Underwater glider fault diagnosis method based on improved dynamic grey prediction

The invention relates to an underwater glider fault diagnosis method based on improved dynamic grey prediction, and the method comprises the steps: carrying out the online prediction of sensor data through employing an improved dynamic grey prediction method for a sensor fault of an underwater glider, and judging whether the sensor has a fault or not through analyzing the error between a predicted value and an actual value. Therefore, faults can be found in time and corresponding emergency measures can be taken. According to the method, fault alarm can be carried out in time at the early stage of fault occurrence, and various losses caused by sensor faults are reduced to the maximum extent. 本发明涉及一种基于改进动态灰色预测的水下滑翔机故障诊断方法，针对水下滑翔机的传感器故障，采用改进后的动态灰色预测方法对传感器数据进行在线预测，通过分析预测值与实际值的误差大小来判断传感器是否发生故障，以便及时发现故障并采取相应的应急措施。该方法可以在故障发生初期及时进行故障报警，在最大程度上降低传感器故障带来的各种损失。