The continuous expansion of expressway network scale and the increase of electromechanical system faults put forward higher requirements for spare parts inventory demand forecast, while the lack of maintenance data becomes one of the major challenges for factor analysis. To address this issue, this work presents a forecast method of spare parts demand for expressway electromechanical equipment. The approach firstly utilizes a segmented fault model to characterize the fault characteristics of equipment. Then, the failure rate, which is transformed by the running state of equipment using the above fault model, combined with other related factors is used to construct an influence factor sequence. Finally, the factor sequence is fed into the improved random forests model to predict the spare parts demand. Experiments performed on an automatic railing machine of the charging system indicate that the failure rate plays an important role in the demand forecast. The proposed method shows significant performance gains when compared with the current state-of-art methods.

CICTP