Time-phased refined short-term load prediction method based on BOA-SVR and fuzzy clustering

Abstract

The invention discloses a time-phased refined short-term load prediction method based on BOA-SVR and fuzzy clustering. The method comprises the following steps: S1, analyzing factors affecting the load change according to the historical load data and characteristics of a region, and proposing a time-phased comprehensive prediction method through considering the real-time influence factors; s2, clustering the historical data by using a fuzzy C-means clustering method to obtain different types, and evaluating the clustering effect of the historical data; s3, improving a bat algorithm by using adynamic adaptive weight method and a Cauchy distribution inverse accumulation partial function; s4, optimizing the parameters of the SVR by using an improved bat algorithm, searching the optimal parameters to establish an SVR time-phased refinement model, and taking the average load value of each time period to be predicted as the output of the SVR to obtain a prediction result; according to the method, the prediction precision can be effectively improved, the prediction precision for different types of days can reach 96% or above, the scale of the prediction model is effectively reduced through the improved bat algorithm, and the performance of the prediction model is obviously improved.

Description

technical field [0001] The invention relates to the technical field of load forecasting of electric power systems, in particular to a period-by-period refined short-term load forecasting method based on BOA-SVR and fuzzy clustering. Background technique [0002] Scientific load forecasting plays an important role and significance for many departments of the power system. Nowadays, smart grid technology is developing rapidly, and the complex grid scale increases the complexity of power data, which affects the accuracy and real-time performance of load forecasting. There are higher requirements, so this has become the ultimate goal of people's continuous in-depth research and development of power system load forecasting theory. [0003] At present, scholars at home and abroad have conducted in-depth research on the optimal forecasting method of load forecasting, and proposed more optimized load forecasting methods: one is the traditional statistical method, including linear re...

AI Technical Summary

Problems solved by technology

Among these methods, support vector regression (SVR) has the advantages of strong learning ability, good adaptability, and is suitable for small sample data learning, but its prediction effect is greatly affected by model parameters. If the parameters are not selected properly, the prediction accuracy of SVR will be low. low, the prediction efficiency also decreases

Bat Algorithm (BA) is an iterative search optimization swarm intelligence algorithm with the characteristics of "generation + inspection", which has strong local search ability and good global optimization ability, and is suitable for parameter optimization of predictive models. However, it also has Easy to fall into local optimal solution, low convergence accuracy, unbalanced algorithm convergence speed and other shortcomings

On the other hand, when studying the factors affecting load changes, only the similarity of historical data is considered and analyzed, and when the influence of meteorological factors is considered, the similarity of meteorological factors to load is used to deal with daily characteristic meteorological factors, but Only considering the similarity of meteorological factors can not fully analyze the load characteristics, and because the factors affecting the load change are complex, the influence of different factors on the load is nonlinear, which will lead to the reduction of the accuracy of the prediction method or even failure

Images