点击次数：

发表时间：2022-10-10

发表刊物：Journal of Engineering for Thermal Energy and Power

卷号：36

期号：3

页面范围：26-34

ISSN号：1001-2060

关键字："Matlab/Simulink; Matlab/Simulink; compressor characteristic curve; prediction methods; dynamic performance simulation"

CN号：23-1176/TK

摘要：The study of component refinement modeling method has always been a hot topic in the field of gas turbine dynamic performance simulation. Here with the compressor core components of a classic single shaft gas turbine as object, based on modular modeling idea using Matlab/Simulink platform a system simulation platform is set up. The least square method, cubic spline interpolation method and BP neural network method are embedded into the platform for this prediction application study. The results show that in the compressor performance prediction,three methods can effectively predict the component performance, but the predictive results of the BP neural network and the cubic spline interpolation method are superior to the least square method. for the performance prediction of the whole machine,the simulation results of the least square method deviate from the preset value,while the simulation results of the BP neural network method and cubic spline interpolation method have high accuracy. For the timeliness, the time cost of BP neural network method is higher than the other two methods.

备注：新增回溯数据