Data Analytics Towards Predictive Maintenance for Industrial Ovens
Autor: | Vaia Rousopoulou, Peter Haigh, Luigi Giugliano, Dimosthenis Ioannidis, Alexandros Nizamis, Dimitrios Tzovaras, Luis Martins |
---|---|
Rok vydání: | 2019 |
Předmět: | |
Zdroj: | Lecture Notes in Business Information Processing ISBN: 9783030209476 CAiSE Workshops |
DOI: | 10.1007/978-3-030-20948-3_8 |
Popis: | In Industry 4.0, predictive maintenance aims to improve both production and maintenance efficiency. The interconnected machines and IoT devices produce a variety of data that enable the early detection of anomalies and failures by predictive analytic algorithms. Predictive analytics can also reduce the machines downtimes and decrease the production of faulty products. This paper introduces predictive analytics for industrial ovens and their application in a real-world’s oven used by a leading medical devices manufacturer. Two distinct approaches are presented in this work. A technique based on existing sensors for oven failure prediction based on monitoring and log data; and a technique based on deployed sensors for fault diagnosis based on acoustic data. Deep learning techniques have been applied on existing sensor and event log data, especially temperature monitoring, whereas an outlier detection analysis were implemented on acoustic sensor measurements. Both analytics methods create a complete solution able to detect early oven failures from their root. |
Databáze: | OpenAIRE |
Externí odkaz: |