Power distribution network gridding load prediction method and device, electronic equipment and storage medium

Abstract

The invention discloses a grid load prediction method and device for a power distribution network, electronic equipment and a storage medium, and the method comprises the steps: carrying out the grid division of a power distribution network area according to the regional parameters of a transformer of the power distribution network area, and dividing the power distribution network area into a plurality of levels of power utilization unit grids; acquiring load characteristic data in the lowest-level power utilization unit grid; analyzing to obtain power utilization characteristics and power utilization trends of various types of loads in the lowest-level power utilization unit grid; according to the power utilization characteristics and the power utilization trend of each type of load, load prediction is carried out on the under-construction area and the built area in the lowest-level power utilization unit grid, and a load prediction result corresponding to the under-construction area and a load prediction result corresponding to the built area are obtained; and a load prediction result in the lowest-level power utilization unit grid is obtained. According to the invention, the accuracy of grid load prediction of the power distribution network can be improved.

Description

Technical Field

[0001] The present invention relates to the technical field of load forecasting, and particularly to a method, device, electronic device and storage medium for grid-based load forecasting of a distribution network. Background Art

[0002] Currently, the grid-based method of distribution network is widely used in China. Among them, in the grid-based load forecasting of the distribution network, different load characteristics affect different load forecasting results, resulting in different loads in each horizontal year and target year. For the power supply area, the load in the saturation year determines the load density, that is, the level of grid framework planning. For the power supply grid, the load directly determines the power balance within the grid. For the power supply unit, the load directly determines the wiring mode and wiring scale within the power supply unit. Therefore, considering the load characteristics is of great significance for the grid-based planning results of the distribution network. Correctly analyzing the load characteristics of the planning area and reasonably applying the analysis results to the grid-based planning of the distribution network can improve the scientificity and rationality of the distribution network planning. However, when the existing technology conducts grid-based load forecasting of the distribution network, it fails to consider the load characteristics, resulting in a low accuracy of load forecasting. Summary of the Invention

[0003] The present invention provides a method, device, electronic device and storage medium for grid-based load forecasting of a distribution network to solve the technical problem that when the existing technology conducts grid-based load forecasting of the distribution network, it fails to consider the load characteristics, resulting in a low accuracy of load forecasting.

[0004] To solve the above technical problem, an embodiment of the present invention provides a method for grid-based load forecasting of a distribution network, including:

[0005] Obtaining regional parameters of a transformer in a distribution network substation area, and dividing the distribution network substation area into grid units of several levels of power consumption units according to the regional parameters; wherein, the regional parameters include: the regional function of the transformer in the substation area, administrative division, total regional load, and power supply area;

[0006] Obtaining load characteristic data within the grid unit of the lowest level of power consumption unit; wherein, the load characteristic data includes: annual maximum load utilization hours, annual load curve, and daily load characteristics;

[0007] Analyzing the power consumption characteristics and power consumption trends of various types of loads within the grid unit of the lowest level of power consumption unit according to the load characteristic data; wherein, the load types include: residential load, commercial load, industrial load, and administrative load;

[0008] According to the electricity consumption characteristics and trends of various types of loads, load forecasting is carried out separately for the under-construction areas and completed areas within the grid of the lowest-level electricity consumption units, and the load forecasting results corresponding to the under-construction areas and the load forecasting results corresponding to the completed areas are obtained. Then, comprehensive analysis is performed on the load forecasting results corresponding to the under-construction areas and the load forecasting results corresponding to the completed areas to obtain the load forecasting results within the grid of the lowest-level electricity consumption units.

[0009] As an optimal solution, the grid division of the distribution network substation area according to the regional parameters, and the distribution network substation area is divided into several levels of electricity consumption unit grids, including:

[0010] According to the regional functions of the substation transformers, administrative divisions, total regional load, and power supply area, the distribution network substation area is divided into several first-level electricity consumption unit grids and several second-level electricity consumption unit grids; among them, the second-level electricity consumption unit grids belong to the first-level electricity consumption unit grids;

[0011] Adjust the division results of the second-level electricity consumption unit grids so that the total preset daily average load forecast amounts of the second-level electricity consumption unit grids included in each first-level electricity consumption unit grid are equal;

[0012] For each second-level electricity consumption unit grid, according to the preset daily average load forecast amount and preset land use nature within the second-level electricity consumption unit grid, the second-level electricity consumption unit grid is divided into several third-level electricity consumption unit grids;

[0013] Adjust the division results of the third-level electricity consumption unit grids so that the total preset daily average load forecast amounts of the third-level electricity consumption unit grids included in each second-level electricity consumption unit grid are equal;

[0014] For each third-level electricity consumption unit grid, according to the preset land use nature and air humidity within the third-level electricity consumption unit grid, the third-level electricity consumption unit grid is divided into several fourth-level electricity consumption unit grids.

[0015] As an optimal solution, before obtaining the load characteristics within the grid of the lowest-level electricity consumption units, it further includes:

[0016] Adjust the number of grids of the fourth-level electricity consumption unit grids, and update the division results of the fourth-level electricity consumption unit grids according to the adjusted number of grids.

[0017] As an optimal solution, the adjustment of the number of grids of the fourth-level electricity consumption unit grids and the update of the division results of the fourth-level electricity consumption unit grids according to the adjusted number of grids include:

[0018] Decompose the third - level power - consuming unit grid into a first grid including geographical boundaries and a second grid without geographical boundaries;

[0019] Obtain the power - consuming unit data within the second grid; wherein, the power - consuming unit data includes: power - consuming unit longitude, power - consuming unit latitude, land - use nature weight, historical average air humidity, and historical monthly repair data;

[0020] Cluster the second grid according to the power - consuming unit data, and take the number of resulting clustering clusters as the number of grids of the fourth - level power - consuming unit grid. Then, update the division result of the fourth - level power - consuming unit grid according to the number of grids.

[0021] As an optimal solution, the step of respectively performing load forecasting on the under - construction area and the completed area within the lowest - level power - consuming unit grid according to the power - consumption characteristics and power - consumption trends of various types of loads, to obtain the load forecasting result corresponding to the under - construction area and the load forecasting result corresponding to the completed area, includes:

[0022] For the under - construction area within the lowest - level power - consuming unit grid, obtain the load density index corresponding to the under - construction area and the building area within the under - construction area. Then, predict the saturated annual load of the under - construction area according to the power - consumption characteristics of various types of loads, the power - consumption trends of various types of loads, the load density index, and the building area, to obtain the saturated annual load forecasting result corresponding to the under - construction area;

[0023] Predict the horizontal - year load of the under - construction area according to the power - consumption characteristics of various types of loads, the power - consumption trends of various types of loads, and the proportionality constant of the preset Logistic function, to obtain the horizontal - year load forecasting result corresponding to the under - construction area;

[0024] Take the saturated annual load forecasting result and the horizontal - year load forecasting result as the load forecasting result corresponding to the under - construction area;

[0025] For the completed area within the lowest - level power - consuming unit grid, obtain the historical load data of the completed area, and perform trend extrapolation and grey - theory prediction on the future load situation of the completed area according to the historical load data, to obtain the load forecasting result corresponding to the completed area.

[0026] As an optimal solution, the step of performing trend extrapolation and grey - theory prediction on the future load situation of the completed area according to the historical load data, to obtain the load forecasting result corresponding to the completed area, includes:

[0027] According to the historical load data of the built-up area, using the historical load data as the dependent variable and the preset factors related to the historical load data as the independent variables, a corresponding regression model is constructed, and then the future load of the built-up area is predicted according to the regression model to obtain the first load prediction result corresponding to the built-up area;

[0028] According to the historical load data and the preset grey Verhulst prediction model, the historical load data is input into the grey Verhulst prediction model, so that the grey Verhulst prediction model predicts the future load of the built-up area and outputs the second load prediction result corresponding to the built-up area;

[0029] Taking the first load prediction result and the second load prediction result as the load prediction result corresponding to the built-up area.

[0030] As a preferred solution, comprehensively analyzing the load prediction result corresponding to the under-construction area and the load prediction result corresponding to the built-up area to obtain the load prediction result within the grid of the lowest-level power consumption unit, including:

[0031] Combining the load prediction result corresponding to the under-construction area and the load prediction result corresponding to the built-up area to obtain several load prediction result combinations;

[0032] Obtaining the evaluation grade of each load prediction result combination, and selecting a load prediction result combination as the load prediction result within the grid of the lowest-level power consumption unit according to the evaluation grades of each load prediction result combination.

[0033] Based on the above embodiments, another embodiment of the present invention provides a distribution network grid-based load prediction device, including: a distribution network grid division module, a load characteristic data acquisition module, a load characteristic data analysis module, and a load prediction module;

[0034] The distribution network grid division module is used to obtain the regional parameters of the distribution network substation transformers, divide the distribution network substations into grid divisions according to the regional parameters, and divide the distribution network substations into several levels of power consumption unit grids; wherein, the regional parameters include: the regional function of the substation transformer, administrative division, total regional load, and power supply area;

[0035] The load characteristic data acquisition module is used to obtain the load characteristic data within the grid of the lowest-level power consumption unit; wherein, the load characteristic data includes: annual maximum load utilization hours, annual load curve, and daily load characteristics;

[0036] The load characteristic data analysis module is used to analyze the electricity consumption characteristics and trends of various types of loads in the lowest-level power consumption unit grid according to the load characteristic data. Among them, the load types include residential load, commercial load, industrial load, and administrative load;

[0037] The load forecasting module is used to perform load forecasting on the under-construction area and the completed area in the lowest-level power consumption unit grid respectively according to the electricity consumption characteristics and trends of various types of loads, obtain the load forecasting results corresponding to the under-construction area and the load forecasting results corresponding to the completed area, and then comprehensively analyze the load forecasting results corresponding to the under-construction area and the load forecasting results corresponding to the completed area to obtain the load forecasting result in the lowest-level power consumption unit grid.

[0038] Based on the above embodiments, another embodiment of the present invention provides an electronic device, which includes a processor, a memory, and a computer program stored in the memory and configured to be executed by the processor. When the processor executes the computer program, it implements the grid-based load forecasting method for the distribution network described in the above embodiments of the present invention.

[0039] Based on the above embodiments, another embodiment of the present invention provides a storage medium, which includes a stored computer program. When the computer program runs, it controls the device where the storage medium is located to execute the grid-based load forecasting method for the distribution network described in the above embodiments of the present invention.

[0040] Compared with the prior art, the embodiments of the present invention have the following beneficial effects:

[0041] The present invention provides a grid-based load forecasting method for a distribution network. First, the regional parameters of the distribution network substation transformer are obtained, and the distribution network substation is divided into grid cells of several levels according to the regional parameters. Then, the load characteristic data in the lowest-level power consumption unit grid is obtained. According to the load characteristic data, the electricity consumption characteristics and trends of various types of loads in the lowest-level power consumption unit grid are analyzed. According to the electricity consumption characteristics and trends of various types of loads, load forecasting is performed on the under-construction area and the completed area in the lowest-level power consumption unit grid respectively, and the load forecasting results corresponding to the under-construction area and the load forecasting results corresponding to the completed area are obtained. Then, the load forecasting results corresponding to the under-construction area and the load forecasting results corresponding to the completed area are comprehensively analyzed to obtain the load forecasting result in the lowest-level power consumption unit grid.

[0042] After the power distribution network is divided into grids, the load characteristic data within the grids of the lowest-level electricity-consuming units is obtained, and the load characteristic data is analyzed to analyze the electricity consumption characteristics and trends of load types such as residential load, commercial load, industrial load, and administrative load. Then, based on the electricity consumption characteristics and trends, the load within the grids of the lowest-level electricity-consuming units is predicted to obtain the load prediction results within the grids of the lowest-level electricity-consuming units. When predicting the grid load of the power distribution network, the present invention takes into account the load characteristics. By analyzing the load characteristics of the grid area and applying the analysis results to the prediction of the grid load of the power distribution network, the scientificity and rationality of the power distribution network planning can be improved, and the accuracy of the grid load prediction of the power distribution network can be enhanced. BRIEF DESCRIPTION OF THE DRAWINGS

[0043] Figure 1 is a schematic flowchart of a method for predicting the grid load of a power distribution network provided by an embodiment of the present invention;

[0044] Figure 2 is a schematic diagram of the overall idea of dividing the L2-level grid;

[0045] Figure 3 is a schematic diagram of the algorithm for dividing the central line electricity-consuming units;

[0046] Figure 4 is a flowchart of the algorithm for dividing electricity-consuming units based on K-medians;

[0047] Figure 5 is a schematic diagram of the proportion of industrial electricity consumption and the annual maximum load utilization hours in City A in the past 10 years;

[0048] Figure 6 is the annual load curve of City A in the past 10 years;

[0049] Figure 7 is the typical daily load curve of City A;

[0050] Figure 8 is a schematic diagram of the overall idea of load prediction considering load characteristics;

[0051] Figure 9 is a schematic structural diagram of a device for predicting the grid load of a power distribution network provided by an embodiment of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0052] To make the objectives, technical solutions, and advantages of the present application clearer, the technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present application. Apparently, the described embodiments are some, but not all, of the embodiments of the present application. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present application without creative efforts shall fall within the protection scope of the present application.

[0053] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the technical field to which this application belongs; the terms used herein are for the purpose of describing specific embodiments only and are not intended to limit this application; the terms "including" and "having" and any variations thereof in the specification and claims of this application and the above drawings are intended to cover non-exclusive inclusion.

[0054] In the description of the embodiments of this application, technical terms such as "first" and "second" are only used to distinguish different objects and cannot be understood as indicating or implying relative importance or implicitly indicating the quantity, specific order or primary-secondary relationship of the indicated technical features. In the description of the embodiments of this application, the meaning of "a plurality" is more than two, unless otherwise specifically defined.

[0055] Referring to [Figure X] (not shown), which is a schematic flowchart of a method for grid-based load forecasting of a distribution network provided by an embodiment of the present invention, the method includes the following specific steps:

[0056] Embodiment 1

[0057] Please refer to Figure 1 , which is a schematic flowchart of a method for grid-based load forecasting of a distribution network provided by an embodiment of the present invention, and includes the following specific steps:

[0058] S1. Obtain the regional parameters of the distribution network substation transformers, and divide the distribution network substation areas into grid units at several levels according to the regional parameters; wherein, the regional parameters include: the regional functions of the substation transformers, administrative divisions, total regional load, and power supply area.

[0059] Note: In the original text, the reference in step S1 is missing. I added a placeholder "[Figure X]" in the translation for step S8 for reference. You can replace it with the actual figure number. Also, the reference in step S16 is also missing in the original text. I added a placeholder "[Figure X]" in the translation for reference. You can replace it with the actual figure number.Preferably, the grid division of the distribution network substation area is carried out according to the area parameters, and the distribution network substation area is divided into several hierarchical power consumption unit grids, including: the grid division of the distribution network substation area is carried out according to the regional function, administrative division, total regional load and power supply area of the substation transformer, and the distribution network substation area is divided into several first-level power consumption unit grids and several second-level power consumption unit grids; wherein, the second-level power consumption unit grids belong to the first-level power consumption unit grids; the division result of the second-level power consumption unit grids is adjusted so that the total predicted daily average load of the second-level power consumption unit grids included in each first-level power consumption unit grid is equal; for each second-level power consumption unit grid, the grid division of the second-level power consumption unit grid is carried out according to the predicted daily average load and the preset land use nature in the second-level power consumption unit grid, and several third-level power consumption unit grids are obtained; the division result of the third-level power consumption unit grids is adjusted so that the total predicted daily average load of the third-level power consumption unit grids included in each second-level power consumption unit grid is equal; for each third-level power consumption unit grid, the grid division of the third-level power consumption unit grid is carried out according to the preset land use nature and air humidity in the third-level power consumption unit grid, and several fourth-level power consumption unit grids are obtained.

[0060] Preferably, before obtaining the load characteristics of the lowest-level power consumption unit grid, it further includes: adjusting the number of grids of the fourth-level power consumption unit grid, and updating the division result of the fourth-level power consumption unit grid according to the adjusted number of grids.

[0061] Preferably, the adjusting the number of grids of the fourth-level power consumption unit grid and updating the division result of the fourth-level power consumption unit grid according to the adjusted number of grids includes: decomposing the third-level power consumption unit grid into a first grid including a geographical boundary and a second grid without a geographical boundary; obtaining the power consumption unit data in the second grid; wherein, the power consumption unit data includes: power consumption unit longitude, power consumption unit latitude, land use nature weight, historical average air humidity and historical monthly repair data; clustering the second grid according to the power consumption unit data, and taking the number of the obtained clustering clusters as the number of grids of the fourth-level power consumption unit grid, and then updating the division result of the fourth-level power consumption unit grid according to the number of grids.

[0062] Specifically, the present invention provides a grid-based load forecasting method for distribution networks based on multi-dimensional data to achieve grid-based load forecasting for distribution network planning. Compared with the prior art, the technical problems to be solved by the present invention include:

[0063] (1) Comprehensively consider multi-dimensional data such as the load records of electricity-consuming units, repair data, and the climate conditions of the substation areas, and effectively conduct grid division of the distribution network for the to-be-planned substation areas;

[0064] (2) Combine the CLARANS and Chameleon clustering algorithms to achieve automatic update of the lowest-level grid of the distribution network in the built substation areas;

[0065] (3) Conduct research and analysis on the load characteristics from multiple perspectives. Take the annual load characteristics and daily load characteristics as the objects, analyze the electricity consumption characteristics and trends of residential loads, commercial loads, industrial loads, administrative loads, and other loads, make horizontal and vertical comparisons of different types of loads, find the characteristics of different types of loads, and use them to guide load forecasting, so as to optimize grid division and grid framework planning, and put forward a load forecasting idea considering load characteristics;

[0066] (4) Select appropriate load forecasting methods for different regions, and sort out the load forecasting methods for the under-construction areas and the built areas.

[0067] By solving the above technical problems, the present invention will be able to effectively achieve accurate load forecasting of the distribution network, thereby improving the power supply reliability and economy of the distribution network construction. The specific steps are as follows:

[0068] Step 1: Manually divide the grid of the distribution network substation area:

[0069] The specific steps for manual grid division of the substation area are as follows:

[0070] 1. Adjustment of grid division at L1 and L2 levels: Initially divide the L1 and L2 level grids of the distribution network according to the regional functions of the substation transformers, administrative divisions, regional loads, and power supply areas. Specifically as follows:

[0071] a. Divide according to administrative regions: Plan the distribution network of cities, towns, and rural areas at all levels respectively. The grid demarcation points are clear, but the electricity consumption demands of each grid vary greatly.

[0072] b. Divide according to regional functions: Plan different industries respectively. The electricity consumption characteristics within the grid are similar and convenient for management, but it is easy to cause the situation of "peak on peak", exacerbating the peak-valley difference.

[0073] c. Divide according to the total regional load: Plan the areas where the total load reaches the set value. The total loads of each area are close and the balance degree is relatively high, but the load characteristics within the grid vary greatly.

[0074] d. Divide according to the power supply area: Plan the areas where the power supply area reaches the set value. It is convenient for the management and maintenance of each grid, but it may cause load imbalance.

[0075] Adjust the grid to evenly distribute the predicted daily average load of the preset electricity consumption units within the L1 and L2 levels, so that the total predicted daily average load of the preset electricity consumption units in each L2-level grid within each L1-level grid is roughly similar. The specific adjustment process is as follows:

[0076] According to the characteristics of the electricity consumption units, use the K-medoids algorithm to divide the electricity consumption units, and divide the power supply grid into several electricity consumption units within the L2 level. Please refer to Figure 2 , which is a schematic diagram of the overall idea of dividing the L2-level grid.

[0077] The electricity consumption unit is essentially an area divided according to the distribution of the loads carried on the outgoing lines of a substation within the power supply area of a certain substation. Since the power supply area of a substation is generally approximately circular, and the feeder generally starts from the substation and is a straight line to the farthest load point in the power supply area, the area formed by the loads carried by the feeder is an approximately fan-shaped area. Therefore, from a geometric perspective, the division of electricity consumption units is essentially a problem of dividing multiple sectors within a circular area.

[0078] Since the traditional load clustering algorithm can only divide circular or elliptical areas and cannot meet the requirements of dividing electricity consumption units, therefore, the present invention applies the idea of the K-means clustering algorithm, combines the actual requirements of dividing electricity consumption units, considers the load characteristic factors and introduces a weighting factor, and proposes a new load clustering algorithm, that is, the K-medoids load clustering algorithm. Please refer to Figure 3 , which is a schematic diagram of the center line electricity consumption unit division algorithm. In each iteration, continuously update the center lines of the K divided fan-shaped areas until the updated center lines overlap with the center lines of the previous iteration. L1, L2, etc. are the completed electricity consumption units.

[0079] Please refer to Figure 4 , which is a flowchart of the electricity consumption unit division algorithm based on K-medoids. The specific algorithm steps of the above division are as follows:

[0080] (1) Taking the substation as the center, randomly generate K rays starting from the center, that is, the initial center lines, and each initial center line is randomly distributed;

[0081] (2) Calculate the geometric distance and weighted distance from each load point to the center line;

[0082] (3) According to the weighted distance calculated in (2), compare the weighted distances from each load point to the center line, so as to judge which center line the load point belongs to;

[0083] (4) According to the load point attribution in (3), cluster the load points belonging to the same center line and divide them into a sector, and calculate the geometric center line of the divided area;

[0084] (5) Determine whether the geometric centerline calculated in step 4 overlaps with the centerline of the previous iteration. If so, stop the calculation; otherwise, return to step 2.

[0085]

[0086]

[0087] where: d ij is the geometric distance from load point j to centerline i; A i , B i , C i are the linear expression coefficients of centerline i; is the weighted distance from load point j to centerline i; λ i is the load weighting factor of power supply unit i. When the total load value P i in power supply unit i is larger, λ i is also larger, making the load of each power supply unit more balanced; is the peak-valley difference weighting factor. When the peak-valley difference becomes larger after load point j is added to power supply unit i, is also larger, making the peak-valley difference of the power consumption unit smaller, thereby improving the utilization rate of the line; k1 and k2 are the influence coefficients of the load weighting factor and the peak-valley difference weighting factor, with values ranging from 1 to 2.

[0088] 2. L3-level grid division: For each L2-level grid, a reasonable manual grid plan is made according to factors such as the predicted daily average load of each preset power consumption unit in the area and the land use nature. While clustering in areas with the same land use nature, the total predicted daily average load of the preset power consumption units in the L3-level grids belonging to the same L2-level grid is made similar.

[0089] 3. Preliminary setting of L4-level grids: For each L3-level grid, a reasonable manual grid plan is made according to factors such as the land use nature and air humidity in the area. While clustering in areas with similar land use natures, the areas with relatively high air humidity are clustered as independently as possible. Since the range of the L4-level grid is small and the deviation between the predicted daily average load of the preset power consumption unit and the actual value is relatively large, this step is not considered.

[0090] Step 2: Automatically update the lowest-level grids of the distribution network. The specific steps are as follows:

[0091] 1. Preprocessing of L3-level grids: In the grid division of the power grid, the geographical boundary refers to the crossing rejection areas such as roads and rivers that do not contain any power consumption units. Here, those with a width exceeding 15 m are defined as large geographical boundaries. Adjust the L3-level grids containing large geographical boundaries inside, and decompose them into the first grid containing the geographical boundary and the second grid (L3' - level grid) without large geographical boundaries for subsequent clustering processing.

[0092] Assign the land use nature lu of the buildings under the electricity-consuming unit as shown in Table 1 below to obtain the land use nature weight LU. In the table: the land use nature lu is the major land use nature in the "Standard for Classification of Urban Land Use and Planning Construction Land (GB50137 - 2011)".

[0093] Land use nature lu Land use nature quantity B1 / B3 / R 0 A4 / G1 / G3 1 / 6 A1 / A2 / A3 / B4 1 / 3 A5 / B2 1 / 2 W 5 / 6 M / H 1

[0094] Table 1 Land Use Nature Assignment Table

[0095] In Table 1, B1 is land for commercial facilities; B3 is land for entertainment and fitness facilities; R is residential land; A4 is sports land; G1 is park green space; G3 is square land; A1 is administrative office land; A2 is cultural facility land; A3 is educational and scientific research land; B4 is land for public utility business outlets; A5 is medical and health land; B2 is business facility land; W is logistics and warehousing land; M is industrial land; H is urban construction land.

[0096] Then, reasonably segment and assign the total number of repair requests r of the electricity-consuming unit in the previous quarter, and rewrite it as the repair data R in the previous quarter. For the data set used in this article, the assignment is as follows:

[0097]

[0098] Finally, normalize the longitude and latitude long&lat of the electricity-consuming unit, the average air humidity h in the previous quarter, etc. within the L3' hierarchical grid range of the electricity-consuming unit respectively in the maximum-minimum normalization method to construct the L3' hierarchical power grid model:

[0099]

[0100] In the formula: x j ′ is the normalized data, x j is the data to be normalized, x min and x max are respectively the minimum and maximum values of the data to be normalized. The jth electricity-consuming unit is represented by the point x j .

[0101] 2. Chameleon clustering process:

[0102] Take the longitude long, latitude lat, land use nature weight LU, average air humidity h in the previous month, repair data R in the previous month, etc. of the electricity-consuming unit as input quantities, and model each electricity-consuming unit as follows:

[0103] x j =(long j , lat j , LU j , hj ,R j );

[0104]

[0105] Where: S i is the i-th cluster in the Chameleon clustering; is the center of the i-th cluster.

[0106] Considering the actual requirements of the substation area grid, determine the optimal number of the lowest-level grids under the L3' level grid, that is, the number of result clusters C. Perform the Chameleon clustering operation according to the following steps:

[0107] Process 1: Construct the k-nearest neighbor graph G k : The sparse graph is represented by the k-nearest neighbor graph method, and the vertices in G k represent a data object, the lines represent that two data objects are k most similar objects to each other, and the weights of the lines are represented by similarity, that is, the line weights between two data objects with a larger distance are larger and the similarity is smaller. Thus, construct the k-nearest neighbor graph G k . The calculation formula is as follows:

[0108]

[0109] Where: a mn is the similarity between two data; z m is the m-th data; z n is the n-th data.

[0110] Process 2: Divide G k Figure: Divide the G graph according to the sum of the weights of the truncated lines to be minimized, and split it into many unconnected subgraphs. Each subgraph is used as the initial sub-cluster in the second-stage hierarchical clustering.

[0111] Process 3: Merge sub-clusters to form result clusters: Determine the similarity between two clusters in a dynamic modeling manner, and at the same time consider the relative interconnectivity RI and relative approximation RC between the two clusters. When the similarity function value is higher than the given threshold, merge the two sub-clusters. Repeat the above merging process until the number of result clusters is reached.

[0112] Thus, obtain the Chameleon clustering grid analysis result, and obtain the distance from the j-th point to its clustering center as follows: as follows:

[0113]

[0114] 3. CLARANS clustering process: First, set the minimum value of the daily average power consumption load of the electricity consumption unit last month as the standard quantity household p a, the average daily electricity load of all other electricity-consuming units in the previous month is p i , calculated according to the following formula:

[0115] p i / p a ≈n;

[0116] According to the size of the average daily electricity load of each electricity-consuming unit relative to the standard amount, a single electricity-consuming unit is rewritten as n electricity-consuming units with the same longitude and latitude. Since the CLARANS clustering algorithm tends to obtain clusters of similar sizes, this method can be used to control the total average daily electricity load of each minimum substation area grid to be similar.

[0117] Taking the longitude long, latitude lat, land use property weight LU, average air humidity h in the previous month, repair data R in the previous month, etc. of the electricity-consuming unit as input quantities, the model for each electricity-consuming unit is built as follows:

[0118] x j =(long j , lat j , LU j , h j , R j );

[0119]

[0120] In the formula: S′ i is the i-th cluster in the CLARANS clustering, is the center of the i-th cluster in the CLARANS clustering.

[0121] Taking the number C of the result clusters determined in the Chameleon clustering process as the number of clusters in this stage, perform CLARANS clustering, thereby obtaining the CLARANS clustering grid analysis result, and obtaining the distance from the j-th point to its cluster center as follows:

[0122]

[0123] 4. Clustering synthesis: For the j-th point, first compare its cluster centers under the two clustering methods Calculate the distance Δd of the projection of the two centers on the long-lat plane :

[0124]

[0125] If Δd is not greater than the threshold 0.1, it can be considered that the clusters S i , S′ i: Projections on the long-lat plane are similar and can be regarded as the same cluster; if they cannot be regarded as the same cluster, then compare the projection length on the long-lat plane The power distribution equipment belongs to the cluster corresponding to the one with a smaller projection length.

[0126] Thus, an updated L4-level grid is constructed to obtain an automatically updated result of the lowest-level grid that is more suitable for the substation area power grid dispatch.

[0127] S2. Obtain the load characteristic data within the grid of the lowest-level power consumption unit; wherein, the load characteristic data includes: annual maximum load utilization hours, annual load curve, and daily load characteristics;

[0128] Step 3: Accurately predict the load in the power distribution network planning area and analyze the load characteristics from multiple perspectives:

[0129] Load is an important basis in power distribution network planning. Accurately predicting the load in the planning area is of great significance. Load characteristics can be studied from multiple perspectives. This invention uses the load data of City A as a reference, takes the annual load characteristics and daily load characteristics as objects, analyzes the electricity consumption characteristics and trends of residential load, commercial load, industrial load, administrative load, and other loads, makes horizontal and vertical comparisons of different types of loads, finds the characteristics of different types of loads, and uses them to guide load forecasting, so as to optimize grid division and grid structure planning.

[0130] Specifically, in the grid-based planning of the power distribution network, different load characteristics affect different load forecasting results, thus resulting in different loads in each horizontal year and the target year. For the power supply area, the load in the saturation year determines the load density, that is, the level of grid structure planning. For the power supply grid, the load directly determines the power balance within the grid. For the power supply unit, the load directly determines the wiring mode and wiring scale within the power supply unit. Therefore, considering load characteristics is of great significance for the grid-based planning results of the power distribution network. Correctly analyzing the load characteristics of the planning area and reasonably applying the analysis results to the grid-based planning of the power distribution network can improve the scientificity and rationality of the power distribution network planning.

[0131] Generally speaking, the load characteristic analysis in this step makes an auxiliary analysis for the selection of the load forecasting method in the next step (the uncertainty factors in the planning area are relatively high, and it is difficult to have accurate actual data. Generally, only basic plot types are available, as described in the first step). The data such as the annual maximum load utilization hours obtained can provide basic references and verifications for the load forecasting in the planning area.

[0132] 1. Analysis of the annual maximum load utilization hours: The annual maximum load utilization hours is the ratio of the annual electricity consumption to the annual maximum load, which is mainly determined by the regional industrial structure. Generally speaking, the larger the proportion of electricity consumption in the secondary industry in the total electricity consumption, the larger the annual maximum load utilization hours. The statistical table of the electricity consumption of the tertiary industry and residents in City A in the past 10 years is shown in Table 2 below.

[0133]

[0134]

[0135] Table 2 Statistical Table of Electricity Consumption of the Tertiary Industry and Residents in City A in the Past 10 Years

[0136] With the economic development of the region and the improvement of people's living standards, the proportion of electricity consumption in the primary industry and residents in City A has decreased, while the proportion of electricity consumption in the secondary and tertiary industries has increased slightly. Please refer to Figure 5 , which is a schematic diagram of the proportion of industrial electricity consumption and the annual maximum load utilization hours in City A in the past 10 years.

[0137] It can be seen that the growth trend of the proportion of electricity consumption in the secondary industry is similar to the annual maximum load utilization hours. When the proportion of the secondary industry is less than 60%, the annual maximum load utilization hours are between 2,500 hours and 4,000 hours. When the proportion of the secondary industry is higher than 60%, the annual maximum load utilization hours can reach more than 5,000 hours.

[0138] 2. Analysis of the annual load curve: The maximum electricity load of the whole society in City A in 2010 was 32.61 MW, and the maximum electricity load of the whole society reached 74.03 MW in 2015. The average annual growth rate from 2011 to 2015 was 18.04%. The maximum electricity load of the whole society reached 150.7 MW in 2019, and the average annual growth rate from 2016 to 2020 was 19.98%. The annual maximum load values in the past 10 years in City A are shown in Table 3 below.

[0139]

[0140] Table 3 Annual Maximum Load Values in the Past 10 Years in City A

[0141] Please refer to Figure 6, is the annual load curve of City A in the past 10 years. From the annual load curve of City A in the past 10 years, there are two peaks in the load change from January to December, which appear in July, August and January, December of each year respectively, that is, the so-called summer large load and winter large load. City A has a subtropical monsoon climate, with relatively hot summers. Loads of types such as air conditioners and fans are relatively large, resulting in the maximum load appearing in summer. Compared with May and October when the temperature is suitable, the demand for cooling loads is very low, and two troughs appear in the annual load curve. Analyzing the cooling load in summer and the load peak during winter is of great significance for peak shaving in summer and winter, and has a guiding role in the commissioning of new equipment and the transformation of old equipment.

[0142] 3. Analysis of daily load characteristics: The typical daily load curve provides a basis for the load density index of the load density method by analyzing the electricity consumption characteristics of different types of loads at the same moment on a typical day. At the same time, analyzing the load curve of a typical day can guide the selection of the coincidence rate in load forecasting. The coincidence rate is the ratio of the maximum network supply load to the sum of the maximum load values of each user. The size of the coincidence rate is related to factors such as the load structure, the development of social economy, and the development of new energy. In load forecasting, selecting a reasonable coincidence rate can greatly improve the accuracy of load forecasting. Taking the moment of the annual maximum load in City A as a typical day, analyze the typical daily load curves of residential load, commercial load, industrial load, administrative load and other loads. For the convenience of analysis, draw a graph with the percentage of various load values, please refer to Figure 7 , is the typical daily load curve of City A.

[0143] It can be seen from Figure 7 that the residential load is concentrated between 12:00 and 23:00, with two peaks in a day, which are 13:00 and 22:00 respectively, related to the living habits of residents. The commercial load is concentrated between 6:00 and 20:00, and the peak appears at 6:00. The industrial load is mainly concentrated between 0:00 and 8:00 and 22:00 and 23:00. Due to the influence of time-of-use electricity prices, the industrial load is concentrated at night, which is opposite to the residential load curve. Therefore, the coincidence rate between the residential load and the industrial load is relatively low, showing the characteristic of complementary load characteristics. The administrative load is concentrated between 7:00 and 20:00, and the load fluctuation during the day is not large, with a relatively high coincidence rate with the commercial load. Other loads are relatively stable and there is basically no significant fluctuation throughout the day. At the same time, the peak of the commercial load and the trough of the residential load appear at roughly the same time, and the two loads have obvious complementarity. In grid division, reasonably utilizing the characteristic of complementary load characteristics can reduce the peak-valley difference of the power supply unit, thereby improving the utilization rate of a single feeder, reducing the appearance interval of substations, and improving the maximum power supply capacity of substations.

[0144] S3. Analyze the electricity consumption characteristics and trends of various types of loads in the lowest-level power consumption unit grid based on the load characteristic data, where the load types include residential load, commercial load, industrial load, and administrative load.

[0145] After obtaining the load characteristic data of the lowest-level (L4 level) power consumption unit grid, based on the load characteristic data, analyze the electricity consumption characteristics and trends of residential load, commercial load, industrial load, administrative load, and other types of loads in the lowest-level power consumption unit grid.

[0146] S4. According to the electricity consumption characteristics and trends of various types of loads, respectively conduct load forecasting for the under-construction area and the completed area in the lowest-level power consumption unit grid, obtain the load forecasting results corresponding to the under-construction area and the load forecasting results corresponding to the completed area, and then conduct comprehensive analysis on the load forecasting results corresponding to the under-construction area and the load forecasting results corresponding to the completed area to obtain the load forecasting result of the lowest-level power consumption unit grid.

[0147] Preferably, the step of respectively conducting load forecasting for the under-construction area and the completed area in the lowest-level power consumption unit grid according to the electricity consumption characteristics and trends of various types of loads, and obtaining the load forecasting results corresponding to the under-construction area and the load forecasting results corresponding to the completed area includes: for the under-construction area in the lowest-level power consumption unit grid, obtain the load density index corresponding to the under-construction area and the building area within the under-construction area, and then predict the saturated annual load of the under-construction area based on the electricity consumption characteristics of various types of loads, the electricity consumption trends of various types of loads, the load density index, and the building area to obtain the saturated annual load forecasting result corresponding to the under-construction area; predict the horizontal-year load of the under-construction area based on the electricity consumption characteristics of various types of loads, the electricity consumption trends of various types of loads, and the proportional constant of the preset Logistic function to obtain the horizontal-year load forecasting result corresponding to the under-construction area; use the saturated annual load forecasting result and the horizontal-year load forecasting result as the load forecasting results corresponding to the under-construction area; for the completed area in the lowest-level power consumption unit grid, obtain the historical load data of the completed area, and conduct trend extrapolation and grey theory forecasting on the future load situation of the completed area based on the historical load data to obtain the load forecasting result corresponding to the completed area.

[0148] Preferably, trend extrapolation and grey theory prediction are performed on the future load conditions of the built-up area based on the historical load data to obtain the load prediction results corresponding to the built-up area, including: based on the historical load data of the built-up area, using the historical load data as the dependent variable and the preset factors related to the historical load data as the independent variables, constructing a corresponding regression model, and then predicting the future load of the built-up area according to the regression model to obtain the first load prediction result corresponding to the built-up area; according to the historical load data and the preset grey Verhulst prediction model, inputting the historical load data into the grey Verhulst prediction model to enable the grey Verhulst prediction model to predict the future load of the built-up area and output the second load prediction result corresponding to the built-up area; using the first load prediction result and the second load prediction result as the load prediction results corresponding to the built-up area.

[0149] Preferably, comprehensive analysis is performed on the load prediction results corresponding to the under-construction area and the load prediction results corresponding to the built-up area to obtain the load prediction results within the grid of the lowest-level electricity consumption unit, including: combining the load prediction results corresponding to the under-construction area and the load prediction results corresponding to the built-up area to obtain several combinations of load prediction results; obtaining the evaluation grades of each combination of load prediction results, and selecting a combination of load prediction results as the load prediction results within the grid of the lowest-level electricity consumption unit according to the evaluation grades of each combination of load prediction results.

[0150] Step 4: Propose a load prediction method considering load characteristics:

[0151] The load prediction method considering load characteristics is mainly divided into two parts, namely the load prediction method for the under-construction area and the load prediction method for the built-up area. The under-construction area refers to an area that has not been fully developed and where there is still great room for the development of power load, generally referring to the suburban areas of a planned city and industrial parks that have not been fully developed; the built-up area refers to an area that has been fully developed and where the power load is basically saturated, generally referring to the central urban area and mature industrial parks. When there is no clear regulatory plan for the planned area, the division standard of the area is determined by the annual maximum load utilization hours and the average annual growth rate. An area with an annual maximum utilization hours greater than 4000 hours and an average annual growth rate less than 5% is a built-up area, otherwise it is an under-construction area.

[0152] The load forecasting method for the under-construction area is based on the regulatory detailed plan and land use plan, referring to the load situation in the current year and the load density in the target year, selecting the load density index by combining various load characteristics such as residential, commercial, industrial, and administrative loads, using the load density method to predict the saturated-year load value, and using the Logistic model to predict the load values for each target year. The load forecasting method for the built-up area is based on the historical power data, and uses the trend extrapolation method and the grey theory method to predict the future load. Thus, three scenarios of high, medium, and low are formed, and according to the actual situation of the forecasting area, a suitable scenario is selected as the final forecasting result. Multiple methods are used for forecasting, and the results of different methods are high or low, so different forecasting scenarios are obtained. For example, using the load density method in the under-construction area and the grey theory method in the built-up area can form a forecasting scenario, and using the load density method in the under-construction area and the trend extrapolation method in the built-up area can obtain another forecasting scenario. In this way, multiple different scenarios can be obtained, and they are divided into high, medium, and low scenarios according to different forecasting results (specifically, by comparing the rationality of the forecasting results and eliminating relatively unreasonable scenarios). Generally, only the middle scenario is adopted.

[0153] Please refer to Figure 8 , which is a schematic diagram of the overall idea of load forecasting considering load characteristics. The specific forecasting steps are as follows:

[0154] (1) Load density method: The load density method is often used for spatial load forecasting and is applicable to areas with regulatory detailed plans. For the land use nature of different plots, combined with the local economic development situation, a suitable load density index is selected, and multiplying the load density index by the building area can obtain the predicted load value. Therefore, the load density method is used to predict the saturated-year load. The model of the load density method is shown in the following formula:

[0155]

[0156] In the formula: P L is the load forecasting value for the saturated year; i is the number of the power supply unit; m is the number of power supply units; K i is the coincidence rate of different types of loads in the i-th power supply unit; j is the number of the plot; n is the number of all plots in the power supply unit; D j is the load density of the j-th plot; S j is the area of the j-th plot; W j is the demand coefficient of the j-th plot.

[0157] Regarding the selection of the load density index, it is stipulated in the implementation rules for the planning and design of the distribution network.

[0158] (2) Logistic Function Method: The Logistic function is a common S-shaped curve function, initially applied to the study of population growth trends. Later, through development, it has been widely used in the agriculture, economy, and power industries. The Logistic curve has four development stages, namely the initial growth stage, the early rapid growth stage, the late rapid growth stage, and the saturation growth stage. The Logistic function is as follows.

[0159]

[0160] In the formula: y is the dependent variable of the function; K is the final value of the dependent variable; P0 is the initial value of the dependent variable; r is the curve growth rate factor; t is the time variable.

[0161] Since the development trend of power in the planned year is similar to the S-shaped curve, the Logistic function can be used to calculate the load value of each planned year. For the sake of simple calculation and convenient practical application in engineering, the formula can be simplified to obtain:

[0162]

[0163] In the formula: y is the dependent variable of the function; a, b, and c are all constants; t is the time variable.

[0164] The following formula is the simplified formula of the Logistics function. As described below, the corresponding abc constants can be fitted using historical data, and then by changing the time variable t, the load value of the corresponding planned year can be planned.

[0165] In practical engineering applications, the constants a, b, and c can be obtained by directly fitting and solving the historical data through the MATLAB toolbox. From a large number of projects that have been practiced, the different development stages of the Logistic function are very consistent with the development trend of China's power industry. The Logistic curve can well adapt to China's power demand forecast and can accurately predict the load value of each planned year in the planning.

[0166] (3) Trend extrapolation method: The trend extrapolation method is applicable to medium- and long-term load forecasting with complete historical data, featuring simple calculation and strong applicability. The trend extrapolation method takes the historical data of various loads as the dependent variable and the factors related to the load data as the independent variable, and uses regression analysis to establish a mathematical model, and repeatedly calculates to obtain the prediction result. The commonly used regression analysis mathematical functions are exponential function, linear function, power function and polynomial. By analyzing the load values predicted by the regression curve model, combining the current economic development situation and the development and construction situation of the planned area, a comprehensive analysis of the prediction results of various regression models is carried out, and finally the load prediction result of the trend extrapolation method is obtained. Among them, the historical data are the annual electricity consumption, annual maximum load, annual maximum load utilization hours, etc. of different regions, and the relevant specific factors are local development policies, large user installation, industrial plant shutdown, etc.

[0167] (4) Grey theory method: The grey theory method makes predictions on the future state of the system through the processing of original data and the establishment of grey models. At present, the GM(1,1) model is widely used, but the GM(1,1) model is only applicable to sequences with single exponential changes and cannot well reflect the development trend of electric power loads. To address this problem, the grey Verhulst model can be used to predict S-type sequences with saturation states. To improve the prediction accuracy, the Markov model can be used to correct the residuals of the predicted data, that is, the grey Markov Verhulst model. The specific prediction steps are as follows:

[0168] Let the original sample sequence be Perform first-order cumulative summation on this sequence to obtain a new data sequence as Where:

[0169] Obtain the adjacent mean generation sequence from the new cumulative sequence Where:

[0170]

[0171] Establish the whiting differential equation of the grey Verhulst model, that is:

[0172] The parameters a and b can be obtained by least squares estimation. Let A=(a,b) T be the parameter column, then the least estimate of the parameter column is A=(B T B) -1 B T Y, where:

[0173]

[0174] Substitute the obtained a and b into the whiting differential equation to obtain the time response equation of the grey Verhulst model:

[0175] Performing cumulative subtraction reduction on the above formula to obtain the grey Verhulst prediction model of the original data sequence:

[0176]

[0177] According to the above formula, for example, from 2000 to 2024, there are 24 original data, k is 24, and the new data sequence can be calculated, which is the new sequence from 2000 to 2025. Then, by using the model, subtracting 24 years from 25 years to obtain the actual data for 25 years.

[0178] Thus, the present invention provides a method for grid-based load forecasting of a distribution network, and the following beneficial effects can be achieved through the present invention:

[0179] (1) Based on multi-dimensional data analysis of the distribution network: comprehensively considering multi-dimensional data such as the load records of power consumption units, repair data, and the climate conditions of the substation area, effectively dividing the distribution network grid for the substation area to be planned. Compared with most existing load forecasting methods, the present invention has higher adaptability and accuracy;

[0180] (2) Automatic update of the lowest-level grid of the substation area distribution network: The present invention uses the Chameleon, CLARANS, and hMetis algorithms to divide the distribution network grid, and obtains the automatic update result of the lowest-level grid for substation area power grid dispatching. Compared with traditional load forecasting methods, this method has better real-time performance and can well reduce the labor cost of manual update;

[0181] (3) Proposing a load forecasting method considering load characteristics: The present invention analyzes the load characteristics of City A from multiple perspectives, and further proposes a load forecasting idea considering load characteristics on this basis. Compared with traditional distribution network load forecasting methods, the present invention can more accurately consider the characteristics of different places and perform load forecasting according to the characteristics, with higher accuracy;

[0182] (4) Proposing a method for optimizing grid division: By using the K-medoids algorithm and combining manual division of the distribution network grid, load characteristics such as the load size, peak-valley difference, and load type of the grid are introduced into the grid division, and each power consumption unit is further divided to optimize the grid division result of the distribution network. Compared with the traditional manual grid division method, the present invention has a more accurate and efficient grid division method.

[0183] Embodiment 2

[0184] Please refer to Figure 9, which is a schematic structural diagram of a distribution network grid-based load forecasting device provided by an embodiment of the present invention. The device includes: a distribution network grid division module, a load characteristic data acquisition module, a load characteristic data analysis module, and a load forecasting module;

[0185] The distribution network grid division module is used to obtain the regional parameters of the distribution network substation transformers, divide the distribution network substation areas according to the regional parameters, and divide the distribution network substation areas into several levels of power consumption unit grids; wherein, the regional parameters include: the regional functions of the substation transformers, administrative divisions, total regional load, and power supply area;

[0186] The load characteristic data acquisition module is used to obtain the load characteristic data within the lowest-level power consumption unit grid; wherein, the load characteristic data includes: annual maximum load utilization hours, annual load curve, and daily load characteristics;

[0187] The load characteristic data analysis module is used to analyze the power consumption characteristics and power consumption trends of various types of loads within the lowest-level power consumption unit grid according to the load characteristic data; wherein, the load types include: residential load, commercial load, industrial load, and administrative load;

[0188] The load forecasting module is used to perform load forecasting on the under-construction area and the completed area within the lowest-level power consumption unit grid respectively according to the power consumption characteristics and power consumption trends of various types of loads, obtain the load forecasting results corresponding to the under-construction area and the load forecasting results corresponding to the completed area, and then comprehensively analyze the load forecasting results corresponding to the under-construction area and the load forecasting results corresponding to the completed area to obtain the load forecasting result within the lowest-level power consumption unit grid.

[0189] It should be noted that the device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed to multiple network units. Some or all of the modules can be selected according to actual needs to achieve the purpose of the solution of this embodiment. In addition, in the attached drawings of the device embodiments provided by the present invention, the connection relationships between the modules indicate that they have communication connections, which can be specifically implemented as one or more communication buses or signal lines. Those of ordinary skill in the art can understand and implement without creative efforts.

[0190] Those skilled in the art can clearly understand that for the sake of convenience and brevity, the specific working process of the device described above can refer to the corresponding process in the foregoing method embodiment and will not be elaborated herein.

[0191] Embodiment Three

[0192] Accordingly, an embodiment of the present invention provides an electronic device, which includes a processor, a memory, and a computer program stored in the memory and configured to be executed by the processor. When the processor executes the computer program, the method for grid-based load forecasting of a distribution network described in the above embodiments of the present invention is implemented.

[0193] The electronic device may be a computing device such as a desktop computer, a notebook, a palm computer, and a cloud server. The device may include, but is not limited to, a processor and a memory.

[0194] The so-called processor may be a central processing unit (CPU), or may also be other general-purpose processors, digital signal processors (DSPs), application specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. The general-purpose processor may be a microprocessor, or the processor may also be any conventional processor, etc. The processor is the control center of the device, and connects various parts of the entire device through various interfaces and lines.

[0195] Embodiment 4

[0196] Accordingly, an embodiment of the present invention provides a storage medium, which includes a stored computer program. When the computer program runs, the device where the storage medium is located is controlled to execute the method for grid-based load forecasting of a distribution network described in the above embodiments of the present invention.

[0197] The memory may be used to store the computer program. The processor realizes various functions of the device by running or executing the computer program stored in the memory and calling the data stored in the memory. The memory may mainly include a program storage area and a data storage area. Among them, the program storage area may store an operating system, application programs required for at least one function, etc.; the data storage area may store data created according to the use of the mobile phone, etc. In addition, the memory may include high-speed random access memory, and may also include non-volatile memory, such as a hard disk, a memory, a plug-in hard disk, a smart media card (SMC), a secure digital (SD) card, a flash card, at least one magnetic disk storage device, a flash device, or other volatile solid-state storage devices.

[0198] The storage medium is a computer-readable storage medium, and the computer program is stored in the computer-readable storage medium. When the computer program is executed by a processor, the steps of the above-described various method embodiments can be implemented. Among them, the computer program includes computer program code, and the computer program code can be in the form of source code, object code, executable file, or some intermediate form, etc. The computer-readable medium can include: any entity or device capable of carrying the computer program code, recording medium, USB flash drive, mobile hard disk, magnetic disk, optical disc, computer memory, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), electrical carrier signal, telecommunication signal, and software distribution medium, etc. It should be noted that the content included in the computer-readable medium can be appropriately increased or decreased according to the requirements of legislation and patent practice in the jurisdiction. For example, in some jurisdictions, according to legislation and patent practice, the computer-readable medium does not include electrical carrier signals and telecommunication signals.

[0199] The above is the preferred embodiment of the present invention. It should be pointed out that for those of ordinary skill in the art, without departing from the principle of the present invention, several improvements and refinements can be made, and these improvements and refinements are also regarded as the protection scope of the present invention.

Claims

1. A distribution network grid load forecasting method, characterized in that: include: Acquire regional parameters of the transformer in the distribution network area, divide the distribution network area into grids according to the regional parameters, and divide the distribution network area into several levels of power consumption unit grids; wherein the regional parameters include: regional functions, administrative divisions, total regional loads, and power supply area of ​​the transformer in the area; Obtaining load characteristic data within the lowest-level power consumption unit grid; wherein the load characteristic data includes: annual maximum load utilization hours, annual load curve, and daily load characteristics; According to the load characteristic data, the power consumption characteristics and power consumption trends of various types of loads in the lowest-level power consumption unit grid are analyzed; wherein the load types include: residential load, commercial load, industrial load and administrative load; According to the power consumption characteristics and trends of various types of loads, load forecasts are performed for the under-construction areas and built-up areas within the lowest-level power consumption unit grid respectively, and the load forecast results corresponding to the under-construction areas and the load forecast results corresponding to the built-up areas are obtained. Then, the load forecast results corresponding to the under-construction areas and the load forecast results corresponding to the built-up areas are comprehensively analyzed to obtain the load forecast results within the lowest-level power consumption unit grid.

2. The distribution network grid load forecasting method according to claim 1, characterized in that: The grid division of the distribution network area according to the regional parameters, dividing the distribution network area into several levels of power consumption unit grids, includes: According to the regional functions, administrative divisions, total regional load and power supply area of ​​the transformer in the distribution network, the distribution network area is divided into a plurality of first-level power consumption unit grids and a plurality of second-level power consumption unit grids; wherein the second-level power consumption unit grids belong to the first-level power consumption unit grids; Adjusting the division result of the second-level power consumption unit grid so that the total amount of the preset daily average load forecast amount of the second-level power consumption unit grid contained in each first-level power consumption unit grid is equal; For each second-level power consumption unit grid, according to the preset daily average load forecast amount and the preset land use nature in the second-level power consumption unit grid, the second-level power consumption unit grid is grid-divided to obtain a plurality of third-level power consumption unit grids; The division result of the third-level power consumption unit grid is adjusted so that the total amount of the preset power consumption unit daily average load forecast amount of the third-level power consumption unit grid contained in each second-level power consumption unit grid is equal; For each third-level power consumption unit grid, the third-level power consumption unit grid is divided into grids according to the preset land use properties and air humidity in the third-level power consumption unit grid to obtain a plurality of fourth-level power consumption unit grids.

3. The distribution network grid load forecasting method according to claim 2, characterized in that: Before obtaining the load characteristics of the lowest-level power consumption unit grid, it also includes: The number of grids of the fourth-level electricity consumption unit grid is adjusted, and the division result of the fourth-level electricity consumption unit grid is updated according to the adjusted number of grids.

4. The distribution network grid load forecasting method according to claim 3, characterized in that: The step of adjusting the number of grids of the fourth-level power consumption unit grid and updating the division result of the fourth-level power consumption unit grid according to the adjusted number of grids includes: Decomposing the third-level electricity consumption unit grid into a first grid including geographical boundaries and a second grid excluding geographical boundaries; Acquire the data of the power consumption units in the second grid; wherein the data of the power consumption units include: the longitude of the power consumption unit, the latitude of the power consumption unit, the weight of the land use nature, the historical average air humidity and the historical monthly repair data; The second grid is clustered according to the power unit data, and the number of obtained clusters is used as the number of grids of the fourth-level power unit grid, and then the division result of the fourth-level power unit grid is updated according to the number of grids.

5. The distribution network grid load forecasting method according to claim 1, characterized in that: According to the power consumption characteristics and power consumption trends of various types of loads, load forecasting is performed on the under-construction area and the built area in the lowest-level power consumption unit grid, and load forecasting results corresponding to the under-construction area and load forecasting results corresponding to the built area are obtained, including: For the under-construction area within the lowest-level power consumption unit grid, the load density index corresponding to the under-construction area and the building area within the under-construction area are obtained, and then the saturated annual load of the under-construction area is predicted according to the power consumption characteristics of each type of load, the power consumption trend of each type of load, the load density index and the building area, and the saturated annual load prediction result corresponding to the under-construction area is obtained; According to the power consumption characteristics of each type of load, the power consumption trend of each type of load and the proportional constant of the preset Logistic function, the horizontal annual load of the area under construction is predicted to obtain the horizontal annual load prediction result corresponding to the area under construction; Using the saturated year load forecast result and the level year load forecast result as the load forecast result corresponding to the area under construction; For the built-up areas within the lowest-level electricity consumption unit grid, historical load data of the built-up areas are obtained, and trend extrapolation and grey theory prediction are performed on the future load conditions of the built-up areas based on the historical load data to obtain load prediction results corresponding to the built-up areas.

6. The distribution network grid load forecasting method according to claim 5, characterized in that: The trend extrapolation and grey theory prediction of the future load situation of the built-up area based on the historical load data to obtain the load prediction result corresponding to the built-up area includes: According to the historical load data of the built-up area, the historical load data is used as a dependent variable, and the preset factors related to the historical load data are used as independent variables to construct a corresponding regression model, and then the future load of the built-up area is predicted according to the regression model to obtain a first load prediction result corresponding to the built-up area; According to the historical load data and a preset grey Verhulst prediction model, the historical load data is input into the grey Verhulst prediction model, so that the grey Verhulst prediction model predicts the future load of the built-up area and outputs a second load prediction result corresponding to the built-up area; The first load forecast result and the second load forecast result are used as the load forecast results corresponding to the built-up area.

7. The distribution network grid load forecasting method according to claim 6, characterized in that: The load forecast results corresponding to the under-construction area and the load forecast results corresponding to the built area are comprehensively analyzed to obtain the load forecast results within the lowest level power consumption unit grid, including: Combining the load forecast results corresponding to the under-construction area and the load forecast results corresponding to the completed area to obtain several load forecast result combinations; The evaluation levels of the load forecast result combinations are obtained, and according to the evaluation levels of the load forecast result combinations, a load forecast result combination is selected as the load forecast result within the lowest level power consumption unit grid.

8. A distribution network grid load forecasting device, characterized in that: include: Distribution network grid division module, load characteristic data acquisition module, load characteristic data analysis module and load forecasting module; The distribution network grid division module is used to obtain the regional parameters of the distribution network transformer, and divide the distribution network into grids according to the regional parameters, so as to divide the distribution network into several levels of power consumption unit grids; wherein the regional parameters include: the regional function of the transformer in the distribution network, the administrative division, the total regional load and the area of ​​the power supply area; The load characteristic data acquisition module is used to acquire the load characteristic data in the lowest level power consumption unit grid; wherein the load characteristic data includes: annual maximum load utilization hours, annual load curve and daily load characteristics; The load characteristic data analysis module is used to analyze the power consumption characteristics and power consumption trends of various types of loads in the lowest-level power consumption unit grid according to the load characteristic data; wherein the load types include: residential load, commercial load, industrial load and administrative load; The load forecasting module is used to perform load forecasting for the under-construction area and the built-up area within the lowest-level power consumption unit grid according to the power consumption characteristics and power consumption trends of various types of loads, obtain the load forecasting results corresponding to the under-construction area and the load forecasting results corresponding to the built-up area, and then conduct a comprehensive analysis of the load forecasting results corresponding to the under-construction area and the load forecasting results corresponding to the built-up area to obtain the load forecasting results within the lowest-level power consumption unit grid.

9. An electronic device, characterized in that: The method comprises a processor, a memory and a computer program stored in the memory and configured to be executed by the processor, wherein when the processor executes the computer program, the method for grid load forecasting of a distribution network as claimed in any one of claims 1 to 7 is implemented.

10. A storage medium, characterized in that: The storage medium includes a stored computer program, wherein when the computer program is executed, the device where the storage medium is located is controlled to execute the distribution network grid load forecasting method according to any one of claims 1 to 7.

Images