*
M.Sc. Mechanical engineering jesus.caratar@correounivalle.edu.co. PhD student in
electrical engineering, Universidad del Valle. ORCID: 0000-0003-0581-7143 . Cali, Colombia.
*Mechanical engineer andres.valencia.restrepo@correounivalle.edu.co.
Master Student in system engineering, Universidad del Valle. ORCID: 0000-0001-5154-0195 . Cali, Colombia.
*PhD in Electrical engineering nayiver.gladys.caicedo@correounivalle.edu.co.
Titular professor, Universidad del Valle. ORCID: 0000-0002-8679-7465 . Cali, Colombia.
*PhD in Electrical engineering cristian.chamorro@correounivalle.edu.co. Titular
professor, Universidad del Valle. ORCID: 0000-0002-1792-4555 . Cali, Colombia.
How to cite: J.F Caratar-Chaux, A.M. Valencia, G. Caicedo-Delgado, C. Chamorro “Evaluation of artificial intelligence techniques used in the diagnosis of failures in power plants”. Respuestas, vol. 25, no. 2, pp. 177-189, 2020.
© Peer review is the responsibility of the Universidad Francisco de Paula Santander. This is an article under the license CC BY-NC 4.0.
Received: February 07, 2020
Approved: April 5, 2020.
Abstract
This article presents an evaluation about the research related to the development of computational tools based on artificial intelligence techniques, which focus on the detection and diagnosis of faults in the different processes associated with a power generation plant such as: hydroelectric, thermoelectric and nuclear power plants. Initially, the main techniques of artificial intelligence that allow the construction of intelligent systems in the area of fault diagnosis is described in a general way, techniques such as: fuzzy logic, neural networks, knowledge-based systems and hybrid techniques Subsequently A summary of the research based on each of these techniques is presented. Subsequently, the different articles found for each of the techniques are presented in tables, illustrating the year of publication and the description of the research carried out. The result of this work is the comparison and evaluation of each technique focused on the diagnosis of failures in power plants. The novelty of this work is that it presents an extensive bibliography of the applications of the different intelligent techniques in solving the problem of detection and diagnosis of failure in power plants.
Keywords
intelligent system, neuronal networks, fuzzy logic, Bayesian networks, expert systems, fault diagnosis, power plants.
Resumen:
Este artículo presenta una evaluación de herramientas computacionales basadas en técnicas de inteligencia artificial, las cuales se enfocan en la detección y diagnóstico de fallas en los diferentes procesos asociados a una central de generación de energía tal como: hidroeléctricas, termoeléctricas y centrales nucleares. Inicialmente, se describen de manera general las principales técnicas de inteligencia artificial que permiten la construcción de sistemas inteligentes para el diagnóstico de fallas en centrales eléctricas, se presentan técnicas como: lógica difusa, redes neuronales, sistemas basados en el conocimiento y técnicas hibridas. Posteriormente se presentan en tablas los diferentes artículos encontrados para cada una de las técnicas, ilustrando el año de publicación y una descripción de cada publicación. El resultado de este trabajo es la comparación y evaluación de cada técnica enfocada al diagnóstico de fallas en centrales eléctricas. Lo novedoso de este trabajo, es que presenta una extensa bibliografía de las aplicaciones de las diferentes técnicas inteligentes en la solución del problema de detección y diagnóstico de falla en centrales de generación eléctrica.
Palabras Claves
sistemas inteligentes, redes neuronales, lógica difusa, redes Bayesianas, sistemas expertos, diagnóstico de fallas, plantas eléctricas.
Introduction
According to the report presented by the United Nations in 2015 "Energy for all", around 90% of the developed or emerging regions worldwide have access to electricity supply. In the world, according to the report in 2017 of International Energy Agency, the projected energy demand for 2040 will grow by 30% with respect to the demand consumed in 2017. These demonstrate the importance that represents the supply of energy for today's society, which is why the generating plants must guarantee reliability in the provision of this service [1]. For this, equipment has been developed to monitor, supervise, control and protect the mechanical and electrical components, among which are the PLC (Programmable Logic Controller) and the relays [2]. These devices protect the plant against abnormal operating conditions [3] and its implementation has reduced the number of accidents like Three Mile Island Pennsylvania, in the US on March 28, 1979 [4], [5] where a short circuit in the plant operated at 97% of maximum capacity (1000 Mv), start an escape of radioactive water. The widespread use of these monitoring and control mechanisms allows knowing and storing information about the state of the devices that make up the generation plants [3]. This information allows the implementation of new fault detection and diagnosis techniques based on the use of intelligent systems. Currently, intelligent systems are a branch of artificial intelligence that allows to provide knowledge and experience on a specific domain to a machine, with the aim that it develops a specific activity. Intelligent systems (SI) are developed using techniques such as fuzzy logic, neural networks, genetic algorithms and rule-based systems. Each of these techniques gives the SI different qualities, which is why it is important to know in detail each technique to determine the characteristics that the developed tool will have [6].
This article presents a description of the main techniques used in the creation of SI dedicated to the diagnosis of failures in power plants. At the end of each technique, a table is presented with the investigations found to perform detection and diagnosis of failures based on SI using this technique.
Methodology for the analysis of articles
The methodology of this article initially describes the concept of faults in electric power plants, then describes the concept of intelligent system and finally describes some of the techniques used in the fault diagnosis in power plants.
Faults in Power Plants
According to what is stated in [1], electric generation plants are those that transform a type of base energy into electric energy. Currently, the main generation systems include thermoelectric, hydroelectric, solar and wind power plants. These systems involve in its operation different elements that can be classified as mechanical and electrical, therefore, the failures that may occur in these power plants are mechanical or electrical, where a fault is defined as a state of operation outside the nominal or admissible values of the system.
Therefore, power plants have implemented protection systems, which allows to clear the fault to reduce or mitigate the negative effects that these can cause on the system. Currently, protection systems use devices called PLCs and relays [7] which have different protection functions (PF) previously parameterized [8].
When a fault occurs, the personnel dedicated to the identification and diagnosis of faults analyze the information generated by the PLC and the Relay, handling large volumes of information [7], which makes evident the need for tools based on intelligent systems that can assimilate the experience of human experts and organize the information associated with the fault to speed up diagnostic processes.
Intelligent Systems (SI)
Intelligent systems are computational tools that implement artificial intelligence techniques with the objective of providing knowledge and experience on a specific procedure of a machine [9]. For the development of an intelligent system there are several techniques that can be implemented according to the type of application, for this reason it is of great importance to study each of these techniques before starting the development of an intelligent system [10].
Below are the main techniques used in the development of intelligent systems focused on the detection and diagnosis of faults in power plants.
Fuzzy logic
"Fuzzy logic is a branch of Artificial Intelligence (AI) that allows a computer to analyze real-world information on a scale between the false and the true", this technique emulates human decision making, which reasons into a realm of assertions partially true thanks to common sense [9].
According to [11] fuzzy logic is widely used in applications where there is no mathematical model that describes the system; Therefore, these types of applications can be modeled by adapting the values of the variables to qualitative values and subsequently defining rules that allow make inferences in these applications. It can be seen in Table 1, the main investigations developed in power generation plants using fuzzy logic for the detection and diagnosis of faults.
This technique can be easily implemented in cases where the rules that make up the knowledge base are subjective or difficult to specify, since it allows the use of natural language in its construction [9]. It also has great potential for managing processes in which there is uncertainty in the input data due to the way in which this technique is constituted; however, the need to work with other techniques to improve the performance of the tools developed is evident.
Neural Networks
Neural networks are computation techniques inspired by biological models that imitate the reasoning process of the human brain, for this the neural network takes solved problems as examples to find relationships and build its own network, which classifies information to make decisions [9], [10].
The neural network is made up of three main layers, as can be seen in Figure 2, the input layer where the data is entered, the hidden layer made up of one or several layers depending on the complexity of the network and the output layer, for more information see [22].
Neural networks offer the advantage of adaptive learning, self-organization, fault tolerance, real-time operation, and easy implementation in current technology
In Table 2 presents the research developed for the detection and diagnosis of failures in power plants using intelligent systems based on neural networks. The main quality of this technique is that it offers a solution to cases where it cannot or is very complicated to extract the knowledge of specialists to formulate a knowledge base. The neural networks present as limiting the great amount of information that the network requires for its training, limiting to a great extent its field of application.
Knowledge-based systems
According to [30], knowledge-based systems are defined as those that "contain the erudition of a human specialist versed in a specific field of application". With this technique it is possible to condense the knowledge of human specialists in such a way that it can be accessed and processed by computers, obtaining computer models with the reasoning and problem-solving capacities of human specialists within an established domain.
It can be seen in Figure 3 the generic architecture of a knowledge-based system [9], in which its main components and actors are identified.
Knowledge-based systems are composed mainly of the knowledge base, place where all the information of human specialists is stored, the knowledge of specialists is represented into rules and objects of type IF ... THEN, so that they can be accessed by the inference engine, which relates user inputs that can be in the form of questions with rules and objects to get an answer [10].
Next, Table 3 presents the main research focused on the detection and diagnosis of power plant failures using knowledge-based techniques. However, there is a difficulty in the application of this technique in cases where there is uncertainty in the diagnosis, since the system is only capable of diagnosing the cases of failure that have been contemplated previously. Another difficulty observed in the development of this technique is the difficulty of extracting all the knowledge of human specialists to cover all cases of failure.
Bayesian Networks
Bayesian networks are graphs that represent information through a set of variables and the dependency relations between them. In this way a Bayesian network can represent the probabilistic relationships between a fault and the symptoms of a system.
These networks are built on DAG (Directed Acyclic Graph) graphs as shown in Figure 4, where each node represents a random variable and the edges represent conditional dependencies.
In the investigations presented in Table 4 it is observed that the intelligent systems developed under this technique present a great deal of uncertainty, which allows decisions to be made with partial input information. According to the above, it is considered that Bayesian networks give intelligent systems the ability to determine possible solutions from fragments of information in their entry. As a difficulty in the application of Bayesian networks is establish all the relationships between the different symptoms and types of failure so that the network can make a more approximate diagnosis.
Hybrid techniques
Up to this moment some of the most used techniques in the development of intelligent systems have been presented, each one of these techniques presents qualities and deficiencies during its application, reason why it can be assured that a perfect technique does not exist to develop intelligent systems, this has taken to develop systems that integrate different techniques of artificial intelligence, obtaining as result the denominated hybrid techniques.
In Table 5 is presented a compilation of the research developed with hybrid techniques that have had the greatest impact in recent years to detect and diagnose failures in power plants.
The hybrid techniques allow to increase the field of action of the proposed projects without worrying about the limiting particularities of each technique.
Results and discussion
Table 6 shows a comparative evaluation between the different techniques presented in this work; with this evaluation the reader can compare the advantages and disadvantages of intelligent systems developed by each of the artificial intelligence techniques.
According to the foregoing, it is considered that for the development of an intelligent system to detect and diagnose faults in a power plant, hybrid techniques must be included to comply with a greater number of attributes that allow for a more complete solution, maintaining a simple design and direct application.
Conclusions
This article presents an evaluation of the main techniques used in the development of intelligent systems, seeking to provide the reader with a primary source of information on these techniques and their application in the problem of failure analysis in power plants. On the other hand, the reader will have criteria to determine the technique that best fits the complexity of his own system.
There are multiple intelligent systems techniques that can be used in the solution of different problems, therefore, the characteristics of each problem and the resources available must be analyzed carefully before selecting the technique.
The fuzzy technique acquires greater relevance when working in hybrid form with other techniques of artificial intelligence that allow to relate the interpretation of data that this one offers.
Neural networks are highly applicable in systems where it is very complex to extract the knowledge of human specialists, or this knowledge evolves over time, its main deficiency is in the large number of data required for training, limiting its application to diagnostic systems where there is extensive documentation of symptoms and consequences.
Systems based on knowledge present an excellent performance in a specific domain when the capture of knowledge is done in an adequate way, since its efficiency depends on the quality of the knowledge base.
The systems that present a better behavior are those where hybrid techniques are implemented, since they compensate for the weaknesses of each technique. For example, this work was used by the authors to select the appropriate techniques for the development of an intelligent hybrid system for fault diagnosis in a hydroelectric power station, obtaining the combination of expert systems and Bayesian networks since these techniques fit the available information.
Acknowledgements
Colciencias for its support in the framework of the call for national doctorates 727 of 2015.
Univalle in the framework of the C.I. 105 of 2017.
References
[1] T. Wildi, R. Navarro Salas, and L. M. Ortega González, Máquinas eléctricas y sistemas de potencia, Sexta edic. Mexico: Pearson Educación, 2007.
[2] F. cembranos N. Jesus, Automatismos Electricos Neumaticos E Hidraulicos, Quinta. Thomson, 2008.
[3] A. R. Penin, Sistemas SCADA, 2nd ed. 2007.
[4] J. Roldán Viloria, Fuentes de energía, 1st ed. Madrid España: Paraninfo, 2008.
[5] M. A. A. Larrahondo and A. J. B. Arias, “Desastres en Plantas Nucleares,” Bucaramanga, 2000.
[6] B. Shan, D. Zhao, X. Zhang, F. Guan, and Z. Liu, “Research on relay protection setting expert system for main equipment in power plant,” 1st Int. Conf. Sustain. Power Gener. Supply, SUPERGEN ’09, pp. 1–4, 2009.
[7] H. Arroyo, E. L. Tigre, L. A. Máquina, D. E. E. Como, and D. E. Diseño, “Sistema de automatización, supervisión y control del ‘aprovechamiento hidroeléctrico arroyo el tigre’. la máquina de estado como herramienta de diseño.,” Av. en Energías Renov. y Medio Ambient., vol. 13, pp. 195–201, 2009.
[8] S. Ramirez, Protección de Sistemas Eléctricos, 1st ed. Universidad Nacional de Colombia Manizales, 2005.
[9] P. P. Cruz, Inteligencia artificial con aplicaciones a la ingeniería, 1st ed. Alfaomega, 2011.
[10] G. R. Joseph Giarratano, Sistemas expertos: principios y programación. Thomson, 2001.
[11] R. P. Marcos, “Fundamentos De La Lógica Difusa,” Ing. e Investig., vol. 3, pp. 101–101, 2000.
[12] A. Evsukoff and S. Gentil, “Recurrent neurofuzzy system for fault detection and isolation in nuclear reactors,” Adv. Eng. Informatics, vol. 19, no. 1, pp. 55–66, 2005.
[13] J. Falqueto and M. S. Telles, “Automation of diagnosis of electric power transformers in Itaipu Hydroelectric Plant with a fuzzy expert system,” IEEE Int. Conf. Emerg. Technol. Fact. Autom. ETFA, pp. 577–584, 2007.
[14] J. a. Calderón, G. Zapata, and D. Ovalle, “Algoritmo Neuro Difuso para la Detección y Clasificación de Fallas en Líneas de Transmisión Eléctrica Usando ANFIS,” Rev. Av. en Sist. e Informática, vol. 4, no. 1, 2007.
[15] E. J. Amaya and A. J. Alvares, “SIMPREBAL: An expert system for real-time fault diagnosis of hydrogenerators machinery,” Proc. 15th IEEE Int. Conf. Emerg. Technol. Fact. Autom. ETFA 2010, 2010.
[16] Y. Ting, W. B. Lu, C. H. Chen, and G. K. Wang, “A fuzzy reasoning design for fault detection and diagnosis of a computer-controlled system,” Eng. Appl. Artif. Intell., vol. 21, no. 2, pp. 157–170, 2008.
[17] K. El-kobbah, I. Conference, and A. M. Aboshosha, “Neurofuzzy Computing aided Fault Diagnosis of Nuclear Power Reactors,” 7th Int. Conf. Electr. Eng. ICEENG 2010, pp. 25–27, 2010.
[18] P.-C. Lin, M.-T. Yang, and J.-C. Gu, “Intelligent maintenance model for condition assessment of circuit breakers using fuzzy set theory and evidential reasoning,” IET Gener. Transm. Distrib., vol. 8, no. 7, pp. 1244–1253, 2014.
[19] P. N. Montes Dorantes, J. P. Nieto Gonzalez, and G. M. Mendez, “Fault Detection Systems via a Novel hybrid Methodology for Fuzzy Logic Systems based on Individual base inference and Statistical Process Control,” IEEE Lat. Am. Trans., vol. 12, no. 4, pp. 706–712, 2014.
[20] C. Octavio, H. Morales, J. Pablo, N. González, E. Gabriel, and C. Siller, “Detección y diagnóstico de fallas en sistemas eléctricos de potencia (SEP) combinando lógica difusa, métricas y una red neuronal probabilística,” Research in Computing Science, vol. 72. pp. 47– 59, 2014.
[21] L. Tarba and P. MacH, “Analysis on quality of diagnostic processes in power electrical engineering using combined methods of lead six sigma and fuzzy approaches,” Proc. Int. Conf. - 2016 Conf. Diagnostics Electr. Eng. Diagnostika, 2016.
[22] D. J. Matich, Redes Neuronales: Conceptos Básicos y Aplicaciones. 2001.
[23] E. J. Amaya Simeón, “Aplicação de Técnicas de Inteligência Artificial no Desenvolvimento de um Sistema de Manutenção Baseada em Condição,” Universidade de Brasília, 2008.
[24] A. A. Bittencourt, M. R. De Carvalho, and J. G. R. M. Ieee, “Adaptive Strategies in Power Systems Protection using Artificial Intelligence Techniques,” 2009 15th Int. Conf. Intell. Syst. Appl. to Power Syst., pp. 1–6, 2009.
[25] W. Chine, A. Mellit, V. Lughi, A. Malek, G. Sulligoi, and A. Massi Pavan, “A novel fault diagnosis technique for photovoltaic systems based on artificial neural networks,” Renew. Energy, vol. 90, pp. 501–512, 2016.
[26] A. D. S. Nicolau, J. P. D. S. C. Augusto, and R. Schirru, “Accident diagnosis system based on real-time decision tree expert system,” AIP Conf. Proc., vol. 1836, 2017.
[27] M. Talaat, M. H. Gobran, and M. Wasfi, “A hybrid model of an artificial neural network with thermodynamic model for system diagnosis of electrical power plant gas turbine,” Eng. Appl. Artif. Intell., vol. 68, no. November 2017, pp. 222–235, 2018.
[28] S. Zhao, C. Jiru, and S. Qian, “Research on artificial intelligent nuclear power plant emergency operating guide,” in International Conference on Nuclear Engineering, Proceedings, ICONE, 2019, vol. 2019-May.
[29] H. A. Saeed, H. Wang, M. Peng, A. Hussain, and A. Nawaz, “Online fault monitoring based on deep neural network & sliding window technique,” Prog. Nucl. Energy, vol. 121, no. January, p. 103236, 2020.
[30] R. P. Díez, A. G. Gómez, and N. de A. Martínez, Introducción a la inteligencia artificial: sistemas expertos, redes neuronales artificiales y computación evolutiva. 2001.
[31] S. D. J. McArthur, J. R. McDonald, S. C. Bell, and G. M. Burt, “Expert systems and modelbased reasoning for protection performance analysis,” Artif. Intell. Appl. Power Syst. IEE Colloq., p. 1/1-1/4, 1995.
[32] K. El-Arroudi, D. McGillis, and G. Joos, “A methodology for power system protection design based on an intelligent system approach,” Electr. Comput. …, pp. 1164–1169, 1999.
[33] F. Filippetti, M. Martelli, G. Franceschini, and C. Tassoni, “Development of expert system knowledge base to on-line diagnosis of\nrotor electrical faults of induction motors,” Conf. Rec. 1992 IEEE Ind. Appl. Soc. Annu. Meet., 1992.
[34] T. K. Saha and P. Purkait, “Investigation of an expert system for the condition assessment of transformer insulation based on dielectric response measurements,” IEEE Trans. Power Deliv., vol. 19, no. 3, pp. 1127–1134, 2004.
[35] X. Luo and M. Kezunovic, “An expert system for diagnosis of digital relay operation,” Proc. 13th Int. Conf. Intell. Syst. Appl. to Power Syst. ISAP’05, vol. 2005, pp. 175–180, 2005.
[36] L. Amendola, “Sistemas expertos monitoreo de condiciones en máquinas rotativas,” Valencia, pp. 1–4, 2008.
[37] B. Rodrigo, N. Felipe, C. Aldo, and P. Rodrigo, “Expert fault detection and diagnosis for the refrigeration process of a hydraulic power plant,” Proc. 27th Chinese Control Conf. CCC, pp. 122–126, 2008.
[38] E. J. Amaya and A. J. Alvares, “Expert system for power generation fault diagnosis using hierarchical meta-rules,” Proc. 2012 IEEE 17th Int. Conf. Emerg. Technol. Fact. Autom. (ETFA 2012), pp. 1–8, 2012.
[39] S. Saludes, L. j. de Miguel, and J. R. Perán, “Sistema experto para el mantenimiento predictivo de una central hidroeléctrica,” ResearchGate, no. December 2013, pp. 148–159, 2013.
[40] I. Buaphan and S. Premrudeepreechacharn, “Development of expert system for fault diagnosis of an 8-MW bulb turbine downstream irrigation hydro power plant,” 2017 6th Int. Youth Conf. Energy, IYCE 2017, p. 8003740, 2017.
[41] A. J. Alvares and R. Gudwin, “Integrated system of predictive maintenance and operation of eletronorte based on expert system,” IEEE Lat. Am. Trans., vol. 17, no. 1, pp. 155–166, 2019.
[42] C. Y. Wu X., Guo C., “A new fault diagnosis approach of powe system based on bayesian network and temporal order information,” Proc. CSEE 25, vol. 13, pp. 14–18, 2005.
[43] Z. Yongli, H. Limin, and L. Jinling, “Bayesian networks-Based approach for power systems fault diagnosis,” IEEE Trans. Power Deliv., vol. 21, no. 2, pp. 634–639, 2006.
[44] M. C. Method, “Uncertain Fault,” pp. 1–6, 2006.
[45] Q. Z. Qin Li, Zhi Bin Li, “Research of Power Transformer Fault Diagnosis System Based on Rough Sets and Bayesian Networks,” Adv. Mater. Res., vol. 320, pp. 524–529, 2011.
[46] Y. Zhao, F. Xiao, and S. Wang, “An intelligent chiller fault detection and diagnosis methodology using Bayesian belief network,” vol. 57. pp. 278–288, 2013.
[47] B. Cai et al., “Multi-source information fusion-based fault diagnosis of ground-source heat pump using Bayesian network,” Applied Energy, vol. 114. 2014.
[48] B. Cai, H. Liu, and M. Xie, “A real-time fault diagnosis methodology of complex systems using object-oriented Bayesian networks,” Mech. Syst. Signal Process., vol. 80, pp. 31–44, 2016.
[49] H.-B. Jun and D. Kim, “A Bayesian networkbased approach for fault analysis,” Expert Syst. Appl., vol. 81, pp. 332–348, 2017.
[50] G. Wu, J. Tong, L. Zhang, Y. Zhao, and Z. Duan, “Framework for fault diagnosis with multi-source sensor nodes in nuclear power plants based on a Bayesian network,” Ann. Nucl. Energy, vol. 122, pp. 297–308, 2018.
[51] J. Rohmer and P. Gehl, “Sensitivity analysis of Bayesian networks to parameters of the conditional probability model using a Beta regression approach,” Expert Syst. Appl., vol. 145, 2020.
[52] S. Sarkar, T. Sharma, A. Baral, B. Chatterjee, D. Dey, and S. Chakravorti, “An expert system approach for transformer insulation diagnosis combining conventional diagnostic tests andPDC, RVM data,” IEEE Trans. Dielectr. Electr. Insul., vol. 21, no. 2, pp. 882–891, 2014.
[53] Y. Liu, C. Xie, M. Peng, and S. Ling, “Improvement of fault diagnosis efficiency in nuclear power plants using hybrid intelligence approach,” Prog. Nucl. Energy, vol. 76, pp. 122– 136, Sep. 2014.
[54] G. Muller and D. Falcão, “A Fuzzy Knowledge-Based System to Assess the Impact of Demand Response on the Long-Term Demand of Electricity: Application to the Brazilian Interconnected Power System,” 2019 IEEE PES Conf. Innov. Smart Grid Technol. ISGT Lat. Am. 2019, p. 8894988, 2019.
[55] F. B. Ismail Alnaimi, R. I. Bin Ismail, P. J. Ker, and S. K. B. Wahidin, “Development of intelligent early warning system for steam turbine,” J. Eng. Sci. Technol., vol. 14, no. 2, pp. 844–858, 2019.
[56] H. Wang, M. jun Peng, J. Wesley Hines, G. yang Zheng, Y. kuo Liu, and B. R. Upadhyaya, “A hybrid fault diagnosis methodology with support vector machine and improved particle swarm optimization for nuclear power plants,” ISA Trans., vol. 95, pp. 358–371, 2019.
[57] Y. Zhao, J. Tong, L. Zhang, and G. Wu, “Diagnosis of operational failures and ondemand failures in nuclear power plants: An approach based on dynamic Bayesian networks,” Ann. Nucl. Energy, vol. 138, p. 107181, 2020.