Equipment cluster health monitoring method based on cloud edge collaboration

The invention discloses an equipment cluster health monitoring method based on cloud edge collaboration. The method comprises the steps that 1, a data set is constructed and preprocessed; and step 2, model training: the model training stage relates to iterative training of an edge end local model, and a cloud global model, a group model and an edge end edge model are generated through iterative aggregation by a model aggregation strategy based on weighted Kappa value design until the number of cloud side communication times reaches a threshold value. And step 3, equipment health monitoring: at an edge end, the edge model of each edge node undertakes a health monitoring task. And the running equipment continuously uploads the equipment data to the edge end. Once it is monitored that the equipment is in a fault state, the edge end records the fault occurrence time and the fault type of the equipment and uploads the information to the cloud end, and finally the cloud end issues an alarm in a unified mode. When t