An Efficient Fault Diagnosis Framework for Digital Twins Using Optimized Machine Learning Models in Smart Industrial Control Systems

In recent times, digital twins (DT) is becoming an emerging and key technology for smart industrial control systems and Industrial Internet of things (IIoT) applications. The DT presently supports a significant tool that can generate a huge dataset for fault prediction and diagnosis in a real-time scenario for critical industrial applications with the support of powerful artificial intelligence (AI). The physical assets of DT can produce system performance data that is close to reality, which delivers remarkable opportunities for machine fault diagnosis for effective measured fault conditions. Therefore, this study presents an intelligent and efficient AI-based fault diagnosis framework using new hybrid optimization and machine learning models for industrial DT systems, namely, the triplex pump model and transmission system. The proposed hybrid framework utilizes a combination of optimization techniques (OT) such as the flower pollination algorithm (FPA), particle swarm algorithm (PSO), Harris hawk optimization (HHO), Jaya algorithm (JA), gray wolf optimizer (GWO), and Salp swarm algorithm (SSA), and machine learning (ML) such as K-nearest neighbors (KNN), decision tree (CART), and random forest (RF). The proposed hybrid OT-ML framework is validated using two different simulated datasets which are generated from both the mechanized triplex pump and transmission system models, respectively. From the experimental results, the hybrid FPA-CART and FPA-RF models within the proposed framework give acceptable results in detecting the most relevant subset of features from the two employed datasets while maintaining fault detection accuracy rates exemplified by the original set of features with 96.8% and 85.7%, respectively. Therefore, the results achieve good and acceptable performance compared to the other existing models for fault diagnosis in real time based on critical IIoT fields.