Sensor placement method based on bond graph models

In order to study the influence of sensor placement on fault diagnosis, a new sensor placement method based on bond graph model was proposed. A group of virtual sensor test points were set in bond graph model, then the information of bond graph structure and causality constraints were used to deduce a set of analytical redundancy relations, namely system residuals. The relationships among residuals, faults and sensor placement configuration were analysed, and this sensor placement method can meet the requirements of system fault detection and isolation performance. The sensor placement configuration with the minimum number of sensors was chosen under the premise of maximizing system diagnosis performance. Finally, the synchronous generator was used as an example to establish the sensor placement algorithm. Residual analysis was used to derive the structure fault feature matrix and the sensor feature matrix. The proposed sensor placement optimization algorithm was verified by this example, and the experimental results show that the final sensor placement configuration can meet the requirements of maximum parameter fault detection and isolation performance.