运用改进的教学-模拟退火算法辨识关节臂式三坐标测量机的结构参数

Structural Parameter Identification for Articulated Arm Coordinate Measuring Machines Using Modified Teaching-learning-based Optimization Optimized by Simulated Annealing Algorithm

为了提高关节臂式三坐标测量机的精度,降低其运动不确定度,提出了一种改进的教学-模拟退火混合优化算法来辨识其结构参数并补偿其误差,从而提高其精度。分析了教学算法(teaching-learning-based optimization,TLBO)的优缺点并对其进行改进从而得到改进的教学算法;提出了一种收敛精度转换准则,将改进的教学算法(modified TLBO,mTLBO)和模拟退火算法(simulated annealing,SA)融合得到改进的教学-模拟退火算法(mTLBO-SA);基于此理论和单点重复率误差实验,分别用TLBO、SA和mTLBO-SA对关节臂式三坐标测量机的结构参数进行了辨识;分别用辨识前后的关节臂式三坐标测量机再次进行单点重复率误差实验,并比较相应的实验结果。实验结果表明,所提算法能有效且高效地辨识关节臂式三坐标测量机的结构参数,进而有效提高其定位精度,降低其运动不确定度。

To improve the accuracy and decease the movement uncertainty of AACMM,a hybrid optimization algorithm(mTLBO-SA)was proposed to calibrate the structural parameters of AACMM,furtherly compensating for the errors and increasing the precision.First,the advantages and disadvantages of TLBO algorithm were analysed and it was then correspondingly improved to obtain mTLBO;next,a transitional criterion for convergence prevision was put forward to combine mTLBO with SA algorithm,resulting in mTLBO-SA;then,the structural parameters of AACMM were calibrated based on this theory and error experiments using TLBO,SA and mTLBO-SA,respectively;at last,the error experiments of single point repeatability were conducted again respectively using the AACMM before calibration and after calibration and the related results were compared resulting in corresponding conclusions.The results show that the proposed algorithm is effective and efficient to calibrate the real parameters of AACMM,which extremely enhances the accuracy and decreases the movement uncertainty of AACMM.