Search

CN-121212678-B - Grid planning analysis system and method based on power distribution network

CN121212678BCN 121212678 BCN121212678 BCN 121212678BCN-121212678-B

Abstract

The invention discloses a grid planning analysis system and method based on a power distribution network, which belong to the technical field of power distribution network planning, and are characterized in that a power consumption load, urban space characteristics and power grid historical data taking a minimum power supply grid unit as a reference in a planning area are collected, a traceability model of load abnormality and space characteristics is established, quantitative association relation is determined, load prediction logic is further constructed and checked, a space-time dynamic load curve is generated, three types of constraints including space time sequence, functional linkage and environmental sensitivity in urban planning are extracted, the constraints and power grid facility parameters are respectively converted into feature vectors, a dynamic matching degree is introduced, a scheme meeting requirements is screened, the power grid adaptability is evaluated through simulating scenes such as space function mutation, constraint adjustment and multi-fault linkage, reverse optimization is carried out on a substandard scheme, and finally a complete planning scheme comprising load prediction, constraint matching, adaptation verification report and dynamic adjustment rules is formed.

Inventors

  • YANG ZIJIE
  • Xie Qiduan
  • CHEN ZHENQING
  • Zhan Nannan
  • JI HONGWEI
  • Feng Qinzhi
  • HUANG ZIANG
  • Xie Congdong

Assignees

  • 福建扬天电能有限公司

Dates

Publication Date
20260505
Application Date
20250923

Claims (8)

  1. 1. A grid planning analysis method based on a power distribution network is characterized by comprising the following steps: The method comprises the steps of S1, collecting power load data, urban space feature data and historical power grid operation data which are used as references in a planning area, establishing a tracing model of load abnormality and space features, carrying out tracing analysis on load abnormality fluctuation based on the tracing model, and determining quantitative association relation between space feature change and load fluctuation, wherein the space features are divided into two types of static features and dynamic features according to data attributes, the static features comprise land property, building density, land area and affiliated administrative division in the grid units, and the dynamic features comprise population flow amount, public facility operation state, building occupancy and use proportion of the grid units; S2, constructing load prediction logic based on quantitative association relation, and calling collected historical load data to check and correct the prediction logic to generate a dynamic load curve containing time and space dimensions; Step 3, extracting space utilization time sequence constraint, functional linkage constraint and environment sensitivity time sequence constraint in urban planning, converting three types of constraint conditions into multidimensional feature vectors, and encoding core parameters of candidate substations and power channels to form facility feature vectors; the step S3 includes: step S3-1, systematically extracting three types of constraint conditions from related data of urban planning, and converting the constraint conditions into multidimensional feature vectors: The space use time sequence constraint feature vector C a comprises specific time node features of land parcel planning purpose change, development intensity index features of different stages and construction progress features of infrastructure engineering, and is supplemented to n dimensions through feature interpolation; The functional linkage constraint feature vector comprises C b , namely power supply reliability difference requirement features, load peak-valley complementary relation features and power supply capacity linkage proportion features of public facilities and surrounding plots between different functional areas, and is supplemented to n dimensions through feature interpolation; The environment-sensitive time sequence constraint characteristic vector C c comprises a construction prohibition period characteristic of an electric power facility, an operation load limiting characteristic of a transformer substation in a noise sensitive area, a construction period adjustment coefficient characteristic in a special environment-friendly period, a construction dust control standard characteristic of an electric power channel, a waterproof and anticorrosion grade requirement characteristic of cable laying, and the main component analysis is used for reducing the dimension to n dimension; S3-2, encoding core parameters of candidate transformer substation sites and power channels to form a facility feature vector F, wherein the core parameters comprise transformer substation capacity, construction period, power supply radius, upper limit of load factor, power channel laying mode, construction period, corridor width and wire cross section; Step S4, setting a space function mutation scene, a constraint condition change scene and a multi-fault linkage scene, constructing a power grid adaptability evaluation model, simulating the load transfer capacity, the power facility adaptability and the construction emergency capacity of the power grid under each scene, outputting an adaptation standard rate, and reversely optimizing a scheme which does not reach a preset adaptation standard; Step S5, summarizing dynamic adjustment rules forming a planning scheme based on a dynamic load curve, a screening scheme and an optimization result, and generating a planning scheme comprising a dynamic load prediction report, a constraint and electric power facility matching scheme, an adaptation verification and optimization report and the dynamic adjustment rules; The step S5 includes: Step S5-1, summarizing the dynamic load curve generated in the step S2, storing the dynamic load curve according to the classification of power supply grid units and time dimension, establishing a load prediction database which comprises the load peak value/valley value time sequence distribution of each grid and the association mapping of key space characteristic events and load fluctuation, sorting the candidate scheme set screened in the step S3 and the scheme data optimized in the step S4, establishing a facility configuration Fang Anzi library, and storing the parameter configuration, constraint matching degree score and scene adaptation capability standard record of a transformer substation and a power channel; step S5-2, constructing a dynamic adjustment rule system, which comprises the following steps: The load response rule is that a stepwise adjustment threshold value of the power grid facility capacity is formulated based on the fluctuation characteristic of the dynamic load curve; a constraint adaptation rule is that a constraint priority adjustment mechanism is established according to the attention weight distribution of different regional planning scenes; And (3) setting up a standardized emergency adjustment flow according to the scene handling rules, namely aiming at the three types of scenes set in the step S4.
  2. 2. The grid planning analysis method based on the power distribution network of claim 1, wherein the step S1 comprises the following steps: s1-1, collecting power consumption load data of each grid unit in a continuous period by taking a minimum power supply grid unit in a planning area as space granularity, dividing a load period into a basic load period, a flat load period and a peak load period according to the period characteristics of the power consumption load, respectively calculating and marking a load mean value and a load fluctuation amplitude in the period for each load period, and simultaneously setting a load abnormal threshold value, namely judging a load value exceeding a reference preset proportion range as an abnormal load by taking the load mean value of each period as a reference; Step S1-2, constructing a tracing model of load abnormality and space feature based on normalized space feature data and load data, wherein normalized values of static features and dynamic features are selected from the normalized space feature data to serve as input variables of the model, a Pearson correlation analysis method is adopted to calculate correlation coefficients of the input variables and the load abnormality, the correlation coefficients are used for representing the linear correlation degree between single space feature variables and the load abnormality, a correlation coefficient screening threshold is set, input variables with the absolute value of the correlation coefficients being larger than or equal to the screening threshold are reserved, weak correlation variables with the absolute value of the correlation coefficients being smaller than the screening threshold are deleted, finally, a feature correlation layer and an output layer of the model are constructed by adopting a decision tree algorithm, the feature correlation layer takes the reserved input variables as decision nodes, the priority of the nodes is determined according to the influence weight of the variables on the load abnormality, the output layer takes the correlation type of the load abnormality as an output result, corresponding logic of an input variable combination and a load abnormality correlation type is established, and finally, a tracing model structure capable of tracing the load abnormality induction is formed through the space feature.
  3. 3. The grid planning analysis method based on the power distribution network of claim 2, wherein the step S1 further comprises the following steps: step S1-3, inputting the abnormal load data marked in the step S1-1 into a traceability model, carrying out traceability analysis on each load abnormal fluctuation through layer-by-layer matching of decision nodes of a traceability model feature association layer, and positioning specific space feature factors which cause the abnormal fluctuation: For a single space characteristic factor, adopting unitary linear regression analysis to construct a quantitative relation between the normalized value variation of the characteristic factor and the load fluctuation amplitude, wherein the specific formula is delta L=k multiplied by delta X+b, delta L represents the load fluctuation amplitude, delta X represents the normalized value variation of the single space characteristic factor, k represents a regression coefficient, k represents the influence intensity of the space characteristic factor on the load fluctuation, b represents a regression constant term, represents the error of a single factor regression model, and k and b are calculated by fitting historical abnormal event data through a least square method; For multi-space characteristic factor combination, adopting multiple linear regression analysis to construct a quantitative relation between the normalized value variation of the multi-characteristic factor and the load fluctuation amplitude, wherein a specific formula is delta L=sigma (ki×delta Xi) +bn, delta Xi represents the normalized value variation of the ith space characteristic factor, ki represents the regression coefficient of the ith space characteristic factor, bn is a multiple linear regression constant item representing the error of a multi-factor regression model, ki and bn are fitted through a least square method to obtain historical abnormal event data caused by the multi-characteristic combination, finally, historical abnormal data with preset proportion is selected to serve as a model verification set, the spatial characteristic data in the verification set is substituted into the quantitative relation, the deviation between a load fluctuation predicted value and an actual load abnormal value is calculated, a deviation judging threshold is set, the current quantitative relation is determined to be effective if the calculated deviation is smaller than or equal to a deviation judging threshold, if the calculated deviation is larger than the judging threshold, S1-2 is returned, the screening threshold or decision tree node priority of the model input variable is readjusted, and the model is constructed again until the calculated and the quantitative relation between the calculated deviation and the urban fluctuation demand and the load fluctuation demand is met by the electric fluctuation threshold.
  4. 4. The method for grid planning analysis based on power distribution network of claim 3, wherein said step S2 comprises: s2-1, constructing an initial load prediction logic frame based on the quantitative association relation between the spatial feature change and the load fluctuation determined in the step S1, taking planning adjustment values of the future urban spatial feature of a target area and trend prediction values of the dynamic feature as inputs, calculating corresponding load fluctuation amounts through the spatial feature change amounts, and then combining with the historical synchronous base load data to generate a preliminary load prediction curve; S2-2, calling historical load data acquired in the step S1 to verify an initial prediction logic, selecting historical time period data with similar spatial characteristic change trend as a verification sample, inputting the spatial characteristic data of the time period into the prediction logic, generating a load prediction verification value, comparing the load prediction verification value with a contemporaneous actual load value, and calculating a prediction deviation rate; And S2-3, integrating space characteristic dynamic change data in a future planning period of a target area based on the corrected prediction logic, generating a dynamic load curve comprising time and space dimensions, outputting the dynamic load curve according to minimum power supply grid units, enabling each grid unit to correspond to a load curve, marking load peaks, valleys and fluctuation intervals of each time node in the curve, and synchronously associating key space characteristic events causing load change.
  5. 5. The grid planning analysis method based on the power distribution network of claim 4, wherein the step S3 further comprises: Step S3-3, introducing an attention mechanism to calculate dynamic matching degree, and classifying to calculate constraint matching degree: Space time sequence matching degree M a =Similarity space (C a , F), time sequence alignment analysis and quantification suitability calculation are carried out on time nodes, development intensity, construction progress characteristics contained in space use time sequence constraint characteristic vectors and corresponding construction period, power supply capacity and construction progress parameters in facility characteristic vectors, a multi-dimensional matching result is integrated into a single numerical value through a preset logic rule, and a range [0,1] is output; the functional linkage matching degree M b =Similarity func (C b , F) establishes a quantization scoring standard with the reliability requirement, the complementary relation and the capacity linkage proportion characteristic in the functional linkage constraint characteristic vector and the corresponding load rate and capacity parameter in the facility characteristic vector, distributes weights according to the importance of each characteristic, accumulates the scores and outputs the range [0,1]; environmental sensitivity matching degree M c =Similarity env (C c , F), namely carrying out compliance verification on forbidden period, load limit and protection level threshold requirements in the environmental sensitivity time sequence constraint feature vector and corresponding construction period, operation parameters and protection standards in the facility feature vector, converting the construction period, the operation parameters and the protection standards into quantized scores according to the standard reaching degree, and outputting a range [0,1]; S3-4, generating attention scores e a ,e b ,e c corresponding to space, function and environment constraints through a linear weighting model based on regional planning scene characteristics including regional function positioning, load density and environment sensitivity level, wherein the linear weighting model takes a standardized quantitative value of the regional planning scene characteristics as input, performs weighted calculation through a preset scene adaptation coefficient, and outputs three types of constraint attention scores; weight normalization of the attention score: ; ; ; Wherein, alpha a +α b +α c =1, which respectively represent the relative importance of space, function and environment constraint in the current planning scene; calculating the comprehensive matching degree M, wherein M=alpha a ×M a +α b ×M b +α c ×M c ; Wherein, M is E [0,1], the higher the numerical value is, the better the constraint satisfaction degree of the candidate transformer station address and the power channel scheme is; And S3-5, taking the dynamic matching degree M as constraint satisfaction degree scores of candidate substation sites and corresponding power channels, sorting all candidate schemes based on the scores, determining a preset score threshold value by combining the planning level of a target area and the dynamic load curve generated in the step S2, setting a higher threshold value for a core area with high load density, properly reducing the threshold value for an area with low load density, and finally screening schemes with matching degree scores not lower than the preset score threshold value to form a candidate scheme set meeting the requirements.
  6. 6. The grid planning analysis method based on the power distribution network of claim 1, wherein the step S4 comprises the following steps: s4-1, constructing a multi-scene simulation system, and setting three types of key test scenes: Simulating sudden adjustment of land parcel planning application in a target area, generating space characteristic parameters after land parcel property change, wherein the space characteristic parameters comprise a load density sudden change coefficient and a power utilization type conversion rate, and calculating a load demand jump value in the scene based on a quantitative association relation determined in the step S1; Dynamic adjustment of simulated planning constraint, including development progress advance and delay in space time sequence constraint, reliability grade improvement in functional linkage constraint and protection standard upgrading in environment sensitive constraint, so as to form a constraint parameter adjustment matrix; Simulating a composite fault state of a power grid facility, and generating fault point distribution and influence range data; Step S4-2, constructing a power grid adaptability evaluation model, wherein the power grid adaptability evaluation model comprises three core evaluation dimensions: Calculating the maximum capacity of transferable load in the existing power grid topology based on a scene simulated load jump value or a fault influence range, and outputting a load transfer completion rate = actual load transfer/total load transfer required; the power facility adaptation capability assessment, namely recalculating the matching degree of the facility feature vector and the adjusted constraint vector aiming at the constraint condition changing scene, and outputting the adaptation standard rate = the adjusted matching degree/the original matching degree; The construction emergency capability assessment comprises the steps of simulating the scheduling efficiency of emergency repair resources according to a multi-fault linkage scene, calculating the ratio of fault recovery time to preset emergency repair time limit, and outputting an emergency response standard rate=max (0, 1- (actual recovery time length-standard recovery time length)/standard recovery time length); S4-3, inputting the candidate schemes screened in the step S3 into an evaluation model, respectively performing simulation under three types of scenes, and outputting the standard rate of each capability; The method comprises the steps of starting reverse optimization for a scheme which does not reach the standard, adjusting the power supply radius of a transformer substation or newly-added connecting lines according to the scheme with insufficient load transfer capability, recalculating topology margin, optimizing the capacity parameters of the transformer substation or the laying mode of a power channel according to the scheme with insufficient facility adaptation capability, improving the matching degree of constraint after adjustment, adding a standby power supply configuration point or optimizing the emergency repair resource layout according to the scheme with insufficient emergency capability, shortening the fault recovery time, and repeating simulation evaluation and optimization processes until the scheme meets comprehensive adaptation standards to form an optimized scheme set subjected to scene verification.
  7. 7. The grid planning analysis method based on the power distribution network of claim 1, wherein the step S5 comprises the following steps: S5-3, displaying load distribution and change trend in a planning period through a space-time thermodynamic diagram, marking a load sensitive area and a time period which need to be focused on, generating a dynamic load prediction report, simultaneously displaying constraint satisfaction degree sequencing of candidate schemes, parameter configuration of an optimal scheme and an attached attention weight distribution thermodynamic diagram to form a constraint and electric power facility matching scheme; And S5-4, setting a data updating period, and automatically starting iterative computation of the steps S2 to S4 when new data trigger any dynamic adjustment rule, so that the planning scheme continuously adapts to the development change of the target area.
  8. 8. The grid planning analysis system based on the power distribution network is applied to the grid planning analysis method based on the power distribution network, which is characterized by comprising a traceability modeling module, a dynamic load prediction verification module, a constraint facility scheme matching module, a multi-scenario scheme optimization evaluation module and a planning scheme generation module; The trace cause modeling module collects power load data, urban space feature data and historical power grid operation data which take a minimum power grid unit as a reference in a planning area, establishes a trace cause model of load abnormality and space feature, carries out trace cause analysis on load abnormality fluctuation based on the trace cause model, and determines a quantitative association relation between space feature change and load fluctuation; The dynamic load prediction verification module constructs load prediction logic based on quantitative association relation, and invokes collected historical load data to verify and correct the prediction logic to generate a dynamic load curve containing time and space dimensions; the constraint facility scheme matching module extracts space use time sequence constraint, functional linkage constraint and environment sensitive time sequence constraint in urban planning, converts three types of constraint conditions into multidimensional feature vectors, and encodes core parameters of candidate transformer station addresses and power channels to form facility feature vectors; The multi-scenario scheme optimization evaluation module sets a space function mutation scenario, a constraint condition change scenario and a multi-fault linkage scenario, builds a power grid adaptability evaluation model, simulates the load transfer capacity, the power facility adaptation capacity and the construction emergency capacity of a power grid in each scenario, outputs an adaptation standard rate, and reversely optimizes a scheme which does not reach a preset adaptation standard; The planning scheme generating module summarizes dynamic adjustment rules forming a planning scheme based on a dynamic load curve, a screening scheme and an optimization result, and generates a planning scheme comprising a dynamic load prediction report, a constraint and electric power facility matching scheme, an adaptation verification and optimization report and the dynamic adjustment rules.

Description

Grid planning analysis system and method based on power distribution network Technical Field The invention relates to the technical field of power distribution network planning, in particular to a grid planning analysis system and method based on a power distribution network. Background The power distribution network is an important component of urban core infrastructure, the planning and construction level of the power distribution network directly influences the safety and economy of power supply, and along with continuous acceleration of urban process and gradual transformation of energy structures, constraint conditions and target dimensions of power distribution network planning need to be dealt with are increased. In order to improve the refinement and operability of planning, the gridding planning method is widely applied gradually, and the method deconstructs a power supply area into grid cells with relatively clear structures and moderate scales by a space division means, so that a foundation is provided for optimizing configuration and standardized construction of power grid resources. However, under the prior art condition, the implementation of grid planning still depends on the experience judgment of planning personnel, the cooperation between the grid planning and urban space planning is mainly qualitative analysis and manual coordination, and a systematic quantitative evaluation and automatic auxiliary means is lacked, so that the current situation causes that the grid planning is difficult to accurately dock key information in urban development, such as land functions, development intensity, construction time sequence and the like, and potential conflict risks among transformer substation sites, power channels and urban planning cannot be effectively identified in the early stage, therefore, part of grid planning schemes face adjustment in the landing stage, not only influence the construction efficiency and investment benefit, but also can restrict the long-term power supply capacity of regional grids and the acceptance level of renewable energy sources. Disclosure of Invention The invention aims to provide a grid planning analysis system and method based on a power distribution network, which are used for solving the problems in the background technology. In order to solve the technical problems, the invention provides a grid planning analysis method based on a power distribution network, which comprises the following steps: Step S1, collecting power load data, urban space feature data and historical power grid operation data which take a minimum power grid unit as a reference in a planning area, and establishing a tracing model of load abnormality and space feature; S2, constructing load prediction logic based on quantitative association relation, and calling collected historical load data to check and correct the prediction logic to generate a dynamic load curve containing time and space dimensions; Step S3, extracting space utilization time sequence constraint, functional linkage constraint and environment sensitivity time sequence constraint in urban planning, converting three types of constraint conditions into multidimensional feature vectors, and encoding core parameters of candidate transformer station addresses and power channels to form facility feature vectors; Step S4, setting a space function mutation scene, a constraint condition change scene and a multi-fault linkage scene, constructing a power grid adaptability evaluation model, simulating the load transfer capacity, the power facility adaptability and the construction emergency capacity of the power grid under each scene, outputting an adaptation standard rate, and reversely optimizing a scheme which does not reach a preset adaptation standard; And S5, summarizing dynamic adjustment rules forming a planning scheme based on the dynamic load curve, the screening scheme and the optimization result, and generating the planning scheme comprising a dynamic load prediction report, a constraint and electric power facility matching scheme, an adaptation verification and optimization report and the dynamic adjustment rules. Further, step S1 includes: The method comprises the steps of S1-1, taking a minimum power supply grid unit in a planning area as space granularity, collecting power consumption load data of each grid unit in a continuous period, dividing a load period into a basic load period, a flat load period and a peak load period according to the period characteristics of the power consumption load, respectively calculating and marking a load mean value and a load fluctuation amplitude in the period for each load period, setting a load abnormal threshold value, judging a load value exceeding a preset proportion range of the reference as abnormal load based on the load mean value of each period, synchronously collecting urban space characteristic data, classifying the urban space characteristic data into two types of static c