Fault diagnosis for maglev system based on improved principal component analysis

The problem of sensor-fault detection and diagnosis (FDD) of maglev system was studied based on principal component analysis (PCA). First, the mathematic model of single electromagnet suspension system was constructed, and then a sensor-FDD strategy was designed for it based on PCA. The performance of the FDD strategy was simulated. At last, tracking-differentiator (TD) was introduced into the sensor-FDD system. The result of simulation shows that the FDD strategy based on PCA combined with TD are superior to that based on PCA only in the precision of FDD and to that based on PCA combined with exponentially weighted moving average (EWMA) in time consuming.