Written by Amir Darbandsari, Mahdi Nozarian, and Alireza Fereidunian
Critical infrastructures of smart cities, such as those that support emergency services and power delivery, should be properly maintained to minimize the risk of failure. Analytical utilization of operational data gathered from various sources and sensing networks are stored by computerized maintenance management systems (CMMSs). When properly implemented, CMMS will drastically improve performance of maintenance efficiency. Data-driven maintenance management of smart cities’ critical infrastructures is thus introduced in this paper with the primary objective of making critical infrastructures more reliable. This paper also provides a demonstrative case study of an energy distribution system's maintenance management scheduling project to illustrate the implementation of CMMS.
Maintenance regimes are expected to ensure the availability and functionality of infrastructure assets at the lowest cost possible while not affecting operational reliability. Maintenance management relates to the planning of inspection, monitoring, repair, and replacement of equipment across the infrastructure in order to maintain or restore it to a state in which it can perform the required function as intended. In the context of smart city development, the role of critical infrastructures, as well as their reliability, in the daily lives of citizens is vitally important when keeping things going. Therefore, maintenance management plays a vital role in reducing the failure rates of critical infrastructures. These infrastructures include electric distribution networks, water supply networks, gas distribution networks, road transportation, and communication networks, as interoperating systems [1,2].
Data-driven approaches have evolved significantly over the last decade with new technologies making smart city infrastructures more efficient and reliable, data-driven applications in maintenance management are emerging in several industry sectors. Essentially, data-driven maintenance (DDM) increases objectivity, reliability, sustainability and cost reduction. DDM affects smart city infrastructures by reducing failures and strengthening its robustness against sever contingencies. In other words, thanks to DDM, the resiliency and sustainability of the smart city are enhanced simultaneously.
Maintenance management systems are investigated in literature, especially the data-driven approach in recent years. Authors have proposed a two-stage robust model to consider the impact of failure rate uncertainty in the Reliability Centered Maintenance (RCM) scheduling problem . Furthermore, an RCM-based method has been proposed to determine optimal maintenance actions on distribution feeders with the aim of minimizing the total customer interruption cost and total energy not supplied cost . An asset management approach to momentary failure risk analysis is presented in a local electricity distribution company in , which uses a feature selection algorithm, for the identification of top momentary failure modes. Comprehensive reviews on data-driven maintenance management for water and gas infrastructures respectively have also been discussed by researchers [6,7].
Maintenance Management Methodology:In this section, the data-driven maintenance management methodology is presented, which aims to optimally allocate the preventive maintenance and repair groups into time slots for the purpose of scheduling maintenance work in the most efficient manner. The decision variables here are the number of preventive and corrective maintenance groups dedicated to each feeder and the constraints of the problem, relating to the reliability indicators, the non-budgetary constraints, and the number of repair personnel in each group. Eventually, the objective function of the problem is to minimize the overall costs, including the costs related to preventive and corrective maintenance, as well as the costs imposed on the customers by service interruption. Figure 1 demonstrates a sample block diagram of the data-driven maintenance management.
Figure 1 The data-driven maintenance management methodology methodlogy
A demonstrative example is presented in this part, as the DDM is first implemented on a local electricity distribution network, and the results of this model are compared to the conventional maintenance management method, followed by the examination of performance indicators to evaluate the effectiveness and efficiency of DDM comparing to that of the conventional method. Table 1 compares normalized results for conventional and data-driven maintenance scheduling methods . The table shows cost of energy not served (ENS) and customer interruption cost (CIC) a one-year period, using conventional and data-driven scheduling methods. Results demonstrates that utilizing data-driven method decreases both the CENS and CIC. In fact, using data-driven maintenance management results in a more accurate maintenance schedule, which reduces failure rates in different parts of the distribution network. Consequently, the imposed costs including, CENS and CIC will be decreased.
Table 1: Comparative normalized results for conventional and data-driven maintenance scheduling method 
This paper introduces the concept of data-driven maintenance management for smart cities critical infrastructures using a general concept of data-driven maintenance management methodology. The efficiency data-driven maintenance management approach has been compared to that of the conventional maintenance management in a demonstrative case study, which is the outcome of utilizing the operational data collected from the critical infrastructure. The paper can be concluded as:
- The data-driven maintenance management increases the sysem reliability and sustainability, by reducing failure rates in critical infrastructures.
- Utilizing accurate maintenance scheduling as a result of data-driven management, the availability and maintainability of the equipment is achieved at an appropriate level. Therefore, the resiliency and robustness of critical infrastructures are enhanced against sever contingencies.
- Using data-driven maintenance scheduling, the imposed costs including, repairs and customers interruption costs are drastically declined.
- K Etemadi, A Fereidunian, “A Comparative Study of Two People-Centric Smart Cities Evaluation Frameworks”, IEEE Smart Cities Newsletter, June 2022, Available at: https://smartcities.ieee.org/newsletter/june-2022/a-comparative-study-of-two-people-centric-smart-cities-evaluation-frameworks
- M. Kojury-Naftchali, A. Fereidunian, “Smart Energy-Aware Cities: Customer Characterization by Energy Data Analytics to Improve Demand Response Performance”, In: Shafie-khah, M., Amini, M.H. (eds) Flexible Resources for Smart Cities. Springer, Cham. https://doi.org/10.1007/978-3-030-82796-0_2
- A. Alizadeh, A. Fereidunian, M. Moghimi and, H. Lesani, "Reliability-Centered Maintenance Scheduling Considering Failure Rates Uncertainty: A Two-Stage Robust Model", IEEE Transactions on Power Delivery, Vol. 37, No. 3, p. 1941, 2022.
- H. Mirsaeedi, A. Fereidunian, S. M. Mohammadi-Hosseininejad, and H. Lesani, "Electricity Distribution System Maintenance Budgeting: A Reliability-Centered Approach”, IEEE Transactions on Power Delivery, Vol. 33, No. 4, p. 1599, 2017.
- M. Haghifam, E. Akhavan-Rezai, and A. Fereidunian. "An Asset Management Approach to Momentary Failure Risk Analysis on MV Overhead Lines”, IEEE 11th International Conference on Probabilistic Methods Applied to Power Systems, p. 542, 2010.
- L. Chik, D. Albrecht, and J. Kodikara, "Modeling Failures in Water Mains Using the Minimum Monthly Antecedent Precipitation Index." Journal of Water Resources Planning and Management, Vol. 144, No. 4, p. 06018004, 2018.
- M.W. Fan, C.C. Ao, and X.R. Wang, “Comprehensive Method of Natural Gas Pipeline Efficiency Evaluation Based on Energy and Big Data Analysis”, Energy, Vol. 188, p.116069, 2019.
- M. H. Tamanaeifar, et al, "Electricity Distribution Network Maintenance Scheduling based on Data-driven Evaluation of Feeders Risk Priority Number", In 8th International Conference on Industrial and Systems Engineering (ICISE 2022), 2022.
This article was edited by Bernard Fong