Smart meter life prediction method based on health index and random forest

Abstract Smart electricity meter is the key instrument to realize accurate measurement of consumption. However, with increase service life meter, its accuracy will deteriorate. Aiming at problems that a single parameter cannot accurately characterize performance degradation process electric and traditional intelligent learning model difficult fit smart new method proposed fuse data construct health index, combine multi-model similarity matching integrated calculate residual energy meter. Methods prediction. Hierarchical clustering silhouette coefficients are used filter parameters them into index. Multi-model for regression prediction optimize results model. Select set verify effectiveness method. The show has advantages fast speed, fusion index greatly improves remaining