A hidden Markov model to predict hot socket issue in smart grid

Smart meters collect sensor data at distribution ends of smart grid. The collection process performs nonstop data bundling and results in ‘hot socket’ issue due to high resistance. This results an abnormal generation of dataset and overall severely affect the operational aspects of smart grid. In this paper, we present a model for Smart Meter Abnormal Data Identification (SMADI) over the communication bridge of Smart grid repository and distribution end units, to redirect abnormal samples to HBase error repository using Message propagation strategy. SMADI predicts possible hot socket smart meter node through HMM and generates a sequence of possible hot socket smart meters over time interval. The simulation results show that SMADI precisely collect error samples and reduce complexity of performing data analytics over giant data repository of a smart grid. Our model predicts hot socket smart meter nodes efficiently and prevent computation cost of performing error analytics over smart grid repository.