根据发电机组轴心轨迹图的特点,提出将振动信号及原始轴心轨迹图同步处理以实现轴心轨迹提纯的新思路.给出了对振动信号进行幅度调制的EMD(empirical mode decomposition)调制方法.将轴心轨迹图上的每一个点看作一个粒子,参照蚂蚁等群体的特性,建立了一种新的粒子群模型,分析了这种粒子群模型实现轨迹提纯的机理,并作了加快粒子收敛速度的改进.给出了EMD调制和粒子群相结合的轴心轨迹提纯新算法,通过模拟实验对算法参数选取进行了研究,并验证了算法的有效性.该算法比低通滤波和形态学滤波效果更好,是一种实用的轴心轨迹提纯新方法.
According to the characteristics of rotor orbit chart of the generator units, a method of synchronously processing the vibration signal and the original rotor orbit chart is proposed to purify rotor orbit purification. The method of EMD (empirical mode decomposition) modulation for the amplitude of vibration signals is presented. Taking every point on the rotor orbit chart as a particle, a new particle swarm model is established referring to the characteristics of some colonies, such as ant colony, and the orbit purification mechanism of the model is analyzed. Some improvements are made to speed up particle's convergence. A new algorithm of orbit purification combining EMD modulation with particle swarm is presented. The selection of the algorithm's parameters is studied through a simulation experiment which also verifies the effectiveness of the algorithm. Filtering results of the algorithm is better than those of the low-pass filtering and morphological filtering methods. It is a new and practical method of orbit purification.