(737e) Condition-Based Maintenance and Operation Optimization of Large-Scale Energy Systems

Commercial buildings consume about 25% of the energy in the United States, and over 40% of the consumption is attributed to building heating, ventilation and air conditioning (HVAC) systems [1]. While various optimization methods have been proposed for HVAC system operation and control, equipment degradation (e.g., chiller tube fouling and boiler scaling problem) and the resulting maintenance optimization have received limited attention. However, it is reported that major HVAC units (e.g., chillers and boilers) suffer from efficiency loss of up to 24% per decade [2,3]. Thus, optimal maintenance scheduling is required to maintain satisfactory performance of the HVAC systems over medium to long horizon.

Condition-based maintenance (CBM) is a rapidly growing maintenance strategy that recommends maintenance decisions based on the current condition of the system. Utilization of CBM allows more efficient maintenance optimization considering the operation (and the resulting degradation), which ultimately leads to better system performance compared to other tradition maintenance strategies [4]. However, CBM optimization of large-scale energy systems, and HVAC systems in particular, remains an open area due to their complexities, such as the complicated demand-side responses to the time-sensitive electricity pricing [5] and the highly nonlinear models that describe equipment performance [6]. Recent studies [7,8] on the maintenance optimization of HVAC systems are limited to more traditional maintenance strategies or CBM for only a single-unit system. However, these methods cannot be applied to multi-unit systems because of the dependencies among the units.

In this work, we present optimization methods for the CBM optimization of HVAC systems using mixed integer programming. The optimization framework we introduce is composed of data preprocessing methods and optimization models. The data preprocessing methods are introduced to provide parameters for constraints that are used to provide an approximation of the detailed operating patterns. The optimization models are composed of several “modules”: the operation module, the maintenance module, and the unit health module. The operation module is an approximate submodel (based on data obtained from a detailed optimization model with hourly granularity), which is used to describe the medium-to-long term impact of operation while maintaining computational tractability. The maintenance module includes constraints related to the timing of maintenance tasks and the recovery of the “health” of the units after maintenance. Finally, the health module predicts unit degradation according to the approximate system operation. Operation, maintenance and health dependencies among units are considered in the optimization methods. The proposed optimization framework can account for various energy systems with degradation. Importantly, the proposed framework is the first to account for the interplay of the three components (operation, maintenance, and equipment status) in systems with high variability on two scales (daily and annual).

We first demonstrate the accuracy of the proposed methods using a cooling system based on a real-world application, while additional case studies, based on large-scale HVAC systems are also provided.

[1] U.S. Department of Energy, Buildings energy databook, Energy Effic. Renew. Energy Dep. (2012) 286. http://buildingsdatabook.eren.doe.gov/DataBooks.aspx.

[2] D.A. Waddicor, E. Fuentes, L. Sisó, J. Salom, B. Favre, C. Jiménez, M. Azar, Climate change and building aging impact on building energy performance and mitigation measures application: A case study in Turin, northern Italy, Build. Environ. 102 (2016) 13–25. doi:10.1016/j.buildenv.2016.03.003.

[3] G. Eleftheriadis, M. Hamdy, Impact of building envelope and mechanical component degradation on the whole building performance: A review paper, Energy Procedia. 132 (2017) 321–326. doi:10.1016/j.egypro.2017.09.739.

[4] G.P. Sullivan, R. Pugh, A.P. Melendez, W.D. Hunt, Operations & Maintenance Best Practices: A Guide to Achieving Operational Efficiency, Fed. Energy Manag. Progr. (2010) 321. doi:10.2172/1034595.

[5] M.H. Albadi, E.F. El-Saadany, Demand response in electricity markets: an overview, IEEE Power Eng. Soc. Gen. Meet. 24-28 June. (2007) 1–5. doi:10.1109/PES.2007.385728.

[6] J.M. Gordon, K.C. Ng, H.T. Chua, Centrifugal chillers: Thermodynamic modeling and a diagnostic case study, Int. J. Refrig. 18 (1995) 253–257. doi:10.1016/0140-7007(95)96863-2.

[7] B. Wang, Z. Wu, X. Xia, A Multistate-Based Control System Approach Toward Optimal Maintenance Planning, 25 (2016) 1–8. doi:10.1109/TCST.2016.2550505.

[8] N. Cauchi, K. Macek, A. Abate, Model-based predictive maintenance in building automation systems with user discomfort, Energy. 138 (2017) 306–315. doi:10.1016/j.energy.2017.07.104.