Power grid load forecasting method, device and related products

Abstract

The present disclosure relates to the technical field of power systems, and discloses a method, device and related products for power grid load forecasting; the method includes obtaining the first single-region load forecast results reported by multiple sub-regions in the area to be predicted; obtaining the index value sets corresponding to each sub-region; calculating the weight data of each sub-region based on each index value set; and calculating the overall load forecast result of the area to be predicted based on each first single-region load forecast result and each weight data. The index value of the present invention can reflect the characteristics of the load forecast quality, impact degree and difficulty of each sub-region. Accordingly, the weight data of each sub-region is calculated based on the first single-region load forecast result and the index value set, and its proportion in the overall load forecast can be reasonably allocated. Finally, combining the first single-region load forecast result and the weight, the overall load forecast result of the area to be predicted can be calculated more accurately and reliably.

Description

Technical Field

[0001] The present disclosure relates to the technical field of power systems, and in particular to a power grid load forecasting method, device, and related products. Background Art

[0002] In the power system sector, power load forecasting is a crucial basis for scientific decision-making by power dispatch centers. Whether it's the overall load forecasting of a grid-level dispatch center for its provincial and municipal regions, or the overall load forecasting of a provincial dispatch center for its municipal regions, the results directly impact key aspects such as energy allocation strategies and supply planning. Therefore, developing an efficient power grid load forecasting method is of great significance.

[0003] Currently, the cumulative method is often used to predict the overall load of multiple subordinate regions. This method simply adds up the load forecast results for each subordinate region. However, this method does not consider the unique characteristics of each subordinate region and ignores the importance of the load forecast results for each individual region. This makes the final forecast difficult to accurately reflect the actual load situation. Therefore, there is an urgent need to solve this technical problem. Summary of the Invention

[0004] In view of the above situation, the embodiments of the present disclosure provide a power grid load forecasting method, device and related products, which aim to solve the above problems or at least partially solve the above problems.

[0005] In a first aspect, an embodiment of the present disclosure provides a method for predicting power grid load, the method comprising:

[0006] Obtaining a first single-region load forecast result reported by multiple sub-regions within the region to be forecasted; the first single-region load forecast result is output by the sub-region using its trained first load forecast model;

[0007] Obtaining a set of index values ​​corresponding to each of the sub-regions; the set of index values ​​includes an index value of at least one of the following indicators: a prediction result credibility index, a sub-region load impact index, and a sub-region load prediction difficulty index;

[0008] Calculate the weight data of each sub-region according to the indicator value sets;

[0009] An overall load forecast result of the area to be forecasted is calculated based on the first single-area load forecast results and the weight data.

[0010] In a second aspect, an embodiment of the present disclosure further provides a power grid load forecasting device, the device comprising:

[0011] A first acquisition module is configured to acquire a first single-region load forecast result reported by multiple sub-regions within the area to be predicted; the first single-region load forecast result is output by the sub-region using its trained first load forecast model;

[0012] The second acquisition module is configured to acquire a set of index values ​​corresponding to each of the sub-regions; the set of index values ​​includes an index value of at least one of the following indicators: a prediction result credibility index, a sub-region load impact index, and a sub-region load prediction difficulty index;

[0013] A calculation module, configured to calculate weight data of each sub-region based on each of the indicator value sets;

[0014] The prediction model is used to calculate the overall load prediction result of the area to be predicted based on the load prediction results of each of the first single areas and the weight data.

[0015] In a third aspect, an embodiment of the present disclosure further provides an electronic device, comprising: a processor; and a memory arranged to store computer-executable instructions, wherein the executable instructions, when executed, cause the processor to execute the steps of the above-mentioned power grid load forecasting method.

[0016] By means of the above technical solution, the power grid load forecasting method, device and related products provided by the embodiments of the present disclosure first obtain the first single-region load forecasting results reported by multiple sub-regions in the area to be predicted, as well as at least one of the prediction result credibility index value, sub-region load influence index value and sub-region load forecasting difficulty index value corresponding to each sub-region. These index values ​​can reflect various characteristics such as the load forecasting quality, influence degree and load forecasting difficulty of each sub-region. Then, the weight data of each sub-region is calculated based on the first single-region load forecasting result and the index value set, so that the proportion of the first single-region load forecasting result of each sub-region in the overall load forecast can be reasonably distributed, reflecting the difference in importance of different sub-regions. Finally, combined with the first single-region load forecasting results and weight data, a more accurate and reliable overall load forecasting result can be calculated for the area to be predicted.

[0017] The above description is only an overview of the technical solution of the present disclosure. In order to more clearly understand the technical means of the present disclosure, it can be implemented in accordance with the contents of the specification. In order to make the above and other purposes, features and advantages of the present disclosure more obvious and easy to understand, the specific implementation methods of the present disclosure are listed below. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] The drawings described herein are used to provide a further understanding of the present disclosure and constitute a part of the present disclosure. The exemplary embodiments of the present disclosure and their descriptions are used to explain the present disclosure and do not constitute an improper limitation of the present disclosure. In the drawings:

[0019] Figure 1 A schematic diagram of a flow chart of a power grid load forecasting method provided by an embodiment of the present disclosure is shown;

[0020] Figure 2 A flowchart of constructing an evaluation index system provided by an embodiment of the present disclosure is shown;

[0021] Figure 3 The figure shows the model architecture of i-Transformer provided by the embodiment of the present disclosure;

[0022] Figure 4 A schematic diagram of the structure of a power grid load prediction device provided by an embodiment of the present disclosure is shown;

[0023] Figure 5 A schematic structural diagram of an electronic device provided by an embodiment of the present disclosure is shown. DETAILED DESCRIPTION

[0024] To make the objectives, technical solutions, and advantages of the present disclosure more clear, the technical solutions of the present disclosure will be clearly and completely described below in conjunction with the specific embodiments of the present disclosure and the corresponding drawings. Obviously, the described embodiments are only part of the embodiments of the present disclosure, not all of the embodiments. Based on the embodiments of the present disclosure, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present disclosure.

[0025] It should be noted that similar reference numerals and letters denote similar items in the following drawings, and therefore, once an item is defined in one drawing, it does not need to be further defined or explained in subsequent drawings.

[0026] It should be noted that the terms "first," "second," and the like in the specification and claims of the present disclosure and the accompanying drawings are used to distinguish similar objects and are not necessarily used to describe a specific order or precedence. It should be understood that such usage is interchangeable where appropriate, so that the embodiments of the present disclosure described herein can be implemented in sequences other than those illustrated or described herein. In addition, the term "including" and its variations are to be interpreted as open-ended terms meaning "including but not limited to."

[0027] As previously mentioned, to predict the overall load of multiple subordinate regions, the existing technology often uses an accumulation method, that is, directly adding the load prediction results of each subordinate region. However, this method does not take into account the characteristics of each subordinate region and ignores the importance of the load prediction results of a single region, making it difficult for the final prediction results to accurately reflect the actual load situation. Based on this, the present invention proposes a power grid load prediction method, device, and related products. The following is a detailed description of the present disclosure through specific embodiments.

[0028] To facilitate understanding of this embodiment, a power grid load forecasting method disclosed in an embodiment of the present disclosure is first introduced in detail. The execution subject of the power grid load forecasting method provided in the embodiment of the present disclosure is generally a computer device with certain computing capabilities. The computer device includes, for example: a terminal device or a server or other processing device. The terminal device can be a user equipment (UE), a mobile device, a user terminal, a terminal, etc. In some possible implementations, the power grid load forecasting method can be implemented by a processor calling computer-readable instructions stored in a memory.

[0029] Figure 1 The flow chart of the power grid load forecasting method provided by the embodiment of the present disclosure is shown. Figure 1 It can be seen that the embodiment of the present disclosure includes at least steps S101-S104:

[0030] S101: Obtain first single-region load forecast results reported by multiple sub-regions in the area to be predicted; the first single-region load forecast results are output by the sub-regions using their trained first load forecast models.

[0031] This embodiment has no strict restrictions on the attributes such as the level of the sub-region. A sub-region can be, for example, a district in a city, a city, or a provincial-level administrative unit.

[0032] The first single-region load forecast result refers to the result obtained by using the first load forecast model to predict the power load conditions of each sub-region within the forecast area within a specific time range. The first single-region load forecast result includes the power load forecast values ​​of the corresponding sub-region at one or more future time points or time periods. For example, the first single-region load forecast result includes 24 load forecast values ​​for the next day (forecasted every hour).

[0033] This embodiment does not limit the means for obtaining the first single-region load forecast result. During implementation, for example, the execution subject of this embodiment can remotely obtain the first single-region load forecast result for each sub-region from a central database or distributed data node via a network interface.

[0034] S102: Obtaining an index value set corresponding to each of the sub-regions; the index value set includes an index value of at least one of the following indicators: a prediction result credibility index, a sub-region load influence index, and a sub-region load prediction difficulty index.

[0035] Here, the prediction result credibility index is used to measure the reliability of the first single-region load forecast result of the sub-region. The sub-region load impact index is used to measure the degree of influence of the sub-region load on the prediction area. The sub-region load prediction difficulty index is used to measure the difficulty of predicting the load of the sub-region.

[0036] During implementation, for example, power experts can be organized to score each indicator for each sub-region based on their understanding of the prediction model, data quality, and prediction process, combined with historical experience. This generates a corresponding set of indicator values ​​and sends them to the execution entity of this embodiment. For example, the set of indicator values ​​corresponding to a sub-region includes: 0.7, 0.5, and 0.2. Of these, 0.7 represents the prediction result credibility indicator value, 0.5 represents the sub-region load impact indicator value, and 0.2 represents the sub-region load prediction difficulty indicator value.

[0037] S103: Calculate weight data of each sub-region according to each indicator value set.

[0038] During implementation, for example, if each indicator value set includes three indicator values, each indicator value can be standardized first; then the subjective weighting method, the objective weighting method or the combined weighting method can be used to determine the weights of each of the three indicators; each standardized indicator value is multiplied by the corresponding indicator weight and then added to obtain the comprehensive score of each sub-region; the comprehensive score of each sub-region is normalized to obtain the weight data of each sub-region.

[0039] S104: Calculate the overall load forecast result of the area to be forecasted based on the first single-area load forecast results and the weight data.

[0040] During specific implementation, the weighted data are used to perform weighted average calculation on the load forecast results of each first single area to obtain the overall load forecast result.

[0041] It can be seen that the embodiment of the present disclosure first obtains the first single-region load forecast results reported by multiple sub-regions in the area to be predicted, as well as at least one of the prediction result credibility index value, sub-region load influence index value and sub-region load forecast difficulty index value corresponding to each sub-region. These index values ​​can reflect various characteristics such as the load forecast quality, influence degree and load forecast difficulty of each sub-region. Then, the weight data of each sub-region is calculated based on the first single-region load forecast result and the index value set, so that the proportion of the first single-region load forecast result of each sub-region in the overall load forecast can be reasonably distributed, reflecting the difference in importance of different sub-regions. Finally, combined with the first single-region load forecast results and weight data, a more accurate and reliable overall load forecast result can be calculated for the area to be predicted.

[0042] Furthermore, in order to better illustrate the process of the above-mentioned power grid load forecasting method, as a refinement and extension of the above-mentioned embodiment, the embodiment of the present invention provides several embodiments, but is not limited thereto, which are specifically as follows.

[0043] For the aforementioned set of indicator values, the embodiment of the present disclosure also provides an indicator value determination method, which constructs an indicator system to achieve quantification of the prediction result credibility index, sub-region load impact index and sub-region load prediction difficulty index. Figure 2 A flowchart of constructing an evaluation index system provided by an embodiment of the present disclosure is shown.

[0044] See also Figure 2 As shown, the evaluation system construction process consists of four main steps. The first step is to define the purpose and requirements. This involves clarifying the evaluation objectives, system usage, scope, and focus. This involves collecting demand information, including the expectations of the regional coordination center and background data on the power industry. The second step is to determine the indicator content. This includes first-level indicators, which are the highest-level indicators that must comprehensively cover the evaluation content; second-level indicators, which are decomposed into specific measurement dimensions; and third-level indicators, which are bottom-level, specific indicators that must be directly quantifiable or calculable. The third step is to establish an evaluation model. First, weights are assigned to each level of indicators based on their importance. Data collection and processing are then performed, including pre-processing based on information from various provinces and cities. The evaluation model is then constructed based on the indicator system, and scoring criteria and comparison benchmarks are established. The final step is pilot evaluation and ongoing monitoring. Pilot evaluations are conducted to test effectiveness, followed by revisions and improvements. During subsequent implementation and evaluation, the effectiveness of the forecast integration is regularly monitored, and the evaluation results are continuously monitored in the form of reports. Based on this, in some embodiments of the present disclosure, the index value of the prediction result credibility index is calculated according to the following method: using a preset index weight determination algorithm, at least one of the obtained first historical prediction accuracy, real-time prediction accuracy, and prediction accuracy under extreme weather conditions is processed to calculate the historical prediction performance index value; using the index weight determination algorithm, at least one of the obtained prediction model update frequency, prediction result continuity data, and abnormal prediction ratio is processed to calculate the prediction model stability index value; using the index weight determination algorithm, at least one of the obtained data completeness rate, data timeliness data, and data consistency data is processed to calculate the data quality level index value of the data required for prediction; using the index weight determination algorithm, at least one of the historical prediction performance index value, the prediction model stability index value, and the data quality level index value is processed to calculate the index value of the prediction result credibility index.

[0045] In this embodiment, the preset indicator weight determination algorithm can be, for example, the entropy weight method, the standard deviation coefficient method, the CRITIC method, etc., which is not limited in the embodiment of the present disclosure. During implementation, the prediction result credibility index value is obtained by using the preset indicator weight determination algorithm to calculate the weight of at least one of the historical prediction performance index value, the prediction model stability index value, and the data quality level index value, and performing weighted averaging on each index value. Here, the historical prediction performance index is used to measure the accuracy of the load forecast of the prediction model in the past. The prediction model stability index is used to measure the stability of the prediction model. The data quality level index is used to measure the quality of the data input into the prediction model.

[0046] For the historical forecast performance indicator, implementation involves using a preset indicator weighting algorithm to calculate the weight of at least one of the following: the first historical forecast accuracy, the real-time forecast accuracy, and the extreme weather forecast accuracy. The weighted average of these accuracy rates is then calculated. The first historical forecast accuracy measures the accuracy of the forecast model's past load forecasts. For example, implementation involves receiving multiple sets of data from the forecast system, each set including the load forecast and actual load values ​​at a certain point in time the day before. The accuracy corresponding to each set of data is then calculated using the formula (1): "Accuracy = 1 - |Forecast Value - Actual Value| / Actual Value × 100%). Finally, the average is taken to obtain the first historical forecast accuracy. The real-time forecast accuracy measures the accuracy of the forecast model's real-time load forecasts. For example, implementation involves receiving the load forecast and actual values ​​at the most recent point in time from the forecast system in real time and calculating the real-time forecast accuracy using the aforementioned formula (1). The extreme weather forecast accuracy measures the forecast model's ability to predict under extreme weather conditions (such as heavy rain, high temperatures, and heavy snow). During implementation, multiple sets of data sent by the prediction system can be received. Each set of data is the load forecast value and actual value of the prediction model when extreme weather occurs. The accuracy corresponding to each set of data is calculated using Formula 1 and the average is taken to obtain the prediction accuracy under extreme weather conditions.

[0047] For the stability index of the prediction model, when it is implemented, the preset indicator weight determination algorithm is used to calculate the weight of at least one of the obtained prediction model update frequency, prediction result continuity, and abnormal prediction ratio, and the weighted average of each indicator value is obtained. Among them, the prediction model update frequency refers to the number of times the prediction model is updated within a period of time (such as one day or one week). The prediction result continuity data is used to measure the degree of smoothness of the change between the load forecast results at adjacent time points, and can reflect the fluctuation of the prediction results of the prediction model. During implementation, for example, the continuity can be quantified by calculating statistics such as the standard deviation of the difference between adjacent prediction results within a period of time. The abnormal prediction ratio refers to the proportion of abnormal prediction results in all prediction results within a period of time. During implementation, for example, you can first customize the judgment criteria for abnormal predictions (such as the deviation from the actual value exceeds a certain threshold), and then count the number of abnormal prediction results within a period of time, divide it by the total number of prediction results, and obtain the abnormal prediction ratio.

[0048] For the data quality level indicator, when it is implemented, the preset indicator weight determination algorithm is used to calculate the weight of at least one of the acquired data completeness rate, data timeliness data, and data consistency data, and the weighted average of each indicator value is obtained. Among them, for the data completeness rate (Data Completeness Rate, DCR), when it is implemented, it can be expressed as the proportion of the total number of valid data points acquired to the total number of data points that should be collected. Here, the evaluation criteria for valid data points can be set according to actual needs and will not be repeated here. For the data timeliness data (Data Timeliness, DT), when it is implemented, it can be calculated, for example, according to the following formula:

[0049]

[0050] Where Δt is the actual data delay time, and τ is the maximum delay threshold allowed by the system. The DT value range is [0, 1], with larger values ​​indicating better data timeliness.

[0051] For data consistency (DC), during implementation, for example, it can be calculated according to the following formula:

[0052] DC=DC format ×DC logic

[0053]

[0054] Where n is the number of samples, N gormat is the number of data points with inconsistent formats, N structure is the number of data points with inconsistent structures, N sampling is the number of data points with inconsistent sampling.i is the current data point, μ i is the mean of historical data, σ i is the standard deviation of historical data.

[0055] In some embodiments, the index value of the sub-area load impact index is calculated according to the following method: using a preset index weight determination algorithm, at least one of the obtained maximum load proportion, average load proportion, and peak load contribution rate is processed to calculate the load scale proportion index value; using the index weight determination algorithm, at least one of the obtained peak-shaving capacity proportion, new energy installed capacity proportion, and demand response capability data is processed to calculate the regulation capability index value; using the index weight determination algorithm, at least one of the load scale proportion index value and the regulation capability index value is processed to calculate the index value of the sub-area load impact index.

[0056] During implementation, the sub-region load impact index is calculated using a preset indicator weighting algorithm to calculate the weight of at least one of the load scale ratio index and the regulation capacity index. Using these weights, the weighted average of these index values ​​is then calculated. The load scale ratio index refers to the proportion of a sub-region's load scale to the total load scale of the region being predicted. The regulation capacity index is a quantitative indicator used to measure a sub-region's ability to regulate and balance load changes, power output fluctuations, and other factors within the power system.

[0057] For the load scale ratio index value, when it is implemented, the preset index weight determination algorithm is used to calculate the weight of at least one of the maximum load ratio, average load ratio, and peak load contribution rate, and the weighted average of each index value is obtained using each weight. Among them, the maximum load ratio refers to the ratio of the maximum load of a sub-region to the total maximum load of the area to be predicted within a specific time period. The average load ratio refers to the ratio of the average load of the sub-region within a specific time period to the total average load of the area to be predicted. The peak load contribution rate refers to the ratio of the peak load of the sub-region to the total peak load of the area to be predicted. During implementation, the peak load judgment standard can be determined first (such as the load exceeds a certain threshold and the duration is short), and the peak load of each sub-region is found based on this standard. Then, the peak load contribution rate of each sub-region is calculated using the formula "peak load contribution rate = sub-region peak load / total peak load × 100%".

[0058] For the regulation capacity index value, during implementation, the index weight determination algorithm is used to calculate the weight of at least one of the peak-shaving capacity ratio, new energy installed capacity ratio, and demand response capability data, and use the weight to perform weighted averaging on the index values.

[0059] The peak-shaving capacity ratio refers to the ratio of the available peak-shaving capacity in the subregion's power grid to the available peak-shaving capacity in the forecasted region. The renewable energy installed capacity ratio refers to the ratio of the subregion's installed renewable energy capacity to the forecasted region's installed renewable energy capacity. The demand response capability data, which measures the ratio of the subregion's user-adjustable load to the total adjustable load in the forecasted region, reflects the subregion's potential to assist in grid regulation through demand-side management.

[0060] In some embodiments, the index value of the sub-region load forecasting difficulty index is calculated according to the following method:

[0061] Using a preset indicator weight determination algorithm, at least one of the obtained daily load fluctuation rate, weekly load fluctuation rate, and seasonal fluctuation intensity is processed to calculate a load fluctuation characteristic indicator value;

[0062] Using the index weight determination algorithm, at least one of the acquired temperature sensitivity, humidity sensitivity, and holiday sensitivity is processed to calculate an external factor sensitivity index value;

[0063] Using the indicator weight determination algorithm, at least one of the obtained industrial electricity consumption proportion, the number of key users, and the number of user types is processed to calculate the power consumption structure complexity index value;

[0064] The index weight determination algorithm is used to process at least one of the load fluctuation characteristic index value, the external factor sensitivity index value, and the power consumption structure complexity index value to obtain the index value of the sub-region load forecasting difficulty index.

[0065] During implementation, the sub-region load forecasting difficulty index value is calculated using a preset index weight determination algorithm to calculate the weight of at least one of the load fluctuation characteristic index value, external factor sensitivity index value, and power consumption structure complexity index value, and then using each weight to perform a weighted average of each index value. Here, the load fluctuation characteristic index is used to measure the severity of the sub-region load changes over time. The external factor sensitivity index is used to measure the extent to which the sub-region load is affected by various external factors. The power consumption structure complexity index is used to measure the composition of different types of electricity users in the sub-region and the complexity of their electricity consumption characteristics.

[0066] For the load fluctuation characteristic index value, when it is implemented, a preset index weight determination algorithm is used to calculate the weight of at least one of the daily load fluctuation rate, weekly load fluctuation rate, and seasonal fluctuation intensity, and each weight is used to perform a weighted average on each index value to obtain the result.

[0067] The daily load fluctuation rate measures the fluctuations in a sub-region's power load over a day. For example, it can be calculated by taking the difference between the sub-region's daily maximum and minimum loads and the ratio of that to the daily average load. The weekly load fluctuation rate measures the fluctuations in a sub-region's power load over a week. For example, the following formula can be used to calculate a sub-region's weekly load fluctuation rate:

[0068]

[0069] Among them, L i represents the average load of the sub-region on day i, Represents the average daily load of the sub-area during the week.

[0070] Seasonal fluctuation intensity is used to measure the degree of change in sub-region load between different seasons, which can reflect the fluctuation characteristics of power load caused by seasonal factors. In implementation, for example, the seasonal fluctuation intensity of a sub-region can be calculated using the following formula:

[0071]

[0072] Among them, SI represents the seasonal fluctuation intensity, L i represents the average load of the sub-region in the i-th season, n represents the number of seasons, Indicates the average annual load of the sub-area.

[0073] For the external factor sensitivity index value, when it is implemented, the index weight determination algorithm is used to calculate the weight of at least one of temperature sensitivity, humidity sensitivity, and holiday sensitivity, and the weighted average of each index value is obtained using each weight.

[0074] Among them, temperature sensitivity is used to measure the impact of temperature factors on the power load of the sub-region. During implementation, for example, it can be expressed as the ratio of the power load change caused by a unit temperature change to the average load. Humidity sensitivity is used to measure the impact of humidity factors on the power load of the sub-region. During implementation, for example, it can be expressed as the ratio of the power load change caused by a unit humidity change to the average load. Holiday sensitivity is used to measure the impact of holidays on the power load of the sub-region. During implementation, for example, the following formula can be used to calculate the holiday sensitivity of the sub-region:

[0075]

[0076] Among them, L 节假日 represents the average load of the sub-area during holidays, L 非节假日 Indicates the average load of the sub-area during the non-holiday period, which is similar in length to the time before and after the holiday.

[0077] When implementing the power consumption structure complexity index value, the index weight determination algorithm is used to calculate the weight of at least one of the following: the proportion of industrial electricity consumption, the number of key users, and the number of user types, and then use each weight to perform a weighted average of each index value.

[0078] Among them, the proportion of industrial electricity consumption is used to measure the important position and scale of industrial electricity consumption in the entire electricity consumption structure in the sub-region, and can be expressed as the proportion of industrial electricity consumption in the total electricity consumption. The number of key users refers to the number of key users in the power system who have an important impact on the stability and reliability of power supply, have a large electricity consumption scale or special electricity consumption characteristics. Key users include but are not limited to large industrial enterprises, important public service institutions, etc. The number of user types is used to measure the richness and diversity of user types in the power system, taking into account the composition of various types of users with different industries, different natures, and different electricity demands in electricity consumption. The more user types and the greater the differences, the more complex the electricity consumption structure, and the difficulty of system operation and management may also increase accordingly.

[0079] In some embodiments, the method further comprises:

[0080] Obtaining load forecast influencing factor data of each sub-region, and preprocessing each load forecast influencing factor data to obtain corresponding processed influencing factor data;

[0081] For any target sub-region, using the corresponding second load forecasting model, the corresponding target processed influencing factor data is predicted to obtain the corresponding second single-region load forecasting result;

[0082] The step of calculating the overall load forecast result of the area to be forecasted based on the first single-area load forecast results and the weight data includes:

[0083] For any sub-region, using the corresponding weight data, the corresponding first single-region load forecast result and the second single-region load forecast result are weighted averaged to obtain the corresponding fused load forecast result;

[0084] The fused load forecast results are added together to obtain the overall load forecast result.

[0085] In this embodiment, the load forecast influencing factor data for each sub-region includes, but is not limited to, at least one of the following: grid operation data (such as load data, industry electricity consumption data, distributed energy generation data, peak-to-valley difference, average load, maximum load, minimum load, load curve trend data, etc.), climate data (such as temperature, humidity, precipitation, wind speed, etc.), social factor data (such as holidays, population, real-time electricity prices, electricity consumption behavior preferences, industrial structure, etc.), etc. The means for obtaining the load forecast influencing factor data can be, for example, obtained through a Supervisory Control and Data Acquisition (SCADA) system, an open source data platform, etc., which is not limited in this embodiment. After obtaining the load forecast influencing factor data for each sub-region, in order to facilitate subsequent efficient data processing, the load forecast influencing factor data can be preprocessed during implementation. The preprocessing means include, but are not limited to, at least one of the following: data cleaning (such as outlier processing, missing value interpolation), feature engineering (such as feature selection, feature encoding), etc.

[0086] Before predicting the processed influencing factor data, a corresponding second load forecasting model is first trained for each sub-region.

[0087] During implementation, an initial model can be selected first. The initial model can be, for example, ARIMA (Autoregressive Integrated Moving Average Model), Holt-Winters, LSTM (Long Short-Term Memory), GRU (Gated Recurrent Unit), Transformer and other models; then, the initial model is trained using a pre-constructed data set to obtain each second load forecasting model.

[0088] In some embodiments, the second load forecasting model corresponding to the target sub-area is generated based on i-transformer model training.

[0089] The i-Transformer model is a generative AI model that reverses the order of modules in the Transformer architecture. After inputting data, it can map the entire time series of the same feature variable into a high-dimensional vector. The resulting high-dimensional feature vector describes the main time series, which can independently reflect the changes in the historical series.

[0090] Figure 3 The model architecture of i-Transformer provided by the embodiment of the present disclosure is shown. Figure 3As can be seen, the input data is first converted into a vector by the embedding layer, then enters the transformation module consisting of a multi-attention layer, an addition and normalization layer, and a feedforward layer. This module can be repeated L times to gradually extract complex features. The multi-attention layer calculates the Q, K, and V vectors to capture element-by-element dependencies. The addition and normalization layer includes residual connections and layer normalization to assist in training. The feedforward layer performs nonlinear transformations to enhance expressiveness. After that, it is transformed by the projection layer and the final output is the result. Table 1 below shows the main parameters of the i-Transformer model.

[0091] Table 1 Main parameters of the i-Transformer model

[0092] Code Name Chinese name Parameter meaning Num of attention heads Number of attention heads Multiple attention heads Model dimension Model Dimensions Attention head dimension Hidden Layer Dimension Hidden layer dimensions Hidden layer dimension size Feed-Forward Network Layers Feedforward Neural Network Number of feedforward neural network layers Num of Encoder Layers Number of encoder layers The number of encoder stacking layers in the model Num of Decoder Layers Number of decoder layers Number of decoder stacking layers in the model Loss Function Loss Function Optimization goals during training Optimizer Optimizer Training model optimization algorithm Learning Rate Learning rate Compensation for weight updates Batch Size Batch size The number of samples used for training Activation Function Activation Function Output mapping inside a neuron

[0093] This embodiment uses the i-Transformer model as the underlying architecture for the second load forecasting model. Through its unique sequence perspective and inverted structure design, the i-Transformer model effectively addresses issues such as noise interference in multivariate time series and difficulty modeling lagged relationships between variables in traditional models, significantly improving load forecasting accuracy. Its attention mechanism accurately captures the dynamic correlation between load and external factors in multi-source heterogeneous data, while the feedforward network enhances time-dimensional feature extraction. While reducing reliance on long historical data, it achieves rapid forecasting through parallelization. This makes it particularly suitable for multi-regional, integrated load forecasting scenarios with transmission delays and high inherent complexity of variables.

[0094] In some embodiments, the second load forecasting model corresponding to the target sub-area is generated according to the following method: obtaining data sets under multiple power grid operation scenarios respectively; the power grid operation scenarios are divided according to at least one dimension of whether it is a holiday or season type; for any target power grid operation scenario, using the corresponding data set, at least two models of ARIMA, Holt-Winters, LSTM, GRU, Transformer and i-Transformer are trained respectively to obtain multiple candidate load forecasting models; the prediction model with the best performance among the multiple candidate load forecasting models is determined as the load forecasting model under the target power grid operation scenario; according to the power grid operation scenario information of the target load prediction influencing factor data, the power grid operation scenario is matched, and the load forecasting model corresponding to the matched power grid operation scenario is used as the second load forecasting model corresponding to the target sub-area.

[0095] In this embodiment, the power grid operation scenarios are divided based on at least one of the following dimensions: whether it is a holiday or a weekday, and the season type. For example, based on whether it is a holiday (e.g., holiday or weekday) and the season type (e.g., spring, summer, autumn, winter), the following eight power grid operation scenarios can be obtained: spring holiday, spring weekday, summer holiday, summer weekday, autumn holiday, autumn weekday, winter holiday, and winter weekday.

[0096] During implementation, a data set is first obtained for each grid operation scenario. It is understandable that the type and acquisition method of the original data corresponding to the characteristic data of the samples in the data set are similar to the aforementioned load forecast influencing factor data, and will not be repeated here.

[0097] Then, for each grid operation scenario, we can use the corresponding dataset to train at least two of the following: ARIMA, Holt-Winters, LSTM, GRU, Transformer, and i-Transformer. For example, we can train ARIMA, Holt-Winters, LSTM, GRU, Transformer, and i-Transformer, respectively, to obtain eight candidate load forecasting models. The specific training methods are based on existing technologies and will not be detailed here.

[0098] Next, for any power grid operation scenario, the best-performing prediction model among the corresponding multiple candidate load prediction models is determined as the corresponding load prediction model. This embodiment does not limit the method for determining the best-performing model. For example, the mean squared error (MSE) of each candidate load prediction model on the corresponding validation set can be calculated, and the model with the smallest MSE is the best-performing prediction model.

[0099] Finally, based on the grid operation scenario information of the target load forecast influencing factor data, a grid operation scenario is matched, and the load forecast model corresponding to the matched grid operation scenario is used as the second load forecast model corresponding to the target sub-region. For example, assuming that the grid operation scenario information corresponding to the target load forecast influencing factor data is "Spring during the May Day holiday," the prediction model with the best performance pre-trained for the "Spring holiday" grid operation scenario can be used as the second load forecast model corresponding to the target sub-region.

[0100] This embodiment divides the power grid operation scenarios according to dimensions such as whether it is a holiday, season type, etc. and obtains corresponding data sets; then, for each target power grid operation scenario, multiple models (such as ARIMA, LSTM, etc.) are trained using the corresponding data sets to obtain multiple candidate load forecasting models. In this way, the advantages of different models are fully utilized and the potential of models suitable for each scenario is explored; then, the best-performing one is selected from multiple candidate load forecasting models as the load forecasting model under the target power grid operation scenario, ensuring that each scenario has the most suitable model and improving the prediction accuracy of the model in a specific scenario; finally, according to the power grid operation scenario information of the target load prediction influencing factor data, the scenario is matched and the second load forecasting model corresponding to the target sub-area is determined, so that the prediction model can accurately correspond to the actual scenario, thereby improving the pertinence and applicability of the model. In summary, this embodiment can achieve accurate matching of the optimal load forecasting model according to different power grid operation scenarios, significantly improving the accuracy and reliability of load forecasting.

[0101] After the second load forecasting model for each sub-region is trained, the processed influencing factor data can be input into the corresponding second load forecasting model through a sliding window, and the model outputs the second single-region load forecasting results.

[0102] Then, for any sub-region, the corresponding weight data is used to perform weighted averaging on the corresponding first single-region load forecast results and the second single-region load forecast results to obtain the corresponding fused load forecast result; finally, the fused load forecast results are added together to obtain the overall load forecast result.

[0103] For example, the overall load forecast result can be calculated according to the following formula:

[0104]

[0105] Among them, Load final,t represents the overall load forecast result of the area to be predicted at time t, n represents the total number of sub-areas, w i Indicates the weight data corresponding to the i-th province (sub-region), Load province,i,t represents the first single-region load forecast result of the i-th province (sub-region) at time t, Load main,i,t It represents the load forecast result of the second single region of the technical solution of this embodiment for the i-th province (sub-region) at time t. If the index value used to calculate the weight data of the sub-region changes with time, then w i,t Instead of w i , w i,t Represents the weight data corresponding to the i-th province (sub-region) at time t.

[0106] The execution entity of this embodiment uses the second load forecasting model deployed by itself to predict the processed data of each sub-region and obtain the second single-region load forecast result. Then, for each sub-region, the single-region load forecast results from the two sources are weighted averaged according to the weight data to obtain the fused load forecast result. Finally, the fused results are added together to obtain the overall load forecast result. It can be seen that this embodiment can comprehensively consider the advantages of the sub-region's own model and the model of this embodiment, reduce the prediction error of a single model, and improve the accuracy, reliability and scientificity of the overall load forecast of the area to be predicted. In addition, the first single-region load result of each sub-region contains more detailed regional characteristics, which can enhance the interpretability of the overall load forecast result to a certain extent.

[0107] In some embodiments, the method further includes: obtaining the second historical prediction accuracy of each first load forecasting model within a preset time period, and multiplying each second historical prediction accuracy by the corresponding weight data to obtain multiple products; calculating a dynamic adjustment coefficient based on the multiple products; for any sub-region, using the corresponding weight data, performing weighted averaging on the corresponding first single-region load forecast result and the second single-region load forecast result to obtain the corresponding fused load forecast result, including: for any sub-region, using the corresponding weight data and the dynamic adjustment coefficient, performing weighted averaging on the corresponding first single-region load forecast result and the second single-region load forecast result to obtain the fused load forecast result.

[0108] In this embodiment, the preset time period can be set according to actual needs and is not limited in this embodiment. For example, the preset time period is the most recent three months. The calculation method for the second historical prediction accuracy rate is similar to the calculation method for the first historical prediction accuracy rate described above and will not be repeated here.

[0109] During implementation, each second historical prediction accuracy rate can be multiplied by the corresponding weight data to obtain multiple products, and then the products are added together to obtain the dynamic adjustment coefficient.

[0110] During implementation, a basic adjustment coefficient (which can be used to adjust the influence of the two single-region load forecast results) can also be introduced. The dynamic adjustment coefficient is multiplied by the basic adjustment coefficient, and the product is used as the new dynamic adjustment coefficient. The basic adjustment coefficient can range from 0.4 to 0.6.

[0111] During implementation, if the time resolution of the index value and dynamic adjustment coefficient used to calculate the weight data of the sub-region is consistent with the time resolution of the load forecast result of a single region, w i,t Instead of w i, in this case, the second historical prediction accuracy can be the average of all prediction accuracy rates at the same time within the preset time period. Specifically, the dynamic adjustment coefficient can be calculated according to the following formula:

[0112]

[0113] Among them, β is the basic adjustment coefficient, accuracy i,t is the historical prediction accuracy of the i-th province (sub-region) at time t (such as the average accuracy at the same time in the past three months), w i,t is the weight data of the i-th province (sub-region) at time t.

[0114] After obtaining the dynamic adjustment coefficient, for any sub-region, the corresponding weight data and dynamic adjustment coefficient can be used to perform weighted averaging on the corresponding first single-region load forecast result and the second single-region load forecast result to obtain a fused load forecast result.

[0115] In specific implementation, the overall load forecast result of the area to be forecasted at time t can be calculated according to the following formula:

[0116]

[0117] The overall load forecast result of the area to be forecasted at time t can also be calculated according to the following formula:

[0118]

[0119] This embodiment evaluates the performance of the model in a specific period of time by obtaining the second historical prediction accuracy of each first load forecast model over a period of time in history. Then, the accuracy is multiplied by the weight of the first load forecast result of the corresponding sub-region, and the dynamic adjustment coefficient is calculated based on the product. For the sub-region, the weight and the dynamic adjustment coefficient that plays a secondary adjustment function are used to perform a weighted average of the first and second single-region load forecast results, integrating different forecast advantages and dynamically adjusting the proportion. Finally, the fusion results of each sub-region are added to obtain the overall load forecast result. This embodiment can effectively improve the accuracy of the overall load forecast and the flexibility to cope with load changes, model performance changes, and time changes.

[0120] The present disclosure also provides a method for predicting power grid load, comprising the following steps:

[0121] Step S1: Obtain the first single-region load forecast results reported by multiple sub-regions within the region to be forecasted. The first single-region load forecast results are output by the sub-regions using their trained first load forecasting models. Here, the region to be forecasted is the main network, and the sub-regions are provincial administrative units.

[0122] Step S2: Obtain the index value set corresponding to each sub-region; the index value set includes the index values ​​of the following indicators: prediction result credibility index, sub-region load impact index and sub-region load prediction difficulty index.

[0123] During implementation, the index value of the prediction result credibility index is calculated according to the following method: using the entropy weight method, the obtained first historical prediction accuracy, real-time prediction accuracy, and prediction accuracy under extreme weather conditions are processed to calculate the historical prediction performance index value; using the entropy weight method, the obtained prediction model update frequency, prediction result continuity data, and abnormal prediction ratio are processed to calculate the prediction model stability index value; using the entropy weight method, the obtained data completeness rate, data timeliness data, and data consistency data are processed to calculate the data quality level index value of the data required for prediction; using the entropy weight method, the historical prediction performance index value, the prediction model stability index value, and the data quality level index value are processed to calculate the index value of the prediction result credibility index.

[0124] The index value of the sub-region load impact index is calculated according to the following method: using the entropy weight method, the obtained maximum load proportion, average load proportion, and peak load contribution rate are processed to calculate the load scale proportion index value; using the entropy weight method, the obtained peak-shaving capacity proportion, new energy installed capacity proportion, and demand response capability data are processed to calculate the regulation capacity index value; using the entropy weight method, the load scale proportion index value and the regulation capacity index value are processed to calculate the index value of the sub-region load impact index.

[0125] The index value of the sub-region load forecasting difficulty index is calculated according to the following method: using the entropy weight method, the obtained daily load fluctuation rate, weekly load fluctuation rate, and seasonal fluctuation intensity are processed to calculate the load fluctuation characteristic index value; using the entropy weight method, the obtained temperature sensitivity, humidity sensitivity, and holiday sensitivity are processed to calculate the external factor sensitivity index value; using the entropy weight method, the obtained industrial electricity consumption proportion, the number of key users, and the number of user types are processed to calculate the power consumption structure complexity index value; using the entropy weight method, the load fluctuation characteristic index value, the external factor sensitivity index value, and the power consumption structure complexity index value are processed to obtain the index value of the sub-region load forecasting difficulty index.

[0126] Step S3: Use principal component analysis to process the set of indicator values ​​and calculate the weight data for each sub-region. During implementation, you can also use statistical methods based on historical prediction deviations to determine the confidence interval of the predicted value. The upper and lower bounds of the prediction interval are calculated as follows:

[0127]

[0128] Among them, X M-Average 、X Mean 、X L-Average They are the upper bound, lower bound and mean of the final overall load forecast result, x Mi 、x Li 、x mean-i They are the upper bound of the probability interval, the lower bound of the interval, and the predicted mean of the single-area load forecast results of each sub-area.

[0129] Step S4: Obtain the load forecast influencing factor data of each sub-region, and pre-process the load forecast influencing factor data to obtain the corresponding processed influencing factor data. In specific implementation, the preprocessing process includes three parts: data acquisition, data cleaning and feature engineering. The first is data acquisition, followed by data cleaning, which specifically includes outlier isolation forest algorithm detection, context-related analysis, missing value NaN detection and LS-SVR (least squares support vector regression) completion. Finally, feature engineering, including interval coding and label coding in feature coding, Pearson coefficient and random forest method in feature selection, and Z-score normalization processing.

[0130] Step S5: For any target sub-region, use the corresponding second load forecasting model to predict the corresponding target processed influencing factor data to obtain the corresponding second single-region load forecasting result.

[0131] During implementation, the second load forecasting model corresponding to the target sub-area is generated according to the following method: obtaining data sets under a variety of power grid operation scenarios respectively; dividing the power grid operation scenarios according to whether they are holidays and seasonal types; for any target power grid operation scenario, using the corresponding data sets, respectively training the ARIMA, Holt-Winters, LSTM, GRU, Transformer and i-Transformer models to obtain multiple candidate load forecasting models; determining the best-performing prediction model among the multiple candidate load forecasting models as the load forecasting model under the target power grid operation scenario; matching the power grid operation scenarios according to the power grid operation scenario information of the target load prediction influencing factor data, and using the load forecasting model corresponding to the matched power grid operation scenario as the second load forecasting model corresponding to the target sub-area.

[0132] Step S6: Obtain the second historical prediction accuracy of each first load forecasting model within 3 months, and multiply each second historical prediction accuracy by the corresponding weight data to obtain multiple products.

[0133] Step S7: Add up the multiple products and multiply them with the basic weight coefficient to obtain a dynamic adjustment coefficient.

[0134] Step S8: For any sub-region, using the corresponding weight data and dynamic adjustment coefficient, perform weighted averaging on the corresponding first single-region load forecast result and the second single-region load forecast result to obtain a fused load forecast result.

[0135] Step S9: Add up the fused load forecast results to obtain the overall load forecast result.

[0136] This embodiment proposes an innovative power load forecasting technology with strong feature extraction and data integration capabilities. By constructing a comprehensive evaluation index system, quantitatively evaluating the power load characteristics of each province and city, and using the principal component analysis method to determine the weights, combined with the multi-source data fusion method, the results of the first load forecasting model and the second load forecasting model customized based on the power grid operation scenario are effectively integrated. This method not only adapts to the diversity of the economy, climate, and industrial structure of each province and city, but also solves problems such as data update frequency, transmission delay, and time scale processing, significantly improving the accuracy and credibility of load forecasting. Through feature extraction and fusion, the accuracy and robustness of the model in the face of a constantly changing environment are enhanced, and it can solve the problem of multi-source data fusion and the superposition of forecast results from different provinces, achieving efficient and accurate main grid load forecasting, and providing strong support for the stable operation of the power grid.

[0137] Those skilled in the art will understand that in the above method of the specific embodiment, the writing order of each step does not mean a strict execution order, but does not constitute any limitation on the implementation process. The specific execution order of each step should be determined by its function and possible internal logic.

[0138] It should be noted that in actual applications, all of the above possible implementation methods can be combined in any way to form possible embodiments of the present disclosure, and they will not be described in detail here. The information (including but not limited to device information, user information, etc.) and data (including but not limited to data used for analysis, storage and display, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties.

[0139] Based on the same concept, an embodiment of the present disclosure further provides a power grid load prediction device, which corresponds one-to-one to the power grid load prediction method in the above embodiment. Figure 4 The schematic diagram of the structure of the power grid load prediction device provided by the embodiment of the present disclosure is shown in FIG. Figure 4 As shown, the power grid load prediction device 400 provided by the embodiment of the present disclosure includes:

[0140] The first acquisition module 401 is configured to acquire a first single-region load forecast result reported by multiple sub-regions within the region to be forecasted; the first single-region load forecast result is output by the sub-region using its trained first load forecast model;

[0141] The second acquisition module 402 is configured to acquire a set of index values ​​corresponding to each of the sub-regions; the set of index values ​​includes an index value of at least one of the following indicators: a prediction result credibility index, a sub-region load impact index, and a sub-region load prediction difficulty index;

[0142] A calculation module 403 is configured to calculate weight data of each sub-region based on each of the indicator value sets;

[0143] The prediction module 404 is configured to calculate the overall load prediction result of the area to be predicted based on the first single-area load prediction results and the weight data.

[0144] In some embodiments, the device also includes a first calculation module, which is used to: use a preset indicator weight determination algorithm to process at least one of the obtained first historical prediction accuracy, real-time prediction accuracy, and prediction accuracy under extreme weather conditions to calculate a historical prediction performance index value; use the indicator weight determination algorithm to process at least one of the obtained prediction model update frequency, prediction result continuity data, and abnormal prediction ratio to calculate a prediction model stability index value; use the indicator weight determination algorithm to process at least one of the obtained data completeness rate, data timeliness data, and data consistency data to calculate a data quality level index value of the data required for prediction; use the indicator weight determination algorithm to process at least one of the historical prediction performance index value, the prediction model stability index value, and the data quality level index value to calculate an index value of the prediction result credibility index.

[0145] In some embodiments, the device also includes a second calculation module, which is used to: use a preset indicator weight determination algorithm to process at least one of the obtained maximum load proportion, average load proportion, and peak load contribution rate to calculate the load scale proportion index value; use the indicator weight determination algorithm to process at least one of the obtained peak-shaving capacity proportion, new energy installed capacity proportion, and demand response capability data to calculate the regulation capability index value; use the indicator weight determination algorithm to process at least one of the load scale proportion index value and the regulation capability index value to calculate the index value of the sub-area load impact index.

[0146] In some embodiments, the device also includes a third calculation module, which is used to: use a preset indicator weight determination algorithm to process at least one of the obtained daily load fluctuation rate, weekly load fluctuation rate, and seasonal fluctuation intensity to calculate the load fluctuation characteristic index value; use the indicator weight determination algorithm to process at least one of the obtained temperature sensitivity, humidity sensitivity, and holiday sensitivity to calculate the external factor sensitivity index value; use the indicator weight determination algorithm to process at least one of the obtained industrial electricity consumption proportion, the number of key users, and the number of user types to calculate the power consumption structure complexity index value; use the indicator weight determination algorithm to process at least one of the load fluctuation characteristic index value, the external factor sensitivity index value, and the power consumption structure complexity index value to obtain the index value of the sub-area load forecasting difficulty index.

[0147] In some embodiments, the device also includes a load forecasting module, which is used to: obtain load forecast influencing factor data of each sub-area, and pre-process each load forecast influencing factor data to obtain corresponding processed influencing factor data; for any target sub-area, use the corresponding second load forecasting model to predict the corresponding target processed influencing factor data to obtain the corresponding second single-area load forecast result; the prediction module 404 is specifically used to: for any sub-area, use the corresponding weight data to perform weighted averaging on the corresponding first single-area load forecast result and the second single-area load forecast result to obtain the corresponding fused load forecast result; add each of the fused load forecast results to obtain the overall load forecast result.

[0148] In some embodiments, the device also includes a coefficient calculation module, which is used to: obtain the second historical prediction accuracy of each first load prediction model within a preset time period, and multiply each second historical prediction accuracy by the corresponding weight data to obtain multiple products; calculate the dynamic adjustment coefficient based on the multiple products; the prediction module 404 is specifically used to: for any sub-area, use the corresponding weight data and the dynamic adjustment coefficient to perform weighted averaging on the corresponding first single-area load prediction result and the second single-area load prediction result to obtain the fused load prediction result.

[0149] In some embodiments, in the above-mentioned device, the second load prediction model corresponding to the target sub-area is generated based on i-transformer model training.

[0150] In some embodiments, the device also includes a training module, which is used to: obtain data sets under multiple power grid operation scenarios respectively; the power grid operation scenarios are divided according to at least one dimension of whether it is a holiday or a season type; for any target power grid operation scenario, use the corresponding data set to train at least two models of ARIMA, Holt-Winters, LSTM, GRU, Transformer and i-Transformer respectively to obtain multiple candidate load forecasting models; determine the prediction model with the best performance among the multiple candidate load forecasting models as the load forecasting model under the target power grid operation scenario; match the power grid operation scenario according to the power grid operation scenario information of the target load prediction influencing factor data, and use the load forecasting model corresponding to the matched power grid operation scenario as the second load forecasting model corresponding to the target sub-area.

[0151] The present invention provides a power grid load forecasting device, which first obtains the first single-region load forecast results reported by multiple sub-regions in the area to be predicted, as well as at least one of the prediction result credibility index value, sub-region load influence index value and sub-region load forecast difficulty index value corresponding to each sub-region. These index values ​​can reflect various characteristics such as the load forecast quality, influence degree and load forecast difficulty of each sub-region. Then, based on the first single-region load forecast result and the index value set, the weight data of each sub-region is calculated, so that the proportion of the first single-region load forecast result of each sub-region in the overall load forecast can be reasonably distributed, reflecting the difference in importance of different sub-regions. Finally, by combining the first single-region load forecast results and weight data, a more accurate and reliable overall load forecast result can be calculated for the area to be predicted.

[0152] For the specific definition of the power grid load forecasting device, please refer to the definition of the power grid load forecasting method above, which will not be repeated here. The various modules in the above-mentioned power grid load forecasting device can be implemented in whole or in part by software, hardware, or a combination thereof. The above-mentioned modules can be embedded in or independent of the processor in the computer device in the form of hardware, or can be stored in the memory of the computer device in the form of software, so that the processor can call and execute the operations corresponding to the above modules.

[0153] Figure 5 FIG. 1 shows a schematic diagram of the structure of an electronic device provided by an embodiment of the present disclosure. Figure 5As shown, at the hardware level, the electronic device includes a processor and, optionally, an internal bus, a network interface, and memory. The memory may include internal memory, such as high-speed random-access memory (RAM), and may also include non-volatile memory, such as at least one disk storage device. Of course, the electronic device may also include other hardware required for its services.

[0154] The processor, network interface, and memory can be interconnected through an internal bus, which can be an ISA (Industry Standard Architecture) bus, a PCI (Peripheral Component Interconnect) bus, or an EISA (Extended Industry Standard Architecture) bus. Buses can be divided into address buses, data buses, control buses, etc. For ease of representation, Figure 5 Only one bidirectional arrow is used in the diagram, but this does not mean that there is only one bus or one type of bus.

[0155] The memory is used to store programs. Specifically, the program may include program code, which includes computer operating instructions. The memory may include internal memory and non-volatile memory, and provides instructions and data to the processor.

[0156] The processor reads the corresponding computer program from the non-volatile memory into the internal memory and then runs it, forming a power grid load forecasting device at the logical level. The processor executes the program stored in the memory and is specifically used to perform the aforementioned method.

[0157] The processor may be an integrated circuit chip with signal processing capabilities. During implementation, each step of the above method can be completed by hardware integrated logic circuits in the processor or by software instructions. The above processor may be a general-purpose processor, including a central processing unit (CPU), a network processor (NP), etc.; it may also be a digital signal processor (DSP), an application-specific integrated circuit (ASIC), a field-programmable gate array (FPGA), or other programmable logic devices, discrete gate or transistor logic devices, or discrete hardware components. The various methods, steps, and logic block diagrams disclosed in the embodiments of the present disclosure can be implemented or executed. The general-purpose processor may be a microprocessor or any conventional processor. The steps of the method disclosed in conjunction with the embodiments of the present disclosure can be directly implemented and executed by a hardware decoding processor, or by a combination of hardware and software modules in the decoding processor. The software module can be located in a storage medium mature in the art, such as random access memory, flash memory, read-only memory, programmable read-only memory, electrically erasable programmable memory, registers, etc. The storage medium is located in the memory, and the processor reads the information in the memory and completes the steps of the above method in combination with its hardware.

[0158] The electronic device can execute the power grid load prediction method provided by multiple embodiments of the present disclosure and realize a power grid load prediction device. Figure 4 The functions of the illustrated embodiment will not be described in detail in the embodiments of the present disclosure.

[0159] An embodiment of the present disclosure also proposes a computer-readable storage medium, which stores one or more programs, and the one or more programs include instructions. When the instructions are executed by an electronic device including multiple application programs, the electronic device can execute the power grid load forecasting method provided by multiple embodiments of the present disclosure.

[0160] Those skilled in the art will appreciate that the embodiments of the present disclosure may be provided as methods, systems, or computer program products. Therefore, the present disclosure may take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware. Furthermore, the present disclosure may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to magnetic disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0161] The present disclosure is described with reference to the flowcharts and / or block diagrams of the methods, devices (systems), and computer program products according to the embodiments of the present disclosure. It should be understood that each process and / or box in the flowchart and / or block diagram, as well as the combination of the processes and / or boxes in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 a process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0162] These computer program instructions may also be stored in a computer readable memory that can direct a computer or other programmable data processing device to work in a specific manner, so that the instructions stored in the computer readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 a process or multiple processes and / or boxes Figure 1 The function specified in one or more boxes.

[0163] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operational steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing the instructions executed on the computer or other programmable device for implementing the process. Figure 1 a process or multiple processes and / or boxes Figure 1 A step that specifies a function in one or more boxes.

[0164] In a typical configuration, a computing device includes one or more processors (CPUs), input / output interfaces, network interfaces, and memory.

[0165] Memory may include non-permanent storage in a computer-readable medium, random access memory (RAM) and / or non-volatile memory in the form of read-only memory (ROM) or flash RAM. Memory is an example of a computer-readable medium.

[0166] Computer-readable media includes permanent and non-permanent, removable and non-removable media that can be implemented by any method or technology to store information. The information can be computer-readable instructions, data structures, program modules or other data. Examples of computer storage media include, but are not limited to, phase change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technology, compact disc read-only memory (CD-ROM), digital versatile disc (DVD) or other optical storage, magnetic cassettes, magnetic tape, disk storage or other magnetic storage devices or any other non-transmission media that can be used to store information that can be accessed by a computing device. As defined herein, computer-readable media does not include transitory computer-readable media (transitory media), such as modulated data signals and carrier waves.

[0167] It should also be noted that the terms "comprises," "includes," or any other variations thereof are intended to encompass non-exclusive inclusion, such that a process, method, commodity, or apparatus that includes a series of elements includes not only those elements but also other elements not explicitly listed, or includes elements inherent to such process, method, commodity, or apparatus. In the absence of further limitations, an element defined by the phrase "comprises a ..." does not exclude the presence of other identical elements in the process, method, commodity, or apparatus that includes the element.

[0168] Those skilled in the art will appreciate that embodiments of the present disclosure may be provided as methods, systems, or computer program products. Thus, the present disclosure may take the form of a fully hardware embodiment, a fully software embodiment, or an embodiment combining software and hardware. Furthermore, the present disclosure may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to magnetic disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0169] The above are merely examples of the present disclosure and are not intended to limit the present disclosure. Various modifications and variations are possible for those skilled in the art. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present disclosure are intended to be included within the scope of the claims of the present disclosure.

Claims

1. A method for predicting power grid load, characterized in that: The method comprises: Obtaining a first single-region load forecast result reported by multiple sub-regions within the region to be forecasted; the first single-region load forecast result is output by the sub-region using its trained first load forecast model; Obtaining a set of index values ​​corresponding to each of the sub-regions; the set of index values ​​includes an index value of at least one of the following indicators: a prediction result credibility index, a sub-region load impact index, and a sub-region load prediction difficulty index; Calculate the weight data of each sub-region according to the indicator value sets; Calculating an overall load forecast result for the area to be forecasted based on the first single-area load forecast results and the weight data; The method further comprises: Obtaining load forecast influencing factor data of each sub-region, and preprocessing each load forecast influencing factor data to obtain corresponding processed influencing factor data; For any target sub-region, using the corresponding second load forecasting model, the corresponding target processed influencing factor data is predicted to obtain the corresponding second single-region load forecasting result; The step of calculating the overall load forecast result of the area to be forecasted based on the first single-area load forecast results and the weight data includes: For any sub-region, using the corresponding weight data, the corresponding first single-region load forecast result and the second single-region load forecast result are weighted averaged to obtain the corresponding fused load forecast result; Adding the fused load forecast results to obtain the overall load forecast result; The method further comprises: Obtaining a second historical prediction accuracy rate of each of the first load forecasting models within a preset time period, and multiplying each of the second historical prediction accuracy rates by corresponding weight data to obtain a plurality of products; Calculating a dynamic adjustment coefficient based on the multiple products; For any sub-region, using corresponding weight data, weighted averaging the corresponding first single-region load forecast result and the second single-region load forecast result is performed to obtain a corresponding fused load forecast result, including: For any sub-region, the corresponding weight data and the dynamic adjustment coefficient are used to perform weighted averaging on the corresponding first single-region load forecast result and the second single-region load forecast result to obtain the fused load forecast result.

2. The method according to claim 1, characterized in that The index value of the prediction result credibility index is calculated according to the following method: Using a preset indicator weight determination algorithm, at least one of the obtained first historical forecast accuracy, real-time forecast accuracy, and extreme weather forecast accuracy is processed to calculate a historical forecast performance indicator value; Using the indicator weight determination algorithm, at least one of the obtained prediction model update frequency, prediction result continuity data, and abnormal prediction ratio is processed to calculate a prediction model stability index value; Using the indicator weight determination algorithm, at least one of the acquired data completeness rate, data timeliness data, and data consistency data is processed to calculate a data quality level indicator value of the data required for prediction; The indicator weight determination algorithm is used to process at least one of the historical prediction performance indicator value, the prediction model stability indicator value, and the data quality level indicator value to calculate the indicator value of the prediction result credibility indicator.

3. The method according to claim 1, characterized in that The index value of the sub-region load impact index is calculated according to the following method: Using a preset indicator weight determination algorithm, at least one of the obtained maximum load proportion, average load proportion, and peak load contribution rate is processed to calculate a load scale proportion indicator value; Using the indicator weight determination algorithm, at least one of the obtained peak-shaving capacity ratio, new energy installed capacity ratio, and demand response capability data is processed to calculate a regulation capability indicator value; The indicator weight determination algorithm is used to process at least one of the load scale proportion indicator value and the regulation capacity indicator value to calculate the indicator value of the sub-region load impact indicator.

4. The method according to claim 1, wherein The index value of the sub-region load forecast difficulty index is calculated according to the following method: Using a preset indicator weight determination algorithm, at least one of the obtained daily load fluctuation rate, weekly load fluctuation rate, and seasonal fluctuation intensity is processed to calculate a load fluctuation characteristic indicator value; Using the index weight determination algorithm, at least one of the acquired temperature sensitivity, humidity sensitivity, and holiday sensitivity is processed to calculate an external factor sensitivity index value; Using the indicator weight determination algorithm, at least one of the obtained industrial electricity consumption proportion, the number of key users, and the number of user types is processed to calculate the power consumption structure complexity index value; The index weight determination algorithm is used to process at least one of the load fluctuation characteristic index value, the external factor sensitivity index value, and the power consumption structure complexity index value to obtain the index value of the sub-region load forecasting difficulty index.

5. The method according to claim 1, wherein The second load forecasting model corresponding to the target sub-area is generated based on i-transformer model training.

6. The method according to claim 1, characterized in that The second load forecasting model corresponding to the target sub-area is generated according to the following method: Acquire data sets under various power grid operation scenarios respectively; the power grid operation scenarios are divided according to at least one dimension of whether it is a holiday and a seasonal type; For any target grid operation scenario, using the corresponding data set, at least two models from ARIMA, Holt-Winters, LSTM, GRU, Transformer, and i-Transformer are trained to obtain multiple candidate load forecasting models. Determining the best-performing prediction model among the multiple candidate load prediction models as the load prediction model for the target power grid operation scenario; According to the grid operation scenario information of the load forecast influencing factor data of the target sub-area, the grid operation scenario is matched, and the load forecast model corresponding to the matched grid operation scenario is used as the second load forecast model corresponding to the target sub-area.

7. A power grid load forecasting device, characterized in that: The device comprises: A first acquisition module is configured to acquire a first single-region load forecast result reported by multiple sub-regions within the area to be predicted; the first single-region load forecast result is output by the sub-region using its trained first load forecast model; The second acquisition module is configured to acquire a set of index values ​​corresponding to each of the sub-regions; the set of index values ​​includes an index value of at least one of the following indicators: a prediction result credibility index, a sub-region load impact index, and a sub-region load prediction difficulty index; A calculation module, configured to calculate weight data of each sub-region based on each of the indicator value sets; A prediction module, configured to calculate an overall load prediction result for the area to be predicted based on the first single-area load prediction results and the weight data; The load forecasting module is configured to obtain load forecast influencing factor data for each sub-region and pre-process each load forecast influencing factor data to obtain corresponding processed influencing factor data; for any target sub-region, use the corresponding second load forecasting model to forecast the corresponding target processed influencing factor data to obtain a corresponding second single-region load forecast result; the forecasting module is specifically configured to perform a weighted average of the corresponding first single-region load forecast result and the second single-region load forecast result for any sub-region using the corresponding weight data to obtain a corresponding fused load forecast result; and add the fused load forecast results to obtain the overall load forecast result. The coefficient calculation module is used to: obtain the second historical prediction accuracy of each of the first load forecasting models within a preset time period, and multiply each of the second historical prediction accuracy rates by the corresponding weight data to obtain multiple products; calculate the dynamic adjustment coefficient based on the multiple products; the prediction module is specifically used to: for any sub-area, use the corresponding weight data and the dynamic adjustment coefficient to perform weighted averaging on the corresponding first single-area load forecast result and the second single-area load forecast result to obtain the fused load forecast result.

8. An electronic device comprising: processor; as well as A memory arranged to store computer-executable instructions, wherein when the executable instructions are executed, the processor is caused to perform the steps of the method according to any one of claims 1 to 6.

Images