CN-121981456-A - Multi-objective optimization method and device for transformer substation engineering

CN121981456ACN 121981456 ACN121981456 ACN 121981456ACN-121981456-A

Abstract

The invention provides a multi-objective optimization method and device for transformer substation engineering, which comprises the steps of acquiring a working procedure set, a working procedure front-and-back logic relation, optional construction modes of each working procedure and corresponding parameter information, introducing the logic relation into constraint, taking a construction mode selection result as a discrete decision variable, constructing a multi-objective optimization model comprising a construction period, a cost present value, quality and a carbon emission target, adopting an improved group intelligent algorithm to carry out iterative solution, generating an initial population by layered sampling, establishing an external file, carrying out iterative calculation of an objective function and feasibility judgment, carrying out non-dominant sorting and crowding degree after merging, updating the external file, selecting a leading body, generating candidate positions based on non-linear convergence control and trapping, random search and spiral bubble network updating, discretizing and restoring constraint after merging, and outputting an external file non-dominant solution set. The technical scheme of the invention realizes multi-target scheme optimization, and improves scheme generation efficiency and evaluation quality.

Inventors

ZHAO YONGFEI
GUO ZIJIAN
YANG ZHEN
HU BOWEN
ZHAO XIN
WU XIN
ZHANG ZHIQIANG
LIU YANGTAO
QI ZHENYU
LI RAN
RUAN YUHUI

Assignees

国网北京市电力公司

Dates

Publication Date: 20260505
Application Date: 20260115

Claims (10)

1. A multi-objective optimization method for substation engineering, the method comprising: acquiring a working procedure set, a front-to-back logic relation of working procedures, optional construction modes of each working procedure and parameter information corresponding to each construction mode of a transformer substation engineering project group; the method comprises the steps of taking a front-back logic relation of the working procedure as constraint input, taking the optional construction mode and parameter information thereof as parameter input, and taking a construction mode selection result of each working procedure as a discrete decision variable to construct the multi-objective optimization model, wherein the multi-objective optimization model at least comprises a construction period target, a cost present value target, a quality target and a carbon emission target, and setting constraint conditions at least comprising the constraint of a value range of the discrete decision variable; solving the multi-objective optimization model with an improved swarm intelligence algorithm to obtain a non-dominant solution set, the improved swarm intelligence algorithm performing the following: setting population scale and termination conditions, generating an initial population by adopting a hierarchical sampling initialization strategy, and establishing an external file for storing non-dominant solutions; In each iteration, calculating an objective function value of a candidate construction scheme corresponding to the current population and determined by the construction mode selection result, and judging feasibility according to the constraint condition; combining the candidate construction schemes meeting the constraint conditions in the current population with an external file, then executing non-dominant sorting, calculating the crowding degree, updating the external file based on the crowding degree, and selecting a leading individual from the external file, wherein the leading individual is used as a position reference individual for capturing update and updating a spiral bubble network; Generating a position update control parameter according to a convergence control strategy which varies nonlinearly along with an iterative process, generating a candidate update position by adopting at least one of three position update mechanisms including trapping update, random search update and spiral bubble network update based on the position update control parameter, and obtaining a fusion position by adopting a self-adaptive weight fusion strategy; discretizing the fusion position into a construction mode selection result and executing feasibility restoration to generate a next generation population, repeating iteration until the termination condition is met, and generating a non-dominant solution set in an external file as a multi-objective optimization result.
2. The multi-objective optimization method of substation engineering according to claim 1, wherein the hierarchical sampling initialization strategy is a Latin hypercube sampling strategy, and the Latin hypercube sampling strategy comprises performing hierarchical sampling on a value range of the discrete decision variable to obtain a sampling result, and generating the initial population based on the sampling result.
3. The substation engineering multi-objective optimization method according to claim 1, wherein the convergence control strategy that varies non-linearly with the iterative process comprises: And calculating a convergence control parameter according to the iteration times, and carrying out nonlinear update on the convergence control parameter in the iteration process, wherein the convergence control parameter is used for generating the position update control parameter.
4. The substation engineering multi-objective optimization method according to claim 1, wherein the spiral bubble network update comprises an exponential function factor and a cosine function factor, and the candidate update location is generated by performing spiral approximation calculation on the distance between the current individual and the lead individual based on the exponential function factor and the cosine function factor.
5. The substation engineering multi-objective optimization method according to claim 1, wherein when a leader is selected from the external archive, a congestion level of non-dominant solutions in the external archive is calculated, and the leader is selected from the external archive according to the congestion level.
6. The multi-objective optimization method of substation engineering according to claim 1, wherein when updating the external archive, non-dominant ranking is performed on the combined candidate individuals and a crowdedness is calculated, and when the capacity of the external archive exceeds a preset capacity, a part of non-dominant solutions are deleted from the external archive according to the crowdedness.
7. The substation engineering multi-objective optimization method according to claim 1, wherein, Mapping the fusion position into a construction mode index of each procedure; the feasibility restoration comprises the steps of carrying out back projection or replacement on the construction mode index exceeding the discrete decision variable value range, and adjusting the construction mode index violating the front-back logic constraint of the working procedure.
8. The method of claim 1, wherein calculating the cost present value targets includes calculating a cost present value based on a discount rate for direct and indirect costs of the candidate construction solution, and wherein the cost present value is one of the targets.
9. The substation engineering multi-objective optimization method according to claim 1, wherein the calculation of the carbon emission objective includes determining a resource consumption amount of a candidate construction scheme based on the parameter information, calculating a carbon emission amount corresponding to the resource consumption amount based on a carbon emission factor, and taking the carbon emission amount as one of the optimization objectives.
10. A substation engineering multi-objective optimization device, characterized in that the device comprises a computer arrangement comprising a processor and a memory, the processor having stored therein computer instructions which, when executed, implement the substation engineering multi-objective optimization method according to any one of claims 1 to 9.

Description

Multi-objective optimization method and device for transformer substation engineering Technical Field The invention belongs to the technical field of power system dispatching automation, and particularly relates to a multi-objective optimization method and device for transformer substation engineering. Background Substation engineering projects are typical complex systems with multiple constraints, multiple processes and multiple participants in cooperation, and are generally composed of multiple sub-projects and multiple procedures, wherein each sub-project often has a close logic relationship and construction sequence dependency relationship. The realization of the whole project target depends on the comprehensive achievement of each sub-target, and obvious interaction exists among targets such as construction period, cost, quality, carbon emission and the like, for example, construction period compression can lead to cost increase, carbon emission increase and the like. This conflict between multiple objectives makes construction scheme optimization a complex multi-objective equalization problem. In the prior art, construction optimization analysis is often unfolded around a small amount of indexes such as construction period, cost, quality and the like, the coverage of an index system is limited, and under the background that the green low-carbon management requirements are continuously improved, the engineering scheme is difficult to comprehensively evaluate. Meanwhile, cost accounting is usually based on explicit costs such as labor, materials, machinery, and the like, and the cost items related to the construction organization and the construction period change are not covered enough, so that the cost result is deviated from the actual execution situation. On the other hand, the engineering construction process has the characteristics of multiple procedures, complex relationship, frequent field condition change and the like, the existing model is difficult to combine the complex situations, and the problems of incomplete logic relationship description or inconsistency with reality are easy to occur, so that the accuracy and stability of scheme evaluation are affected. When the traditional intelligent optimization algorithm processes complex construction networks, nonlinear objective functions and high-dimensional constraint conditions, the problems of low calculation efficiency, weak convergence capacity, easiness in sinking into local optimization and the like still exist, and the optimization requirement of an actual construction project group under the green and high-efficiency targets is difficult to meet. Therefore, an optimization technology system which can adapt to a complex construction network structure, give consideration to the whole dimension targets of construction period/cost/quality/carbon emission and solve the multi-target conflict problem with high precision is needed to support the green development and lean management requirements of the building industry. Disclosure of Invention According to the scheme, a multi-objective optimization model comprising a working procedure set, a working procedure front-back logic relation and construction mode parameter information is constructed, a construction period objective, a cost present value objective, a quality objective and a carbon emission objective are adopted in the model, a construction mode selection result is used as a discrete decision variable, a candidate construction scheme is solved, and a non-dominant solution set is output, so that the multi-objective conflict problem among construction period, cost, quality and carbon emission under a complex construction network of the transformer substation engineering is effectively solved. The invention provides a multi-objective optimization method for transformer substation engineering, which comprises the following steps: acquiring a working procedure set, a front-to-back logic relation of working procedures, optional construction modes of each working procedure and parameter information corresponding to each construction mode of a transformer substation engineering project group; the method comprises the steps of taking a front-back logic relation of the working procedure as constraint input, taking the optional construction mode and parameter information thereof as parameter input, and taking a construction mode selection result of each working procedure as a discrete decision variable to construct the multi-objective optimization model, wherein the multi-objective optimization model at least comprises a construction period target, a cost present value target, a quality target and a carbon emission target, and setting constraint conditions at least comprising the constraint of a value range of the discrete decision variable; solving the multi-objective optimization model with an improved swarm intelligence algorithm to obtain a non-dominant solution set, the improved swarm intell