基于新的广义粒子群方法的发电机组轴心轨迹提纯

Orbit Purification of Generator Units Based on a New Generalized Particle Swarm Optimization Method

针对发电机组等旋转机械轴心轨迹图的特点,结合案例提出一种新的广义粒子群思想,并将这一新的广义粒子群优化方法用于轴心轨迹提纯。将轴心轨迹图作为粒子群,建立粒子运动模型,阐述了广义粒子群方法的提纯和去噪机制,并给出了广义粒子群–模拟退火算法。通过转子振动试验平台上的模拟测试实例验证了该方法提纯轨迹的效果;对提纯后的轨迹图反演振动信号的分析表明,该方法去噪效果优于数学形态学滤波方法。该方法实现简单,是一种有效实现轴心轨迹提纯的新方法。

Targeting at the characteristics of orbit chart of the rotating machinery, such as generator units, one new idea of generalized particle swarm optimization （PSO） method was described with a case, and orbits purification using this method was proposed. Taking the orbit chart as swarm, the model of particle＇s motion was established, and the mechanism of purification and de-noising were discussed. Then one algorithm of generalized PSO-SAA （simulated annealing algorithm） was proposed to purify the orbit. By some simulation examples on the test bed of rotor vibration, the effectiveness of the method to purify the orbit is demonstrated. The analysis of inversion signal after purification also shows that the method is more effective to signal de-noising than mathematical morphology filters. The algorithm is very simple, and is a novel effective method of the orbit purification.