Abstract:

With large-scale wind and solar energy power units put into operation, it brings great influence on the stability and safety of the power network due to the high fluctuation and randomness features of the wind and solar resources. In order to absorb more new energy power, regional load dispatch center puts up very high requirement for the operation flexibility, load fast-following and deep peak load regulation ability. Therefore, it is very significant to adopt advanced control strategies to optimize the coordinated control system (CCS) to improve its control quality. In this work, the load prediction neural network model considering "condensate throttling" and the main steam pressure prediction model were developed by taking a 1000 MW ultra-supercritical power unit as the object investigated. Based on the established models, a classic optimization theory-improved simplex method was selected as the CCS optimization algorithm. An intelligent coordination optimization scheme with condensate throttling was designed, and the real-time optimization program was developed with MATLAB software. Detailed simulation tests were carried out in the full-scope simulator of the 1000 MW power unit. Simulation results indicate that the proposed scheme can effectively improve the coordinated control quality with faster load response, less main steam pressure fluctuation during dynamic load-changing process.

Key words: ultra-supercritical power unit, coordinated control system, neural network modelling, simplex method, predictive optimization control, condensate throttling