A study on modeling using big data and deep learning method for failure diagnosis of system

Power data of the customers are generated in real time and in large quantities according to the activation of the low voltage meter reading service and accumulated and stored in the central FEP, NMS and SMS server. The power data has characteristics of big data and unlike big data of other types, it has the advantage of applying efficient and useful services by applying various analysis and prediction techniques without any additional data processing. Despite these advantages, the field of automated meter reading still focuses on the construction of the system, so data analysis and application aren't actively implemented. However, the auto metering infrastructure is being installed continuously and the amount of data is continuously increasing. Therefore, the importance of utilization of power data based on the auto metering infrastructure becomes an issue, and research on power big data analysis and data mining is actively being carried out. In this paper, we have performed a probabilistic analysis, diagnosis, and prediction of the fault condition of the system through application of the artificial intelligent deep learning algorithm using the system state data stored in real time in the low voltage meter reading system. In this paper, we propose an optimal method for designing and operating a reasonable automated meter reading system with stability and reliability.