Knowledge-Data-Model Fusion Driven Fault Prediction Method for Snow Groomer

To address the problems of frequent unpredictable failures and difficult maintenance of snow groomer operating in complex and harsh environments such as night, steep slope, low temperature, and strong wind, a knowledge-data-model fusion driven fault prediction method for snow groomer is proposed by combining digital twin and deep learning technology. A snow groomer fault prediction system framework is established by analyzing the constituent elements of digital twin technology and the working characteristics of snow groomer. The digital twin model of snow groomer is established based on the multi-attribute elements of physical space. Subsequently, the data interconnection of multi-source information is established based on the data perception of operating condition data and environmental parameters to realize the virtual-real interaction between digital virtual body and physical entity across time scales. Furthermore, combined with the long-short-term memory network (LSTM) in deep learning, the fault prediction of snow groomer under real-time status monitoring is realized. Finally, the digital twin system of snow groomer is designed and built to verify the effectiveness of the proposed method, which provides a novel idea for the fault prediction of snow groomer.