Progress in smart industrial control applied to renewable energy system

Abstract

Abstract The industrial Supervising Control and Data Acquisition, referred by SCADA system, tends to improve its accuracy in detecting faults. In that, it uses fault diagnosis models based mostly on probabilistic methods with close uncertainties. These models are based on a subjective evaluation by comparing the obtained signal to its reference. Therefore, SCADA precision fault detection varies depending on the operation environment, system design and analysis approach among other factors. The contribution of this research work is to propose a smart strategy that will enrich and enhance failure recognition in SCADA systems by integrating two additional models into the classic technique. The first model is a SOM map reduce simple classifier and the second model is an evolutionary recurrent self-organizing neural filter for final decision-making. This integrated paradigm improves results accuracy and robustness against signal interference. The proposed idea involves best details around any remotely listed defect. This study has been conducted on Simulink-Matlab, through the analysis of multi signals emitted by sensors and received by corresponding antennas.