论文部分内容阅读
研究发动机部件性能参数变化规律,对于减少维修次数和推动视情维修具有重要意义。针对测量参数个数少于待估性能参数的情况,给出了一种通过构建代价函数和优化算法的参数估计方法。原代价函数只考虑当前点参数,缺少与前面点参数的联系,因此结合自组织神经网络,构造了包含以前与当前点参数的距离代价函数。并提出了一种快速的参数估计方法。由于准确的部件性能参数很难获取,并且参数趋势估计不同于单纯的点估计问题,以对应的测量参数为基础,利用信息熵方法评定部件性能参数估计效果。进一步得到距离代价函数对应的参数信息熵为0.6805,优于原代价函数的估计结果。最后通过实例验证了参数估计方法的有效性。
Research on the variation of performance parameters of engine components is of great significance to reduce the number of maintenance and promote the maintenance of the situation. Aiming at the situation that the number of measurement parameters is less than the estimated performance parameters, a parameter estimation method is proposed by constructing cost function and optimization algorithm. The original cost function only considers the current point parameter, and lacks the connection with the previous point parameter. Therefore, the self-organizing neural network is used to construct the distance cost function including the previous and the current point parameters. And a fast parameter estimation method is proposed. Because accurate parameters of part performance are difficult to obtain, and parameter trend estimation is different from simple point estimation, the information entropy method is used to evaluate the performance of component performance parameters based on the corresponding measurement parameters. Further, the information entropy corresponding to the cost function is 0.6805, which is better than the estimation result of the original cost function. Finally, the effectiveness of the parameter estimation method is verified by an example.