A power grid net load prediction method and device, electronic equipment and storage medium

Abstract

The application discloses a power grid net load prediction method and device, electronic equipment and a storage medium. The method and device are applied to the electronic equipment, specifically, historical data of a power grid are acquired, and the historical data are filtered to obtain effective historical data; a plurality of typical scenes are generated based on the effective historical data; a gray scale prediction is performed according to the effective historical data to obtain a net load annual growth rate in a future time and a maximum net load in each prediction year in the future time, and the maximum net load is processed according to the plurality of typical scenes to obtain the power consumption in each prediction year. The scheme uses the typical scenes to depict the uncertainty of distributed new energy, indirectly predicts the net load annual growth rate through a gray prediction algorithm to depict the growth of the net load with time, and thus accurate prediction of the net load is realized. Over-redundancy or deficiency of capacity configuration planned when the power grid is planned can be avoided.

Description

Technical Field

[0001] This application relates to the field of power grid construction technology, and more specifically, to a method, apparatus, electronic device, and storage medium for predicting net load of a power grid. Background Technology

[0002] The annual increase in grid-connected capacity nationwide will reach over 100 million kilowatts, and by around 2030, the total installed capacity of new energy sources is expected to exceed 1.2 billion kilowatts, becoming the main power source of the power system. The high proportion of new energy integration will have a significant impact on the stability of the distribution network, power quality, and the planning and operation of the distribution network.

[0003] In the load forecasting stage of distribution network planning and operation, since the high proportion of distributed renewable energy connected to the distribution network reduces the distribution network load to a certain extent, the target of load forecasting will change from traditional load forecasting to forecasting the net load after subtracting renewable energy output from the load demand, which is equivalent to the load exchange between the local power grid and the upper-level power grid.

[0004] Due to the significant differences in the forecast time span, net load forecasting is typically categorized into two methods: grid dispatch and grid planning. The former's time span varies from 1 hour to 1 month depending on the objective; the latter forecasts on an annual basis, requiring predictions of load values ​​for several to several decades. Existing net load forecasting research mostly focuses on short-term load forecasts of a few days to a dozen days. For example, deterministic point forecasts based on deep neural networks and wavelet transforms, and non-deterministic probabilistic forecasts based on backpropagation neural networks or Gaussian process regression, have achieved good results. However, the forecasting performance for medium- and long-term loads based on grid planning is very poor. This may lead to excessive redundancy or undercapacity in the planned grid capacity configuration, resulting in resource waste or grid accidents. Summary of the Invention

[0005] In view of this, this application provides a method, apparatus, electronic device and storage medium for predicting the net load of a power grid, for predicting the net load of the power grid in the medium and long term, so as to avoid excessive redundancy or insufficiency in the capacity configuration planned when planning the power grid.

[0006] To achieve the above objectives, the following solution is proposed:

[0007] A method for predicting net load on a power grid, applied to electronic equipment, the method comprising the steps of:

[0008] Obtain historical power grid data;

[0009] The historical data of the power grid is filtered to obtain valid historical data;

[0010] Multiple typical scenarios are generated based on the aforementioned valid historical data;

[0011] Based on the effective historical data, grayscale prediction is performed to obtain the annual net load growth rate for future time and the maximum net load for each predicted year within the future time. The maximum net load is then processed according to the multiple typical scenarios to obtain the electricity consumption for each predicted year.

[0012] Optionally, the filtering of the historical power grid data includes the following steps:

[0013] The Lada criterion algorithm is used to filter out abnormal data in the historical power grid data to obtain the valid historical data.

[0014] Optionally, generating typical scenarios based on the valid historical data includes the following steps:

[0015] The effective historical data is processed using the k-means clustering algorithm to obtain the typical scenario.

[0016] Optionally, the process of processing the effective historical data based on the k-means clustering algorithm to obtain the typical scenario includes the following steps:

[0017] The valid historical data is normalized to obtain multiple samples;

[0018] K samples are randomly selected from the plurality of samples as initial cluster centers;

[0019] Clustering is performed based on the Euclidean distance between each sample and the initial cluster center to obtain K clusters;

[0020] For each of the classes, the cluster centers are recalculated, and the clusters are re-clustered based on the cluster centers. If the clustering result is different from the K classes, the process returns to the step of clustering based on the Euclidean distance between each sample and the initial cluster centers.

[0021] If the clustering result is the same as the K classes, then calculate the average distance of all the samples to the nearest cluster center, and increase the number of classes in the clustering result. If the number of classes in the clustering result is less than or equal to a preset threshold, then return to the step of randomly selecting K samples from the multiple samples as the initial cluster centers.

[0022] If the number of clusters in the clustering result is greater than the preset threshold, then the appropriate number of clusters is determined according to the elbow method, and the corresponding clustering result is output.

[0023] The clustering results are normalized and restored according to the historical power grid data of the most recent year to obtain the multiple typical scenarios.

[0024] A power grid net load forecasting device, applied to electronic equipment, the forecasting device comprising:

[0025] The data acquisition module is configured to acquire historical power grid data;

[0026] The data filtering module is configured to filter the historical data of the power grid to obtain valid historical data;

[0027] The scene generation module is configured to generate multiple typical scenes based on the effective historical data;

[0028] The prediction execution module is configured to perform grayscale prediction based on the effective historical data to obtain the annual net load growth rate for future time and the maximum net load for each predicted year within the future time, and to process the maximum net load according to the multiple typical scenarios to obtain the electricity consumption for each predicted year.

[0029] Optionally, the data filtering module is configured to use the Laida criterion algorithm to filter out abnormal data in the power grid historical data to obtain the valid historical data.

[0030] Optionally, the scene generation module is configured to process the effective historical data based on the k-means clustering algorithm to obtain the typical scene.

[0031] Optionally, the scene generation module includes:

[0032] The normalization processing unit is configured to normalize the valid historical data to obtain multiple samples;

[0033] The center selection unit is configured to randomly select K samples from the plurality of samples as initial cluster centers;

[0034] The first clustering unit is configured to cluster according to the Euclidean distance between each sample and the initial cluster center to obtain K classes;

[0035] The second clustering unit is configured to recalculate the cluster center for each of the classes and re-cluster based on the cluster center. If the clustering result is different from the K classes, the first clustering unit is controlled to re-cluster based on the Euclidean distance between each sample and the initial clustering center.

[0036] The mean calculation unit is configured to calculate the mean distance of all samples to the nearest cluster center if the clustering result is the same as the K classes, and increase the number of classes in the clustering result; if the number of classes in the clustering result is less than or equal to a preset threshold, control the center selection unit to randomly select K samples from the plurality of samples again as initial cluster centers.

[0037] The clustering determination unit is configured to determine an appropriate number of clusters based on the elbow method and output the corresponding clustering result if the number of clusters in the clustering result is greater than the preset threshold.

[0038] The generation execution unit is configured to normalize and restore the clustering results according to the historical power grid data of the most recent year to obtain the multiple typical scenarios.

[0039] An electronic device includes at least one processor and a memory connected to the processor, wherein:

[0040] The memory is used to store computer programs or instructions;

[0041] The processor is used to execute the computer program or instructions to enable the electronic device to implement the power grid net load prediction method as described above.

[0042] A storage medium is applied to an electronic device, characterized in that the storage medium carries one or more computer programs that can be executed by the electronic device, and when the electronic device executes the one or more computer programs, it can realize the power grid net load prediction method as described above.

[0043] As can be seen from the above technical solution, this application discloses a method, apparatus, electronic device, and storage medium for predicting grid net load. The method and apparatus are applied to electronic devices, specifically involving acquiring historical grid data and filtering it to obtain valid historical data; generating multiple typical scenarios based on the valid historical data; performing gray-scale prediction based on the valid historical data to obtain the annual net load growth rate for future times and the maximum net load for each predicted year within the future timeframe; and processing the maximum net load according to multiple typical scenarios to obtain the electricity consumption for each predicted year. This solution uses typical scenarios to characterize the uncertainty of distributed renewable energy sources and indirectly predicts the annual net load growth rate through a gray prediction algorithm to characterize the growth of net load over time, thereby achieving accurate prediction of net load. This can avoid excessive redundancy or insufficiency in the capacity configuration planned during grid planning. Attached Figure Description

[0044] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0045] Figure 1This is a flowchart illustrating a method for predicting net load of a power grid according to an embodiment of this application.

[0046] Figure 2a This is historical data for net load in 2015, broken down by hour.

[0047] Figure 2b An image showing how the mean distance L from each sample to the nearest cluster center varies with the number of categories k when generating typical scenarios for k-means clustering;

[0048] Figure 2c Historical data on average annual net load from 2015 to 2021;

[0049] Figure 2d The projected average annual net load for 2022-2030;

[0050] Figure 2e This is a graph showing the annual continuous net load for 2022.

[0051] Figure 3 This is a block diagram of a power grid net load prediction device according to an embodiment of this application;

[0052] Figure 4 This is a block diagram of an electronic device according to an embodiment of this application. Detailed Implementation

[0053] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the scope of protection of this application.

[0054] For current power grids, historical output data for distributed renewable energy sources over the past few years is often unavailable in practical engineering projects. Therefore, many proposed novel net load forecasting methods lack a realistic basis for implementation. Power grid companies' planning reports still rely on traditional load forecasting methods. Existing planning studies often fail to consider the impact of distributed renewable energy sources or directly use the average output of distributed renewable energy sources as the predicted output of renewable energy units. This may lead to excessive redundancy or undercapacity in power grid equipment configuration, resulting in resource waste or accidents. Based on the above analysis, this application proposes the following technical solutions to achieve medium- and long-term forecasting of power grid load, specifically including the following embodiments.

[0055] Example 1

[0056] Figure 1 This is a flowchart illustrating a method for predicting net load of a power grid, as described in an embodiment of this application.

[0057] like Figure 1 As shown, the prediction method provided in this embodiment is applied to electronic devices to predict electricity consumption for each predicted year in the future, such as 5 to 10 years. Here, the electronic device can be understood as a computer or server with data computing and information processing capabilities. The prediction method of this application includes the following steps:

[0058] S1. Obtain historical power grid data.

[0059] Obtain historical power grid data for the region to be predicted within the power grid to be forecasted over several years. Historical data is the foundation of load forecasting, and its accuracy and reliability directly affect the forecast accuracy. An accurate load forecast is the result of the combined effect of an effective and feasible forecasting method and high-quality historical data.

[0060] This embodiment takes the net load of Hengjie 110kV substation in Ningbo City, Zhejiang Province as the case study. Using the technical solution of this invention, based on the historical net load data of the substation from 2015 to 2021, the annual maximum net load, annual continuous load curve, and annual electricity consumption from 2022 to 2030 are predicted.

[0061] S2. Filter the historical data of the power grid to obtain valid historical data.

[0062] Load forecasting for power grid planning utilizes a large amount of historical data, which may contain anomalies caused by communication equipment failures or human error. Therefore, failure to verify the validity of historical data will impact forecasting accuracy. To improve load forecasting accuracy, this application employs the Laida criterion algorithm (also known as the 3σ criterion) to filter out anomalies, thus obtaining valid historical data excluding them. Assuming the historical data contains only random errors, the standard deviation is calculated and determined. An interval is then defined with a certain probability; errors exceeding this interval are considered gross errors, not random errors, and data containing such errors are considered suspicious. Suspicious data is verified, and if confirmed to be erroneous, it is discarded.

[0063] Let n be the sample size, then the sample data of the net load of the studied area changing over time can be expressed as:

[0064] x={x(t)}(t=1,2,...,n) (1)

[0065] If the mean of x is μ, the standard deviation is σ, and x approximately follows a normal distribution, then the probability that a sample in x falls within the range (μ-3σ, μ+3σ) is 0.9974. Therefore, it is assumed that if:

[0066] |x(t)-μ|≥3σ (2)

[0067] Then x(t) is considered suspicious data and its accuracy needs to be verified before it is discarded.

[0068] According to the Laida criterion, the data that satisfies equation (2) in the original data were screened year by year, and no suspicious data was found.

[0069] Next, using Matlab, the original data from 2015, μ-3σ, and μ+3σ were plotted on [the graph]. Figure 2a From Figure 2a As can be seen, all data points are within the range of (μ-3σ, μ+3σ). Therefore, it is considered that there is no suspicious data in the original data of 2015, and all of them can be used. The same applies to the data of other years.

[0070] S3. Generate multiple typical scenarios based on valid historical data.

[0071] The uncertainty of distributed renewable energy output poses significant challenges to net load forecasting in new distribution networks. Therefore, scenario analysis is employed to characterize the randomness and volatility of distributed renewable energy output. Initial scenarios are generated based on historical net load data for the studied region, and k-means clustering is used to reduce these scenarios, resulting in typical scenarios.

[0072] In this application, the elbow method is used to determine the number of categories k. The mean distance from each sample point to the nearest cluster center is calculated when k varies from 3 to 10, and the appropriate number of categories k is determined accordingly. b Let the specified number of categories be K, with an initial value of 3. For a given historical net load data x = {x(t)} (t = 1, 2, ..., n), the specific scheme for generating a typical scenario is as follows:

[0073] S31. Normalize the valid historical data to obtain multiple samples.

[0074] Because the historical annual net load average increases with each year, directly clustering multi-year historical data will result in clustering results influenced by annual differences in net load, failing to reflect the distribution pattern of net load throughout the year. Therefore, it is necessary to normalize the effective historical data on an annual basis to eliminate this influence. The normalization formula is:

[0075]

[0076] In the formula, This is the result after normalization.

[0077] S32. Randomly select K samples as initial cluster centers {C1, C2, ..., C...} K}, .

[0078] S33. Calculate the Euclidean distance from each sample to each initial cluster center, as shown in the following formula:

[0079]

[0080] By sequentially comparing the Euclidean distance of each sample to each initial cluster center, the sample is assigned to the cluster containing the nearest initial cluster center, resulting in K clusters {Q1, Q2, ..., Q...}. K}

[0081] S34. Recalculate the cluster centers. The formula for calculating the new cluster centers is:

[0082]

[0083] In the formula, |Q i | represents the number of samples in the i-th class.

[0084] If the clustering result is different from the previous clustering result, return to S33; otherwise, proceed to S35.

[0085] S35. Calculate the mean distance of each sample point to the nearest cluster center, then let K = K + 1. If K is less than or equal to 10, return to S32; otherwise, proceed to S36.

[0086] S36. Determine the appropriate number of categories k based on the elbow method. b Then output the corresponding clustering results.

[0087] S37, the obtained k b The generated k cluster centers can be obtained by normalizing and restoring them using the historical net load data of the most recent year. b A typical scenario The proportion of samples in each class to the total number of samples is the probability of the corresponding typical scenario occurring, which can be expressed by the formula:

[0088]

[0089] Typical scenes are generated using the k-means clustering algorithm, and the number of clusters k is determined using the elbow method. The graph showing the mean distance L from each sample point to the nearest cluster center as a function of the number of clusters k is shown below. Figure 2b As shown:

[0090] according to Figure 2b The optimal number of clusters can be determined by the elbow method to be 9, and the clustering results at this time are shown in Table 1:

[0091] Table 1 Clustering Results

[0092]

[0093]

[0094] S4. Make predictions based on valid historical data and typical scenarios.

[0095] Based on valid historical data, gray-scale prediction is performed to obtain the annual growth rate of net load in the future and the maximum net load for each predicted year in the future. The maximum net load is then processed according to multiple typical scenarios to obtain the electricity consumption for each predicted year.

[0096] Scenario analysis method characterizes the uncertainty of distributed renewable energy output by generating typical scenarios, while the net load is also increasing year by year. Therefore, based on historical data of annual average net load, a grey prediction algorithm is used to predict the future annual average net load, and the annual growth rate of net load is calculated from the predicted value to characterize the growth of net load over time.

[0097] Let the historical annual average net load series be:

[0098] L(0)={L(0)(t)}(t=1,2,...,N) (6)

[0099] The new sequence generated after one accumulation is:

[0100]

[0101] The whitening differential equation is:

[0102]

[0103] In the formula, a is called the development coefficient, and u is called the gray action quantity.

[0104] The obtained prediction model is:

[0105]

[0106] In the formula:

[0107]

[0108]

[0109] Then the cumulative sequence of predicted values ​​L (1) (t) is:

[0110]

[0111] The predicted value sequence L can be obtained by cumulative subtraction and restoration. (0) (t) is:

[0112] L(0) (t+1)=L (1) (t+1)-L (1) (13)

[0113] Let L(i) be the average net load for the i-th forecast year, then the annual growth rate α of the net load for the i-th forecast year is... i It can be represented as:

[0114]

[0115] The annual maximum net load L in the i-th forecast year m (i) is:

[0116] L m (i)=(1+α i )L m (i-1),( (15)

[0117] In the formula, L m (i-1) represents the maximum annual net load for the (i-1)th forecast year.

[0118] Furthermore, based on the typical scenarios obtained in (2) The typical scenario corresponding to the i-th prediction year can be obtained as follows: in:

[0119]

[0120] From this, the annual continuous net load curve for the i-th forecast year can be obtained, and thus the electricity consumption W for the i-th forecast year can be calculated. i For a typical scenario C in the i-th prediction year (i) j Its duration in hours t (i) j for:

[0121]

[0122] The electricity consumption for the i-th predicted year is:

[0123]

[0124] Average annual net load data from 2015 to 2021 are as follows: Figure 2c As shown:

[0125] Based on the raw data from 2015 to 2021, the grey prediction method was used to predict the data for 2022 to 2030. The prediction results are as follows. Figure 2d As shown:

[0126] from Figure 2dThe forecast results show that the average net load will continue to grow steadily from 2022 to 2030. Based on this, the annual net load growth rate from 2022 to 2030 was calculated, and the maximum net load for each year was further calculated, as shown in Table 2:

[0127] Table 2 Annual Net Load Growth Rate and Maximum Net Load Forecast

[0128]

[0129]

[0130] Furthermore, based on the generated typical scenarios, annual continuous net load curves for each forecast year can be plotted, and the total annual electricity consumption can be estimated. Taking 2022 as an example, the annual continuous net load curve for 2022 is as follows: Figure 2e As shown in Table 3, electricity consumption from 2022 to 2030 will be as follows:

[0131] Table 3 Electricity Consumption Forecast for 2022-2030

[0132]

[0133] As can be seen from the above technical solution, this embodiment provides a method for predicting the net load of a power grid. This method is applied to electronic equipment and specifically involves acquiring historical power grid data, filtering it to obtain valid historical data, generating multiple typical scenarios based on the valid historical data, performing gray-scale prediction based on the valid historical data to obtain the annual growth rate of net load in the future and the maximum net load for each predicted year in the future, and processing the maximum net load according to multiple typical scenarios to obtain the electricity consumption for each predicted year. This solution uses typical scenarios to characterize the uncertainty of distributed renewable energy, and indirectly predicts the annual growth rate of net load through a gray prediction algorithm to characterize the growth of net load over time, thereby achieving accurate prediction of net load and avoiding excessive redundancy or insufficiency in the capacity configuration planned during power grid planning.

[0134] The flowcharts and block diagrams in the accompanying drawings illustrate the architecture, functionality, and operation of possible implementations of systems, methods, and computer program products according to various embodiments of this disclosure. In this regard, each block in a flowchart or block diagram may represent a module, segment, or portion of code containing one or more executable instructions for implementing a specified logical function. It should also be noted that in some alternative implementations, the functions indicated in the blocks may occur in a different order than those indicated in the drawings. For example, two consecutively indicated blocks may actually be executed substantially in parallel, and they may sometimes be executed in reverse order, depending on the functions involved. It should also be noted that each block in the block diagrams and / or flowcharts, and combinations of blocks in the block diagrams and / or flowcharts, can be implemented using a dedicated hardware-based system that performs the specified function or operation, or using a combination of dedicated hardware and computer instructions.

[0135] Although the operations are described in a specific order, this should not be construed as requiring these operations to be performed in the specific order shown or in a sequential order. In certain environments, multitasking and parallel processing may be advantageous.

[0136] It should be understood that the steps described in the method embodiments of this disclosure may be performed in different orders and / or in parallel. Furthermore, the method embodiments may include additional steps and / or omit the steps shown. The scope of this disclosure is not limited in this respect.

[0137] Computer program code for performing the operations of this disclosure can be written in one or more programming languages ​​or a combination thereof, including but not limited to object-oriented programming languages ​​such as Java, Smalltalk, and C++, as well as conventional procedural programming languages ​​such as C or similar languages. The program code can be executed entirely on the user's computer, partially on the user's computer, as a standalone software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In cases involving remote computers, the remote computer can be connected to the user's computer via any type of network—including a local area network (LAN) or a wide area network (WAN)—or can be connected to an external computer.

[0138] Example 2

[0139] Figure 3 This is a block diagram of a power grid net load prediction device according to an embodiment of this application.

[0140] like Figure 3As shown, the prediction method provided in this embodiment is applied to electronic devices to predict electricity consumption for each predicted year in the future, such as 5 to 10 years. Here, the electronic device can be understood as a computer or server with data computing and information processing capabilities. The prediction device of this application includes a data acquisition module 10, a data filtering module 20, a scene generation module 30, and a prediction execution module 40.

[0141] The data acquisition module is used to acquire historical power grid data.

[0142] Obtain historical power grid data for the region to be predicted within the power grid to be forecasted over several years. Historical data is the foundation of load forecasting, and its accuracy and reliability directly affect the forecast accuracy. An accurate load forecast is the result of the combined effect of an effective and feasible forecasting method and high-quality historical data.

[0143] The data filtering module is used to filter historical power grid data to obtain valid historical data.

[0144] Load forecasting for power grid planning utilizes a large amount of historical data, which may contain anomalies caused by communication equipment failures or human error. Therefore, failure to verify the validity of historical data will impact forecasting accuracy. To improve load forecasting accuracy, this application employs the Laida criterion algorithm (also known as the 3σ criterion) to filter out anomalies, thus obtaining valid historical data excluding them. Assuming the historical data contains only random errors, the standard deviation is calculated and determined. An interval is then defined with a certain probability; errors exceeding this interval are considered gross errors, not random errors, and data containing such errors are considered suspicious. Suspicious data is verified, and if confirmed to be erroneous, it is discarded.

[0145] The scene generation module is used to generate multiple typical scenes based on valid historical data.

[0146] The uncertainty of distributed renewable energy output poses significant challenges to net load forecasting in new distribution networks. Therefore, scenario analysis is employed to characterize the randomness and volatility of distributed renewable energy output. Initial scenarios are generated based on historical net load data for the studied region, and k-means clustering is used to reduce these scenarios, resulting in typical scenarios.

[0147] In this application, the elbow method is used to determine the number of categories k. The mean distance from each sample point to the nearest cluster center is calculated when k varies from 3 to 10, and the appropriate number of categories k is determined accordingly. bLet the specified number of categories be K, with an initial value of 3. For a given historical net load data x = {x(t)} (t = 1, 2, ..., n), the scene generation module in this embodiment includes a normalization processing unit, a center selection unit, a first clustering unit, a second clustering unit, a mean calculation unit, a cluster determination unit, and a generation execution unit.

[0148] The normalization processing unit is used to normalize the valid historical data to obtain multiple samples.

[0149] Because the historical annual net load average increases with each year, directly clustering multi-year historical data will result in clustering results influenced by annual differences in net load, failing to reflect the distribution pattern of net load throughout the year. Therefore, it is necessary to normalize the effective historical data on an annual basis to eliminate this influence. The normalization formula is:

[0150]

[0151] In the formula, This is the result after normalization.

[0152] The center selection unit is used to randomly select K samples as initial cluster centers {C1, C2, ..., C...}. K}

[0153] The first clustering unit is used to calculate the Euclidean distance from each sample to each initial cluster center, as shown in the following formula:

[0154]

[0155] By sequentially comparing the Euclidean distance of each sample to each initial cluster center, the sample is assigned to the cluster containing the nearest initial cluster center, resulting in K clusters {Q1, Q2, ..., Q...}. K}

[0156] The second clustering unit is used to recalculate the cluster centers. The formula for calculating the new cluster centers is:

[0157]

[0158] In the formula, |Q i | represents the number of samples in the i-th class.

[0159] If the clustering result is different from the previous clustering result, the first clustering unit is controlled to recalculate the above Euclidean distance.

[0160] The mean calculation unit calculates the mean distance from each sample point to the nearest cluster center when the above clustering result is the same as the previous clustering result. Then, it sets K = K + 1. If K is less than or equal to 10, the control center selection unit selects the initial cluster center again.

[0161] The clustering determination unit is used to determine the appropriate number of clusters k based on the elbow method. b Then output the corresponding clustering results.

[0162] The generation execution unit is used to generate the obtained k b The generated k cluster centers can be obtained by normalizing and restoring them using the historical net load data of the most recent year. b A typical scenario The proportion of samples in each class to the total number of samples is the probability of the corresponding typical scenario occurring, which can be expressed by the formula:

[0163]

[0164] The prediction execution module is used to make predictions based on valid historical data and typical scenarios.

[0165] Based on valid historical data, gray-scale prediction is performed to obtain the annual growth rate of net load in the future and the maximum net load for each predicted year in the future. The maximum net load is then processed according to multiple typical scenarios to obtain the electricity consumption for each predicted year.

[0166] Scenario analysis method characterizes the uncertainty of distributed renewable energy output by generating typical scenarios, while the net load is also increasing year by year. Therefore, based on historical data of annual average net load, a grey prediction algorithm is used to predict the future annual average net load, and the annual growth rate of net load is calculated from the predicted value to characterize the growth of net load over time.

[0167] Let the historical annual average net load series be:

[0168] L (0) ={L (0) (t)}(t=1,2,...,N) (6)

[0169] The new sequence generated after one accumulation is:

[0170]

[0171] The whitening differential equation is:

[0172]

[0173] In the formula, a is called the development coefficient, and u is called the gray action quantity.

[0174] The obtained prediction model is:

[0175]

[0176] In the formula:

[0177]

[0178]

[0179] Then the cumulative sequence of predicted values ​​L (1) (t) is:

[0180]

[0181] The predicted value sequence L can be obtained by cumulative subtraction and restoration. (0) (t) is:

[0182] L (0) (t+1)=L (1) (t+1)-L (1) (13)

[0183] Let L(i) be the average net load for the i-th forecast year, then the annual growth rate α of the net load for the i-th forecast year is... i It can be represented as:

[0184]

[0185] The annual maximum net load L in the i-th forecast year m (i) is:

[0186] L m (i)=(1+α i )L m (i-1),( (15)

[0187] In the formula, L m (i-1) represents the maximum annual net load for the (i-1)th forecast year.

[0188] Furthermore, based on the typical scenarios obtained in (2) The typical scenario corresponding to the i-th prediction year can be obtained as follows: in:

[0189]

[0190] From this, the annual continuous net load curve for the i-th forecast year can be obtained, and thus the electricity consumption W for the i-th forecast year can be calculated. i For a typical scenario C in the i-th prediction year (i) j Its duration in hours t (i)j for:

[0191]

[0192] The electricity consumption for the i-th predicted year is:

[0193]

[0194] As can be seen from the above technical solution, this embodiment provides a power grid net load prediction device. This device is applied to electronic equipment and specifically acquires historical power grid data, filters it to obtain valid historical data, generates multiple typical scenarios based on the valid historical data, performs gray-scale prediction based on the valid historical data to obtain the annual net load growth rate in the future and the maximum net load for each predicted year in the future, and processes the maximum net load according to multiple typical scenarios to obtain the electricity consumption for each predicted year. This solution uses typical scenarios to characterize the uncertainty of distributed renewable energy, and indirectly predicts the annual net load growth rate through a gray prediction algorithm to characterize the growth of net load over time, thereby achieving accurate prediction of net load and avoiding excessive redundancy or insufficiency in the capacity configuration planned during power grid planning.

[0195] The units described in the embodiments of this disclosure can be implemented in software or in hardware. The name of a unit does not necessarily limit the unit itself; for example, the first acquisition unit can also be described as "a unit that acquires at least two Internet Protocol addresses".

[0196] The functions described above in this document can be performed, at least in part, by one or more hardware logic components. For example, exemplary types of hardware logic components that can be used, without limitation, include: Field Programmable Gate Arrays (FPGAs), Application-Specific Integrated Circuits (ASICs), Application Standard Products (ASSPs), System-on-Chip (SoCs), Complex Programmable Logic Devices (CPLDs), and so on.

[0197] Example 3

[0198] This embodiment provides an electronic device, referenced... Figure 4 The diagram illustrates a structural schematic suitable for implementing the electronic device in the embodiments of this disclosure. The terminal device in the embodiments of this disclosure may include, but is not limited to, mobile terminals such as mobile phones, laptops, digital broadcast receivers, PDAs (personal digital assistants), PADs (tablet computers), PMPs (portable multimedia players), in-vehicle terminals (e.g., in-vehicle navigation terminals), and fixed terminals such as digital TVs and desktop computers. This electronic device is merely an example and should not impose any limitations on the functionality and scope of use of the embodiments of this disclosure.

[0199] The electronic device may include a processing unit (e.g., a central processing unit, a graphics processing unit, etc.) 401, which can perform various appropriate actions and processes according to a program stored in a read-only memory (ROM) or a program loaded from an input device 406 into a random access memory (RAM) 403. The RAM also stores various programs and data required for the operation of the electronic device. The processing unit, ROM, and RAM are interconnected via a bus 404. An input / output (I / O) interface 405 is also connected to the bus 404.

[0200] Typically, the following devices can be connected to the I / O interface: input devices including, for example, touchscreens, touchpads, keyboards, mice, cameras, microphones, accelerometers, gyroscopes, etc.; output devices 407 including, for example, liquid crystal displays (LCDs), speakers, vibrators, etc.; storage devices including, for example, magnetic tapes, hard disks, etc.; and communication devices 409. Communication device 409 allows the electronic device to communicate wirelessly or wiredly with other devices to exchange data. Although electronic devices with various devices are shown in the figures, it should be understood that it is not required to implement or possess all of the devices shown. More or fewer devices may be implemented or possessed alternatively.

[0201] Example 4

[0202] This embodiment provides a computer-readable storage medium carrying one or more programs. When these programs are executed by an electronic device, the device acquires historical power grid data, filters it to obtain valid historical data, generates multiple typical scenarios based on the valid historical data, performs gray-scale prediction based on the valid historical data to obtain the annual net load growth rate for future time and the maximum net load for each predicted year within the future timeframe, and processes the maximum net load according to the multiple typical scenarios to obtain the electricity consumption for each predicted year. This solution uses typical scenarios to characterize the uncertainty of distributed renewable energy, and indirectly predicts the annual net load growth rate through a gray prediction algorithm to characterize the growth of net load over time, thereby achieving accurate prediction of net load and avoiding excessive redundancy or insufficiency in the capacity configuration planned during power grid planning.

[0203] It should be noted that the computer-readable medium described in this disclosure can be a computer-readable signal medium or a computer-readable storage medium, or any combination thereof. A computer-readable storage medium can be, for example,—but not limited to—an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any combination thereof. More specific examples of a computer-readable storage medium may include, but are not limited to: an electrical connection having one or more wires, a portable computer disk, a hard disk, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fiber, portable compact disk read-only memory (CD-ROM), optical storage device, magnetic storage device, or any suitable combination thereof. In this disclosure, a computer-readable storage medium can be any tangible medium containing or storing a program that can be used by or in connection with an instruction execution system, apparatus, or device. In this disclosure, a computer-readable signal medium can include a data signal propagated in baseband or as part of a carrier wave, carrying computer-readable program code. Such propagated data signals can take various forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination thereof. A computer-readable signal medium can be any computer-readable medium other than a computer-readable storage medium, which can send, propagate, or transmit a program for use by or in connection with an instruction execution system, apparatus, or device. The program code contained on the computer-readable medium can be transmitted using any suitable medium, including but not limited to: wires, optical fibers, RF (radio frequency), etc., or any suitable combination thereof.

[0204] The various embodiments in this specification are described in a progressive manner, with each embodiment focusing on the differences from other embodiments. The same or similar parts between the various embodiments can be referred to each other.

[0205] Although preferred embodiments of the present invention have been described, those skilled in the art, upon learning the basic inventive concept, can make other changes and modifications to these embodiments. Therefore, the appended claims are intended to be interpreted as including the preferred embodiments as well as all changes and modifications falling within the scope of the embodiments of the present invention.

[0206] Finally, it should be noted that in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or terminal device that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or terminal device. Without further limitations, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or terminal device that includes said element.

[0207] The technical solution provided by the present invention has been described in detail above. Specific examples have been used to illustrate the principle and implementation of the present invention. The description of the above embodiments is only for the purpose of helping to understand the method and core idea of ​​the present invention. At the same time, for those skilled in the art, there will be changes in the specific implementation and application scope based on the idea of ​​the present invention. Therefore, the content of this specification should not be construed as a limitation of the present invention.

Claims

1. A method for predicting net load on a power grid, applied to electronic equipment, characterized in that, The prediction method includes the following steps: Obtain historical power grid data; The Laida criterion algorithm is used to filter out abnormal data in the power grid historical data to obtain valid historical data; The effective historical data was processed using the k-means clustering algorithm to obtain several typical scenarios; Based on the effective historical data, grayscale prediction is performed to obtain the annual net load growth rate for future time and the maximum net load for each predicted year within the future time. The maximum net load is then processed according to the multiple typical scenarios to obtain the electricity consumption for each predicted year. The prediction model for grayscale prediction is as follows: ； a is the development coefficient, and u is the gray action quantity; Then the cumulative sequence of predicted values ​​L (1) (t) is: The predicted value sequence L is obtained by cumulative subtraction and restoration. (0) (t) is: Let L(i) be the average net load for the i-th forecast year, then the annual growth rate α of the net load for the i-th forecast year is... i It can be represented as: The annual maximum net load L in the i-th forecast year m (i) is: In the formula, L m (i-1) represents the maximum annual net load for the (i-1)th forecast year; Based on typical scenarios The typical scenario corresponding to the i-th prediction year is obtained as follows: ,in: By obtaining the annual continuous net load curve for the i-th forecast year, the electricity consumption W for the i-th forecast year can be calculated. i For a typical scenario C of the i-th prediction year (i) j Its duration in hours t (i) j for: The electricity consumption for the i-th predicted year is: 。 2. The prediction method as described in claim 1, characterized in that, The process of processing the effective historical data using the k-means clustering algorithm to obtain the typical scenario includes the following steps: The valid historical data is normalized to obtain multiple samples; K samples are randomly selected from the plurality of samples as initial cluster centers; Clustering is performed based on the Euclidean distance between each sample and the initial cluster center to obtain K clusters; For each of the classes, the cluster centers are recalculated, and the clusters are re-clustered based on the cluster centers. If the clustering result is different from the K classes, the process returns to the step of clustering based on the Euclidean distance between each sample and the initial cluster centers. If the clustering result is the same as the K classes, then calculate the average distance of all the samples to the nearest cluster center, and increase the number of classes in the clustering result. If the number of classes in the clustering result is less than or equal to a preset threshold, then return to the step of randomly selecting K samples from the multiple samples as the initial cluster centers. If the number of clusters in the clustering result is greater than the preset threshold, then the appropriate number of clusters is determined according to the method, and the corresponding clustering result is output. The clustering results are normalized and restored according to the historical power grid data of the most recent year to obtain the multiple typical scenarios.

3. A device for predicting net load on a power grid, applied to electronic equipment, characterized in that, The prediction device includes: The data acquisition module is configured to acquire historical power grid data; The data filtering module is configured to use the Laida criterion algorithm to filter out abnormal data in the power grid historical data to obtain valid historical data; The scene generation module is configured to process the effective historical data based on the k-means clustering algorithm to obtain multiple typical scenes; The prediction execution module is configured to perform grayscale prediction based on the effective historical data to obtain the annual net load growth rate for future time and the maximum net load for each predicted year within the future time, and to process the maximum net load according to the multiple typical scenarios to obtain the electricity consumption for each predicted year. The prediction model used by the prediction execution module for grayscale prediction is as follows: ； a is the development coefficient, and u is called the gray action quantity; Then the cumulative sequence of predicted values ​​L (1) (t) is: The predicted value sequence L is obtained by cumulative subtraction and restoration. (0) (t) is: Let L(i) be the average net load for the i-th forecast year, then the annual growth rate α of the net load for the i-th forecast year is... i It can be represented as: The annual maximum net load L in the i-th forecast year m (i) is: In the formula, L m (i-1) represents the maximum annual net load for the (i-1)th forecast year; Based on typical scenarios The typical scenario corresponding to the i-th prediction year is obtained as follows: ,in: By obtaining the annual continuous net load curve for the i-th forecast year, the electricity consumption W for the i-th forecast year can be calculated. i For a typical scenario C of the i-th prediction year (i) j Its duration in hours t (i) j for: The electricity consumption for the i-th predicted year is: 。 4. The prediction device as described in claim 3, characterized in that, The scene generation module includes: The normalization processing unit is configured to normalize the valid historical data to obtain multiple samples; The center selection unit is configured to randomly select K samples from the plurality of samples as initial cluster centers; The first clustering unit is configured to cluster according to the Euclidean distance between each sample and the initial cluster center to obtain K classes; The second clustering unit is configured to recalculate the cluster center for each of the classes and re-cluster based on the cluster center. If the clustering result is different from the K classes, the first clustering unit is controlled to re-cluster based on the Euclidean distance between each sample and the initial cluster center. The mean calculation unit is configured to calculate the mean distance of all samples to the nearest cluster center if the clustering result is the same as the K classes, and increase the number of classes in the clustering result; if the number of classes in the clustering result is less than or equal to a preset threshold, control the center selection unit to randomly select K samples from the plurality of samples again as initial cluster centers. The clustering determination unit is configured to determine an appropriate number of clusters and output the corresponding clustering result if the number of clusters in the clustering result is greater than the preset threshold. The generation execution unit is configured to normalize and restore the clustering results according to the historical power grid data of the most recent year to obtain the multiple typical scenarios.

5. An electronic device, characterized in that, It includes at least one processor and a memory connected to the processor, wherein: The memory is used to store computer programs or instructions; The processor is used to execute the computer program or instructions to enable the electronic device to implement the power grid net load prediction method as described in any one of claims 1 to 2.

6. A storage medium used in electronic devices, characterized in that, The storage medium carries one or more computer programs that can be executed by the electronic device. When the electronic device executes the one or more computer programs, it can implement the power grid net load prediction method as described in any one of claims 1 to 2.

Images