Taxonomy of Fault Detection and Diagnosis Methods

According to the technical review of the IAEA [1], “a fault is an abnormal deviation in the condition of at least one piece of equipment or at least one process variable.” Fault detection is a particular case of the Anomaly Detection problem in the data that refers to the monitoring of industrial systems and components such as rotary equipment, heat exchangers, etc. Fault detection is a critical link in the diagnostic circuit since only in case of a deviation detection in the equipment operation, the processes of solving the remaining diagnostic problems (identification or isolation, diagnosis, prognosis or process recovery) are started.

Related literature for an interested reader

The book [3] gives a detailed overview of the most common fault detection and diagnosis techniques for industrial applications.

The book [5] gives a detailed overview of the machine learning methods used for process monitoring and fault diagnosis.

Three part series of papers [6–8] classifies fault diagnosis methods into three general categories (quantitative model-based methods, qualitative model-based methods, and process history based methods) and reviews them.

A review [10] examined several anomaly detection techniques for real-time big data processing: Nearest Neighbours, Bayesian network, support vector machine, decision trees, Random forest, Fuzzy logic, principal component analysis, Ant colony optimization, Hierarchical Temporal Memory. In turn, a review [11] examined Nearest Neighbor, Clustering and Statistical based Anomaly Detection algorithms in detail.

The surveys [2;9] and technical review [1] present methods used for NPP equipment fault detection and diagnosis and current state of fault diagnostics in the nuclear industry.

Finally, the book [4] gives a full review of modern statistical methods for process monitoring, quality control, and improvement.

References

1. https://www.iaea.org/publications/8763/advanced-surveillance-diagnostic-and-prognostic-techniques-in-monitoring-structures-systems-and-components-in-nuclear-power-plants
2. Ayo-Imoru R.M., Cilliers A.C. A survey of the state of condition-based maintenance (CBM) in the nuclear power industry // Annals of Nuclear Energy. — 2018. — feb. — Vol. 112. — Pp. 177–188.
3. Chiang Leo H., Russell Evan L., Braatz Richard D. Fault Detection and Diagnosis in Industrial Systems. — Springer London, 2001.
4. Montgomery Douglas C. Introduction to statistical quality control. — John Wiley & Sons, 2007.
5. Aldrich Chris, Auret Lidia. Unsupervised process monitoring and fault diagnosis with machine learning methods. — Springer, 2013.
6. A review of process fault detection and diagnosis: Part I: Quantitative model-based methods / Venkat Venkatasubramanian, Raghunathan Rengaswamy, Kewen Yin, Surya N. Kavuri // Com- puters & Chemical Engineering. — 2003. — mar. — Vol. 27, no. 3. — Pp. 293–311.
7. Venkatasubramanian Venkat, Rengaswamy Raghunathan, Kavuri Surya N. A review of process fault detection and diagnosis: Part II: Qualitative Models and Search Strategies // Computers & Chemical Engineering. — 2003. — mar. — Vol. 27, no. 3. — Pp. 313–326.
8. A review of process fault detection and diagnosis: Part III: Process history based methods / Venkat Venkatasubramanian, Raghunathan Rengaswamy, Surya N Kavuri, Kewen Yin // Computers & chemical engineering. — 2003. — Vol. 27, no. 3. — Pp. 327–346.
9. Ma Jianping, Jiang Jin. Applications of fault detection and diagnosis methods in nuclear power plants: A review // Progress in Nuclear Energy. — 2011. — apr. — Vol. 53, no. 3. — Pp. 255–266.
10. Real-time big data processing for anomaly detection: A Survey / Riyaz Ahamed Ariyalu- ran Habeeb, Fariza Nasaruddin, Abdullah Gani et al. // International Journal of Information Management. — 2019. — apr. — Vol. 45. — Pp. 289–307.
11. An Investigation of Performance Analysis of Anomaly Detection Techniques for Big Data in SCADA Systems / Mohiuddin Ahmed, Adnan Anwar, Abdun Naser Mahmood et al. // EAI Endorsed Transactions on Industrial Networks and Intelligent Systems. — 2015. — may. — Vol. 2, no. 3. — P. e5.