基于粒子群算法的FFT信号分析方法及应用

Method and Application of FFT Signal Analysis Based on Particle Swarm Optimization Algorithm

针对快速傅里叶变换在频谱分析时易出现栅栏效应和频谱泄漏问题,首先提出基于粒子群算法的FFT信号分析方法对信号进行频谱分析,优化信号采样过程中采样持续时间,实现整周期采样,减少频谱泄漏现象;其次,通过对模拟信号的仿真分析,验证方法的可行性。采集某厂650六辊可逆冷轧机的缸位移信号且针对现场采集信号提出一种基于插值算法的数据处理方法,并对信号进行频谱分析,证明该方法对生产实际具有指导意义。

For the problems of picket fence effect and spectrum leakage in the spectrum analysis by fast Fourier transform, FFT signal analysis method was firstly proposed based on particle swarm optimization algorithm for signal spectrum analysis, and the signal sampling duration in sampling process was then optimized, which could realize integral period sampling and reduce the spectrum leakage phenomenon. After that, the simulation of the analog signal has verified the feasibility of this method. The cylinder displacement signal of 650 six-roller reversible cold roll of a factory was gathered. For such kind of on-site signal acquisition, a data processing method based on interpolation algorithm was put forward, which proved to be instructive to actual production by the following signal spectrum analysis.