容栅传感器多因素耦合误差的自辨识机理与分离方法研究

Encoder-like capacitive sensors (ECS) offer an intrinsic advantage as large measuring scale, however due to the effect of diverse coupling errors, including nonlinearity errors caused by fringe effect, static and dynamic errors brought in during setting up or measuring, ECS can not yet live up to the requirement of ultra precision measurement. With all the considerations, an approach includes self-recognition and fast separation of ECS diverse coupling errors is presented in this project. 1) The full capacitance field is constructed based on the MAXWELL equations, fast separation of nonlinearity fringe effect errors is realized using the empirical mode decomposition (EMD) method; 2) diverse coupling errors are directly separated owing to a novel self-recognition method, measuring accuracy of X-Y displacements are drastically improved, besides measuring errors of roll, yawn and pitch angles are all recognized without any extra probes. Research of this project starts from diminishing the influences of ECS diverse coupling errors, mainly focusing on resolving the conflicts of large measuring scale and ultra precision measurement, a sound theoretical basis will be established for the development of precision measurement sensors with relatively simple structure and effective costs.

容栅传感器测量行程大，但对多因素耦合误差较为敏感，包括边缘效应引起的非线性误差、加工制造与安装过程中的静态误差、振动及电磁干扰引起的动态误差等相互耦合，实现高精度测量的难度较大，针对上述问题，本项目提出建立容栅传感器多因素耦合误差的自辨识机理和快速分离方法。1）基于MAXWELL方程式计算其电容场模型，针对因电容传感器边缘效应导致的非线性误差，拟基于经验模态分解实现快速分离；2）构建一种新式的耦合误差自辨识方法，采用对称反正切解耦算法直接从传感器原输出信号中分离有效的耦合信息，提高X-Y二维位移测量的精度，同时在不增加传感器数量的基础上实现翻滚、俯仰与偏摆角等三个误差量的辨识。项目研究从消减容栅传感器的多因素耦合误差切入，着重解决大行程和高精度测量这对矛盾，为开发结构简洁、低成本的现代精密位移检测技术与系统提供理论支撑，具有重要的科研价值和工程意义。

随着先进制造技术的不断发展与进步，现代精密位移检测技术对于大行程、高精度、多自由度的要求愈发强烈。本项目通过开展多因素耦合误差的自辨识机理和快速分离方法研究，充分利用容栅传感器的优势，解决大行程和高精度测量这对矛盾，满足了现代精密位移检测技术的要求。（1）项目首先基于MAXWELL方程组，通过建立二维容栅位移传感器的标准电容场模型，量化了边缘效应对输出信号的影响，揭示边缘效应导致非线性误差的规律，并针对因边缘效应导致的非线性误差，应用经验模态分解法（EMD）将信号被分解为一组固有模态函数（IMF）组和残留分量的组合，并将与传感器信号具有相同的频率一组IMF信号定义为基波分量，该分量包含了传感器信号的所有特征信息，通过从基波分量中提取代替从原始信号中提取传感器信息的方式，大大减少了电场边缘效应的影响，实验结果表明，信号波形的非线性误差从4.18%降低为0.89%。（2）建立了容栅位移传感器的误差模型，明确偏摆角位移θz引起波形相位偏移误差，翻滚、俯仰角位移θx-θy引起灵敏度误差，项目通过一种新式的对称反正切解耦算法直接从X-Y二维位移输出信号中分离出耦合信息，在不增加传感器数量的基础上实现多自由度测量误差的自辨识。实验结果表明，当三个旋转角位移的行程不超过2°时，测量误差不超过0.01°。