data. The larger the value is, the higher the abnormal possibility index of the i-th power load data in the load monitoring data sequence is.

[0106] Similarly, according to the method for obtaining the abnormal possibility index of the i-th power load data in the load monitoring data sequence, the abnormal possibility index of each power load data in the load monitoring data sequence is obtained respectively. After that, the load abnormality monitoring of the power system can be performed according to the abnormal possibility index of each power load data in the load monitoring data sequence. The specific process is:

[0107] The abnormal possibility index threshold is set to 0.4. There is no restriction here and it can be set according to the specific implementation scenario. When the abnormal possibility index of all power load data in the load monitoring data sequence is less than 0.4, it is considered that the power system is in a normal state within the preset time period (that is, 2 minutes); when the abnormal possibility index of any power load data in the load monitoring data sequence is greater than or equal to 0.4, it is considered that the power system has a load abnormality within the preset time period (that is, 2 minutes).

[0108] At this point, the load anomaly monitoring of the load monitoring data sequence in the power system whose anomaly degree is greater than the set anomaly degree threshold is completed.

[0109] In summary, this embodiment targets any power system monitored by a virtual power plant, collects power load data of the power system at each sampling moment within a preset time period, and forms a load monitoring data sequence; obtains an initial load forecasting model, and uses the initial load forecasting model to predict the power load data within the preset time period to obtain an initial load forecasting data sequence, and obtains the abnormality of the load monitoring data sequence according to the data difference between the initial load forecasting data sequence and the load monitoring data sequence; when the abnormality of the load monitoring data sequence is greater than a set abnormality threshold, the stability of the load monitoring data sequence is obtained according to the fluctuation of the power load data of the load monitoring data sequence, and the initial load forecasting model is optimized according to the stability of the load monitoring data sequence to obtain an optimized load forecasting model; uses the optimized load forecasting model to predict the power load data within the preset time period to obtain an optimized load forecasting data sequence, and obtains an abnormality possibility index of each power load data in the load monitoring data sequence according to the data difference between the optimized load forecasting data sequence and the load monitoring data sequence, and performs load abnormality monitoring on the power system according to the abnormality possibility index of each power load data in the load monitoring data sequence. This embodiment analyzes the stability and change characteristics of abnormal load data and artificial scheduling data, distinguishes between power load data changes caused by artificial scheduling data and real abnormal power load data, and optimizes the load forecasting model according to the stability change characteristics of load monitoring data to obtain more accurate forecasting results, thereby improving the adaptability of the forecasting model to actual changes in electricity demand and making the monitoring results of the power system more accurate.

[0110] The above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit the same. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that the technical solutions described in the aforementioned embodiments may still be modified, or some of the technical features may be replaced by equivalents. Such 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, and should be included in the protection scope of the present invention.

Claims

1. A virtual power plant control terminal capable of realizing a new type of load monitoring function, comprising a memory, a processor, and a computer program stored in the memory and running on the processor, characterized in that: When the processor executes the computer program, the following method is implemented: For any power system monitored by the virtual power plant, collect the power load data of the power system at each sampling time within a preset period to form a load monitoring data sequence; Obtaining an initial load forecasting model, using the initial load forecasting model to forecast the power load data within a preset time period to obtain an initial load forecasting data sequence, and obtaining the abnormality degree of the load monitoring data sequence according to the data difference between the initial load forecasting data sequence and the load monitoring data sequence; When the abnormality of the load monitoring data sequence is greater than a set abnormality threshold, the stability of the load monitoring data sequence is obtained according to the power load data fluctuation of the load monitoring data sequence, and the initial load forecasting model is optimized by using the stability of the load monitoring data sequence to obtain an optimized load forecasting model; The power load data within a preset time period is predicted using an optimized load prediction model to obtain an optimized load prediction data sequence, an abnormality possibility index of each power load data in the load monitoring data sequence is obtained according to the data difference between the optimized load prediction data sequence and the load monitoring data sequence, and load abnormality monitoring of the power system is performed according to the abnormality possibility index of each power load data in the load monitoring data sequence; The obtaining the stability of the load monitoring data sequence according to the power load data fluctuation of the load monitoring data sequence comprises: For any power load data in the load monitoring data sequence, a time window of a preset length with the power load data as the starting moment is constructed, and a stability index of the time window corresponding to the power load data is obtained according to the power load data in the time window; According to the fluctuation of the power load data in the time window corresponding to the power load data, obtaining a normal possibility index of the power load data; Obtaining a normal possibility index of each electric load data in the load monitoring data sequence and a stability index of a corresponding time window, and obtaining the stability degree of the load monitoring data sequence according to the normal possibility index of each electric load data in the load monitoring data sequence and the stability index of the corresponding time window; The optimizing the initial load forecasting model by utilizing the stability of the load monitoring data sequence to obtain an optimized load forecasting model comprises: The lag coefficient of the initial load forecasting model is obtained as the first lag coefficient, and a real-time forecasting model is constructed according to the power load data in the load monitoring data sequence, and a corresponding second lag coefficient is obtained. The stability of the load monitoring data sequence is used as the weight coefficient of the second lag coefficient, and the difference between a constant 1 and the stability of the load monitoring data sequence is used as the weight coefficient of the first lag coefficient. The first lag coefficient and the second lag coefficient are weightedly summed, and the weighted sum result is used as the optimized lag coefficient. The lag coefficient of the initial load forecasting model is adjusted to the optimized lag coefficient, and an optimized load forecasting model is obtained.

2. The virtual power plant control terminal capable of realizing a new type of load monitoring function according to claim 1 is characterized in that: After obtaining the abnormality degree of the load monitoring data sequence, the method further includes: When the abnormality degree of the load monitoring data sequence is less than or equal to the set abnormality degree threshold, the abnormality possibility index of each power load data in the load monitoring data sequence is obtained according to the data difference between the initial load forecast data sequence and the load monitoring data sequence, and the power system is subjected to load abnormality monitoring according to the abnormality possibility index of each power load data in the load monitoring data sequence.

3. The virtual power plant control terminal capable of realizing a new type of load monitoring function according to claim 1 is characterized in that: The obtaining the abnormality degree of the load monitoring data sequence according to the data difference between the initial load forecasting data sequence and the load monitoring data sequence comprises: For any power load data in the load monitoring data sequence, obtain the initial predicted power load data corresponding to the sampling time of the power load data in the initial load prediction data sequence, calculate the absolute value of the difference between the power load data and the initial predicted power load data, and record it as the prediction difference index of the power load data; Obtain a prediction difference index for each electric load data in the load monitoring data sequence, accumulate all the prediction difference indexes to obtain a corresponding accumulated value, normalize the accumulated value, and obtain the abnormality degree of the load monitoring data sequence.

4. The virtual power plant control terminal capable of realizing a new type of load monitoring function according to claim 1 is characterized in that: The step of obtaining the stability index of the time window corresponding to the power load data according to the power load data in the time window includes: Obtain the standard deviation of the power load data in the time window corresponding to the power load data, calculate a first addition result of the standard deviation and a constant 1, and obtain the inverse of the first addition result as a stability indicator of the time window corresponding to the power load data.

5. The virtual power plant control terminal capable of realizing a new type of load monitoring function according to claim 1 is characterized in that: The obtaining of a normal possibility index of the power load data according to the fluctuation of the power load data in the time window corresponding to the power load data comprises: When the stability index of the time window corresponding to the power load data is greater than or equal to a preset stability index threshold, the normal possibility index of the power load data is a preset value; When the stability index of the time window corresponding to the power load data is less than a preset stability index threshold, the difference between every two adjacent power load data in the time window corresponding to the power load data is calculated respectively to obtain a corresponding difference sequence, the absolute value of the difference between every two adjacent differences in the difference sequence is obtained respectively to obtain the accumulated value of the absolute value of the difference, the second addition result of the accumulated value of the absolute value of the difference and the constant 1 is calculated, and the inverse of the second addition result is normalized to obtain the normal possibility index of the power load data.

6. The virtual power plant control terminal capable of realizing a new type of load monitoring function according to claim 1 is characterized in that: The step of obtaining the stability of the load monitoring data sequence according to the normal possibility index of each power load data in the load monitoring data sequence and the stability index of the corresponding time window includes: The addition results between the normal possibility index of each power load data in the load monitoring data sequence and the stability index of the corresponding time window are respectively obtained, all the addition results are accumulated to obtain the cumulative value of the addition results, the cumulative value of the addition results is normalized to obtain the stability of the load monitoring data sequence.

7. The virtual power plant control terminal capable of realizing a new type of load monitoring function according to claim 1 is characterized in that: The step of obtaining the abnormal possibility index of each electric load data in the load monitoring data sequence according to the data difference between the optimized load forecasting data sequence and the load monitoring data sequence comprises: For any power load data in the load monitoring data sequence, obtain the optimized predicted power load data corresponding to the sampling time of the power load data in the optimized load prediction data sequence, calculate the absolute value of the difference between the power load data and the optimized predicted power load data, normalize the absolute value of the difference, and use the normalized result as the abnormal possibility indicator of the power load data.

8. The virtual power plant control terminal capable of realizing a new type of load monitoring function according to claim 1 is characterized in that: The performing load anomaly monitoring on the power system according to the anomaly possibility index of each power load data in the load monitoring data sequence comprises: When the abnormality possibility indicators of all the power load data in the load monitoring data sequence are less than the set abnormality possibility indicator threshold, it is determined that the power system is in a normal state within the preset time period; when the abnormality possibility indicator of any one of the power load data in the load monitoring data sequence is greater than or equal to the set abnormality possibility indicator threshold, it is determined that the power system has a load abnormality within the preset time period.

Images