2020/04/20 by Daniel Gutiérrez-Rojas, Gutierrez-Rojas, Daniel, Mehar Ullah +19
Computer Science · Engineering · #Advanced Chemical Sensor Technologies #Anomaly Detection Techniques and Applications #FOS: Electrical engineering #Fault Detection and Control Systems #Signal Processing (eess.SP) #Systems and Control (eess.SY) #electronic engineering #information engineering
paper · pdf · doi:10.48550/arxiv.2004.09097
openalex publication_date 2020/04/20 · openalex created_date 2022/07/29 · openalex updated_date 2026/07/28
This paper introduces a general approach to design a tailored solution to\ndetect rare events in different industrial applications based on Internet of\nThings (IoT) networks and machine learning algorithms. We propose a general\nframework based on three layers (physical, data and decision) that defines the\npossible designing options so that the rare events/anomalies can be detected\nultra-reliably. This general framework is then applied in a well-known\nbenchmark scenario, namely Tennessee Eastman Process. We then analyze this\nbenchmark under three threads related to data processes: acquisition, fusion\nand analytics. Our numerical results indicate that: (i) event-driven data\nacquisition can significantly decrease the number of samples while filtering\nmeasurement noise, (ii) mutual information data fusion method can significantly\ndecrease the variable spaces and (iii) quantitative association rule mining\nmethod for data analytics is effective for the rare event detection,\nidentification and diagnosis. These results indicates the benefits of an\nintegrated solution that jointly considers the different levels of data\nprocessing following the proposed general three layer framework, including\ndetails of the communication network and computing platform to be employed.\n