Method and device for determining substation load

Abstract

The application discloses a kind of determination method and device of substation load, it is related to new energy and energy-saving technical field.Therein, the method comprises: obtaining N load data sets and M influence factor data sets of substation in first time period;Obtain the correlation between each load data set and each influence factor data set;N load data sets, M influence factor data sets and the correlation between each load data set and each influence factor data set are input into target model;Data in N load data set, data in M influence factor data set and correlation are all converted into feature vector by target model, and N target load data set of substation in second time period is determined according to feature vector.The application solves the technical problem that the accuracy of the main transformer total load prediction result is low in the prior art, which focuses on the analysis of the overall load prediction of the power load and does not consider other influencing factors.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of new energy and energy-saving technology, in particular to a method and device for determining substation load. BACKGROUND

[0002] Power system load forecasting is an important link to ensure stable power supply, and plays a crucial role in the safe and economic operation of the entire power system. Accurate load forecasting can provide decision support for dispatching departments, enabling them to proactively develop power generation plans and optimize the allocation of power resources, thereby effectively reducing operating costs.

[0003] The accuracy of power load forecasting is influenced by weather, holidays, special events and other factors, and the randomness of load changes is relatively large. Due to the large-scale integration of distributed photovoltaic and wind power, as well as the widespread popularity of electric vehicles, the randomness and volatility of load have been further increased, making it more difficult to accurately predict load. Traditional power load forecasting research often focuses on macro analysis of overall load, while neglecting the micro dissection of power load components, especially the components with high randomness such as air conditioning load, electric vehicle load, and new energy output, which limits the accuracy of the prediction results.

[0004] In view of the above problems, no effective solutions have been proposed so far. SUMMARY

[0005] The present application provides a method and device for determining substation load, to at least solve the technical problem that the analysis of power load forecasting in the prior art often focuses on the analysis of overall load without considering other influencing factors, resulting in low accuracy of the main transformer total load prediction results.

[0006] In order to achieve the above object, according to one aspect of the present application, a method and device for determining substation load are provided, comprising: obtaining N load data sets and M influence factor data sets of a substation in a first time period, wherein the N load data sets at least include a basic load data set, a load data set of an electricity-using device and a load data set of a power supply device, the M influence factor data sets at least include a holiday data set and a climate data set, and the basic load data set is a collection of self-load data of the substation; obtaining a correlation degree between each load data set and each influence factor data set, wherein the correlation degree is used to quantify the correlation between each load data set and each influence factor data set; inputting the N load data sets, the M influence factor data sets and the correlation degrees between each load data set and each influence factor data set into a target model; converting the data in the N load data sets, the data in the M influence factor data sets and the correlation degrees into feature vectors by the target model, and determining N target load data sets of the substation in a second time period according to the feature vectors, wherein the N target load data sets correspond to the N load data sets one by one, and the second time period is after the first time period.

[0007] Optionally, the load data set of the electricity-using device includes at least one of the following load data sets: electric vehicle load data, used to represent load data generated by an electric vehicle powered by the substation in the first time period; target seasonal device load data, used to represent load data generated by a target device powered by the substation in a target season included in the first time period, wherein the target device has a higher usage frequency in the target season than in a non-target season.

[0008] Optionally, the load data set of the power supply device includes at least one of the following load data sets: photovoltaic power generation device load data, used to represent output load data of a photovoltaic power generation device connected to the substation in the first time period; wind power generation device load data, used to represent output load data of a wind power generation device connected to the substation in the first time period.

[0009] Optionally, the basic load data set is obtained by the following steps: obtaining a historical load data set of the substation; calculating the difference between the historical load data set and load data of the same time point in the load data set of the electricity-using device, and forming a first load data set by all the calculated differences; and performing summation calculation on load data of the same time point in the first load data set and the load data set of the power supply device, and forming the basic load data set by all the summation results.

[0010] Optionally, in the process of obtaining the correlation between each load dataset and each influencing factor dataset, the method for determining the substation load further includes: normalizing the i-th load dataset among N load datasets to obtain the i-th target load dataset; normalizing the j-th influencing factor dataset among M influencing factor datasets to obtain the j-th target influencing factor dataset; calculating the correlation coefficient between the i-th target load dataset and the j-th target influencing factor dataset, wherein the correlation coefficient is used to evaluate the similarity between the i-th target load dataset and the j-th target influencing factor dataset; and obtaining the correlation between the i-th target load dataset and the j-th target influencing factor dataset based on the correlation coefficient.

[0011] Optionally, calculating the correlation coefficient between the i-th target load dataset and the j-th target influencing factor dataset includes: dividing the first time period into K time points; calculating the absolute difference between the i-th target load dataset and the j-th target influencing factor dataset at the target time point, where the target time point is any of the K time points; calculating the absolute difference between the i-th target load dataset and the j-th target influencing factor dataset at each of the K time points to obtain K absolute difference values; obtaining the minimum and maximum values ​​among the K absolute difference values; and determining the correlation coefficient at the target time point based on the absolute difference, minimum, and maximum values ​​at the target time point.

[0012] Optionally, the correlation degree between the i-th target load dataset and the j-th target influencing factor dataset is obtained based on the correlation coefficient, including: calculating the average of all correlation coefficients between the i-th target load dataset and the j-th target influencing factor dataset to obtain the correlation degree.

[0013] Optionally, the target model is trained through the following steps: obtaining a training dataset, wherein the training dataset includes N load datasets and M influencing factor datasets of the substation in a first historical time period, and N load datasets of the substation in a second historical time period, wherein the second historical time period is after the first historical time period, and the interval between the second historical time period and the first historical time period is the same as the interval between the first time period and the second time period; using the N load datasets and M influencing factor datasets of the substation in the first historical time period as training samples; using the N load datasets of the substation in the second historical time period as training labels; and iteratively training the neural network based on the training samples and training labels to obtain the target model.

[0014] According to another aspect of this application, a substation load determination apparatus is also provided, comprising: a first acquisition unit for acquiring N load datasets and M influencing factor datasets of the substation in a first time period, wherein the N load datasets include at least a base load dataset, a load dataset of electrical equipment, and a load dataset of power supply equipment, and the M influencing factor datasets include at least a holiday dataset and a climate dataset, wherein the base load dataset is a collection of the substation's own load data; a second acquisition unit for acquiring the correlation degree between each load dataset and each influencing factor dataset, wherein the correlation degree is used to quantify the correlation between each load dataset and each influencing factor dataset; a first input unit for inputting the N load datasets, the M influencing factor datasets, and the correlation degree between each load dataset and each influencing factor dataset into a target model; and a first determination unit for converting the data in the N load datasets, the data in the M influencing factor datasets, and the correlation degree into feature vectors through the target model, and determining N target load datasets of the substation in a second time period based on the feature vectors, wherein the N target load datasets correspond one-to-one with the N load datasets, and the second time period is after the first time period.

[0015] According to another aspect of this application, an electronic device is also provided, comprising: one or more processors; and a storage device for storing one or more programs, which, when executed by the one or more processors, cause the one or more processors to perform a program execution, wherein the program is configured to execute the substation load determination method described in any of the preceding claims. In this application, firstly, N load datasets and M influencing factor datasets of the substation in the first time period are obtained. The N load datasets include at least a basic load dataset, a load dataset of electrical equipment, and a load dataset of power supply equipment. The M influencing factor datasets include at least a holiday dataset and a climate dataset. The basic load dataset is a collection of the substation's own load data. Secondly, the correlation degree between each load dataset and each influencing factor dataset is obtained. The correlation degree is used to quantify the correlation between each load dataset and each influencing factor dataset. Then, the N load datasets, the M influencing factor datasets, and the correlation degree between each load dataset and each influencing factor dataset are input into the target model. Finally, the target model converts the data in the N load datasets, the data in the M influencing factor datasets, and the correlation degree into feature vectors. Based on the feature vectors, the N target load datasets of the substation in the second time period are determined. The N target load datasets correspond one-to-one with the N load datasets. The second time period is after the first time period.

[0016] As described above, this application determines the degree of influence of different influencing factors on the load dataset by acquiring multiple load datasets and multiple influencing factor datasets from historical time periods and calculating the correlation between these datasets. Based on these three dimensions—multiple load datasets, multiple influencing factor datasets, and the correlation—the application predicts the substation load for future time periods, ultimately obtaining the total substation load data for future time periods. Compared to existing technologies that only analyze the total load of historical time periods to predict the total load of future time periods, this application considers the impact of multiple influencing factors on various types of loads, thereby improving the technical effect of increasing the accuracy of load prediction. This solves the technical problem in existing technologies where power load prediction often focuses on the analysis of the overall load without considering other influencing factors, resulting in low accuracy of the total load prediction results for the main transformer. Attached Figure Description

[0017] The accompanying drawings, which are included to provide a further understanding of this application and form part of this application, illustrate exemplary embodiments and are used to explain this application, but do not constitute an undue limitation of this application. In the drawings:

[0018] Figure 1 This is a flowchart of an optional method for determining substation load according to an embodiment of this application;

[0019] Figure 2 This is a schematic diagram of an optional method for determining substation load according to an embodiment of this application;

[0020] Figure 3 This is a schematic diagram of an optional basic structure of an LSTM neural network according to an embodiment of this application;

[0021] Figure 4 This is a schematic diagram of an optional substation load determination device according to an embodiment of this application. Detailed Implementation

[0022] To enable those skilled in the art to better understand the present application, the technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present application, and not all embodiments. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without creative effort should fall within the scope of protection of the present application.

[0023] It should be noted that the terms "first," "second," etc., in the specification, claims, and accompanying drawings of this application are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments of this application described herein can be implemented in orders other than those illustrated or described herein. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover non-exclusive inclusion; for example, a process, method, system, product, or apparatus that comprises a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or apparatus.

[0024] It should also be noted that the information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, stored data, and displayed data) collected in this application are information and data authorized by the user or fully authorized by all parties. Furthermore, the collection, storage, use, processing, transmission, provision, disclosure, and application of related data all comply with the relevant laws, regulations, and standards of the relevant regions, necessary confidentiality measures have been taken, and they do not violate public order and good morals. Corresponding operation entry points are provided for users to choose to authorize or refuse. For example, this system has interfaces with relevant users or organizations. Before obtaining relevant information, a request to obtain the information needs to be sent to the aforementioned user or organization through the interface, and the relevant information is obtained only after receiving consent from the aforementioned user or organization.

[0025] According to an embodiment of this application, an embodiment of a method for determining substation load is provided. It should be noted that the steps shown in the flowchart in the accompanying drawings can be executed in a computer system such as a set of computer-executable instructions. Furthermore, although a logical order is shown in the flowchart, in some cases, the steps shown or described may be executed in a different order than that shown here.

[0026] It should be noted that an intelligent prediction system can serve as the execution subject of the substation load determination method in the embodiments of this application. It is understood that the substation load determination method provided in the embodiments of this application can also be executed by other systems or devices, and the embodiments of this application do not specifically limit this.

[0027] Figure 1 This is a flowchart of an optional method for determining substation load according to an embodiment of this application, such as... Figure 1 As shown, the method includes the following steps:

[0028] Step S101: Obtain N load datasets and M influencing factor datasets for the substation in the first time period.

[0029] In step S101, the N load datasets include at least the basic load dataset, the load dataset of electrical equipment, and the load dataset of power supply equipment, and the M influencing factor datasets include at least the holiday dataset and the climate dataset.

[0030] In step S101, the basic load dataset is a collection of the substation's own load data.

[0031] Optionally, the load dataset of electrical equipment includes, but is not limited to, electric vehicle load dataset and air conditioning load dataset.

[0032] Optionally, the load dataset of the power supply equipment includes, but is not limited to, photovoltaic power output dataset and wind power output dataset.

[0033] Optionally, the climate dataset includes, but is not limited to, temperature, humidity, wind speed, and solar radiation intensity.

[0034] Optionally, data on holidays, temperature, humidity, wind speed, and solar radiation intensity can be collected from sources such as economic statistics departments and meteorological stations to form a holiday data set D. time and climate data set D weather :

[0035] D time ={(hour,week,holiday),hour∈H,week∈W,holiday∈{0,1}}

[0036] D weather ={(temp,humidity,wind,solar direct ,solar scatter ),t∈T}

[0037] Where H represents the set of hours 1-24, W represents the set of weeks 1-7, holiday is the holiday identifier, where 1 represents a holiday and 0 represents a non-holiday, and temp, humidity, wind, solar are also included. direct and solar scatter These represent the temperature, humidity, wind speed, direct solar radiation intensity, and diffuse solar radiation intensity at time point t, respectively.

[0038] Optionally, the obtained N load datasets can be decomposed into multiple load datasets, including a basic load dataset, an electric vehicle load dataset, a photovoltaic and wind power output dataset, and a target seasonal equipment load dataset. The target seasonal equipment load data includes summer air conditioning load data and winter heating load data.

[0039] Step S102: Obtain the correlation between each load dataset and each influencing factor dataset.

[0040] In step S102, the correlation degree is used to quantify the correlation between each load dataset and each influencing factor dataset.

[0041] Optionally, grey relational analysis can be used to analyze the correlation between each influencing factor dataset and each load dataset, determine the degree of influence of different influencing factors on different loads, and calculate the correlation between each load dataset and each influencing factor dataset.

[0042] Optionally, grey relational analysis is a multivariate analysis method commonly used to analyze the correlation and degree of influence between multiple factors. It is primarily used when sample data is incomplete, information is uncertain, or there is significant variability, to assess and rank the correlation between factors. Specifically, grey relational analysis standardizes the data sequence and then calculates the correlation degree of each factor under different conditions, thereby identifying the extent to which each factor influences the outcome. Its basic idea is to transform the original data sequence into a grey series and then use the regularity of these series to analyze the relationships between factors.

[0043] Step S103: Input the N load datasets, M influencing factor datasets, and the correlation between each load dataset and each influencing factor dataset into the target model.

[0044] Optionally, an LSTM (Long Short-Term Memory) model can be used as the target model.

[0045] Step S104: Convert the data from the N load datasets, the data from the M influencing factor datasets, and the correlation degree into feature vectors using the target model, and determine the N target load datasets of the substation in the second time period based on the feature vectors.

[0046] In step S104, the N target load datasets correspond one-to-one with the N load datasets, and the second time period is after the first time period.

[0047] Optionally, the data from the N load datasets, the data from the M influencing factor datasets, and the correlation are used to construct a feature vector as input to the trained prediction model to predict the N target load datasets for a future time period (i.e., the second time period).

[0048] Alternatively, for example, Figure 2 This is a schematic diagram of an optional method for determining substation load according to an embodiment of this application, as shown below. Figure 2As shown, multiple load datasets and influencing factor datasets are first obtained. The load datasets include the total load of major substations, summer load, winter load, typical daily load in spring and autumn, photovoltaic and wind power output, and electric vehicle load. The influencing factor datasets include holidays and weather. The obtained load data is decomposed into basic load, winter heating load, summer air conditioning load, electric vehicle load, and photovoltaic and wind power output. Based on grey relational analysis, the key factors affecting the load are accurately identified to provide a scientific basis for the prediction model. Finally, the LSTM algorithm is used to predict various types of loads.

[0049] Optionally, for each type of load L k Where k represents the load type, such as base load, summer air conditioning load, etc., and its feature vector X is constructed. k (t) serves as the input to the LSTM model. The feature vector includes historical load data and related influencing factor data, specifically expressed as formula (1):

[0050] X k (t)=[L k (tn),…,L k (t-1),F1(t),F2(t),…,F m (t)] (1)

[0051] Among them, L k (tn) to L k (t-1) represents historical load data, F1(t) to F m (t) represents the selected influencing factor data, such as temperature, humidity, time information, etc., n is the time window length of historical data, and m is the number of influencing factors.

[0052] For each type of load L k Using its corresponding feature vector X k (t) Train the LSTM model and perform load forecasting. The output of the LSTM model is the load forecast for the next time step t+1. The prediction process can be expressed as formula (2):

[0053]

[0054] Among them, LSTM k This represents the LSTM model trained for the k-th type of load.

[0055] base load The predicted values ​​are shown in formula (3):

[0056]

[0057] Summer air conditioning load The predicted values ​​are shown in formula (4):

[0058]

[0059] Winter heating load L heating,winter The predicted value for (t+1) is shown in formula (5):

[0060] L heating,winter (t+1)=LSTM heating,winter (X heating,winter (t)) (5)

[0061] Electric vehicle load The predicted values ​​are shown in formula (6):

[0062]

[0063] Photovoltaic power output and wind power output The predicted values ​​are shown in formulas (7) and (8), respectively:

[0064]

[0065] The total load forecast of the substation is obtained by combining the LSTM forecast results of all load categories with the forecast results of photovoltaic and wind power output. The calculation formula is shown in (9):

[0066]

[0067] As can be seen from steps S101 to S104, in this application, firstly, N load datasets and M influencing factor datasets of the substation in the first time period are obtained. The N load datasets include at least a basic load dataset, a load dataset of electrical equipment, and a load dataset of power supply equipment. The M influencing factor datasets include at least a holiday dataset and a climate dataset. The basic load dataset is a collection of the substation's own load data. Secondly, the correlation degree between each load dataset and each influencing factor dataset is obtained. The correlation degree is used to quantify the correlation between each load dataset and each influencing factor dataset. Then, the N load datasets, the M influencing factor datasets, and the correlation degree between each load dataset and each influencing factor dataset are input into the target model. Finally, the target model converts the data in the N load datasets, the data in the M influencing factor datasets, and the correlation degree into feature vectors. Based on the feature vectors, the N target load datasets of the substation in the second time period are determined. The N target load datasets correspond one-to-one with the N load datasets. The second time period is after the first time period.

[0068] As described above, this application determines the degree of influence of different influencing factors on the load dataset by acquiring multiple load datasets and multiple influencing factor datasets from historical time periods and calculating the correlation between these datasets. Based on these three dimensions—multiple load datasets, multiple influencing factor datasets, and the correlation—the application predicts the substation load for future time periods, ultimately obtaining the total substation load data for future time periods. Compared to existing technologies that only analyze the total load of historical time periods to predict the total load of future time periods, this application considers the impact of multiple influencing factors on various types of loads, thereby improving the technical effect of increasing the accuracy of load prediction. This solves the technical problem in existing technologies where power load prediction often focuses on the analysis of the overall load without considering other influencing factors, resulting in low accuracy of the total load prediction results for the main transformer.

[0069] In one optional embodiment, the load dataset of electrical equipment includes at least one of the following load datasets: electric vehicle load data, used to characterize the load data generated by electric vehicles powered by the substation in a first time period; and target season equipment load data, used to characterize the load data generated by target equipment powered by the substation in the target season included in the first time period, wherein the target equipment is used more frequently in the target season than in non-target seasons.

[0070] Optionally, operational data from smart meters and charging stations can be used to record the charging load of electric vehicles, forming an electric vehicle load dataset D. EV :D EV ={L EV L(t), t∈T}, where L EV (t) represents the load value of the electric vehicle at time point t.

[0071] Optionally, historical load data for spring and autumn can be collected, and days with strong representativeness and small fluctuations can be selected as typical days to form a set of typical days D. typical :D typical ={L spring (t),L autumn (t), t∈T typical}, where L spring (t) and L autumn (t) represents the load value at time t on a typical day in spring and autumn, respectively. typical This is the set of observation times for a typical day.

[0072] Optionally, the target seasonal equipment load data includes summer air conditioning load data and winter heating load data. Summer air conditioning load data and winter heating load data are special load types caused by seasonal changes. The difference between the total summer load and the typical daily load in spring is used as the summer air conditioning load data, and the difference between the total winter load and the typical daily load in autumn is used as the winter heating load data. The summer air conditioning load data is shown in Equation (10), and the winter heating load data is shown in Equation (11).

[0073] L AC,summer =L total,summer -L typical,spring (10)

[0074] Among them, L AC,summer Indicates summer air conditioning load, L total,summer For the total summer load, L typical,spring This represents the typical daily load in spring and serves as the baseline load for the non-air-conditioned season.

[0075] L heating,winter =L total,winter -L typical,autumn (11)

[0076] Among them, L heating,winter Indicates winter heating load, L total,winter For the total load in winter, L typical,autumn This represents the typical daily load in autumn and also serves as the baseline load for the non-heating season.

[0077] As can be seen from the above, the load dataset of electrical equipment includes electric vehicle load dataset and target seasonal equipment load dataset. By refining and decomposing the load dataset of electrical equipment, we can more accurately analyze the correlation between each type of load dataset and various influencing factors, and obtain the impact of different influencing factors on different types of load datasets. This can help the intelligent prediction system to more accurately understand and predict the load demand of substations.

[0078] In one optional embodiment, the load dataset of the power supply equipment includes at least one of the following load datasets: photovoltaic power generation equipment load data, used to characterize the output load data of photovoltaic power generation equipment connected to the substation in a first time period, and wind power generation equipment load data, used to characterize the output load data of wind power generation equipment connected to the substation in a first time period.

[0079] Optionally, historical power output data can be obtained from the operational data platforms of photovoltaic power plants and wind power plants to form a photovoltaic power output dataset D. PV :D PV ={P PV (t), t∈T} and wind power output dataset D win:D wind ={P wind (t), t∈T}, where P PV (t) and P wind (t) represents the photovoltaic and wind power output values ​​at time point t, respectively.

[0080] As can be seen from the above, power supply equipment contains multiple output load datasets. Treating these multiple output load datasets as a single type of load dataset and studying the influence of various influencing factors on the output load dataset can help intelligent prediction systems predict the output load dataset data more accurately.

[0081] In one optional embodiment, the intelligent prediction system acquires the historical load dataset of the substation, then calculates the difference between the historical load dataset and the load data of the power equipment at the same time, and combines all the calculated differences into a first load dataset. Finally, the system sums the first load dataset with the load data of the power supply equipment at the same time, and combines all the summation results into a basic load dataset.

[0082] Optionally, by using the database interface of the power system operator, time-series data of the historical total load of the substation can be obtained to form a historical load dataset D of the substation. total :D total ={L total L(t), t∈T}, where L total (t) represents the total load value of the main transformer at time point t, where T is the set of observation times.

[0083] Optionally, the base load refers to the load obtained by removing the target seasonal equipment load (such as summer air conditioning load and winter heating load) and electric vehicle load, and then adding renewable energy output (photovoltaic and wind power). The calculation method is shown in (12):

[0084] L base =L total -(L AC,summer +L heating,winter )-L EV +(P PV +P wind (12)

[0085] Among them, L base L represents the base load. total For the historical total load of the substation, L EV For electric vehicle load, P PV and P wind These are the outputs of photovoltaic and wind power, respectively.

[0086] As can be seen from the above, the intelligent prediction system calculates the basic load dataset based on the historical total load dataset of the substation, the load dataset of the electrical equipment, and the load dataset of the power supply equipment. This can help manage the substation load, and further subdividing the historical load dataset of the substation into various sub-loads is beneficial for studying the degree of influence of the factors affecting each type of sub-load.

[0087] In one optional embodiment, the intelligent prediction system normalizes the i-th load dataset from N load datasets to obtain the i-th target load dataset, normalizes the j-th influencing factor dataset from M influencing factor datasets to obtain the j-th target influencing factor dataset, calculates the correlation coefficient between the i-th target load dataset and the j-th target influencing factor dataset, wherein the correlation coefficient is used to evaluate the similarity between the i-th target load dataset and the j-th target influencing factor dataset, and finally obtains the correlation degree between the i-th target load dataset and the j-th target influencing factor dataset based on the correlation coefficient.

[0088] Optionally, grey relational analysis can be used to analyze the correlation between influencing factors such as holidays and climate and various loads, thereby determining the degree of influence of different influencing factors on different loads. First, the N load datasets and M influencing factor datasets are normalized to ensure that the data are compared on the same order of magnitude. The normalization formula is shown in (13):

[0089]

[0090] Where X is the load dataset or influencing factor dataset, X * It is a standardized load dataset or influencing factor dataset, X min and X max These are the minimum and maximum values ​​in the load dataset or the influencing factor dataset, respectively.

[0091] As can be seen from the above, this process combines data normalization and correlation coefficient calculation to evaluate and analyze the complex relationships between multiple load datasets and multiple influencing factor datasets, thereby providing more accurate forecasting capabilities and decision support for intelligent forecasting systems.

[0092] In one optional embodiment, the intelligent prediction system divides the first time period into K time points, calculates the absolute difference between the i-th target load dataset and the j-th target influencing factor dataset at the target time, where the target time is any of the K time points, and then calculates the absolute difference between the i-th target load dataset and the j-th target influencing factor dataset at each of the K time points to obtain K absolute differences. The minimum and maximum values ​​among the K absolute differences are then obtained, and the correlation coefficient of the target time is determined based on the absolute difference, minimum value, and maximum value of the target time.

[0093] Optionally, for each load dataset D k And influencing factors dataset D f Calculate the correlation coefficient ξ between them. k,f To assess the similarity between them, the correlation coefficient is calculated according to formula (14):

[0094]

[0095] Where, Δ k,f (t)=|X k (t)-X f (t)|,X k (t) is the load dataset D k At time t, X f (t) is the dataset of influencing factors D. f At time t, Δ k,f (t) represents D k and D f The absolute difference at time point t, |Δ k,f | represents D k and D f The set of absolute differences at every moment in history, minmin|Δ k,f | represents the minimum value in this set of absolute differences, maxmax|Δ k,f | represents the maximum value in this set of absolute differences, and ρ is the resolution coefficient, which is usually taken as 0.5.

[0096] As can be seen from the above, by comparing the differences between the target load dataset and the influencing factor dataset at different times, and combining the minimum and maximum differences to determine the correlation coefficient, the degree of correlation between them at a specific time can be evaluated. Such analysis can help the prediction system understand and quantify the relationship between the load dataset and the influencing factor dataset, providing a more accurate basis for prediction and decision-making.

[0097] In one alternative embodiment, the intelligent prediction system obtains the correlation degree by calculating the average of all correlation coefficients between the i-th target load dataset and the j-th target influencing factor dataset.

[0098] Optionally, based on the correlation coefficient, the grey relational degree γ between various types of loads and various influencing factors is calculated. k,f To quantify the overall degree of correlation between them, the degree of correlation is calculated as shown in (15):

[0099]

[0100] Where N is the length of the time series.

[0101] Optionally, by comparing the grey relational degree of different influencing factor datasets with various load datasets, influencing factors with higher grey relational degree will be regarded as important factors of the corresponding load changes, and the top three influencing factors with the highest degree of influence on the load will be given priority in model prediction for model construction.

[0102] As can be seen from the above, by calculating the degree of correlation, the influence of various loads and influencing factors can be determined, thereby identifying the key factors affecting the load and providing a reference for load forecasting.

[0103] In one optional embodiment, the intelligent prediction system acquires a training dataset, which includes N load datasets and M influencing factor datasets of the substation in a first historical time period, and N load datasets of the substation in a second historical time period. The second historical time period is after the first historical time period, and the interval between the second historical time period and the first historical time period is the same as the interval between the first time period and the second time period. The N load datasets and M influencing factor datasets of the substation in the first historical time period are used as training samples, and the N load datasets of the substation in the second historical time period are used as training labels. Finally, the neural network is iteratively trained based on the training samples and training labels to obtain the target model.

[0104] Optionally, various types of loads can be predicted based on the LSTM model, and the prediction results can be superimposed to obtain the total load at the next time step.

[0105] Optionally, an LSTM network consists of multiple LSTM units, each including three gates: an input gate, a forget gate, and an output gate. These gates capture and remember long-term dependencies in time-series data by controlling the flow of information. Figure 3 This is a schematic diagram of an optional basic structure of an LSTM neural network according to an embodiment of this application, such as... Figure 3 As shown, the memory unit, along with the hidden state, stores historical information about the sequence data. The information in the memory unit is controlled by three gating units. The forget gate is determined according to h. t-1 and x t Delete the information in the memory cell, as shown in equation (16):

[0106] f t =σ(W f ·[h t-1 ,x t ]+b f (16)

[0107] Where σ is the Sigmoid activation function, W f Let b be the weight matrix of the forget gate. f This is the bias term for the forget gate.

[0108] Input gate based on h t-1 and x t Add information to the memory unit, as shown in equations (17) and (18):

[0109] i t =σ(W i ·[h t-1 ,x t ]+b i (17)

[0110]

[0111] Among them, i t It is the activation value of the input gate, i.e., the information that needs to be memorized. W serves as a candidate memory unit for updating memory units. i W C Let b be the weight matrix of the input gate. i b C This is the bias term for the input gate.

[0112] The current memory unit C is updated by combining information from the forget gate and the input gate. t As shown in equation (19):

[0113]

[0114] Among them, C t-1 This represents the state of the memory unit at the previous moment.

[0115] The output gate depends on h t-1 and x t The output information of the current memory unit is determined as shown in equation (20):

[0116] o t =σ(W o ·[h t-1 ,x t ]+b o (20)

[0117] Among them, W o Let b be the weight matrix of the output gate. o This is the bias term for the output gate.

[0118] Generate the hidden state h at the current moment by combining the information from the current memory unit and the output gate. t As shown in equation (21):

[0119] h t =o t ·tanh(C t ) (twenty one)

[0120] Optionally, historical data can be used as input for model training. Once the model has undergone multiple iterations of training and achieves satisfactory performance metrics on the validation set, the training is considered complete. At this point, the resulting neural network model can be used to predict the load data of the substation in future time periods.

[0121] As can be seen from the above, through the above steps, the intelligent prediction system has carried out a supervised learning training process using historical data, thereby obtaining a target model that can predict future load data.

[0122] According to an embodiment of this application, an embodiment of a substation load determination device is also provided. Figure 4 This is a schematic diagram of an optional substation load determination device according to an embodiment of this application, as shown below. Figure 4 As shown, the substation load determination device includes: a first acquisition unit 401, a second acquisition unit 402, a first input unit 403, and a first determination unit 404.

[0123] Optionally, the first acquisition unit 401 is used to acquire N load datasets and M influencing factor datasets of the substation in the first time period, wherein the N load datasets include at least a basic load dataset, a load dataset of electrical equipment, and a load dataset of power supply equipment, and the M influencing factor datasets include at least a holiday dataset and a climate dataset, wherein the basic load dataset is a collection of the substation's own load data; the second acquisition unit 402 is used to acquire the correlation degree between each load dataset and each influencing factor dataset, wherein the correlation degree is used to quantify the correlation between each load dataset and each influencing factor dataset; the first input unit 403 is used to input the N load datasets, the M influencing factor datasets, and the correlation degree between each load dataset and each influencing factor dataset into the target model; the first determination unit 404 is used to convert the data in the N load datasets, the data in the M influencing factor datasets, and the correlation degree into feature vectors through the target model, and determine the N target load datasets of the substation in the second time period based on the feature vectors, wherein the N target load datasets correspond one-to-one with the N load datasets, and the second time period is after the first time period.

[0124] Optionally, the first acquisition unit 401 includes: a first data subunit and a second data subunit. The first data subunit contains electric vehicle load data, used to characterize the load data generated by electric vehicles powered by the substation in a first time period; the second data subunit contains target season-specific equipment load data, used to characterize the load data generated by target equipment powered by the substation within the first time period under the target season, wherein the target equipment is used more frequently in the target season than in non-target seasons.

[0125] Optionally, the first acquisition unit 401 includes a third data subunit and a fourth data subunit. The third data subunit contains photovoltaic power generation equipment load data, used to characterize the output load data of the photovoltaic power generation equipment connected to the substation during a first time period; the fourth data subunit contains wind power generation equipment load data, used to characterize the output load data of the wind power generation equipment connected to the substation during a first time period.

[0126] Optionally, the first acquisition unit 401 includes: a first acquisition subunit, a first calculation subunit, and a second calculation subunit. The first acquisition subunit is used to acquire the historical load dataset of the substation; the first calculation subunit is used to calculate the difference between the historical load dataset and the load data of the power supply equipment at the same time, and to combine all the calculated differences into a first load dataset; the second calculation subunit is used to sum the first load dataset and the load data of the power supply equipment at the same time, and to combine all the calculated sums into a basic load dataset.

[0127] Optionally, the second acquisition unit 402 includes: a first processing subunit, a second processing subunit, a third calculation subunit, and a first determination subunit. The first processing subunit is used to normalize the i-th load dataset from N load datasets to obtain the i-th target load dataset; the second processing subunit is used to normalize the j-th influencing factor dataset from M influencing factor datasets to obtain the j-th target influencing factor dataset; the third calculation subunit is used to calculate the correlation coefficient between the i-th target load dataset and the j-th target influencing factor dataset, wherein the correlation coefficient is used to evaluate the similarity between the i-th target load dataset and the j-th target influencing factor dataset; and the first determination subunit is used to obtain the correlation degree between the i-th target load dataset and the j-th target influencing factor dataset based on the correlation coefficient.

[0128] Optionally, the third calculation subunit includes: a first processing module, a first calculation module, a second calculation module, a first acquisition module, and a first determination module. The first processing module is used to divide the first time period into K time points; the first calculation module is used to calculate the absolute difference between the i-th target load dataset and the j-th target influencing factor dataset at a target time point, where the target time point is any one of the K time points; the second calculation module is used to calculate the absolute difference between the i-th target load dataset and the j-th target influencing factor dataset at each of the K time points, obtaining K absolute difference values; the first acquisition module is used to acquire the minimum and maximum values ​​among the K absolute difference values; and the first determination module is used to determine the correlation coefficient at the target time point based on the absolute difference, minimum, and maximum values ​​at the target time point.

[0129] Optionally, the first determining subunit includes: a third calculation module, used to calculate the average value of all correlation coefficients between the i-th target load dataset and the j-th target influencing factor dataset, to obtain the correlation degree.

[0130] Optionally, the first input unit 403 includes: a second acquisition subunit, a third processing subunit, a fourth processing subunit, and a second determination subunit. The second acquisition subunit is used to acquire a training dataset, which includes N load datasets and M influencing factor datasets of the substation in a first historical time period, and N load datasets of the substation in a second historical time period, wherein the second historical time period follows the first historical time period, and the interval between the second historical time period and the first historical time period is the same as the interval between the first and second historical time periods. The third processing subunit is used to use the N load datasets and M influencing factor datasets of the substation in the first historical time period as training samples. The fourth processing subunit is used to use the N load datasets of the substation in the second historical time period as training labels. The second determination subunit is used to iteratively train the neural network based on the training samples and training labels to obtain the target model.

[0131] According to embodiments of this application, an electronic device is also provided, comprising one or more processors; and a storage device for storing one or more programs, wherein when the one or more programs are executed by the one or more processors, the one or more processors implement a method for running the programs, wherein the programs are configured to execute the substation load determination method described above. The sequence numbers of the embodiments in this application are for descriptive purposes only and do not represent the superiority or inferiority of the embodiments.

[0132] In the above embodiments of this application, the descriptions of each embodiment have different focuses. For parts not described in detail in a certain embodiment, please refer to the relevant descriptions of other embodiments.

[0133] In the several embodiments provided in this application, it should be understood that the disclosed technical content can be implemented in other ways. The device embodiments described above are merely illustrative; for example, the division of units can be a logical functional division, and in actual implementation, there may be other division methods. For instance, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the displayed or discussed mutual coupling, direct coupling, or communication connection may be through some interfaces; the indirect coupling or communication connection between units or modules may be electrical or other forms.

[0134] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0135] Furthermore, the functional units in the various embodiments of this application can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit.

[0136] If the integrated unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of this application. The aforementioned storage medium includes various media capable of storing program code, such as a USB flash drive, read-only memory (ROM), random access memory (RAM), portable hard drive, magnetic disk, or optical disk.

[0137] The above description is only a preferred embodiment of this application. It should be noted that for those skilled in the art, several improvements and modifications can be made without departing from the principle of this application, and these improvements and modifications should also be considered within the scope of protection of this application.

Claims

1. A method for determining the load of a substation, characterized in that, include: Obtain N load datasets and M influencing factor datasets for the substation in the first time period. The N load datasets include at least a basic load dataset, a load dataset of electrical equipment, and a load dataset of power supply equipment. The M influencing factor datasets include at least a holiday dataset and a climate dataset. The basic load dataset is a collection of the substation's own load data. Obtain the correlation degree between each load dataset and each influencing factor dataset, wherein the correlation degree is used to quantify the correlation between each load dataset and each influencing factor dataset; The N load datasets, the M influencing factor datasets, and the correlation between each load dataset and each influencing factor dataset are input into the target model; The target model converts the data from the N load datasets, the data from the M influencing factor datasets, and the correlation degree into feature vectors, and determines the N target load datasets of the substation in the second time period based on the feature vectors. The N target load datasets correspond one-to-one with the N load datasets, and the second time period is after the first time period.

2. The method for determining substation load according to claim 1, characterized in that, The load dataset of the electrical equipment includes at least one of the following load datasets: Electric vehicle load data, used to characterize the load data generated by electric vehicles powered by the substation during the first time period; The target seasonal equipment load data is used to characterize the load data generated by the target equipment powered by the substation during the target season within the first time period, wherein the target equipment is used more frequently in the target season than in non-target seasons.

3. The method for determining substation load according to claim 1, characterized in that, The load dataset of the power supply equipment includes at least one of the following load datasets: Photovoltaic power generation equipment load data, used to characterize the output load data of photovoltaic power generation equipment connected to the substation during the first time period; Wind power generation equipment load data is used to characterize the output load data of the wind power generation equipment connected to the substation during the first time period.

4. The method for determining substation load according to claim 1, characterized in that, The basic load dataset is obtained through the following steps: Obtain historical load datasets for substations; Calculate the difference between the historical load dataset and the load dataset of the electrical equipment at the same time, and combine all the calculated differences into a first load dataset; The load data corresponding to the same moment in the first load data set and the load data set of the power supply equipment are summed, and all the summation results are combined to form the basic load data set.

5. The method for determining substation load according to claim 1, characterized in that, The method for determining substation load further includes the following steps in obtaining the correlation between each load dataset and each influencing factor dataset: The i-th load dataset among the N load datasets is normalized to obtain the i-th target load dataset; The j-th influencing factor dataset in the M influencing factor datasets is normalized to obtain the j-th target influencing factor dataset; Calculate the correlation coefficient between the i-th target load dataset and the j-th target influencing factor dataset, wherein the correlation coefficient is used to evaluate the similarity between the i-th target load dataset and the j-th target influencing factor dataset; The correlation coefficient is used to obtain the correlation degree between the i-th target load dataset and the j-th target influencing factor dataset.

6. The method for determining substation load according to claim 5, characterized in that, Calculate the correlation coefficient between the i-th target load dataset and the j-th target influencing factor dataset, including: Divide the first time period into K time periods; Calculate the absolute difference between the i-th target load dataset and the j-th target influencing factor dataset at the target time, where the target time is any of the K times; Calculate the absolute difference between the i-th target load dataset and the j-th target influencing factor dataset at each of the K time points to obtain K absolute differences; Obtain the minimum and maximum values ​​among the K absolute differences; The correlation coefficient of the target time is determined based on the absolute difference of the target time, the minimum value, and the maximum value.

7. The method for determining substation load according to claim 6, characterized in that, The correlation degree between the i-th target load dataset and the j-th target influencing factor dataset is obtained based on the correlation coefficient, including: The correlation degree is obtained by calculating the average of all correlation coefficients between the i-th target load dataset and the j-th target influencing factor dataset.

8. The method for determining substation load according to claim 1, characterized in that, The target model is trained through the following steps: Obtain a training dataset, wherein the training dataset includes N load datasets and M influencing factor datasets of the substation in a first historical time period, and N load datasets of the substation in a second historical time period, wherein the second historical time period is after the first historical time period, and the interval between the second historical time period and the first historical time period is the same as the interval between the first time period and the second time period. The N load datasets and M influencing factor datasets of the substation in the first historical time period are used as training samples. The N load datasets of the substation in the second historical time period are used as training labels; The neural network is iteratively trained based on the training samples and the training labels to obtain the target model.

9. A device for determining the load of a substation, characterized in that, include: The first acquisition unit acquires N load datasets and M influencing factor datasets of the substation in the first time period. The N load datasets include at least a basic load dataset, a load dataset of electrical equipment, and a load dataset of power supply equipment. The M influencing factor datasets include at least a holiday dataset and a climate dataset. The basic load dataset is a collection of the substation's own load data. The second acquisition unit acquires the correlation degree between each load dataset and each influencing factor dataset, wherein the correlation degree is used to quantify the correlation between each load dataset and each influencing factor dataset; The first input unit inputs the N load datasets, the M influencing factor datasets, and the correlation between each load dataset and each influencing factor dataset into the target model; The first determining unit converts the data in the N load datasets, the data in the M influencing factor datasets, and the correlation degree into feature vectors through the target model, and determines the N target load datasets of the substation in the second time period based on the feature vectors. The N target load datasets correspond one-to-one with the N load datasets, and the second time period is after the first time period.

10. An electronic device, characterized in that, The electronic device includes one or more processors; and a storage device for storing one or more programs, which, when executed by the one or more processors, cause the one or more processors to perform the program, wherein the program is configured to execute the substation load determination method according to any one of claims 1 to 8.

Images