An Efficient Fault Diagnosis Framework for Digital Twins Using Optimized Machine Learning Models in Smart Industrial Control Systems
Abstract 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 re...
| Main Authors: | , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Springer
2023-05-01
|
| Series: | International Journal of Computational Intelligence Systems |
| Subjects: | |
| Online Access: | https://doi.org/10.1007/s44196-023-00241-6 |
| _version_ | 1827952555702878208 |
|---|---|
| author | Samar M. Zayed Gamal Attiya Ayman El-Sayed Amged Sayed Ezz El-Din Hemdan |
| author_facet | Samar M. Zayed Gamal Attiya Ayman El-Sayed Amged Sayed Ezz El-Din Hemdan |
| author_sort | Samar M. Zayed |
| collection | DOAJ |
| description | Abstract 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. |
| first_indexed | 2024-04-09T14:00:34Z |
| format | Article |
| id | doaj.art-f2bd4ca243e444df81ecf4f3fae0d344 |
| institution | Directory Open Access Journal |
| issn | 1875-6883 |
| language | English |
| last_indexed | 2024-04-09T14:00:34Z |
| publishDate | 2023-05-01 |
| publisher | Springer |
| record_format | Article |
| series | International Journal of Computational Intelligence Systems |
| spelling | doaj.art-f2bd4ca243e444df81ecf4f3fae0d3442023-05-07T11:23:25ZengSpringerInternational Journal of Computational Intelligence Systems1875-68832023-05-0116111810.1007/s44196-023-00241-6An Efficient Fault Diagnosis Framework for Digital Twins Using Optimized Machine Learning Models in Smart Industrial Control SystemsSamar M. Zayed0Gamal Attiya1Ayman El-Sayed2Amged Sayed3Ezz El-Din Hemdan4Higher Institute of Engineering and Technology (HIET)Computer Science and Engineering Department, Faculty of Electronic Engineering, Menoufia UniversityComputer Science and Engineering Department, Faculty of Electronic Engineering, Menoufia UniversityIndustrial Electronics and Control Engineering Department, Faculty of Electronic Engineering, Menoufia UniversityComputer Science and Engineering Department, Faculty of Electronic Engineering, Menoufia UniversityAbstract 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.https://doi.org/10.1007/s44196-023-00241-6Digital twins (DT)Flower pollination algorithm (FPA)OptimizationMachine learningFault diagnosisControl systems |
| spellingShingle | Samar M. Zayed Gamal Attiya Ayman El-Sayed Amged Sayed Ezz El-Din Hemdan An Efficient Fault Diagnosis Framework for Digital Twins Using Optimized Machine Learning Models in Smart Industrial Control Systems International Journal of Computational Intelligence Systems Digital twins (DT) Flower pollination algorithm (FPA) Optimization Machine learning Fault diagnosis Control systems |
| title | An Efficient Fault Diagnosis Framework for Digital Twins Using Optimized Machine Learning Models in Smart Industrial Control Systems |
| title_full | An Efficient Fault Diagnosis Framework for Digital Twins Using Optimized Machine Learning Models in Smart Industrial Control Systems |
| title_fullStr | An Efficient Fault Diagnosis Framework for Digital Twins Using Optimized Machine Learning Models in Smart Industrial Control Systems |
| title_full_unstemmed | An Efficient Fault Diagnosis Framework for Digital Twins Using Optimized Machine Learning Models in Smart Industrial Control Systems |
| title_short | An Efficient Fault Diagnosis Framework for Digital Twins Using Optimized Machine Learning Models in Smart Industrial Control Systems |
| title_sort | efficient fault diagnosis framework for digital twins using optimized machine learning models in smart industrial control systems |
| topic | Digital twins (DT) Flower pollination algorithm (FPA) Optimization Machine learning Fault diagnosis Control systems |
| url | https://doi.org/10.1007/s44196-023-00241-6 |
| work_keys_str_mv | AT samarmzayed anefficientfaultdiagnosisframeworkfordigitaltwinsusingoptimizedmachinelearningmodelsinsmartindustrialcontrolsystems AT gamalattiya anefficientfaultdiagnosisframeworkfordigitaltwinsusingoptimizedmachinelearningmodelsinsmartindustrialcontrolsystems AT aymanelsayed anefficientfaultdiagnosisframeworkfordigitaltwinsusingoptimizedmachinelearningmodelsinsmartindustrialcontrolsystems AT amgedsayed anefficientfaultdiagnosisframeworkfordigitaltwinsusingoptimizedmachinelearningmodelsinsmartindustrialcontrolsystems AT ezzeldinhemdan anefficientfaultdiagnosisframeworkfordigitaltwinsusingoptimizedmachinelearningmodelsinsmartindustrialcontrolsystems AT samarmzayed efficientfaultdiagnosisframeworkfordigitaltwinsusingoptimizedmachinelearningmodelsinsmartindustrialcontrolsystems AT gamalattiya efficientfaultdiagnosisframeworkfordigitaltwinsusingoptimizedmachinelearningmodelsinsmartindustrialcontrolsystems AT aymanelsayed efficientfaultdiagnosisframeworkfordigitaltwinsusingoptimizedmachinelearningmodelsinsmartindustrialcontrolsystems AT amgedsayed efficientfaultdiagnosisframeworkfordigitaltwinsusingoptimizedmachinelearningmodelsinsmartindustrialcontrolsystems AT ezzeldinhemdan efficientfaultdiagnosisframeworkfordigitaltwinsusingoptimizedmachinelearningmodelsinsmartindustrialcontrolsystems |