计算参数对数值修正载荷识别算法的影响

针对在载荷识别计算中经常遇到的累积误差问题，提出了一种在每个时间步长内迭代修正的 载荷识别算法。首先利用拟静态算法得到载荷初值，再使用数值迭代算法对其进行修正计算，仿真结果表明，该修正算法可以有效地减小由于累积误差导致的发散，得到收敛的识别结果。针对上述算法，本文以多输入多输出简支梁为模型，分别分析了区间放大系数、区间分割系数和精确度指标3个计算参数对于识别结果的影响。计算结果显示，参数的选择对算法的效率和精度影响很大，不当的参数甚至可能引起识别结果严重发散，所以选择合适的计算参数对于数值修正算法十分重要。

Influence of Calculation Parameters on Numerical Correction Algorithm of Load Identification

The cumulative error problem is very common in the calculation of load identification. This paper puts forward a new load identification algorithm with numerical iteration in every time step. Firstly,the quasi-static method is used to obtain the initial value of load, then the numerical iteration algorithm is used for correction computation. Simulation results show that the correction algorithm can effectively decrease the divergence caused by the cumulative error and get convergent identification results.Moreover, a multiple-input multiple-output (MIMO) beam model is built to analyze the influence of interval amplification coefficient, interval partition coefficient and accuracy index on identification results.Calculation results indicate that these parameters have a great effect on the efficiency and precision of the algorithm, and improper parameters may even lead to serious divergence. Therefore it is very important to choose appropriate calculation parameters for the numerical correction algorithm.