基于C8051F120的PI控制过程的快速运算  

Fast Computing Method of PI Control Process Based on C8051F120

作　　者：李浩男 王显军[1] 

机构地区：[1]中国科学院长春光学精密机械与物理研究所,吉林长春130033 [2]中国科学院研究生院,北京100039

出　　处：《测控技术》2014年第5期69-72,78,共5页Measurement & Control Technology

基　　金：中国科学院长春光学精密机械与物理研究所三期创新工程资助项目

摘　　要：在控制系统的校正过程中,使用了增量式PI算法。在分析系统过程后,考虑到在PI校正中运算量集中在浮点数的乘法,运算速度有提升空间。为了提高校正程序运算速度,设计一种基于增量式PI算法的浮点数运算程序,将浮点数运算转化为整数乘法和移位,利用C8051 F120中的MACO(乘法和累加引擎),实现整数乘法和移位的快速运行。分析算法速度,运算时间缩短到原时间的27%。根据系统实际情况,分析数据运算精度,控制增量误差小于1%,能够保证系统控制精度,并得到实验验证。快速运算方法能够满足控制要求,硬件成本低,缩短了单周期内系统运算时间。The incremental PI algorithm is adopted in the calibration process of a control system.To improve the computation speed of the calibration,the system process is analyzed,considering that floating point multiplication is encountered frequently in the PI algorithm,which takes longer time compared to integer multiplication.Therefore,a floating point acceleration program based on incremental PI algorithm is designed to transform the floating point multiplication to integer multiplication associated with shifting operation,and the fast run of integer multiplication and shifting is realized by leveraging the MAC0 in C8051F120.Analyzing the efficiency of the algorithm,the time needed is reduced to 27％ of the original work.By analyzing the accuracy of calculation,the novel method is able to guarantee the system control accuracy,which is verified by experiments.The fast computing method can satisfy the requirement of the control system with low hardware cost and shorten the system computing time per cycle.

关 键 词：MAC0 PI控制 浮点数乘法 快速运算 

分 类 号：TP273[自动化与计算机技术—检测技术与自动化装置]