Power system court forecasting method

Abstract

The present invention discloses a power system court forecasting method, which comprises: optimizing the back-propagation BP neural network connection weight and threshold by an adaptive particle swarm optimization algorithm with dynamically changing inertia weights until the weight and threshold meet the fitness requirements; based on the optimized weight and threshold, optimally training the BP neural network parameters to obtain an optimal combination of parameters; and performing the court forecasting based on the optimal combination of parameters. The embodiments of the present invention have a faster convergence speed and a higher convergence accuracy. The improved particle swarm optimization algorithm replaces the initial optimization, which effectively improves the speed and precision of the BP neural network optimization, thereby improving the power system station court forecasting accuracy.

Description

technical field [0001] The invention relates to the technical field of power systems, in particular to a load forecasting method for a power system station area. Background technique [0002] In the power system, short-term load forecasting focused on substations and busbars in the early days. With the increasing importance of distribution networks, the opening of the power market, and the development of various technologies in the power industry, the demand for short-term load forecasting in the power grid has emerged in many ways. With the trend of portability and multi-application scenarios, short-term load forecasting is extended to the low-voltage level of the distribution network, and the station area is regarded as the basic forecasting unit, which can meet various requirements of each department of the power grid for load forecasting, expand the use of smart distribution networks and improve It has a very important engineering practical value and commercial value. ...

AI Technical Summary

Problems solved by technology

[0008] (2) It is difficult to respond to climate change and human disturbance, which makes the prediction model have low stability, and the prediction accuracy drops sharply when disturbed;

[0009] (3) There is no reasonable and agreed way to predict loads with different types of days, and the prediction accuracy of holidays is particularly poor

[0015] In the existing power system, there are more random factors affecting the load in the low-voltage station area than in the ordinary distribution network, and the nonlinearity of the load is very strong

Therefore, it is necessary to conduct in-depth research on the load forecasting model applicable to the station area, but the existing load forecasting implementation methods mentioned above cannot well meet the load forecasting requirements of the station area

Images