为了建立燃机整体的性能预测模型，采用BP神经网络和基于思维进化算法的神经网络（MEA-BP算法）对压气机特性曲线进行预测和分析，得到了各部件的特性和基于热力学原理的数学表达式，搭建了燃机的整体仿真模型，建立了GE9F 型重型燃气轮机的性能监测及耗差分析模型，分析不同工况条件下机组的性能参数和经济性指标。对机组的运行数据进行了多元线性回归分析，可以实现压气机和燃机的性能预测，为运行人员的运行调整提供参考。

In order to establish the gas turbine performance prediction model,BP neural network and MEABP algorithm are used to predict and analyze the compressor characteristic curve.The characteristics of each component and the mathematical expression based on the principle of thermodynamics are obtained,the overall simulation model of gas turbine is built,the performance monitoring and consumption difference analysis model of GE9F type heavy gas turbine are established,and the performance parameters and economic indexes of several groups under different working conditions are analyzed and studied.Multiple linear regression analysis on the operating data of the unit is carried out,and the performance of the compressor and gas turbine can be predicted.

TK47

上海市科委发电过程智能管控工程技术研究中心基金资助项目（14DZ2251100）；上海市“科技创新行动计划”地方院校能力建设专项项目（19020500700）