工频信号测量中参数自寻优智能采样法的研究

软件同步是数字化测量工频信号常用的采样方法.参数自寻优智能采样用爬山搜索法使误差判据最小,得到每个测量周期的采样次数N和采样间隔tS,最大限度地减小了同步误差或周期截断误差.进一步校正参数自寻优,让测量值乘上一个针对实时周期Tp寻优得到的校正参数kc,使其能最大限度地逼近信号的实际值.遍历频点的仿真结果表明:最优采样参数使工频范围内67%以上频率点的同步误差为零,其余点上的失步也在1～3个td(微处理器中定时器的分辨率),采样精度达到10-5数量级;传统采样法的测量值经校正后向零误差逼近,采样精度达10-6数量级;在基波条件下寻优得来的参数N、ts和kc对未知谐波同样有效.分析了自寻优智能采样法的实时性,提出了"软件硬化"的思想,使微处理器开销压缩,提高了在线寻优效率.

Intelligent sampling with parameter self-optimizing in power frequency measuring

The software synchronization is always used in digital measuring of power frequency.The intelligent sampling with parameter self-optimizing uses mountain-climb searching with minimal error criterion to obtain the sampling number N and the interval tS of each measurement period,which reduces the synchronization error or cycle truncation error.The real value is approached by multiplying the measured value by the correction parameter kc,which is self-optimized according to real-time period Tp.Simulation results prove:the synchronous errors of over 67 % of covered periods are zero,and the remainders are only(1 ~ 3)td(timer resolution of microprocessor),the sampling precision is about 10-5;after the correction,the measured value by traditional sampling approaches zero error,and the sampling precision is about 10-6;the optimized parameters for fundamental are also effective on unknown harmonics.The real-time performance of intelligent sampling with parameter self-optimizing is analyzed and the idea of"software hardening" is proposed to lower microprocessor load and increase online optimizing efficiency.