论文部分内容阅读
对航空发动机压气机原始二维等转速线的数据点进行连分式扩充,通过两次网络训练,增加转速特性数据,在三维空间中进行BP(back propagation)网络模型重构.根据压气机特性数据空间分布的特点,引入压力比函数,调整计算区域,定义网络的输入输出数据,利用试探法确定隐含层维数.采用基于趋利避害原则的粒子群算法对网络的初始权值和阈值进行优化,建立了压气机压比和效率特性的整体代理模型.最后以某型发动机的低压压气机为例进行了压气机特性模型的重构.通过模型的校核与验证表明:采用这种方法建立的模型精度较高,优于普遍采用的传统二维插值方法和普通BP神经网络模型.最终建立的重构模型对于采用选配法、坐标法和部件法等以压气机通用特性曲线为基础的发动机模型的求解,可提高计算精度和迭代速度,具有一定的工程应用价值.
The data points of the original two-dimensional iso-speed line of aeroengine compressors are expanded by using fractional expansion, and the speed characteristics data are increased by two network trainings to reconstruct the BP (back propagation) network model in three-dimensional space. According to the characteristics of the compressor The data space distribution characteristics, the introduction of the pressure ratio function, adjust the calculation of the region, the definition of network input and output data, using heuristic method to determine the hidden layer dimension.Using the principle of profit and avoidance particle swarm optimization algorithm for the network initial weight and The threshold is optimized to establish the overall proxy model of the compressor pressure ratio and efficiency characteristics.Finally, the compressor characteristics model is reconstructed by the low pressure compressor of a certain type of engine.According to the verification and verification of the model, The accuracy of the proposed model is better than that of the traditional two-dimensional interpolation method and the ordinary BP neural network model.Finally, the reconfiguration model is established by using the matching method, the coordinate method, the component method and so on, Based on the engine model to solve, can improve the calculation accuracy and iteration speed, has a certain engineering value.