Generative Adversarial Network based Autoencoder: Application to fault\n detection problem for closed loop dynamical systems

2018/04/15 by Indrasis Chakraborty, Rudrasis Chakraborty, Chakraborty, Indrasis +3
Engineering · #Artificial Intelligence (cs.AI) #FOS: Computer and information sciences #Fault Detection and Control Systems #Machine Learning (cs.LG) #Machine Learning (stat.ML) #Oil and Gas Production Techniques #Water Systems and Optimization

paper · pdf · doi:10.48550/arxiv.1804.05320

openalex publication_date 2018/04/15 · openalex created_date 2022/10/01 · openalex updated_date 2026/07/28

Abstract

Fault detection problem for closed loop uncertain dynamical systems, is\ninvestigated in this paper, using different deep learning based methods.\nTraditional classifier based method does not perform well, because of the\ninherent difficulty of detecting system level faults for closed loop dynamical\nsystem. Specifically, acting controller in any closed loop dynamical system,\nworks to reduce the effect of system level faults. A novel Generative\nAdversarial based deep Autoencoder is designed to classify datasets under\nnormal and faulty operating conditions. This proposed network performs\nsignificantly well when compared to any available classifier based methods, and\nmoreover, does not require labeled fault incorporated datasets for training\npurpose. Finally, this aforementioned network's performance is tested on a high\ncomplexity building energy system dataset.\n

Related

• Deep learning architectures for data-driven damage detection in nonlinear dynamic systems
• Control theoretically explainable application of autoencoder methods to fault detection in nonlinear dynamic systems
• Deep Generative Models in Condition and Structural Health Monitoring: Opportunities, Limitations and Future Outlook
• Deep Learning-Enabled System Diagnosis in Microgrids: A Feature-Feedback GAN Approach
• Towards Dynamic Fault Tolerance for Hardware-Implemented Artificial Neural Networks: A Deep Learning Approach