Multiple regression analysis based approach for condition monitoring of industrial rotating machinery using multi-sensors

Heterogeneous multi-sensors based change detection of machine running status is an important element to ensure the safety of industrial system. Considering the correlation between different sensors, this paper proposes a novel approach used for the detection of changes in machine condition. First, a prediction model is established using multiple regression, where a novel data description and prediction strategy is developed to make full use of the correlation. Residual analysis is then implemented to quantify the anomaly of investigated machinery. At last, decision making is made based on hypothesis testing. The proposed approach is evaluated by a representative condition monitoring application. Encouraging experimental results demonstrate that proposed approach has great potential in practical applications.