Support vector ensemble for incipient fault diagnosis in nuclear plant components

Abiodun Ayodeji,Yong-kuo Liu 한국원자력학회 2018 Nuclear Engineering and Technology Vol.50 No.8

The randomness and incipient nature of certain faults in reactor systems warrant a robust and dynamicdetection mechanism. Existing models and methods for fault diagnosis using different mathematical/statistical inferences lack incipient and novel faults detection capability. To this end, we propose a faultdiagnosis method that utilizes the flexibility of data-driven Support Vector Machine (SVM) forcomponent-level fault diagnosis. The technique integrates separately-built, separately-trained, specializedSVM modules capable of component-level fault diagnosis into a coherent intelligent system, witheach SVM module monitoring sub-units of the reactor coolant system. To evaluate the model, marginalfaults selected from the failure mode and effect analysis (FMEA) are simulated in the steam generator andpressure boundary of the Chinese CNP300 PWR (Qinshan I NPP) reactor coolant system, using a bestestimatethermal-hydraulic code, RELAP5/SCDAP Mod4.0. Multiclass SVM model is trained withcomponent level parameters that represent the steady state and selected faults in the components. Foroptimization purposes, we considered and compared the performances of different multiclass models inMATLAB, using different coding matrices, as well as different kernel functions on the representative dataderived from the simulation of Qinshan I NPP. An optimum predictive model - the Error CorrectingOutput Code (ECOC) with TenaryComplete coding matrix - was obtained from experiments, and utilizedto diagnose the incipient faults. Some of the important diagnostic results and heuristic model evaluationmethods are presented in this paper

A new perspective towards the development of robust data-driven intrusion detection for industrial control systems

Ayodeji, Abiodun,Liu, Yong-kuo,Chao, Nan,Yang, Li-qun Korean Nuclear Society 2020 Nuclear Engineering and Technology Vol.52 No.12

Most of the machine learning-based intrusion detection tools developed for Industrial Control Systems (ICS) are trained on network packet captures, and they rely on monitoring network layer traffic alone for intrusion detection. This approach produces weak intrusion detection systems, as ICS cyber-attacks have a real and significant impact on the process variables. A limited number of researchers consider integrating process measurements. However, in complex systems, process variable changes could result from different combinations of abnormal occurrences. This paper examines recent advances in intrusion detection algorithms, their limitations, challenges and the status of their application in critical infrastructures. We also introduce the discussion on the similarities and conflicts observed in the development of machine learning tools and techniques for fault diagnosis and cybersecurity in the protection of complex systems and the need to establish a clear difference between them. As a case study, we discuss special characteristics in nuclear power control systems and the factors that constraint the direct integration of security algorithms. Moreover, we discuss data reliability issues and present references and direct URL to recent open-source data repositories to aid researchers in developing data-driven ICS intrusion detection systems.

A multi-layer approach to DN 50 electric valve fault diagnosis using shallow-deep intelligent models

Liu, Yong-kuo,Zhou, Wen,Ayodeji, Abiodun,Zhou, Xin-qiu,Peng, Min-jun,Chao, Nan Korean Nuclear Society 2021 Nuclear Engineering and Technology Vol.53 No.1

Timely fault identification is important for safe and reliable operation of the electric valve system. Many research works have utilized different data-driven approach for fault diagnosis in complex systems. However, they do not consider specific characteristics of critical control components such as electric valves. This work presents an integrated shallow-deep fault diagnostic model, developed based on signals extracted from DN50 electric valve. First, the local optimal issue of particle swarm optimization algorithm is solved by optimizing the weight search capability, the particle speed, and position update strategy. Then, to develop a shallow diagnostic model, the modified particle swarm algorithm is combined with support vector machine to form a hybrid improved particle swarm-support vector machine (IPs-SVM). To decouple the influence of the background noise, the wavelet packet transform method is used to reconstruct the vibration signal. Thereafter, the IPs-SVM is used to classify phase imbalance and damaged valve faults, and the performance was evaluated against other models developed using the conventional SVM and particle swarm optimized SVM. Secondly, three different deep belief network (DBN) models are developed, using different acoustic signal structures: raw signal, wavelet transformed signal and time-series (sequential) signal. The models are developed to estimate internal leakage sizes in the electric valve. The predictive performance of the DBN and the evaluation results of the proposed IPs-SVM are also presented in this paper.

Path planning in nuclear facility decommissioning: Research status, challenges, and opportunities

Adibeli, Justina Onyinyechukwu,Liu, Yong-kuo,Ayodeji, Abiodun,Awodi, Ngbede Junior Korean Nuclear Society 2021 Nuclear Engineering and Technology Vol.53 No.11

During nuclear facility decommissioning, workers are continuously exposed to high-level radiation. Hence, adequate path planning is critical to protect workers from unnecessary radiation exposure. This work discusses recent development in radioactive path planning and the algorithms recommended for the task. Specifically, we review the conventional methods for nuclear decommissioning path planning, analyze the techniques utilized in developing algorithms, and enumerate the decision factors that should be considered to optimize path planning algorithms. As a major contribution, we present the quantitative performance comparison of different algorithms utilized in solving path planning problems in nuclear decommissioning and highlight their merits and drawbacks. Also, we discuss techniques and critical consideration necessary for efficient application of robots and robotic path planning algorithms in nuclear facility decommissioning. Moreover, we analyze the influence of obstacles and the environmental/radioactive source dynamics on algorithms' efficiency. Finally, we recommend future research focus and highlight critical improvements required for the existing approaches towards a safer and cost-effective nuclear-decommissioning project.

Radioactive gas diffusion simulation and inhaled effective dose evaluation during nuclear decommissioning

Li-qun Yang,Yong-kuo Liu,Min-jun Peng,Abiodun Ayodeji,Zhi-tao Chen,Ze-yu Long 한국원자력학회 2022 Nuclear Engineering and Technology Vol.54 No.1

During the decommissioning of the nuclear facilities, the radioactive gases in pressure vessels may leakdue to the demolition operations. The decommissioning site has large space, slow air circulation, andmany large nuclear facilities, which increase the difficulty of workers' inhalation exposure assessment. Inorder to dynamically evaluate the activity distribution of radionuclides and the committed effective dosefrom inhalation in nuclear decommissioning environment, an inhalation exposure assessment methodbased on the modified eddy-diffusion model and the inhaled dose conversion factor is proposed in thispaper. The method takes into account the influence of building, facilities, exhaust ducts, etc. on thedistribution of radioactive gases, and can evaluate the influence of radioactive gases diffusion on workersduring the decommissioning of nuclear facilities.

DL-RRT* Algorithm for Least Dose Path Re-planning in Dynamic Radioactive Environments

Nan Chao,Yong-kuo Liu,HONG XIA,Min-jun Peng,Abiodun Ayodeji 한국원자력학회 2019 Nuclear Engineering and Technology Vol.51 No.3

One of the most challenging safety precautions for workers in dynamic, radioactive environments isavoiding radiation sources and sustaining low exposure. This paper presents a sampling-based algorithm,DL-RRT*, for minimum dose walk-path re-planning in radioactive environments, expedient for occupationalworkers in nuclear facilities to avoid unnecessary radiation exposure. The method combines theprinciple of random tree star (RRT*) and D* Lite, and uses the expansion strength of grid search strategyfrom D* Lite to quickly find a high-quality initial path to accelerate convergence rate in RRT*. The algorithminherits probabilistic completeness and asymptotic optimality from RRT* to refine the existingpaths continually by sampling the search-graph obtained from the grid search process. It can not only beapplied to continuous cost spaces, but also make full use of the last planning information to avoid globalre-planning, so as to improve the efficiency of path planning in frequently changing environments. Theeffectiveness and superiority of the proposed method was verified by simulating radiation field undervarying obstacles and radioactive environments, and the results were compared with RRT* algorithmoutput.

An intelligent hybrid methodology of on-line system-level fault diagnosis for nuclear power plant

Min-jun Peng,Hang Wang,Shan-shan Chen,Genglei Xia,Yong-kuo Liu,Xu Yang,Abiodun Ayodeji 한국원자력학회 2018 Nuclear Engineering and Technology Vol.50 No.3

To assist operators to properly assess the current situation of the plant, accurate fault diagnosis methodologyshould be available and used. A reliable fault diagnosis method is beneficial for the safety ofnuclear power plants. The major idea proposed in this work is integrating the merits of different faultdiagnosis methodologies to offset their obvious disadvantages and enhance the accuracy and credibilityof on-line fault diagnosis. This methodology uses the principle component analysis-based model andmulti-flow model to diagnose fault type. To ensure the accuracy of results from the multi-flow model, amechanical simulation model is implemented to do the quantitative calculation. More significantly,mechanism simulation is implemented to provide training data with fault signatures. Furthermore, oneof the distance formulas in similarity measurementdMahalanobis distancedis applied for on-line failuredegree evaluation. The performance of this methodology was evaluated by applying it to the reactorcoolant system of a pressurized water reactor. The results of simulation analysis show the effectivenessand accuracy of this methodology, leading to better confidence of it being integrated as a part of thecomputerized operator support system to assist operators in decision-making.