The health condition of hoisting gearbox bearings of quay cranes is of great importance for the safety of port production.A remaining useful life(RUL)predicting framework for lifting gearbox bearings of quay crane under time-varying operating conditions is proposed.First,the working load is discretized and the condition boundaries are determined.Then,the long short-term memory(LSTM)network model is adopted to predict the load and the corresponding operating conditions.Afterwards,considering the degradation rates and jump coefficients under different operating conditions,the state degradation function is established based on the Wiener process.Finally,the condition-activated particle filter(CAPF)is used to predict the degradation state and RUL of bearings.The proposed prediction framework is verified by the full-life data of the hoisting gearbox bearings in a port in Shanghai collected by the NetCMAS system.A comparison with the other three prediction methods shows that the proposed framework is able to obtain more accurate degradation states and RUL predictions under time-varying operating conditions.
维普
万方数据