摘要
聚焦于运动学的视阈,探讨一种基于克里格模型的机械系统优化策略。利用MSC.A D A M S建立一个简单而又具有普遍意义的多机构机械系统初始模型,修改连杆比值为[1,1.5,2]、间距为[0,5,10]cm,形成9个模型,再加密采样间隔,形成连杆比值为[1,1.25,1.5,1.75,2]、间距为[0,2.5,5,7.5,1 0]c m的2 5个模型,计算两套模型的机械增益和加速度等运动学参数。在MATLAB下基于克里格模型算法形成空间面曲线,预测其他连杆比值和间距下的运动学参数,与真实值进行比对,讨论模型相对误差。当连杆比值为1.6,间距为6 cm时,9个模型/25个模型的相对误差为:机械增益14.32%/4.97%,加速度47.80%/20.90%;当连杆比值为1.4,间距为3 cm时,9个模型/25个模型的相对误差为:机械增益18.26%/13.06%,加速度22.98%/8.07%。随着模型规模的增大,系统各参数的预测精确度也会随之提高。应借助大数据技术等新兴产业,实时把握数据变化,补充约束条件,调节数据趋势,优化全局模型,获取最优参数。
An optimization strategy for Kriging model based mechanical system is discussed focusing on kinematics. An initial but universal model of multi-mechanism mechanical system is established using MSC.ADAMS. 9 models are formed by modifying the link ratios of [1, 1.5, 2] and spacings of [0, 5, 10] cm, and through sampling interval refining, 25 models are formed using the link ratios of [1, 1.25, 1.5, 1.75, 2] and spacings of [0, 2.5, 5, 7.5, 10] cm. Kinematic parameters including mechanical advantage and acceleration are calculated for the two suites of models. Spatial curve is for medusing MAT LAB based on the Kriging model, to predict kinematic parameters under other link ratios and spacings, and further make comparisons with actual value to discuss relative errors of models. When the link ratio is 1.6 and the spacing is 6 cm, the relative errors for the 9 models/25 models of mechanical advantages are 14.32%/4.97%, and accelerations 47.80%/20.90%; when the link ratio is 1.4 and the spacing is 3 cm, the relative errors for the 9 models/25 models of mechanical advantages are 18.26%/13.06%, and accelerations 22.98%/8.07%. With the increase of the model scale, the prediction accuracy of the system parameters will be improved. With the help of emerging industries including big data technology, one should grasp the real-time data changes, supplement constraints, adjust the data trends, optimize the global model, and ultimately obtain the most optimized parameters.
出处
《工业技术创新》
2017年第3期5-11,共7页
Industrial Technology Innovation
