动态扭矩传感器直接连接于旋转动力设备和荷重装置之间,可以输出与实际负载相关联的力矩信号.为了从动态扭矩传感器输出的力矩信号中得到实际负载,在荷重模型建立阶段,通过给荷重装置施加不同的负载,并对动态扭矩传感器输出力矩信号对应的AD转换值进行灰色关联校正,得到趋于稳定的AD转换校正值,在此基础上,利用移动最小二乘回归(MLSR)实现对荷重装置实际负载的非线性回归,得到实际负载与AD转换校正值间的非线性模型.在荷重检测阶段,进一步利用灰色关联校正并结合实际负载回归模型,得到荷重装置最终的实际负载.试验结果表明,结合灰色关联校正及移动最小二乘回归模型,荷重检测误差低于±0.3%.%Directly connected the load equipment to the rotary power one,the dynamic torque sensor could output the torque signal caused by both the actual load and the other factors such as frictional force of the load equipment. The actual load could be extracted from the sensor's output signal of the torque to use two phases shown as fellows. In load modeling phase,a kind of corrected model based on gray correlation analysis was established to adjust the respective analog-to-digital converter( ADC) value caused by different actual loads. The corrected model could effec-tively decrease the fluctuation and random for the ADC value. The load model was gotten by moving least square re-gress( MLSR) through taking the corrected ADC values as the input variables and the individual actual load as the output ones. After the load regress model was founded,the actual load for the load equipment would be attained by combining the load regress model with the gray correlation corrected one in the load detection phrase. Experiment showed that the error of the load detection based on dynamic torque sensor was less than ±0.3%.
展开▼
