Data-based Remaining Lifetime Prediction Method And Its Application

Security issues in industrial production and transportation has always been the focus of worldattention. With the industrial and transportation systems become increasingly complex，it is harder tokeep safety running and lower maintainace costs. Thus, the concept of lifetime prediction has beenproposed.This paper presents a data-driven lifetime prediction method. Compared with the conventionalmethod, which is easy to use, fast to predict, and can be applid to a variety of predicting object. Torecognize the accuracy of lifetime prediction method, computer simulation and physical experimentare conducted. Specific work and research in this thesis are summarized below.1. A data-driven lifetime prediction method is introduced, which includes two parts: Analysisalgorithm and Prediction algorithm. The analysis algorithm calculates degradation factors by workingcurves, and the prediction algorithm predicts lifetime on the basis of the degradation factors.2. A micro quadrotor experimental platform is established for physical experiment. The flightcontroller is STM32chip and MPU6050IMU is the feedback of the flight control system, so that aclosed-loop flight control system is obtained. Meanwhile, battery monitor program is installed tocollect battery data.3. A physical experiment is conducted on the basis of the battery data from micro quadrotorexperimental platform. In order to get the precise open-circuit voltage, the measuring voltage data iscompensated in the calculation. Finally, the accuracy of the lifetime prediction method is drawn fromthe difference between the predicting result and the true result.