Enhancing Interpretability of Data-Driven Fault Detection and Diagnosis Methodology with Maintainability Rules in Smart Building Management

Chew, Michael Yit Lin; Yan, Ke

doi:https://doi.org/10.1155/2022/5975816

Journal of Sensors

On this page

Abstract Introduction Data Availability Conflicts of Interest Authors’ Contributions Acknowledgments References Copyright Related Articles

Special Issue

Intelligent Sensing, Monitoring, and Optimization of Advanced Manufacturing Systems

View this Special Issue

Research Article | Open Access

Volume 2022 | Article ID 5975816 | https://doi.org/10.1155/2022/5975816

Enhancing Interpretability of Data-Driven Fault Detection and Diagnosis Methodology with Maintainability Rules in Smart Building Management

Michael Yit Lin Chew¹and Ke Yan¹

Academic Editor: Haidong Shao

Received23 Oct 2021

Accepted19 Nov 2021

Published04 Jan 2022

Abstract

Data-driven fault detection and diagnosis (FDD) methods, referring to the newer generation of artificial intelligence (AI) empowered classification methods, such as data science analysis, big data, Internet of things (IoT), industry 4.0, etc., become increasingly important for facility management in the smart building design and smart city construction. While data-driven FDD methods nowadays outperform the majority of traditional FDD approaches, such as the physically based models and mathematically based models, in terms of both efficiency and accuracy, the interpretability of those methods does not grow significantly. Instead, according to the literature survey, the interpretability of the data-driven FDD methods becomes the main concern and creates barriers for those methods to be adopted in real-world industrial applications. In this study, we reviewed the existing data-driven FDD approaches for building mechanical & electrical engineering (M&E) services faults and discussed the interpretability of the modern data-driven FDD methods. Two data-driven FDD strategies integrating the expert reasoning of the faults were proposed. Lists of expert rules, knowledge of maintainability, international/local standards were concluded for various M&E services, including heating, ventilation air-conditioning (HVAC), plumbing, fire safety, electrical and elevator systems based on surveys of 110 buildings in Singapore. The surveyed results significantly enhance the interpretability of data-driven FDD methods for M&E services, potentially enhance the FDD performance in terms of accuracy and promote the data-driven FDD approaches to real-world facility management practices.

1. Introduction

1.1. Motivation

Aligning with the fast development of artificial intelligence (AI) technology, data-driven fault detection and diagnosis (FDD) plays an essential role in modern smart building maintenance and management systems [1]. However, while the data-driven FDD models are often viewed as black-box models, the interpretability of FDD models hinders the methods to be widely applied to real-world applications [2, 3]. Expert rules and standards are helpful for data-driven FDD methods to be adapted to real-world scenarios. The expert rules and standards not only increase the interpretability level of the data-driven FDD methods but also improve the FDD performance in terms of diagnosis accuracy rates. According to our literature survey, the relevant expert knowledge is considered as a research gap in the field and is highly demanded to detect and diagnose possible faults in building equipment and services [4–6]. In this study, we are interested to concretize the expert knowledge, using maintainability rules and standards for FDD, both regionally and globally, of different building mechanical & electrical engineering (M&E) services, including HVAC systems, plumbing & sanitary, fire safety, electrical and elevators & escalators systems, and their critical components.

1.2. Background

Building fault detection and diagnosis (FDD) methods automatically recognize potential and existing building facility faults based on existing standards, expert knowledge and sensor information, which are important techniques ensuring the safety, efficiency and quality services of building infrastructure and development [7, 8]. According to the different approaches replying to different types of evident information, FDD methods are categorized into data-driven FDD, physical model based FDD and mathematical model based FDD methods [9]. Data-driven FDD builds computational models based on historical sensor data, while different types of building faults are recognized as classes [10]. A physical model based FDD understands the whole building system using physics-based models and usually requires a significant amount of prior knowledge for faults identifications [11, 12]. A mathematical model based FDD methods also requires prior physics knowledge to define a rule space for an inferencing method searching for the corresponding faults [13, 14].

Data-driven fault detection and diagnosis methods represent the next-generation facility management and maintenance techniques adopting modern AI techniques, such as sensor networks [15, 16], data analytics [17], big data [18, 19], machine learning (ML) [20, 21], cybernetic intelligence (CI) [22, 23] and Internet of things (IoT) [24, 25] and etc. For different building infrastructures, such as heating, ventilation air-conditioning (HVAC), plumbing, fire safety, electrical and elevator systems. In the era of big data, smart building and smart city, data-driven FDD usually serves as one of the most important applications utilizing big data and one of the hottest research topics in the fields of smart city and industry 4.0 [26–28].

Compared with traditional physical model based and mathematical model based methods, data-driven FDD methods are usually more efficient, robust and accurate in detecting and diagnosing various building faults, while the machine learning (ML) techniques, such as the neural networks, are constructed for predictive analysis. The ML techniques are generally much more efficient and effective than traditional PM-FDD, MM-FDD and manual classification methods. For example, for HVAC FDD, the existing works showed FDD accuracy rates over 99% for typical chiller faults and 93% for air handling unit (AHU) faults [29–32]. Traditional approaches, such as the sensitivity test, can only achieve accuracy rates close to 83% for chiller faults and around 80% for AHU faults [33–37]. The 10% to 15% improvement on different FDD approaches saves the energy wasted in buildings significantly, enhances the overall building performance and maintains a sustainable environment for building infrastructure maintenance.

However, the interpretability of the data-driven FDD method has always been the problem for data-driven FDD methods and hinders the data-driven FDD techniques to be widely adopted in real-world applications. While the prediction accuracy and efficiency of the data-driven FDD methods improved significantly in recent years, the internal structures of the AI approaches become more complex, resulting in more challenges for model interpretation [38–41]. The data-driven FDD models were also tentatively called black-box models in many existing publications [42–44], which we believe is not accurate. Many data-driven FDD models are indeed interpretable. For example, Yan et al. [45] presented a decision tree model for FDD of air handling units (AHUs). The decision tree structure is interpretable with if-else rules. However, the if-else rules were not easily recognizable for experts deriving standards for AHU maintenance.

It is evident that the expert knowledge, experience, rules, ISO standards and maintenance guidelines are valuable information and can deeply influence the performance of data-driven FDD methods. Zhao et al. [46] demonstrated that the additional expert knowledge inputs can greatly enhance a Bayesian belief network (BBN) data-driven FDD model’s performance by increasing the FDD accuracy for various chiller faults. Li et al. [47] improved [46] by integrated expert knowledge into a diagnostic Bayesian network (DBN) for AHU fault FDD. The reasonings of the AHU FDD were plotted by local casual graphs. The main shortcoming of the works [46, 47] is that the expert knowledge inputs were generally generated based on the authors’ hypotheses.

1.3. Approach

In this study, we reviewed the recent publications on data-driven FDD for building mechanical & electrical engineering (M&E) services, including HVAC, plumbing, fire safety, electrical and elevator systems. Different M&E faults were surveyed over 110 buildings in Singapore, including commercial, hotels, industrial, institutional, clinical and residential buildings, for all three stages of infrastructure management life-cycles, in all design, construction and management stages. The expert knowledge of M&E FDD is converted into maintainability rules and international/local standards in Singapore. It is evident that the conveyed maintainability rules greatly enhance the interpretability of the data-driven FDD approach and potentially improve the diagnosis accuracy.

We propose two data-driven FDD methods integrating the maintainability rules for general facility management in buildings, particularly focusing on FDD. The two specific data-driven FDD methods integrating maintainability rules are 1. data-driven expert rules for decision making in smart building facility FDD; and 2. maintainability rules as inputs for data-driven FDD systems. The actual implementations of the two proposed approaches were omitted, while there were existing implementations such as [46, 47]. The main aim of this study is to specify the expert knowledge pool of M&E FDD using maintainability rules shown in Section 4. The specification greatly enhances the interpretability of the existing M&E FDD methods.

1.4. Contributions

The current work involves the following contributions to the state-of-art. (i)Extending the existing data-driven FDD from HVAC systems to the infrastructure of the whole building. The majority of the existing work of data-driven FDD integrating expert knowledge, e.g., maintainability rules, focuses on HVAC FDD. In this study, we extend the above-mentioned data-driven FDD framework to the whole building system. The targeted facilities include almost all M&E services for smart building design(ii)Identifying expert rules and standards for various M&E faults in buildings. A total of 110 buildings in Singapore, including commercial, hotels, industrial, institutional, clinical and residential buildings were surveyed, over all three stages of infrastructure management life-cycles, in all design, construction and management stages, collecting necessary FDD information based on experts’ knowledge and international/local standards in Singapore. In this way, typical faults for the major M&E equipment are surveyed with detailed experts’ rules and standards stated in tables. This main contribution impacts the literature for data-driven FDD approaches targeting building M&E services significantly(iii)Enhancing the interpretability of the existing data-driven FDD methods for building infrastructure faults. The interpretability of the data-driven FDD methods has been a bottleneck problem for a long time. The surveyed expert rules and standards bridge the gap between theoretical FDD strategy and real-world practices. The interpretability enhancement greatly improves the practicality of the data-driven methods in Industry 4.0 [48] and Construction 4.0 [49].(iv)Potentially improving the diagnosis accuracy of the existing data-driven FDD methods for building infrastructure faults. According to the literature study, such as the works of [46, 47], expert knowledge, e.g., the maintainability rules enhance the FDD performance significantly in terms of accuracy. The diagnosis accuracy improvements are justified by various publications [46, 47, 50–52].

2. Literature Review for Interpretability Study of the Existing M&E Services FDD Methods

Intelligent facilities management is one of the important topics for smart city design, smart building maintenance system development, Industry 4.0 and Construction 4.0. Techniques based on AI and data-driven approaches attract increasing attention from various perspectives. Besides the effectiveness and robustness of the data-driven approaches for data-driven FDD, the shortcomings and issues, such as the interpretability of the data-driven model and the efficiency for data-driven FDD algorithms, were raised in recent years.

Yan et al. [45] introduced a decision tree induction (DTI) based FDD method for detecting and diagnosing AHU faults. The proposed method is data-driven, and interpretable with a post-pruned binary tree structure. The main concern of [45] is that the derived rules do not explicitly map to expert reasoning available in the HVAC system design. Most of the DTI rules were still unreadable from the perspective of HVAC engineers. Mulumba et al. [50] worked on a Kalman filter-based FDD reasonings for AHU faults. The method works for various AHU faults and is also considered a data-driven approach. The shortcoming is again that the Kalman filter rules do not map correspondingly to HVAC experts. Srinivasan et al. [51] showed the importance of explainable AI (XAI) for chiller fault detection systems to gain human trust. Li et al. [52] developed an explainable one-dimensional convolutional neural networks (CNN)-based fault diagnosis method for building HVAC systems.

Besides the interpretability study of FDD for HVAC systems, there are existing data-driven FDD approaches proposed for other M&E service systems. Kumar et al. [53] developed a deep learning detecting defects in sewerage systems. The deep learning structure relies on the CNN for object detection in images. The image processing technology using CNN is more interpretable using expert knowledge compared to other ML techniques. Gonzalez-Jimenez et al. [54] surveyed the existing fault diagnostic methods to examine faults for electric drives and revisited the general workflow using ML techniques for electric drive FDD. The main drawback of the data-driven FDD method as concluded in [54] is the lack of interpretability and the lack of explanations for specific phenomena in every particular electric drive. Gavan et al. [55] proposed to integrate expert rules and data-driven FDD methods to develop a positive energy building in France. The project has a nice workflow chart utilizing expert rules for building data analysis and FDD practices. However, it is an ongoing project and the performance of the proposed workflow is yet to be verified.

All the above-surveyed existing works showed that there are already quite many efforts on integrating expert rules and reasonings into the existing data-driven FDD methods to enhance the interpretability of the methods as well as improve the FDD performance on the classification accuracy for building maintenance problems. However, there still exist gaps between expert rules and data-driven methods, such as neural networks. The gap is mainly from the reasoning of AI and the ordinary reasoning of human beings. The most appropriate matching and fitting using the expert rules with the modern data-driven FDD methods remain unknown and desired further explorations, such as the current study. The current study expands the scope of the FDD methods and greatly enhances the applicable area of data-driven FDD methods in building services.

3. Integrating the Maintainability Rules into Data-Driven FDD for M&E Services

Two types of data-driven FDD strategies are available in general for the concept of next-generation AI-technology integrated smart buildings for facility management and maintenance. The first strategy is named post-caution maintenance. This strategy is widely adopted for modern buildings when expert knowledge of precautions is lacking. Without sufficient rules and guidance in the stages of design and construction, the only option left is monitoring the facilities regularly using physical, mathematical or AI-driven models and detecting potential errors with frequent data analysis. The expert knowledge and rules are added as an additional layer of the ML model for performance enhancement. Existing examples of post-caution maintenance include [46, 47].

The second type of the data-driven FDD strategy is to involve the expert knowledge in the FDD monitoring of the entire life-cycle of all facilities, or precaution maintenance. With the experts in the fields of, e.g., project management, construction, interior design and quantity survey, relevant regional and global standards, such as SS, BS, ISO, EN, AS and ASTM (Table 1) can be adopted in the precaution of potential risks in the early stage of the FDD. However, there are generally gaps between those standards and the real-world maintenance strategy, i.e., lacking clear guidance of different maintenance rules for different elements of the M&E system. The main contribution of this study is providing systematic and comprehensive maintainability rules for all kinds of M&E elements.

In this section, we demonstrate two data-driven FDD strategies with the maintainability rules for precaution and post-caution maintenance, which apply the maintainability rules as the inputs and the knowledge pool, respectively. These two strategies serve as examples of the usage of the maintainability rules listed in the Tables in Section 4.

3.1. Knowledge-Based Rule System Integrated Data-Driven FDD for M&E Faults

A knowledge-based system is a fundamental AI system that makes decisions purely based on rules. A traditional knowledge-based system comprises a large set of if-else rules that builds a decision tree and processes FDD queries efficiently. A semantic of a typical knowledge-based FDD system is shown in Figure 1. The collected sensor data is evaluated by maintainability rules. The evaluation results lead to the various maintenance decisions following a tree-alike structure.

For example, following the escalator maintainability rules stated in Table 2: ‘The landing area of escalators and passenger conveyors should have a surface that provides a secure foothold for a minimum distance of 0.85 m (measured from the root of the comb teeth)’, the measurement data collected from the sensor can be easily evaluated as ‘satisfactory’ or ‘unsatisfactory’. Different evaluation results will arrive to different decisions for automated maintenance.

A knowledge-based FDD system integrating traditional data-driven FDD framework treats the knowledge pool (Figure 1) as an expert system [9, 56]., where if-else rules are derived from maintainability rules. Following the existing rule-based system structures proposed in the related fields, such as [45, 57, 58], decisions for labeling various faults can be reached. The accuracy and performance of such FDD systems depend on the precision and reasonings of the rules. Compared with the existing rule-based systems, the maintainability rules proposed in this study are more precise and reasonable, consequently providing better results in terms of diagnostic accuracy.

3.2. Maintainability Rules as Inputs for Data-Driven FDD Systems

Data-driven FDD applied machine learning techniques to sensor data and performs automated classification with a pre-defined training process on the collected data. A typical data-driven FDD process is shown in Figure 2, where historical data containing both normal operational and faulty conditional data is received by the machine learning (ML) models. Two particular ML models are trained. The binary ML model handles the fault detection for facilities management, which classifies the future sensor data into normal or faulty classes. The multi-class ML model handles the fault diagnosis part, which classifies the faulty sensor data into different types of faults.

Traditional FDD methods, as shown in Figure 2, assume completely no background knowledge of the maintainability of the facilities. The maintainability rules that we proposed in this study provide a great opportunity to improve the existing FDD approaches. The simplest way of extending the current FDD framework with the maintainability rules is to treat them as inputs for the ML models. We formalize the proposed extension of the existing FDD framework in Figure 3.

In Figure 3, the traditional FDD framework has been improved by adding maintainability rules as inputs for both training and testing phases. Since ML models, in general, do not require background knowledge for classifications, the maintainability rules are served as additional inputs for both training and testing of the ML models. The maintainability rules have the potentials of enhancing the interpretation capability of the ML models as well as the prediction performance.

A concrete example of the proposed framework shown in Figure 3 is the three-layer Bayesian Belief Network (BBN) adapting the maintainability rules as an additional layer for FDD. The three-layer BBN is a three-layer neural network, calculating the probabilities of label assignment based on evidence and conditional probability. The details of the BBN construction can be found in [46, 47]. The internal structure of the BBN is illustrated in Figure 4, where expert knowledge is interpreted using maintainability rules as introduced in the Introduction Section (Section 1) and Section 4. For prediction probabilities calculated by neural networks, the maintainability rules provide evidence that influence the probability calculation. Therefore, the FDD accuracy will be improved significantly.

4. Maintainability Rules Study for Facility Management in M&E Services

In this section, we summarize the maintainability rules following the expert knowledge of typical components in M&E services, namely, HVAC system, plumbing and sanitary system, fire safety, electrical system and elevator & escalator system collected through survey and interview results over 110 buildings in Singapore, including commercial, hotels, industrial, institutional, clinical and residential buildings. The maintainability rules summarize the preventive checklist based on expert knowledge as well as standards regionally or globally in all design, construction and operational stages of buildings. The maintainability rules are useful serving as the knowledge pool for the post-caution FDD approach or as the additional maintainability layer for a precaution FDD approach, as explained in Section 3.

The maintainability rules for the chiller plant, the cooling tower, the air handling unit (AHU) and the air distribution, terminal system of the HVAC system are summarized in Tables 3–6, respectively. The maintainability guidance for general pumping issues, the water supply system and the water tank of the plumbing and sanitary system ae summarized in Tables 7–9, respectively. The maintainability issues for the fire detection, the fire hydrant system, the sprinkler system and the fire extinguishers of the fire safety (Table 10) are listed in Tables 11–13, respectively. The maintainability rules for the switchgear, the standby generator, the artificial lighting, the lightning protection system (LPS) and earthing are summarized in Tables 14–17, respectively. The general rules for the elevators and escalators, common faults for the elevators and escalators, the elevator safety, energy efficiency for the elevators and escalators and the maintenance for escalators, in general, are summarized in Tables 18–21, respectively.

The details of the regional (Singapore-based) and global standards, such as SS, BS, ISO, EN, AS and ASTM, are listed in Table 1.

5. Conclusions, Limitation & Future Works

Maintainability rules for M&E systems based on the survey and interview results of 110 buildings including commercial, hotels, industrial, institutional, healthcare and residential buildings are summarized. The maintainability rules are useful to be integrated into the existing data-driven FDD approaches for 1) an extension of the existing FDD algorithm to all M&E facilities in buildings, 2) enhancing the interpretability of the existing AI models and 3) improving the performances of the AI models. In Section 3, we demonstrate two data-driven FDD strategies integrating the maintainability rules, including 1) data-driven expert rules for decision making in smart building facility FDD; and 2) maintainability rules as inputs for data-driven FDD systems.

Based on the literature study, the surveyed maintainability rules will greatly enhance the interpretability of the existing data-driven FDD methods for M&E services and consequently promote the FDD methods to other building facilities and to other industrial areas, such as the Industry 4.0 evolution solutions. Furthermore, existing works show that the expert knowledge potentially improves the data-driven FDD results by adding the rules to the machine learning models, such as the decision trees.

The limitation of this study includes not showing the actual implementation of the maintainability rules integrated FDD framework, which we believe is a repetitive work to the existing publications. The main contribution of this study is first, to further extend the existing studies and concretize the maintainability rules that are used in existing interpretable data-driven FDD methods based on expert knowledge and existing standards. The second main contribution is to extend the existing FDD methods to a broader scope of facility management.

Future study of this work includes the experiments on the accuracy and efficiency improvement on existing BMS system adding the maintainability rules for additional supports as well as a wider the scope of applications for maintainability rules in smart city design.

Data Availability

The research data used in this study is confidential and only accessible internally for employees of National University of Singapore.

Conflicts of Interest

The authors declare no conflict of interest.

Authors’ Contributions

Conceptualization, M.Y.L.C.; methodology, M.Y.L.C.; investigation, M.Y.L.C.; resources, M.Y.L.C.; data curation, K.Y.; writing—original draft preparation, M.Y.L.C.& K.Y.; writing—review and editing, M.Y.L.C.& K.Y.; visualization, M.Y.L.C.& K.Y.; supervision, M.Y.L.C.; project administration, M.Y.L.C.; funding acquisition, K.Y.

Acknowledgments

This work was supported by the faculty research grant of the National University of Singapore under grant number R-296-000-208-133 (K.Y.).

References

A. Dexter and J. Pakanen, Eds.“Demonstrating automated fault detection and diagnosis methods in real buildings,” in Technical Research Centre of Finland (VTT), A. Dexter and J. Pakanen, Eds., Finland, 2001.
View at: Google Scholar
H. C. Peitsman and V. E. Bakker, Application of Black-Box Models to HVAC Systems for Fault Detection, vol. No. CONF-960254-, American Society of Heating, Refrigerating and Air-Conditioning Engineers, Inc., Atlanta, GA (United States), 1996.
D. Jacob, S. Dietz, S. Komhard, C. Neumann, and S. Herkel, “Black-box models for fault detection and performance monitoring of buildings,” Journal of Building Performance Simulation, vol. 3, no. 1, pp. 53–62, 2010.
View at: Publisher Site | Google Scholar
Y. Zhao, T. Li, X. Zhang, and C. Zhang, “Artificial intelligence-based fault detection and diagnosis methods for building energy systems: advantages, challenges and the future,” Renewable and Sustainable Energy Reviews, vol. 109, pp. 85–101, 2019.
View at: Publisher Site | Google Scholar
X. Dai and Z. Gao, “From model, signal to knowledge: a data-driven perspective of fault detection and diagnosis,” IEEE Transactions on Industrial Informatics, vol. 9, no. 4, pp. 2226–2238, 2013.
View at: Publisher Site | Google Scholar
M. Galagedarage Don and F. Khan, “Dynamic process fault detection and diagnosis based on a combined approach of hidden Markov and Bayesian network model,” Chemical Engineering Science, vol. 201, pp. 82–96, 2019.
View at: Publisher Site | Google Scholar
M. S. Mirnaghi and F. Haghighat, “Fault detection and diagnosis of large-scale HVAC systems in buildings using data-driven methods: a comprehensive review,” Energy and Buildings, vol. 229, article 110492, 2020.
View at: Publisher Site | Google Scholar
Z. Shi and W. O'Brien, “Development and implementation of automated fault detection and diagnostics for building systems: a review,” Automation in Construction, vol. 104, pp. 215–229, 2019.
View at: Publisher Site | Google Scholar
S. Katipamula and M. R. Brambley, “Review Article: Methods for Fault Detection, Diagnostics, and Prognostics for Building Systems—A Review, Part I,” Hvac&R Research, vol. 11, no. 1, pp. 3–25, 2005.
View at: Publisher Site | Google Scholar
K. Yan, W. Shen, T. Mulumba, and A. Afshari, “ARX model based fault detection and diagnosis for chillers using support vector machines,” Energy and Buildings, vol. 81, pp. 287–295, 2014.
View at: Publisher Site | Google Scholar
A. Chong and K. P. Lam, “Uncertainty analysis and parameter estimation of HVAC systems in building energy models,” in Proceedings of the 14th IBPSA building simulation conference, vol. 1, no. 1.4, p. 3, Hyderabad, India, 2015.
View at: Google Scholar
A. Chong and K. Menberg, “Guidelines for the Bayesian calibration of building energy models,” Energy and Buildings, vol. 174, pp. 527–547, 2018.
View at: Publisher Site | Google Scholar
W. Kim and S. Katipamula, “A review of fault detection and diagnostics methods for building systems,” Science and Technology for the Built Environment, vol. 24, no. 1, pp. 3–21, 2018.
View at: Publisher Site | Google Scholar
G. Hackmann, W. Guo, G. Yan, Z. Sun, C. Lu, and S. Dyke, “Cyber-physical codesign of distributed structural health monitoring with wireless sensor networks,” IEEE Transactions on Parallel and Distributed Systems, vol. 25, no. 1, pp. 63–72, 2014.
View at: Google Scholar
Y. Cao, X. Zhou, and K. Yan, “Deep Learning Neural Network Model for Tunnel Ground Surface Settlement Prediction Based on Sensor Data,” Mathematical Problems in Engineering, vol. 2021, Article ID 9488892, 14 pages, 2021.
View at: Publisher Site | Google Scholar
H. Kramer, G. Lin, C. Curtin, E. Crowe, and J. Granderson, “Building analytics and monitoring-based commissioning: industry practice, costs, and savings,” Energy Efficiency, vol. 13, no. 3, pp. 537–549, 2020.
View at: Publisher Site | Google Scholar
Y. Shu, L. Ming, F. Cheng, Z. Zhang, and J. Zhao, “Abnormal situation management: challenges and opportunities in the big data era,” Computers & Chemical Engineering, vol. 91, pp. 104–113, 2016.
View at: Publisher Site | Google Scholar
S. Lazarova-Molnar, H. R. Shaker, and N. Mohamed, “Fault detection and diagnosis for smart buildings: state of the art, trends and challenges,” in 2016 3rd MEC international conference on big data and Smart City (ICBDSC), pp. 1–7, Muscat, Oman, 2016, March.
View at: Google Scholar
C. Yang, W. Shen, Q. Chen, and B. Gunay, “A practical solution for HVAC prognostics: failure mode and effects analysis in building maintenance,” Journal of Building Engineering, vol. 15, pp. 26–32, 2018.
View at: Publisher Site | Google Scholar
K. Yan, “Chiller Fault Detection and Diagnosis with Anomaly Detective Generative Adversarial Network,” Building and Environment, vol. 201, article 107982, 2021.
View at: Publisher Site | Google Scholar
Z. Chen, “Facilities intelligence and evaluation: a multi-criteria assessment approach,” Energy and Buildings, vol. 42, no. 5, pp. 728–734, 2010.
View at: Publisher Site | Google Scholar
K. Yan, L. Liu, Y. Xiang, and Q. Jin, “Guest editorial: AI and machine learning solution cyber intelligence technologies: new methodologies and applications,” IEEE Transactions on Industrial Informatics, vol. 16, no. 10, pp. 6626–6631, 2020.
View at: Publisher Site | Google Scholar
R. Shohet, M. S. Kandil, Y. Wang, and J. J. McArthur, “Fault detection for non-condensing boilers using simulated building automation system sensor data,” Advanced Engineering Informatics, vol. 46, article 101176, 2020.
View at: Publisher Site | Google Scholar
S. A. Gbadamosi, G. P. Hancke, and A. M. Abu-Mahfouz, “Building upon NB-IoT networks: a roadmap towards 5G new radio networks,” IEEE Access, vol. 8, pp. 188641–188672, 2020.
View at: Publisher Site | Google Scholar
M. Dey, S. P. Rana, and S. Dudley, “A case study based approach for remote fault detection using multi-level machine learning in a smart building,” Smart Cities, vol. 3, no. 2, pp. 401–419, 2020.
View at: Publisher Site | Google Scholar
M. Dey, S. P. Rana, and S. Dudley, “Smart building creation in large scale HVAC environments through automated fault detection and diagnosis,” Future Generation Computer Systems, vol. 108, pp. 950–966, 2020.
View at: Publisher Site | Google Scholar
N. Mohamed, J. Al-Jaroodi, and S. Lazarova-Molnar, “Industry 4.0: Opportunities for Enhancing Energy Efficiency in Smart Factories,” in 2019 IEEE International Systems Conference (SysCon), pp. 1–7, Orlando, Florida, USA, 2019.
View at: Google Scholar
T. Mulumba, A. Afshari, K. Yan, W. Shen, and L. K. Norford, “Robust model-based fault diagnosis for air handling units,” Energy and Buildings, vol. 86, pp. 698–707, 2015.
View at: Publisher Site | Google Scholar
K. Yan, Z. Ji, H. Lu, J. Huang, W. Shen, and Y. Xue, “Fast and accurate classification of time series data using extended ELM: application in fault diagnosis of air handling units,” IEEE Transactions on Systems, Man, and Cybernetics: Systems, vol. 49, no. 7, pp. 1349–1356, 2017.
View at: Google Scholar
K. Yan, Z. Ji, and W. Shen, “Online fault detection methods for chillers combining extended kalman filter and recursive one-class SVM,” Neurocomputing, vol. 228, pp. 205–212, 2017.
View at: Publisher Site | Google Scholar
K. Yan, L. Ma, Y. Dai, W. Shen, Z. Ji, and D. Xie, “Selection des fonctions sequentielles et sensibles aux couts pour la detection et le diagnostic des defaillances du refroidisseur,” International Journal of Refrigeration, vol. 86, pp. 401–409, 2018.
View at: Publisher Site | Google Scholar
S. Bendapudi and J. E. Braun, “A Review of Literature on Dynamic Models of Vapor Compression Equipment. Final report for ASHRAE RP-1043,” in American Society of Heating, Refrigerating and Air-Conditioning Engineers, Atlanta, 2002.
View at: Google Scholar
J. M. House, H. Vaezi-Nejad, and J. M. Whitcomb, “An expert rule set for fault detection in air-handling units/discussion,” ASHRAE Transactions, vol. 107, p. 858, 2001.
View at: Google Scholar
W. Y. Lee, J. M. House, C. Park, and G. E. Kelly, “Fault diagnosis of an air-handling unit using artificial neural networks,” Transactions-American society of heating refrigerating and air conditioning engineers, vol. 102, pp. 540–549, 1996.
View at: Google Scholar
S. Li, J. Wen, X. Zhou, and C. J. Klaassen, “Development and validation of a dynamic air handling unit model, part 2,” ASHRAE Transactions, vol. 116, no. 1, pp. 57–73, 2010.
View at: Google Scholar
S. Pourarian, J. Wen, D. Veronica, A. Pertzborn, X. Zhou, and R. Liu, “A tool for evaluating fault detection and diagnostic methods for fan coil units,” Energy and Buildings, vol. 136, pp. 151–160, 2017.
View at: Publisher Site | Google Scholar
C. Zhong, K. Yan, Y. Dai, N. Jin, and B. Lou, “Energy efficiency solutions for buildings: automated fault diagnosis of air handling units using generative adversarial networks,” Energies, vol. 12, no. 3, p. 527, 2019.
View at: Publisher Site | Google Scholar
A. Ragab, M. El-Koujok, B. Poulin, M. Amazouz, and S. Yacout, “Fault diagnosis in industrial chemical processes using interpretable patterns based on logical analysis of data,” Expert Systems with Applications, vol. 95, pp. 368–383, 2018.
View at: Publisher Site | Google Scholar
A. Ragab, M. El Koujok, H. Ghezzaz, M. Amazouz, M. S. Ouali, and S. Yacout, “Deep understanding in industrial processes by complementing human expertise with interpretable patterns of machine learning,” Expert Systems with Applications, vol. 122, pp. 388–405, 2019.
View at: Publisher Site | Google Scholar
J. Liu, G. Li, B. Liu, K. Li, and H. Chen, “Knowledge discovery of data-driven-based fault diagnostics for building energy systems: a case study of the building variable refrigerant flow system,” Energy, vol. 174, pp. 873–885, 2019.
View at: Publisher Site | Google Scholar
B. Butzin, F. Golatowski, and D. Timmermann, “A Survey on Information Modeling and Ontologies in Building Automation,” in IECON 2017-43rd Annual Conference of the IEEE Industrial Electronics Society, pp. 8615–8621, Beijing, China, 2017.
View at: Google Scholar
A. Kroll, “Grey-box models: concepts and application,” New frontiers in computational intelligence and its applications, vol. 57, pp. 42–51, 2000.
View at: Google Scholar
H. Yoshida and S. Kumar, “Development of ARX model based off-line FDD technique for energy efficient buildings,” Renewable Energy, vol. 22, no. 1-3, pp. 53–59, 2001.
View at: Publisher Site | Google Scholar
A. Costa, M. M. Keane, J. I. Torrens, and E. Corry, “Building operation and energy performance: monitoring, analysis and optimisation toolkit,” Applied Energy, vol. 101, pp. 310–316, 2013.
View at: Publisher Site | Google Scholar
R. Yan, Z. Ma, Y. Zhao, and G. Kokogiannakis, “A decision tree based data-driven diagnostic strategy for air handling units,” Energy and Buildings, vol. 133, pp. 37–45, 2016.
View at: Publisher Site | Google Scholar
Y. Zhao, F. Xiao, and S. Wang, “An intelligent chiller fault detection and diagnosis methodology using Bayesian belief network,” Energy and Buildings, vol. 57, pp. 278–288, 2013.
View at: Publisher Site | Google Scholar
T. Li, Y. Zhao, C. Zhang, J. Luo, and X. Zhang, “A knowledge-guided and data-driven method for building HVAC systems fault diagnosis,” Building and Environment, vol. 198, article 107850, 2021.
View at: Publisher Site | Google Scholar
S. Vaidya, P. Ambad, and S. Bhosle, “Industry 4.0 - A Glimpse,” Procedia manufacturing, vol. 20, pp. 233–238, 2018.
View at: Publisher Site | Google Scholar
E. Forcael, I. Ferrari, A. Opazo-Vega, and J. A. Pulido-Arcas, “Construction 4.0: a literature review,” Sustainability, vol. 12, no. 22, p. 9755, 2020.
View at: Publisher Site | Google Scholar
T. Mulumba, A. Afshari, and L. Friedrich, “Kalman filter-based fault detection and diagnosis for air handling units,” in Proceedings of the 15th International Refrigeration and Air Conditioning Conference, West Lafayette, Indiana, USA, 2014.
View at: Google Scholar
S. Srinivasan, P. Arjunan, B. Jin, A. L. Sangiovanni-Vincentelli, Z. Sultan, and K. Poolla, “Explainable AI for chiller fault-detection systems: gaining human trust,” Computer, vol. 54, no. 10, pp. 60–68, 2021.
View at: Publisher Site | Google Scholar
G. Li, Q. Yao, C. Fan et al., “An explainable one-dimensional convolutional neural networks based fault diagnosis method for building heating, ventilation and air conditioning systems,” Building and Environment, vol. 203, article 108057, 2021.
View at: Publisher Site | Google Scholar
S. S. Kumar, D. Abraham, and M. Rosenthal, “Leveraging Visualization Techniques to Develop Improved Deep Neural Network Architecture for Sewer Defect Identification,” in Construction Research Congress 2020: Infrastructure Systems and Sustainability, pp. 827–835, American Society of Civil Engineers, Reston, VA, 2020.
View at: Google Scholar
D. Gonzalez-Jimenez, J. del-Olmo, J. Poza, F. Garramiola, and P. Madina, “Data-driven fault diagnosis for electric drives: a review,” Sensors, vol. 21, no. 12, p. 4024, 2021.
View at: Publisher Site | Google Scholar
V. Gavan, A. Perehinec, S. Agapoff, and S. Derouineau, “Rule based Fault Detection & Diagnosis for high performance buildings: application to a positive energy building in France,” in 12th REHVA World Congress-CLIMA 2016, Aalborg, Denmark, 2016.
View at: Google Scholar
T. M. Rossi and J. E. Braun, “A statistical, rule-based fault detection and diagnostic method for vapor compression air conditioners,” Hvac&R Research, vol. 3, no. 1, pp. 19–37, 1997.
View at: Publisher Site | Google Scholar
F. Arévalo, C. Tito, M. R. Diprasetya, and A. Schwung, “Fault Detection Assessment Using an Extended FMEA and a Rule-Based Expert System,” in 2019 IEEE 17th International Conference on Industrial Informatics (INDIN), vol. 1, pp. 740–745, Helsinki-Espoo, Finland, 2019.
View at: Google Scholar
S. Frank, G. Lin, X. Jin, R. Singla, A. Farthing, and J. Granderson, “A performance evaluation framework for building fault detection and diagnosis algorithms,” Energy and Buildings, vol. 192, pp. 84–92, 2019.
View at: Publisher Site | Google Scholar

Copyright

Copyright © 2022 Michael Yit Lin Chew and Ke Yan. This is an open access article distributed under the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

PDF Download Citation

Download other formats

Order printed copies

Views

1043

Downloads

1079

Citations