Abstract: It can be found through the analysis of the failure of the feeding system in the process of leaf moistening and feeding,the abnormal operation of the feeding pump often occurs before the failure of the feeding system, because this parameter was not the key assessment index, and the failure symptom was not obvious, with transient jump, this phenomenon has not been caught by the relevant personnel, resulting in the gradual deterioration of the equipment operation status, resulting in the failure of the feeding system, or even the shutdown. Using big data analysis to summarize and summarize the equipment management of the original feeding module of the feeder, so as to establish the early warning model, develop an intelligent equipment early warning system, find the fault point in time and quickly, and achieve accurate prediction and maintenance. The results show that the time of fault diagnosis can be reduced to 20 minutes, which greatly shortens the time of fault diagnosis and improves the production efficiency of the workshop.

Key words: feeding system, big data technology, early warning model establishment, diagnosis time