Performance improvement method of power load medium-and-long-term prediction model

Abstract

The invention discloses a performance improvement method for a medium and long term prediction model of a power load, and the method comprises the following steps: S10, inputting an actual load data sequence X(0), and carrying out the accumulation once to generate an accumulated load data sequence X(1); S20, establishing a GM(1,1) gray system model; S30, establishing an IALO optimization GM(1,1) model; S40, reconstructing a fitness function in the IALO based on grey correlation analysis, and performing secondary optimization on model parameters a and b of the GM(1,1); and S50, continuously correcting values of parameters a and b of the GM(1,1) model through multiple rounds of IALO dynamic rolling processes to obtain a latest load data sequence, and finally obtaining all target prediction points according to the continuously updated load data sequence. Compared with some common swarm intelligence algorithms, the method has higher accuracy, and the prediction performance of the power load prediction model is effectively guaranteed.

Description

technical field [0001] The invention relates to the technical field of electric power system load forecasting, in particular to a method for improving the performance of a medium- and long-term electric load forecasting model. Background technique [0002] Accurate medium and long-term load forecasting is the premise of making a reasonable power system development plan, and it is also an important guarantee for the safe and economical operation of the power grid. The development of my country's medium and long-term power load has both the certainty of year-by-year growth and the uncertainty of random fluctuations, which can be regarded as a typical gray system. Therefore, the gray model GM(1,1) proposed by Professor Deng Julong and its optimization The model has been widely used in medium and long-term load forecasting. [0003] However, due to the influence of economic, climate and other factors, the medium and long-term power load often presents a certain degree of mutatio...

AI Technical Summary

Problems solved by technology

The principle of the ALO algorithm is to imitate the antlion hunting mechanism in nature. Compared with other optimization algorithms, it has the characteristics of faster convergence speed and higher solution efficiency. However, the ALO algorithm has unbalanced exploration and development in some complex optimization problems. The shortcomings of the ability, so in the actual solution process often fall into the problem of local optimum

In addition, the ALO algorithm uses the minimum average relative prediction error as the goal to construct a fitness function to screen the optimal model parameter values, which often leads to the problem of high prediction accuracy and low fitting accuracy of the model, and the one-sidedness and limitations of the prediction results will also make the The reliability of gray model predictions is further reduced

Images