Research Area:  Internet of Things

Abstract:

Industrial Internet of things technology relying on artificial intelligence (AI) and big data is booming. Many enterprises tend to deliver heavy load Artificial Intelligence Computing to the cloud. However, cloud centered AI has many problems, such as insufficient real-time performance, insufficient bandwidth and large energy consumption. In order to solve the above problems, this paper proposes the edge intelligence which is composed of edge computing and artificial intelligence. Our idea is that under the edge intelligence driven fault diagnosis architecture, we give guidance on the selection of edge intelligence training model for fault diagnosis in different fault situations, and propose an edge intelligent fault diagnosis method which does not depend on threshold value. By using machine learning method to mine the spatiotemporal correlation between multi-dimensional operation data of equipment, and label the running status of equipment. The possible sources of anomalies are analyzed. The experimental results show that the proposed algorithm has high accuracy of anomaly detection, low average false alarm rate and good robustness.

Keywords:  

Author(s) Name:  Haitao Sang, Bo Chen

Publisher name:  Springer

DOI:  10.1007/978-981-33-6318-2_82

Volume Information:  pp 658-665

Paper Link:   https://link.springer.com/chapter/10.1007/978-981-33-6318-2_82