Machine-Learning-Driven Digital Twin for Lifecycle Management of Complex Equipment

Abstract

The full life cycle management of complex equipment is considered fundamental to the intelligent transformation and upgrading of the modern manufacturing industry. Digital twin technology and machine learning have been emerging technologies in recent years. The application of these two technologies in the full life cycle management of complex equipment can make each stage of the life cycle more responsive, predictable, and adaptable. This paper first proposes a technical system that embeds machine learning modules into digital twins. Next, on this basis, a full life cycle digital twin for complex equipment is constructed, and joint application of sub-models and machine learning is explored. Then, the application of a combination of the digital twin in maintenance with machine learning in predictive maintenance of diesel locomotives is presented. The effectiveness of the proposed management method is verified by experiments. The abnormal axle temperature can be alarmed about one week in advance. Lastly, possible application advantages of the combination of digital twin and machine learning in addressing future research direction in this field are introduced.