Steam load prediction method

Abstract

The invention relates to a steam load prediction method. The steam load prediction method includes the steps: S1, collecting historical steam load data to form a steam load time sequence, conducting noise reduction on the steam load time sequence; S2, obtaining the embedded dimension and the delay time of the time sequence by means of the chaos theory, conducting phase-space reconstruction on the time sequence on which noise reduction is conducted, obtaining the sample data of a phase spatial domain to the sample data of the phase spatial domain before the reconstruction, obtaining a prediction model by means of a least square support vector machine; optimizing training parameters of the least squares support vector machine through an SA_PSO algorithm; predicting the steam loading value of the future 24 hours by means of the obtained training model, and obtaining and evaluating the obtained data. By means of the steam load prediction method, a satisfying result can be obtained under the condition that the external factors such as the data type and weather are not considered, and thus the steam load prediction method is simple and effective.

Description

technical field [0001] The invention belongs to the technical field of short-term load forecasting, and relates to a steam load forecasting method, in particular to a steam load forecasting method based on phase space reconstruction and least square support vector machine. Background technique [0002] The prediction of steam load is of great significance to the economical and safe operation of thermal power plants. In recent years, the state has actively advocated energy conservation and emission reduction, so higher requirements have been placed on load forecasting accuracy and safe production. The steam heating system is a very complex dynamical system. The heating process has the characteristics of large time delay, large inertia, nonlinearity, time variation, and uncertainty. In addition, the steam load is also related to various natural factors such as weather, and also Affected by holidays, major events, etc., these all determine that it is difficult to carry out loa...

AI Technical Summary

Problems solved by technology

However, neural network prediction also has inevitable shortcomings such as easy to fall into local minimum and difficult to determine the number of neural network nodes.

Images