Two phase failure detection using fuzzy logic

Abstract

Control systems certainly deserve increased attention, as they can be found in many industrial environments, they are exploitable for military purposes, in transportation, and in many other areas. In this paper we propose the implementation of a control system for sensor failures detection using fuzzy logic approach for data analysis. We divide the system into two separate subsystems to ensure better suitability and adaptation to specific sensor data. One of these subsystems will be implemented exactly in the nodes which collect data from sensors. The second one will be implemented on the (cloud) server where all the signals are collected. The system implemented on the nodes will ensure that all the data are in the correct range of the monitored sensor and it could initiate the measurement repeatedly when the first measurement failed or its value is corrupted. This phase is able to check all the values in real time, therefore it can be considered as the main method for identification of major problems in the system. In the server part, all the collected values will be analyzed in order to identify other failures like gradual variation of sensor data due to polluted sensor parts, an others. The outcomes identified by this method could be removed or marked as the value that is not corresponding to the real world situation. These data can be ignored by the control system or can be used in some less critical application later on. The proposed system could be of importance in any control or manufacturing process were any redundant sensors and collected data directly affect parameters of the process.