Transformer substation load prediction method and device and terminal equipment

Abstract

The invention discloses a transformer substation load prediction method and device and terminal equipment. The method comprises the following steps: acquiring a predicted temperature value, a predicted residential electricity price, a predicted industrial electricity price, current weather information, historical user travel information, historical enterprise production information, a current user number and a current enterprise number of a to-be-processed substation area in a to-be-predicted time period; inputting the predicted temperature value, the predicted residential electricity price, the current weather information and the determined first user number into a preset user load prediction model to obtain a user electricity load value; inputting the predicted temperature value, the predicted industrial electricity price, the current weather information and the determined first enterprise number into a preset enterprise load prediction model to obtain an enterprise electricity load value; and calculating the total load value of the to-be-processed transformer substation area according to the user electrical load value and the enterprise electrical load value. According to the invention, the accuracy of load prediction is improved.

Description

Technical Field

[0001] The present invention relates to the technical field of power systems, and in particular to a method, a device and a terminal device for predicting load in a substation. Background Art

[0002] In modern power grid systems, kilovolt substations are an important component, and their load forecasting is of great significance for power dispatching and equipment configuration of power grid systems. At present, the main method for load forecasting of kilovolt substations is time series analysis. The time series analysis method is based on historical load data and uses statistical methods to establish a time series model of load changes for load forecasting. However, historical load data are easily affected by random fluctuations, and the prediction effect is poor in emergencies or seasonal changes, resulting in low accuracy of load forecasting of kilovolt substations, which leads to poor stability and reliability of power grid system operation. Summary of the invention

[0003] The embodiments of the present invention provide a substation load prediction method, device and terminal equipment, which can effectively solve the problem that the load prediction accuracy of the kilovolt substation in the prior art is low, thereby causing poor stability and reliability of the operation of the power grid system.

[0004] An embodiment of the present invention provides a substation load prediction method, comprising:

[0005] Obtain the forecast temperature value, forecast residential electricity price, forecast industrial electricity price, current meteorological information, historical user travel information, historical enterprise production information, current number of users and current number of enterprises in the substation area to be processed within the forecast time period;

[0006] Determine the number of first users in the time period to be predicted based on the historical user travel information and the current number of users; and determine the number of first enterprises in the time period to be predicted based on the historical enterprise production information and the current number of enterprises;

[0007] Inputting the predicted temperature value, the predicted residential electricity price, the current meteorological information and the number of first users into a preset user load prediction model to perform user load prediction and obtain a user electricity load value;

[0008] Inputting the predicted temperature value, the predicted industrial electricity price, the current meteorological information and the number of the first enterprise into a preset enterprise load forecasting model to perform enterprise load forecasting and obtain the enterprise electricity load value;

[0009] The total load value of the substation area to be processed within the time period to be predicted is calculated based on the user power load value, the enterprise power load value and the predicted temperature value.

[0010] Furthermore, the historical user travel information includes historical travel proportion and historical travel time;

[0011] Determining the number of first users in the time period to be predicted according to the historical user travel information and the current number of users includes:

[0012] The number of people traveling is calculated by multiplying the historical travel ratio and the current number of users.

[0013] The number of first users in the time period to be predicted is calculated by subtracting the current number of users from the number of people traveling;

[0014] Among them, if the week system corresponding to the time period to be predicted is different from the week system of the historical travel time, the historical travel proportion corresponding to the historical travel time closest to the time period to be predicted and with the same week system as the time period to be predicted will be used as the final historical travel proportion.

[0015] Furthermore, the historical enterprise production information includes historical enterprise production proportion;

[0016] Determining the first number of enterprises in the time period to be predicted according to the historical enterprise production information and the current number of enterprises includes:

[0017] The number of production enterprises is calculated by multiplying the historical production proportion of enterprises and the current number of enterprises;

[0018] Based on the current number of enterprises and the number of production enterprises, the first number of enterprises in the forecasted time period is obtained by subtracting them.

[0019] Furthermore, the training of the user load prediction model includes:

[0020] Obtain historical temperature values, historical meteorological information, historical residential electricity prices, historical number of users, the initial first bird's nest position of the model, and the initial first optimization parameters;

[0021] Input the historical temperature value, historical meteorological information, historical residential electricity price, historical number of users, current first optimization parameter and current first bird's nest position into the user load prediction model to be trained for optimization, and obtain the first target bird's nest position; wherein the current first bird's nest position optimized for the first time is the initial first bird's nest position; the current first optimization parameter optimized for the first time is the initial first optimization parameter;

[0022] The first bird's nest fitness is calculated according to the current first bird's nest position and the first target bird's nest position; and the current first bird's nest position and the current first optimization parameter are updated according to the first bird's nest fitness until a preset number of iterations is reached to obtain the final first bird's nest position;

[0023] According to the first penalty coefficient corresponding to the final first bird's nest position and the first width of the radial basis kernel function, a trained user load prediction model is constructed.

[0024] Furthermore, the training of the enterprise load forecasting model includes:

[0025] Obtain historical temperature values, historical meteorological information, historical industrial electricity prices, historical number of enterprises, and the initial second bird's nest position and initial second optimization parameters of the model;

[0026] Input the historical temperature value, historical meteorological information, historical industrial electricity price, historical enterprise number, current second optimization parameter and current second bird's nest position of the model into the enterprise load forecasting model to be trained for optimization, and obtain the second target bird's nest position; wherein the current second bird's nest position optimized for the first time is the initial second bird's nest position; the current second optimization parameter optimized for the first time is the initial second optimization parameter;

[0027] The second bird's nest fitness is calculated according to the current second bird's nest position and the second target bird's nest position; and the current second bird's nest position and the current second optimization parameter are updated according to the second bird's nest fitness until a preset number of iterations is reached to obtain a final second bird's nest position;

[0028] According to the second penalty coefficient corresponding to the final second bird's nest position and the second width of the radial basis kernel function, a trained enterprise load forecasting model is constructed.

[0029] Further, the user load forecasting model includes: a first user power load forecasting layer for representing the influence of temperature, a second user power load forecasting layer for representing the influence of residential power price, and a user total power load forecasting layer;

[0030] The predicted temperature value, the predicted residential electricity price, the current meteorological information and the number of first users are input into a preset user load prediction model to perform user load prediction to obtain the user electricity load value, including:

[0031] Determine a first user load prediction function of a first user power load prediction layer by comparing the predicted temperature value with a preset first temperature threshold;

[0032] Calculate the first user power load forecast value of the first user power load forecast layer according to the predicted temperature value, the number of first users and the determined first user load forecast function;

[0033] Determine a second user load prediction function of a second user power load prediction layer by comparing the predicted residential power price with a preset residential power price threshold;

[0034] Convert the current meteorological information into corresponding values ​​to obtain seasonal factors;

[0035] Calculate the second user power load forecast value of the second user power load forecast layer according to the predicted residential power price, seasonal factors and the second user load forecast function;

[0036] The first user power load prediction value and the second user power load prediction value are input into the user power total load prediction layer, added and summed to obtain the user power load value.

[0037] Further, the enterprise load forecasting model includes: a first enterprise power load forecasting layer for representing the influence of temperature, a second enterprise power load forecasting layer for representing the influence of industrial power price, and a total enterprise power load forecasting layer;

[0038] The predicted temperature value, the predicted industrial electricity price, the current meteorological information and the number of the first enterprise are input into a preset enterprise load forecasting model to perform enterprise load forecasting, and obtain the enterprise electricity load value, including:

[0039] Determine a first enterprise load forecasting function of a first enterprise power load forecasting layer by comparing the forecast temperature value with a preset second temperature threshold;

[0040] Calculate the first enterprise power load forecast value of the first enterprise power load forecast layer according to the forecast temperature value, the number of first enterprises and the determined first enterprise load forecast function;

[0041] Determine a second enterprise load forecasting function of a second enterprise power load forecasting layer by comparing the predicted industrial power price with a preset industrial power price threshold;

[0042] Calculate the second enterprise power load forecast value of the second enterprise power load forecast layer according to the forecast industrial power price, seasonal factors and the second enterprise load forecast function;

[0043] The first enterprise power load forecast value and the second enterprise power load forecast value are input into the enterprise power total load forecast layer, and are added together to obtain the enterprise power load value.

[0044] Furthermore, according to the user power load value, the enterprise power load value and the predicted temperature value, the total load value of the substation area to be processed in the time period to be predicted is calculated, including:

[0045] Calculating a temperature error load value according to a preset basic error load value, the predicted temperature value, and a preset temperature threshold;

[0046] The total load value of the substation area to be processed within the time period to be predicted is calculated based on the user power load value, the enterprise power load value, the temperature error load value and the preset power weight value.

[0047] As an improvement of the above solution, another embodiment of the present invention provides a substation load prediction device, including:

[0048] The data acquisition module is used to obtain the predicted temperature value, predicted residential electricity price, predicted industrial electricity price, current meteorological information, historical user travel information, historical enterprise production information, current number of users and current number of enterprises in the substation area to be processed within the predicted time period;

[0049] A predicted quantity determination module, used to determine the first number of users in the time period to be predicted based on the historical user travel information and the current number of users; and to determine the first number of enterprises in the time period to be predicted based on the historical enterprise production information and the current number of enterprises;

[0050] A user load prediction module, used for inputting the predicted temperature value, the predicted residential electricity price, the current meteorological information and the number of first users into a preset user load prediction model to perform user load prediction and obtain a user electricity load value;

[0051] An enterprise load forecasting module, used for inputting the predicted temperature value, the predicted industrial electricity price, the current meteorological information and the number of the first enterprise into a preset enterprise load forecasting model to perform enterprise load forecasting and obtain an enterprise electricity load value;

[0052] The total load prediction module is used to calculate the total load value of the substation area to be processed within the time period to be predicted based on the user power load value, the enterprise power load value and the predicted temperature value.

[0053] Another embodiment of the present invention provides a terminal device, including a processor, a memory, and a computer program stored in the memory and configured to be executed by the processor, and when the processor executes the computer program, a substation load forecasting method as described in the above embodiment is implemented.

[0054] Another embodiment of the present invention provides a computer-readable storage medium, which includes a stored computer program, wherein when the computer program is running, the device where the computer-readable storage medium is located is controlled to execute a substation load forecasting method described in the above embodiment.

[0055] By implementing the present invention, at least the following beneficial effects are achieved:

[0056] The present invention provides a substation load prediction method, device and terminal equipment, wherein the method can obtain the forecast temperature value, forecast residential electricity price, forecast industrial electricity price, current meteorological information, historical user travel information, historical enterprise production information, current number of users and current number of enterprises of a substation area to be processed within a time period to be predicted; determine the first number of users within the time period to be predicted based on the historical user travel information and the current number of users; and determine the first number of enterprises within the time period to be predicted based on the historical enterprise production information and the current number of enterprises; input the forecast temperature value, forecast residential electricity price, current meteorological information and the first number of users into a preset user load prediction model to perform user load prediction and obtain the user power load value; input the forecast temperature value, forecast industrial electricity price, current meteorological information and the first number of enterprises into a preset enterprise load prediction model to perform enterprise load prediction and obtain the enterprise power load value; calculate the total load value of the substation area to be processed within the time period to be predicted based on the user power load value, the enterprise power load value and the forecast temperature value. By comprehensively considering multiple factors such as temperature, electricity price, meteorological information, number of users and number of enterprises, the comprehensive impact of different factors on the predicted load is captured, the impact of historical load data on load forecasting is reduced, and when emergencies occur or seasonal changes occur, the impact of meteorological information is taken into account. It is possible to respond to the impact of emergencies or seasonal changes, accurately simulate and predict load changes in the substation area, so as to improve the accuracy of load forecasting for kilovolt substations; at the same time, by separately predicting the electricity load of users and enterprises, the accuracy of load forecasting is improved, thereby improving the stability and reliability of the power grid system operation. BRIEF DESCRIPTION OF THE DRAWINGS

[0057] Figure 1 It is a flow chart of a method for substation load prediction provided by one embodiment of the present invention;

[0058] Figure 2 It is a structural schematic diagram of a substation load prediction device provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0059] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0060] See also Figure 1 , is a flow chart of a substation load prediction method provided by an embodiment of the present invention, comprising:

[0061] S1. Obtain the forecast temperature value, forecast residential electricity price, forecast industrial electricity price, current meteorological information, historical user travel information, historical enterprise production information, current number of users and current number of enterprises in the substation area to be processed within the forecast time period;

[0062] S2. Determine the number of first users in the time period to be predicted based on the historical user travel information and the current number of users; and determine the number of first enterprises in the time period to be predicted based on the historical enterprise production information and the current number of enterprises;

[0063] S3, inputting the predicted temperature value, the predicted residential electricity price, the current meteorological information and the number of the first users into a preset user load prediction model to perform user load prediction and obtain a user electricity load value;

[0064] S4, inputting the predicted temperature value, the predicted industrial electricity price, the current meteorological information and the number of the first enterprise into a preset enterprise load forecasting model to perform enterprise load forecasting and obtain the enterprise electricity load value;

[0065] S5. Calculate the total load value of the substation area to be processed within the time period to be predicted based on the user power load value, the enterprise power load value and the predicted temperature value.

[0066] Specifically, the first number of users represents the number of users in the time period to be predicted; the first number of enterprises represents the number of enterprises in the time period to be predicted; the current meteorological information includes the current temperature, the current humidity and the current season. The current meteorological information and electricity price information are key factors affecting the load of the kilovolt substation. For example, when the temperature is high, the possibility of users going out is relatively low, and the number of users in the community covered by the kilovolt substation at home increases, which may lead to an increase in the load of the kilovolt substation. For example, the current season is autumn, the weather is relatively cool, and the possibility of users going out is relatively high. Therefore, the number of users in the community covered by the kilovolt substation at home decreases, thereby reducing the load of the kilovolt substation. The electricity price information includes the forecast residential electricity price and the forecast industrial electricity price. The electricity price is relatively floating. For example, if the forecast residential electricity price becomes higher, the frequency of users in the community covered by the kilovolt substation using high-power electrical appliances may be reduced, thereby reducing the load of the kilovolt substation. If the forecast industrial electricity price becomes higher, the number of production enterprises in the area covered by the kilovolt substation that start production may be reduced, thereby reducing the load of the kilovolt substation, and so on. All of the above information can be used as factors affecting the load of the kilovolt substation. For the coverage area of ​​the kilovolt substation (substation area to be processed), the predicted temperature value can be obtained directly from the weather forecast, or can be predicted based on historical temperature information. The predicted residential electricity price and the predicted industrial electricity price can be obtained through the information released by the power grid.

[0067] Preferably, the historical user travel information includes historical travel proportion and historical travel time;

[0068] Determining the number of first users in the time period to be predicted according to the historical user travel information and the current number of users includes:

[0069] The number of people traveling is calculated by multiplying the historical travel ratio and the current number of users.

[0070] The number of first users in the time period to be predicted is calculated by subtracting the current number of users from the number of people traveling;

[0071] Among them, if the week system corresponding to the time period to be predicted is different from the week system of the historical travel time, the historical travel proportion corresponding to the historical travel time closest to the time period to be predicted will be used as the final historical travel proportion.

[0072] In a preferred embodiment of the present invention, the historical travel proportion is generated according to the historical travel time proportion of the user in a period of time. For example, the travel proportion of users in coverage area A from July 1 to July 31, 2023 is 40%, 25%, 70%, 68%, 70%, 67%, 69%, 45%, 23%, 70%, 70%, 70%, 67%, 70%, 42%, 27%, 70%, 68%, 70%, 67%, 69%, 40%, 25%, 70%, 70%, 70%, 67%, 69%, 40%, 25%, 70%, so the travel proportion of users from July 1 to July 31, 2023 can be obtained. If the week system corresponding to the time period to be predicted is different from the week system of the historical travel time, the historical travel proportion corresponding to the historical travel time that is the same as the week system of the time period to be predicted and is closest to the time period to be predicted will be used as the final historical travel proportion. For example, the time to be predicted is July 14, 2024. Since July 14, 2023 is Friday and July 14, 2024 is Sunday, the travel proportion of the historical time that is close to July 14 and is Sunday (July 16, 2023) can be determined as the travel proportion of July 14, 2024. Therefore, the travel proportion of July 14, 2024 is 27%. Then, the number of people traveling is calculated by multiplying the historical travel proportion and the current number of users; and the number of the first users in the time period to be predicted is calculated by subtracting the current number of users and the number of people traveling.

[0073] Preferably, the historical enterprise production information includes historical enterprise production proportion;

[0074] Determining the first number of enterprises in the time period to be predicted according to the historical enterprise production information and the current number of enterprises includes:

[0075] The number of production enterprises is calculated by multiplying the historical production proportion of enterprises and the current number of enterprises;

[0076] Based on the current number of enterprises and the number of production enterprises, the first number of enterprises in the forecasted time period is obtained by subtracting them.

[0077] Schematically, the training of the user load prediction model includes:

[0078] Obtain historical temperature values, historical meteorological information, historical residential electricity prices, historical number of users, the initial first bird's nest position of the model, and the initial first optimization parameters;

[0079] Input the historical temperature value, historical meteorological information, historical residential electricity price, historical number of users, current first optimization parameter and current first bird's nest position into the user load prediction model to be trained for optimization, and obtain the first target bird's nest position; wherein the current first bird's nest position optimized for the first time is the initial first bird's nest position; the current first optimization parameter optimized for the first time is the initial first optimization parameter;

[0080] The first bird's nest fitness is calculated according to the current first bird's nest position and the first target bird's nest position; and the current first bird's nest position and the current first optimization parameter are updated according to the first bird's nest fitness until a preset number of iterations is reached to obtain the final first bird's nest position;

[0081] According to the first penalty coefficient corresponding to the final first bird's nest position and the first width of the radial basis kernel function, a trained user load prediction model is constructed.

[0082] In a preferred embodiment of the present invention, the user load forecasting model is obtained by training a cuckoo optimized support vector machine (CS-SVM). The initial optimization parameters include an initial penalty coefficient, an initial width of a radial basis kernel function, and an initial probability. The following three ideal states are assumed to simulate the behavior of a cuckoo looking for a nest to demonstrate the cuckoo optimized search CS algorithm, as follows:

[0083] (1) Assume that each cuckoo lays only one egg each time, and the eggs are placed randomly in any nest;

[0084] (2) A set of bird nest locations are randomly selected using a random method and passed on to the next generation as the optimal bird nest (optimal solution);

[0085] (3) Assume that the number of cuckoo nests is N, and the probability that a cuckoo nest owner finds a cuckoo egg during the nest search is P a .

[0086] Based on the above assumed ideal state, the specific process of updating the location of the cuckoo's nest is as follows: Assume that there are n cuckoos and a D-dimensional search space. The position of each cuckoo i can be represented as a D-dimensional vector x i , representing its position in the search space.

[0087] 1. Initialize the position of the cuckoo. The initial position of each cuckoo can be randomly generated:

[0088]

[0089] in, represents the initial position of the i-th cuckoo, representing its coordinates in the search space; LB represents the lower bound of the search space, representing the minimum value of each dimension; UB represents the upper bound of the search space, representing the maximum value of each dimension; rand(0,1) means generating a random number between 0 and 1.

[0090] 2. Update the cuckoo's position based on the following formula:

[0091]

[0092] in, represents the updated position of the i-th cuckoo (the position of the i-th cuckoo after the t+1th iteration); represents the position of the i-th cuckoo after the t-th iteration; α0 represents the step size control amount; Indicated in The fitness gradient of the position; r represents random noise, which increases the randomness of the search, r = rand (1, D); L (λ) represents the search path, Represents the optimization objective function of the CS-SVM model.

[0093] To ensure that the cuckoo's position remains within the search space, constraints can be added:

[0094]

[0095] The search path L(λ) in the embodiment of the present invention obeys the Levy distribution:

[0096] L(λ)~Levy(μ,σ)

[0097] Among them, λ represents the jump size, and its value is 0<λ≤5; μ represents the location parameter, which is usually 0; σ represents the scale parameter, which affects the jump size.

[0098] The Levy distribution in the embodiment of the present invention is calculated based on the Mantegna algorithm, so the search path L(λ) can be expressed as:

[0099]

[0100] Among them, U represents a uniformly distributed random vector, V represents an exponentially distributed random variable, α and β represent parameters that control the shape of the distribution, and Γ represents the standard Gamma function.

[0101] According to the third ideal state of the hypothesis, the probability that the cuckoo nest owner finds the bird eggs during the nest search is P a , the nest owner can abandon or build a new nest, which is equivalent to abandoning part of the solution, that is, the new solution generated by iteration is as follows:

[0102]

[0103] Among them, γ represents the scaling factor, which is a standard (0,1) distribution and is used to monitor the probability that the newly generated solution is replaced; are two random solutions in the t-th generation iteration process.

[0104] Determine the current optimal nest position according to the best fitness value of the nest, and update other nest positions through the current optimal nest position to generate a new set of nest positions, and compare the current optimal nest position with the new set of nest positions to obtain the prediction error. Obtain the better nest position according to the prediction error, and compare the probability of eliminating the cuckoo egg with the value of the random number. When the probability of eliminating the cuckoo egg is less than the random number, replace the current optimal nest position with a new set of nest positions, and judge whether the iteration end condition is reached. If so, jump out to find the value of the optimal nest position, otherwise continue to find the optimal nest position. The iteration end condition includes the maximum number of iterations or the optimal fitness value determined by the width of the radial basis kernel function. Update the parameters in the cuckoo optimization support vector machine CS-SVM model to the optimal nest position to establish a load prediction model, and complete the training of the user load prediction model.

[0105] In another preferred embodiment of the present invention, obtain the historical temperature value, historical meteorological information, historical residential electricity price, historical user quantity, the initial first nest position of the model, and the initial optimization parameters; input the historical temperature value, historical meteorological information, historical residential electricity price, historical user quantity, the current optimization parameters, and the current first nest position into the user load prediction model to be trained for optimization to obtain the first target nest position (the current optimal nest position, that is, the current optimal solution); among them, the current first nest position for the first optimization is the initial first nest position; the current optimization parameters for the first optimization are the initial optimization parameters; calculate the first nest fitness according to the current first nest position and the first target nest position; and update the current first nest position and the current optimization parameters according to the first nest fitness until the preset number of iterations is reached to obtain the final first nest position; construct the trained user load prediction model according to the first penalty coefficient and the first width of the radial basis kernel function corresponding to the final first nest position. Among them, compare the magnitude of P a (the probability of eliminating the cuckoo egg) with the value of the random number X. When P a < X, then replace the original (discarded) position with the newly generated cuckoo nest position, compare the fitness function values before and after this adjustment, and finally select the optimal cuckoo nest position and fitness value.

[0106] Specifically, the training process of the enterprise load forecasting model is the same as the training process of the user load forecasting model.

[0107] First, obtain historical temperature values, historical meteorological information, historical industrial electricity prices, historical number of enterprises, and the initial second bird's nest position and initial second optimization parameters of the model;

[0108] Input the historical temperature value, historical meteorological information, historical industrial electricity price, historical enterprise number, current second optimization parameter and current second bird's nest position of the model into the enterprise load forecasting model to be trained for optimization, and obtain the second target bird's nest position; wherein the current second bird's nest position optimized for the first time is the initial second bird's nest position; the current second optimization parameter optimized for the first time is the initial second optimization parameter;

[0109] The second bird's nest fitness is calculated according to the current second bird's nest position and the second target bird's nest position; and the current second bird's nest position and the current second optimization parameter are updated according to the second bird's nest fitness until a preset number of iterations is reached to obtain a final second bird's nest position;

[0110] According to the second penalty coefficient corresponding to the final second bird's nest position and the second width of the radial basis kernel function, a trained enterprise load forecasting model is constructed.

[0111] Specifically, the user load forecasting model includes: a first user power load forecasting layer for representing the influence of temperature, a second user power load forecasting layer for representing the influence of residential electricity price, and a user total power load forecasting layer;

[0112] The predicted temperature value, the predicted residential electricity price, the current meteorological information and the number of first users are input into a preset user load prediction model to perform user load prediction to obtain the user electricity load value, including:

[0113] Determine a first user load prediction function of a first user power load prediction layer by comparing the predicted temperature value with a preset first temperature threshold;

[0114] Calculate the first user power load forecast value of the first user power load forecast layer according to the predicted temperature value, the number of first users and the determined first user load forecast function;

[0115] Determine a second user load prediction function of a second user power load prediction layer by comparing the predicted residential power price with a preset residential power price threshold;

[0116] Convert the current meteorological information into corresponding values ​​to obtain seasonal factors;

[0117] Calculate the second user power load forecast value of the second user power load forecast layer according to the predicted residential power price, seasonal factors and the second user load forecast function;

[0118] The first user power load prediction value and the second user power load prediction value are input into the user power total load prediction layer, added and summed to obtain the user power load value.

[0119] In a preferred embodiment of the present invention, the first user power load prediction layer, the second user power load prediction layer and the user total power load prediction layer can all be obtained by fitting a linear regression model. The first user load prediction function of the first user power load prediction layer is determined by comparing the predicted temperature value and the preset first temperature threshold. The preset first temperature threshold is determined according to the geographical location of the coverage area. For example, the preset first temperature threshold of the coverage area in Beijing is different from the preset first temperature threshold in Guangzhou. Therefore, considering the impact of temperature changes on users, for the case where the predicted temperature value is less than the preset first temperature threshold, the first user load prediction function of the first user power load prediction layer is:

[0120]

[0121] Optionally, when the predicted temperature value is greater than or equal to a preset first temperature threshold, the first user load prediction function called by the first user power load prediction layer is:

[0122]

[0123] Then, according to the predicted temperature value, the number of first users and the determined first user load prediction function, the first user power load prediction value of the first user power load prediction layer is calculated. Wherein, Load(U1(t),T(t)) represents the first user power load prediction value, t represents the time to be predicted, U1(t) represents the number of first users at the time to be predicted t, T(t) represents the predicted temperature value at the time to be predicted t, e represents the base of the exponential function, T base represents the preset reference temperature value, and a1, a2, a3 and a4 represent the impact factors of the predicted temperature value on the user. Therefore, it can be understood that the first user power load prediction layer calls different first user load prediction functions according to the predicted temperature value, combines the predicted temperature value and the number of first users for prediction, and obtains the first user power load prediction value.

[0124] The second user load prediction function of the second user electricity load prediction layer is determined by comparing the predicted residential electricity price with the preset residential electricity price threshold. The second user electricity load prediction layer includes multiple second user load prediction functions. In order to improve the accuracy of user electricity load prediction, for different predicted residential electricity prices, the second user electricity load prediction layer performs electricity load prediction through different second user load prediction functions. The preset residential electricity price is determined according to the location of the coverage area. For example, the preset residential electricity price in the city center and the preset residential electricity price in the suburbs are different. Furthermore, taking into account the impact of electricity price fluctuations and seasonal changes on users, for the case where the predicted residential electricity price is less than the preset residential electricity price threshold, the second user load prediction function called by the second user electricity load prediction layer is:

[0125]

[0126] When the predicted residential electricity price is greater than or equal to the preset residential electricity price threshold, the second user load prediction function called by the second user load prediction layer is:

[0127]

[0128] Among them, Load1(P(t), S(t)) represents the predicted value of the electricity load of the second user, t represents the time to be predicted, P(t) represents the predicted residential electricity price at the time to be predicted t, S(t) represents the seasonal factor corresponding to the current season at the time to be predicted t, e represents the base of the exponential function, β1, β2, β3 and β4 represent the influencing factors of the predicted electricity price.

[0129] For the seasonal factor, the current meteorological information can be converted into a corresponding value. For example, if the current season of the current meteorological information is spring, the seasonal factor of spring is 1, the seasonal factor of summer is 2, the seasonal factor of autumn is 3, and the seasonal factor of winter is 4. Therefore, it can be understood that the second user power load prediction layer calls different second user load prediction functions according to the predicted residential electricity price, combines the current season and the predicted residential electricity price for prediction, and obtains the second user power load prediction value.

[0130] The user power total load prediction layer adds the first user power load prediction value and the second user power load prediction value to obtain the user power total load value, thereby obtaining the user power load value output by the user power total load prediction layer. Therefore, the user power total load value L(U1(t)) at the time to be predicted can be expressed as:

[0131] L(U1(t))=Load(U1(t),T(t))+Load1(P(t),S(t))

[0132] The embodiment of the present invention can accurately predict the user power load value of the kilovolt substation at the time to be predicted through the user load prediction model combined with the predicted temperature value, current meteorological information, predicted residential electricity price and the load influencing factors of the first user number, thereby accurately predicting the total load value of the kilovolt substation at the time to be predicted based on the user power load value, thereby improving the accuracy of load prediction for the kilovolt substation and further improving the stability and reliability of the operation of the power grid system.

[0133] Preferably, the enterprise load forecasting model includes: a first enterprise power load forecasting layer for representing the influence of temperature, a second enterprise power load forecasting layer for representing the influence of industrial power price, and a total enterprise power load forecasting layer;

[0134] The predicted temperature value, the predicted industrial electricity price, the current meteorological information and the number of the first enterprise are input into a preset enterprise load forecasting model to perform enterprise load forecasting, and obtain the enterprise electricity load value, including:

[0135] Determine a first enterprise load forecasting function of a first enterprise power load forecasting layer by comparing the forecast temperature value with a preset second temperature threshold;

[0136] Calculate the first enterprise power load forecast value of the first enterprise power load forecast layer according to the forecast temperature value, the number of first enterprises and the determined first enterprise load forecast function;

[0137] Determine a second enterprise load forecasting function of a second enterprise power load forecasting layer by comparing the predicted industrial power price with a preset industrial power price threshold;

[0138] Calculate the second enterprise power load forecast value of the second enterprise power load forecast layer according to the forecast industrial power price, seasonal factors and the second enterprise load forecast function;

[0139] The first enterprise power load forecast value and the second enterprise power load forecast value are input into the enterprise power total load forecast layer, and are added together to obtain the enterprise power load value.

[0140] In a preferred embodiment of the present invention, the first enterprise power load forecasting layer, the second enterprise power load forecasting layer and the enterprise total power load forecasting layer can all be obtained by decision tree fitting. The first enterprise power load forecasting layer includes multiple first enterprise load forecasting functions. In order to improve the accuracy of enterprise production power load forecasting, for different forecast temperature values, the first enterprise power load forecasting layer performs enterprise power load forecasting through different first enterprise load forecasting functions.

[0141] According to the comparison between the forecast temperature value and the preset second temperature threshold, the first enterprise load forecast function of the first enterprise power load forecast layer is determined. Considering the impact of temperature change on the enterprise, when the forecast temperature value is less than the preset second temperature threshold, the first enterprise load forecast function called by the first enterprise power load forecast layer is:

[0142]

[0143] When the predicted temperature value is greater than or equal to the preset second temperature threshold, the first enterprise load forecasting function called by the first enterprise power load forecasting layer is:

[0144]

[0145] Among them, Load(U2(t),T(t)) represents the predicted value of the first enterprise's electricity load, t represents the time to be predicted, U2(t) represents the number of the first enterprise at the time to be predicted t, T(t) represents the predicted temperature value at the time to be predicted t, T avg Indicates the historical annual average temperature value of the covered area, T base represents the preset reference temperature value, e represents the base of the exponential function, and γ1 and γ2 represent the impact factors of the predicted temperature value on the enterprise. Therefore, it can be understood that the first enterprise power load forecasting layer calls different first enterprise load forecasting functions according to the predicted temperature value, combines the predicted temperature value and the number of first enterprises for forecasting, and obtains the first enterprise power load forecasting value.

[0146] According to the comparison between the predicted industrial electricity price and the preset industrial electricity price threshold, the second enterprise load forecast function of the second enterprise electricity load forecast layer is determined, and the second enterprise electricity load forecast value of the second enterprise electricity load forecast layer is calculated according to the predicted industrial electricity price, seasonal factors and the second enterprise load forecast function. The second enterprise electricity load forecast layer makes a forecast based on the current season and the predicted industrial electricity price, and obtains the second enterprise electricity load forecast value output by the second enterprise electricity load forecast layer. Similarly, the second enterprise electricity load forecast layer includes multiple second enterprise load forecast functions. In order to improve the accuracy of enterprise electricity load forecast, for different predicted industrial electricity prices, the second enterprise electricity load forecast layer uses different second enterprise load forecast functions to forecast enterprise electricity load. Therefore, considering the impact of electricity price fluctuations and seasonal changes on users, for the case where the predicted industrial electricity price is less than the preset industrial electricity price threshold, where the preset industrial electricity price threshold is determined according to the type of production enterprise, the second enterprise load forecast function called by the second enterprise electricity load forecast layer is:

[0147]

[0148] When the predicted industrial electricity price is greater than or equal to the preset industrial electricity price threshold, the second enterprise load forecasting function called by the second enterprise electricity load forecasting layer is:

[0149]

[0150] Among them, Load2(P(t),S(t)) represents the predicted value of the power load of the second enterprise, t represents the time to be predicted, P(t) represents the predicted industrial power price at the time to be predicted t, S(t) represents the seasonal factor of the current season at the time to be predicted t, e represents the base of the exponential function, P avg represents the historical average annual electricity price in the coverage area, and η1 and η2 represent the influencing factors of the forecast electricity price. Therefore, it can be understood that the second enterprise power load forecasting layer calls different second enterprise load forecasting functions according to the forecast industrial power price, combines the current season and the forecast industrial power price to make a forecast, and obtains the second enterprise power load forecast value.

[0151] The first enterprise power load forecast value and the second enterprise power load forecast value are input into the enterprise power total load forecast layer and added and summed to obtain the enterprise power load value. The enterprise power total load forecast layer adds and sums the first enterprise power load forecast value and the second enterprise power load forecast value to obtain the enterprise power load value, thereby obtaining the enterprise power load value output by the enterprise power total load forecast layer. The enterprise power load value L(U2(t)) at the time to be predicted can be expressed as:

[0152] L(U2(t))=Load(U2(t), T(t))+Load2(P(t), S(t))

[0153] The embodiment of the present invention can accurately predict the enterprise electricity load value of the kilovolt substation at the time to be predicted through the enterprise load forecasting model combined with the predicted temperature value, the current season, the predicted industrial electricity price and the load influencing factors of the first enterprise quantity, thereby accurately predicting the total load value of the kilovolt substation at the time to be predicted based on the enterprise electricity load value, thereby improving the accuracy of the load forecast of the kilovolt substation, and further improving the stability and reliability of the operation of the power grid system.

[0154] In schematic form, the total load value of the substation area to be processed within the time period to be predicted is calculated based on the user power load value, the enterprise power load value and the predicted temperature value, including:

[0155] Calculating a temperature error load value according to a preset basic error load value, the predicted temperature value, and a preset temperature threshold;

[0156] The total load value of the substation area to be processed within the time period to be predicted is calculated based on the user power load value, the enterprise power load value, the temperature error load value and the preset power weight value.

[0157] In a preferred embodiment of the present invention, the calculation formula of the total load value of the substation area to be processed within the time period to be predicted is as follows:

[0158] L(t)=α*L(U1(t))+λ*L(U2(t))+∈(t)

[0159]

[0160] Wherein, t represents the time to be predicted, L(t) represents the total load value of the kilovolt substation at the time to be predicted t, U1(t) represents the current number of users at the time to be predicted t, L(U1(t)) represents the user power load value at the time to be predicted t, U2(t) represents the current number of enterprises at the time to be predicted t, L(U2(t)) represents the enterprise power load value at the time to be predicted t, α and λ represent the preset weight values, ε(t) represents the temperature error load value at the time to be predicted t, ∈0 represents the preset basic error load value, e represents the base of the exponential function, μ represents the adjustment parameter, T(t) represents the predicted temperature value at the time to be predicted t, T threshold Indicates the preset temperature threshold.

[0161] In another preferred embodiment of the present invention, the substation load forecasting method is applicable to a 35 kV substation, and for the coverage area of ​​the 35 kV substation, the forecast temperature value, the forecast residential electricity price, the forecast industrial electricity price, the current meteorological information, the historical user travel information, the historical enterprise production information, the current number of users and the current number of enterprises are obtained; the first number of users in the time period to be predicted is determined based on the historical user travel information and the current number of users; and the first number of enterprises in the time period to be predicted is determined based on the historical enterprise production information and the current number of enterprises; the forecast temperature value is used as the forecast value of the substation load forecasting method; the forecast residential electricity price, the forecast industrial electricity price, the current meteorological information, the historical user travel information, the historical enterprise production information, the current number of users ... The temperature value, the predicted residential electricity price, the current meteorological information and the first number of users are input into a preset user load forecasting model to perform user load forecasting and obtain the user electricity load value; the predicted temperature value, the predicted industrial electricity price, the current meteorological information and the first number of enterprises are input into a preset enterprise load forecasting model to perform enterprise load forecasting and obtain the enterprise electricity load value; finally, the total load value of the 35 kV substation in the time period to be predicted is calculated based on the user electricity load value, the enterprise electricity load value and the predicted temperature value, thereby improving the stability and reliability of the operation of the power grid system.

[0162] By implementing this embodiment, the forecast temperature value, forecast residential electricity price, forecast industrial electricity price, current meteorological information, historical user travel information, historical enterprise production information, current number of users and current number of enterprises of the substation area to be processed in the time period to be predicted are obtained; the first number of users in the time period to be predicted is determined based on the historical user travel information and the current number of users; and the first number of enterprises in the time period to be predicted is determined based on the historical enterprise production information and the current number of enterprises; the forecast temperature value, forecast residential electricity price, current meteorological information and the first number of users are input into a preset user load prediction model to perform user load prediction and obtain the user electricity load value; the forecast temperature value, forecast industrial electricity price, current meteorological information and the first number of enterprises are input into a preset enterprise load prediction model to perform enterprise load prediction and obtain the enterprise electricity load value; the total load value of the substation area to be processed in the time period to be predicted is calculated based on the user electricity load value, the enterprise electricity load value and the forecast temperature value. By comprehensively considering multiple factors such as temperature, electricity price, meteorological information, number of users and number of enterprises, the comprehensive impact of different factors on the predicted load is captured, the impact of historical load data on load forecasting is reduced, and when emergencies occur or seasonal changes occur, the impact of meteorological information is taken into account. It is possible to respond to the impact of emergencies or seasonal changes, accurately simulate and predict load changes in the substation area, so as to improve the accuracy of load forecasting for kilovolt substations; at the same time, by separately predicting the electricity load of users and enterprises, the accuracy of load forecasting is provided, thereby improving the stability and reliability of the operation of the power grid system.

[0163] See also Figure 2 , is a schematic diagram of the structure of a substation load prediction device provided by an embodiment of the present invention, comprising:

[0164] The data acquisition module is used to obtain the predicted temperature value, predicted residential electricity price, predicted industrial electricity price, current meteorological information, historical user travel information, historical enterprise production information, current number of users and current number of enterprises in the substation area to be processed within the predicted time period;

[0165] A predicted quantity determination module, used to determine the first number of users in the time period to be predicted based on the historical user travel information and the current number of users; and to determine the first number of enterprises in the time period to be predicted based on the historical enterprise production information and the current number of enterprises;

[0166] A user load prediction module, used for inputting the predicted temperature value, the predicted residential electricity price, the current meteorological information and the number of first users into a preset user load prediction model to perform user load prediction and obtain a user electricity load value;

[0167] An enterprise load forecasting module, used for inputting the predicted temperature value, the predicted industrial electricity price, the current meteorological information and the number of the first enterprise into a preset enterprise load forecasting model to perform enterprise load forecasting and obtain an enterprise electricity load value;

[0168] The total load prediction module is used to calculate the total load value of the substation area to be processed within the time period to be predicted based on the user power load value, the enterprise power load value and the predicted temperature value.

[0169] The present invention provides a substation load prediction device. According to a data acquisition module, the predicted temperature value, predicted residential electricity price, predicted industrial electricity price, current meteorological information, historical user travel information, historical enterprise production information, current number of users and current number of enterprises in the substation area to be processed within the predicted time period are obtained; in a predicted number determination module, the first number of users within the predicted time period is determined according to the historical user travel information and the current number of users; and the first number of enterprises within the predicted time period is determined according to the historical enterprise production information and the current number of enterprises; through the user load prediction module, The predicted temperature value, the predicted residential electricity price, the current meteorological information and the first number of users are input into the preset user load prediction model to predict the user load and obtain the user load value; through the enterprise load prediction module, the predicted temperature value, the predicted industrial electricity price, the current meteorological information and the first number of enterprises are input into the preset enterprise load prediction model to predict the enterprise load and obtain the enterprise load value; finally, in the total load prediction module, according to the user load value, the enterprise load value and the predicted temperature value, the total load value of the substation area to be processed in the predicted time period is calculated. By comprehensively considering multiple factors such as temperature, electricity price, meteorological information, number of users and number of enterprises, the comprehensive impact of different factors on the predicted load is captured, the impact of historical load data on load prediction is reduced, and when emergencies occur or seasonal changes occur, the impact of meteorological information is taken into account, the impact of emergencies or seasonal changes can be dealt with, and the load changes in the substation area can be accurately simulated and predicted, so as to improve the accuracy of load prediction of the kilovolt substation; at the same time, by separately predicting the user and enterprise loads, the accuracy of load prediction is provided, thereby improving the stability and reliability of the operation of the power grid system.

[0170] It should be noted that the device embodiments described above are merely schematic, wherein the units described as separate components may or may not be physically separated, and the components displayed as units may or may not be physical units, that is, they may be located in one place, or they may be distributed on multiple network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the scheme of this embodiment. In addition, in the accompanying drawings of the device embodiments provided by the present invention, the connection relationship between the modules indicates that there is a communication connection between them, which may be specifically implemented as one or more communication buses or signal lines. A person of ordinary skill in the art may understand and implement it without paying any creative effort.

[0171] Those skilled in the art can clearly understand that, for the sake of convenience and brevity, the specific working process of the device described above can refer to the corresponding process in the aforementioned method embodiment, and will not be repeated here.

[0172] Another embodiment of the present invention further provides a terminal device, including a processor, a memory, and a computer program stored in the memory and configured to be executed by the processor, wherein the processor implements a substation load forecasting method as described in the above embodiment when executing the computer program. The terminal device may be a computing device such as a desktop computer, a notebook, a PDA, and a cloud server. The terminal device may include, but is not limited to, a processor and a memory.

[0173] The processor may be a central processing unit (CPU), or other general-purpose processors, digital signal processors (DSP), application-specific integrated circuits (ASIC), field-programmable gate arrays (FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. A general-purpose processor may be a microprocessor or any conventional processor, etc. The processor is the control center of the terminal device, and uses various interfaces and lines to connect various parts of the entire terminal device.

[0174] The memory can be used to store the computer program, and the processor realizes various functions of the terminal device by running or executing the computer program stored in the memory and calling the data stored in the memory. The memory can mainly include a program storage area and a data storage area, wherein the program storage area can store an operating system, at least one application required for a function, etc.; the data storage area can store data created according to the use of the mobile phone, etc. In addition, the memory can include a high-speed random access memory, and can also include a non-volatile memory, such as a hard disk, a memory, a plug-in hard disk, a smart memory card (Smart Media Card, SMC), a secure digital (Secure Digital, SD) card, a flash card (Flash Card), at least one disk storage device, a flash memory device or other volatile solid-state storage device.

[0175] Another embodiment of the present invention provides a computer-readable storage medium, which includes a stored computer program, wherein when the computer program is running, the device where the computer-readable storage medium is located is controlled to execute a substation load forecasting method described in the above embodiment.

[0176] The storage medium is a computer-readable storage medium, and the computer program is stored in the computer-readable storage medium. When the computer program is executed by the processor, the steps of each of the above-mentioned method embodiments can be implemented. Among them, the computer program includes computer program code, and the computer program code can be in source code form, object code form, executable file or some intermediate form. The computer-readable medium may include: any entity or device capable of carrying the computer program code, recording medium, USB flash drive, mobile hard disk, magnetic disk, optical disk, computer memory, read-only memory (ROM), random access memory (RAM), electric carrier signal, telecommunication signal and software distribution medium.

[0177] The above is a preferred embodiment of the present invention. It should be pointed out that a person skilled in the art can make several improvements and modifications without departing from the principle of the present invention. These improvements and modifications are also considered to be within the scope of protection of the present invention.

Claims

1. A method for predicting load of a substation, characterized in that: include: Obtain the forecast temperature value, forecast residential electricity price, forecast industrial electricity price, current meteorological information, historical user travel information, historical enterprise production information, current number of users and current number of enterprises in the substation area to be processed within the forecast time period; Determining the number of first users within the time period to be predicted according to the historical user travel information and the current number of users; and determining the number of first enterprises within the time period to be predicted based on the historical enterprise production information and the current number of enterprises; Inputting the predicted temperature value, the predicted residential electricity price, the current meteorological information and the number of first users into a preset user load prediction model to perform user load prediction and obtain a user electricity load value; Inputting the predicted temperature value, the predicted industrial electricity price, the current meteorological information and the number of the first enterprise into a preset enterprise load forecasting model to perform enterprise load forecasting and obtain the enterprise electricity load value; The total load value of the substation area to be processed within the time period to be predicted is calculated based on the user power load value, the enterprise power load value and the predicted temperature value.

2. A method for predicting substation load according to claim 1, characterized in that: The historical user travel information includes historical travel proportion and historical travel time; Determining the number of first users in the time period to be predicted according to the historical user travel information and the current number of users includes: The number of people traveling is calculated by multiplying the historical travel ratio and the current number of users. The number of first users in the time period to be predicted is calculated by subtracting the current number of users from the number of people traveling; Among them, if the week system corresponding to the time period to be predicted is different from the week system of the historical travel time, the historical travel proportion corresponding to the historical travel time closest to the time period to be predicted will be used as the final historical travel proportion.

3. A method for predicting substation load according to claim 1, characterized in that: The historical enterprise production information includes the historical enterprise production proportion; Determining the first number of enterprises in the time period to be predicted according to the historical enterprise production information and the current number of enterprises includes: The number of production enterprises is calculated by multiplying the historical production proportion of enterprises and the current number of enterprises; Based on the current number of enterprises and the number of production enterprises, the first number of enterprises in the forecasted time period is obtained by subtracting them.

4. A method for predicting substation load according to claim 1, characterized in that: The training of the user load prediction model includes: Obtain historical temperature values, historical meteorological information, historical residential electricity prices, historical number of users, the initial first bird's nest position of the model, and the initial first optimization parameters; Input the historical temperature value, historical meteorological information, historical residential electricity price, historical number of users, current first optimization parameter and current first bird's nest position into the user load prediction model to be trained for optimization, and obtain the first target bird's nest position; wherein the current first bird's nest position optimized for the first time is the initial first bird's nest position; the current first optimization parameter optimized for the first time is the initial first optimization parameter; The first bird's nest fitness is calculated according to the current first bird's nest position and the first target bird's nest position; and the current first bird's nest position and the current first optimization parameter are updated according to the first bird's nest fitness until a preset number of iterations is reached to obtain the final first bird's nest position; According to the first penalty coefficient corresponding to the final first bird's nest position and the first width of the radial basis kernel function, a trained user load prediction model is constructed.

5. A method for predicting substation load according to claim 1, characterized in that: The training of the enterprise load forecasting model includes: Obtain historical temperature values, historical meteorological information, historical industrial electricity prices, historical number of enterprises, and the initial second bird's nest position and initial second optimization parameters of the model; Input the historical temperature value, historical meteorological information, historical industrial electricity price, historical enterprise number, current second optimization parameter and current second bird's nest position of the model into the enterprise load forecasting model to be trained for optimization, and obtain the second target bird's nest position; wherein the current second bird's nest position optimized for the first time is the initial second bird's nest position; the current second optimization parameter optimized for the first time is the initial second optimization parameter; The second bird's nest fitness is calculated according to the current second bird's nest position and the second target bird's nest position; and the current second bird's nest position and the current second optimization parameter are updated according to the second bird's nest fitness until a preset number of iterations is reached to obtain a final second bird's nest position; According to the second penalty coefficient corresponding to the final second bird's nest position and the second width of the radial basis kernel function, a trained enterprise load forecasting model is constructed.

6. A method for predicting substation load according to claim 1, characterized in that: The user load prediction model includes: a first user power load prediction layer for representing the influence of temperature, a second user power load prediction layer for representing the influence of residential electricity price, and a user total power load prediction layer; The predicted temperature value, the predicted residential electricity price, the current meteorological information and the number of first users are input into a preset user load prediction model to perform user load prediction to obtain the user electricity load value, including: Determine a first user load prediction function of a first user power load prediction layer by comparing the predicted temperature value with a preset first temperature threshold; Calculate the first user power load forecast value of the first user power load forecast layer according to the predicted temperature value, the number of first users and the determined first user load forecast function; Determine a second user load prediction function of a second user power load prediction layer by comparing the predicted residential power price with a preset residential power price threshold; Convert the current meteorological information into corresponding values ​​to obtain seasonal factors; Calculate the second user power load forecast value of the second user power load forecast layer according to the predicted residential power price, seasonal factors and the second user load forecast function; The first user power load prediction value and the second user power load prediction value are input into the user power total load prediction layer, added and summed to obtain the user power load value.

7. A method for predicting substation load according to claim 6, characterized in that: The enterprise load forecasting model includes: a first enterprise power load forecasting layer for representing the influence of temperature, a second enterprise power load forecasting layer for representing the influence of industrial power price, and a total enterprise power load forecasting layer; The predicted temperature value, the predicted industrial electricity price, the current meteorological information and the number of the first enterprise are input into a preset enterprise load forecasting model to perform enterprise load forecasting, and obtain the enterprise electricity load value, including: Determine a first enterprise load forecasting function of a first enterprise power load forecasting layer by comparing the forecast temperature value with a preset second temperature threshold; Calculate the first enterprise power load forecast value of the first enterprise power load forecast layer according to the forecast temperature value, the number of first enterprises and the determined first enterprise load forecast function; Determine a second enterprise load forecasting function of a second enterprise power load forecasting layer by comparing the predicted industrial power price with a preset industrial power price threshold; Calculate the second enterprise power load forecast value of the second enterprise power load forecast layer according to the forecast industrial power price, seasonal factors and the second enterprise load forecast function; The first enterprise power load forecast value and the second enterprise power load forecast value are input into the enterprise power total load forecast layer, and are added together to obtain the enterprise power load value.

8. A method for predicting substation load according to claim 1, characterized in that: The total load value of the substation area to be processed within the time period to be predicted is calculated according to the user power load value, the enterprise power load value and the predicted temperature value, including: Calculating a temperature error load value according to a preset basic error load value, the predicted temperature value, and a preset temperature threshold; The total load value of the substation area to be processed within the time period to be predicted is calculated based on the user power load value, the enterprise power load value, the temperature error load value and the preset power weight value.

9. A substation load prediction device, characterized in that: include: The data acquisition module is used to obtain the predicted temperature value, predicted residential electricity price, predicted industrial electricity price, current meteorological information, historical user travel information, historical enterprise production information, current number of users and current number of enterprises in the substation area to be processed within the predicted time period; A predicted number determination module, used to determine the number of first users in a time period to be predicted based on the historical user travel information and the current number of users; and determining the number of first enterprises within the time period to be predicted based on the historical enterprise production information and the current number of enterprises; A user load prediction module, used for inputting the predicted temperature value, the predicted residential electricity price, the current meteorological information and the number of first users into a preset user load prediction model to perform user load prediction and obtain a user electricity load value; An enterprise load forecasting module, used for inputting the predicted temperature value, the predicted industrial electricity price, the current meteorological information and the number of the first enterprise into a preset enterprise load forecasting model to perform enterprise load forecasting and obtain an enterprise electricity load value; The total load prediction module is used to calculate the total load value of the substation area to be processed within the time period to be predicted based on the user power load value, the enterprise power load value and the predicted temperature value.

10. A terminal device, characterized in that: The method comprises a processor, a memory, and a computer program stored in the memory and configured to be executed by the processor, wherein when the processor executes the computer program, a substation load prediction method as claimed in any one of claims 1 to 8 is implemented.

Images