Power load prediction algorithm based on genetic algorithm and support vector machine

A technology of support vector machine and electric load, applied in the fields of genetic law, prediction, calculation, etc., can solve the problem of inaccurate prediction, and achieve the effect of accelerating convergence, enriching diversity, and improving accuracy.

CN110598900APending Publication Date: 2019-12-20陕西中石能电力设计集团有限公司
0 Cites 0 Cited by

Patent Information

Authority / Receiving Office
CN · China
Current Assignee / Owner
Publication Date
2019-12-20

Smart Images

  • Figure 1
    Figure 1
  • Figure 2
    Figure 2
  • Figure 3
    Figure 3
Patent Text Reader

Abstract

The invention discloses a power load prediction algorithm based on a genetic algorithm and a support vector machine. The algorithm comprises the following steps of (1) collecting the power load data,carrying out the genetic coding to form an initial population, and sequentially calculating the individual fitness, a selection operator, a crossover operator and a mutation operator of the initial population; (2) setting the maximum genetic algebra i of a fusion layer to be 100, and inputting the populations obtained by respectively calculating the individual fitness, the selection operator, thecrossover operator and the mutation operator into the fusion layer; and (3) inputting the population of which the fusion layer meets the iteration requirement into the support vector machine, and predicting the power load through training and learning of the support vector machine. According to the algorithm, the collected initial data is processed through the genetic algorithm, so that the optimal solution of the data can be realized; then the data is inputted into the support vector machine, so that the data inputted into the support vector machine is corrected and regularized, and the precision of the weight in the support vector machine, the training efficiency, the network performance and the approximation capability of the network are improved.
Need to check novelty before this filing date? Find Prior Art

Description

technical field

[0001] The invention relates to the technical field of human health prediction, in particular to an electric load prediction algorithm based on a genetic algorithm and a support vector machine. Background technique

[0002] Accurate short-term power load forecasting is helpful for fault diagnosis and reduction of power generation cost in industrial production. With the steady advancement of Made in China 2025 and the continuous development of urbanization, the demand for electricity in factory production and people's life is increasing, and it becomes more important to ensure the coordination of the relationship between power supply and consumption. At present, there are mainly traditional forecasting methods and intelligent forecasting methods for short-term power load forecasting at home and abroad. With the rapid development of my country's artificial intelligence technology field, traditional forecasting methods have been gradually banned. Most of the e...

Examples

Embodiment 1

[0018] Such as figure 1 As shown, the power load forecasting algorithm based on genetic algorithm and support vector machine provided by this embodiment includes the following steps:

[0019] (1) Collect power load data and carry out genetic coding to form an initialization population, and calculate the individual fitness, selection operator, crossover operator and mutation operator in turn for the initialization population; the power load data is the power load value, power load The value is the historical collection data, and normalizing the collected large amount of power load data can improve the accuracy of the power load data. The normalized data is genetically encoded according to the real number coding genetic algorithm, and the Real number increments form gene strings, which can improve the quality and precision of offspring, especially in the degree of conformity between the individual transfer direction and the optimization object in the crossover operator and mutat...