固体火箭发动机稳态燃速二维模型参数最优化辨识

1引言发动机参数辨识常用方法是最小二乘法1],即将参数估计问题转化为多元函数求极值问题,通过迭代选择待估参数值,使燃烧室压强计算值和实验值的残差平方和最小。采用最小二乘法进行参数辨识存在局限性2],需要通过台劳级数展开式近似模型替代真实数学模型计算残差平方和对待

Optimization identification for steady burning rate 2-D model of solid rocket motor

To predict the internal ballistic property of the solid rocket motor more accurately, the steady burning rate 2-D model of the composite propellant was set up by taking into account the affect of the hump effect and erosive burning, the burning rate model parameter identification can be converted into nonlinear programming problem which can been solved by hybrid optimization method composed of Powell optimization method and genetic algorithms. The solution efficiency of the optimization problem can been improved by the distributed parallel performance of the genetic algorithms. The global optimum solution of the steady burning rate 2-D model parameter was gained. The optimization identification process and result are more robust with the model parameter initial value. Results show that the effect of the hump and erosive burning must been taking into account for experimental motor because of the high volumetric loading fraction, high length to diameter ratio, and low burning rate composite propellant. It also shows that steady burning rate 2-D model is feasible.