Fault Diagnosis of Intelligent Production Line Based on Digital Twin and Improved Random Forest

Abstract

Digital twin (DT) is a key technology for realizing the interconnection and intelligent operation of the physical world and the world of information and provides a new paradigm for fault diagnosis. Traditional machine learning algorithms require a balanced dataset. Training and testing sets must have the same distribution. Training a good generalization model is difficult in an actual production line operation process. Fault diagnosis technology based on the digital twin uses its ultrarealistic, multisystem, and high-precision characteristics to simulate fault data that are difficult to obtain in an actual production line to train a reliable fault diagnosis model. In this article, we first propose an improved random forest (IRF) algorithm, which reselects decision trees with high accuracy and large differences through hierarchical clustering and gives them weights. Digital twin technology is used to simulate a large number of balanced datasets to train the model, and the trained model can be transferred to a physical production line through transfer learning for fault diagnosis. Finally, the feasibility of our proposed algorithm is verified through a case study of an automobile rear axle assembly line, for which the accuracy of the proposed algorithm reaches 97.8%. The traditional machine learning plus digital twin fault diagnosis method proposed in this paper involves some generalization, and thus has practical value when extended to other fields.

Application of machine learning and statistical modeling to identify sources of air pollutant levels in Kitchener, Ontario, Canada

“…The output of many such trees is analyzed, and a final prediction model is obtained by pooling the trees together. 38–40 Scheme S1† shows an illustration of the procedure followed by the sample code used in the model. When the number of features compared to the number of data points is relatively small, RF has prediction accuracy typically far superior to that of multiple linear regression modeling.…”

Digital Twin-Driven Rear Axle Assembly Torque Prediction and Online Control

“…With the wide application of six-axis robotic arms in production, it has become a common way for the rear axle assembly process to use the robotic arm to drive the tightening gun to output assembly torque [ 35 ]. However, due to the accumulation of errors in the attitude, output angle, and other factors of the tightening gun during the assembly process, the assembly torque cannot meet the quality requirements of the body, and the assembly efficiency is affected.…”

Fault Diagnosis Method for Bearing Based on Digital Twin

“…Jain et al [28] presented a digital twin-based fault diagnosis method for distributed photovoltaic system, which demonstrated higher fault sensitivity. Guo et al [29] simulated a large amount of fault data of a production line through digital twin and trained a reliable fault diagnosis model based on an improved random forest. However, few works have been reported to apply digital twin to bearing fault diagnosis.…”