基于面向对象方法的压气机性能计算

Performance calculation of compressor based on object-oriented method

提出一套预测压气机未知特性的方法,并基于面向对象思想采用变比热容计算方法进行压气机性能计算的分析和编程.结合粒子群优化(PSO)的全局寻优能力和反向传播(BP)神经网络的局部寻优能力提出基于PSO的BP神经网络(PSO-BP神经网络)预测压气机特性,分析了其预测误差和拟合误差:拟合误差基本都小于0.5%,预测误差基本都小于0.8%.其拟合精度和预测精度满足要求.采用变比热容计算方法来计算压气机性能,并采用面向对象方法编写了压气机性能计算程序.对几个压气机变工况点进行验证,各输出参数的最大误差为1.12%.因此,特性预测方法和性能计算的数学模型适用于压气机性能计算,这套方法同样适用于燃气轮机性能计算.

A characteristic prediction method was proposed and variable specific heat calculation was applied to the performance analysis and programming of compressor based on object-oriented theory. Also, a method named particle swarm optimization （PSO） based on back propagation （BP） neural network was presented by combining the global optimization a- bility of the PSO with the local optimization ability of the BP neural network, and the pre diction error and fitting error were analyzed. The fitting error is mostly within 0.5 % while the highest prediction error is within 0.8 % ; and both the fitting accuracy and prediction accuracy could meet the requirements. Variable specific heat calculation method was applied to the compressor performance calculation, and object-oriented method was used to build the compressor performance computing program. Compared with several working condition points of the compressor, the output parameter errors are less than 1.12% . Therefore, the characteristic prediction method and performance mathematical model are suitable for compressor performance calculation, and the compressor calculation procedure is also suitable for gas turbine performance calculation.