论文部分内容阅读
为更好地研究多结构参数耦合变化下减压器PPR(pressure reducing regulator)的稳定性,使用BFGS(Broyden-Fletcher-Goldfarb-Shanno)拟牛顿法替换梯度下降法,实现了基于Wolfe条件的一维线搜索变步长BP(back propagation)算法.结果表明:改进的BP算法使迭代次数减少了1~2个数量级,且易于收敛到最小点.该算法用于逆向卸荷膜片式减压器时,能适应2~3个结构参数的耦合,可预测大于106个数据点的数据集.多结构参数同时变化时,更容易找到使得减压器稳定的结构参数组合.更重要的是这些结构参数同时变化时减压器的稳定性比仅其中一个参数变化时更好.
In order to better study the stability of pressure reducing regulator (PPR) under multi-structural parameters coupling change, a gradient descent method based on quasi-Newton method of BFGS (Broyden-Fletcher-Goldfarb-Shanno) The results show that the improved BP algorithm can reduce the number of iterations by one or two orders of magnitude and easily converge to the minimum point.The algorithm is applied to reverse unload diaphragm decompression Can adapt to the coupling of 2 to 3 structural parameters and predict the data set of more than 106 data points.When the multiple structural parameters change at the same time, it is easier to find the structural parameters that make the pressure reducer stable. More importantly, these The stability of pressure reducer is better than the change of only one of the parameters when the structural parameters change simultaneously.