Avoiding Pitfalls in Energy Storage Modelling
The purpose of this article is to distill some of my observations from modelling energy storage in integrated power system planning and dispatch (IPD) models.
As more energy storage systems become part of the electricity grid and microgrids, there is a need to better understand the drivers of policy and investment decisions and their implications on the operational, environmental, and economic attributes of our changing power system landscape [1]. Integrated planning and dispatch models are mathematical programming tools used to simulate, evaluate, and optimize the performance indicators of anticipated changes in the grid system [2].
There are many types of energy storage technologies and each of these technologies has its own unique attributes. However, the fastest-growing energy storage systems are the battery energy storage systems (BESS) [1, 3]. Consequently, this makes BESS the focus of this post. BESS has applications in electric power systems, electric vehicle systems, and numerous electronic gadgets and devices we use daily [4, 5]. The scale of energy storage in power systems dwarfs other BESS uses.
BESS has very fast response time which makes it suitable for managing fast dynamic system variables such as frequency regulation in power system applications. It can sustain power reliability and availability during the sudden loss of a baseload plant or the slower fluctuations in a renewable generation [5]. Other applications include, but not limited to arbitraging, variable generation firming, and demand-side management for peak load shaving.
As operators, regulators, and investors explore interests in BESS portfolios, IPD models and frameworks are useful to assess potential economic metrics and the performance of BESS under various operational and environmental scenarios. The model can simultaneously search the entire operation and planning horizons of the grid/microgrid system under investigation. Therefore, it must capture the dynamics of the system across multiple timescales ranging from milliseconds to years, as illustrated in the figure below. This enhances the quality and optimality of results from the modelling exercise.
Usually, the objective of an IPD model is to optimize some economic metrics over entire planning and operating horizons while satisfying constraints on reliability, penetration of green/clean technology, GHG emission limits, etc. BESS modelling is often considered challenging; even more so when incorporated into IPD models, and this introduces various pitfalls that can misguide the decision to be made about a proposed investment project, regulatory process, or climate policy.
IPD models can also be used to determine the optimal cycling that balances the reliability requirements of the system and the cost of early replacement or supplementary capacity additions, among other criteria of interest.
Below, three of such common pitfalls (among others) are briefly discussed and sample formulations to avoid them are also presented in equation Panels 1 to 3.
Propagation between states
Charging or discharging operations of the system musts be accounted for over operating, dispatch, and planning periods. Stored energy by the end of each period must be available in the next to accurately reflect the operation of the system. Energy losses in the form of leakages from storage or dissipation during a discharge cycle must also be quantified accordingly. During the implementation of an IPD model, initialization and termination states of the storage system might need to be specified to the existing or expected operational conditions in accordance with the objectives of the modelling assessment. Equation Panel 1 below shows the energy balance equation for BESS which captures the propagation of states.
Degradation between cycles
Nominal capacity degrades over cycles of use and due to the manner of use. Deep discharging during operation, overcharging, and high ambient temperature conditions accelerate degradation rate [4]. This is often ignored in most assessment models. BESS is considered due for replacement upon degradation to 80% of original capacity [4]. Capacity additions may have to occur on an ongoing basis, as represented in the equation Panel 2 below.
Degradation can only be slowed down by adopting best practices such as avoiding deep cycling, maintaining conducive ambient temperature (neither too hot, nor too cold), and adopting an optimal commitment schedule.
Interaction between scales
In an IPD model, some variables in the model change at different timescales. There are multiple timescale interactions between the energy storage level and storage capacity (both pre-existing and new additions), as shown in the first equation in Panel 3. Also, the objective function (system cost) has multiple costs that are at different timescales and need to be aggregated to calculate the entire cost. Panel 3 below shows the multiscale interactions and a formulation of the aggregation of total system cost in the objective function.
Depending on the goal of an assessment, various other objective functions might be used to optimize and analyze the performance of the system. Typically, IPD models maximize or minimize the objective function depending on whether the criterion is a desirable quantity or not. Desirable quantities are maximized but the undesirable ones are minimized. Some examples of other objective functions in IPD models include net revenue, the lifetime of BESS units, environmental impacts, etc. The rest of the equations are constraints on the objective function. There are various other constraints such as non-concurrent charging and discharging of storage units, lower and upper limits on storage level, maximum charging and discharging rates, etc.
As more BESS are added to the electricity grid, other factors such as the cost of end of life disposal or recycling of the systems have to be accounted for as well. IPD models can be built and tailored to include lifecycle costs and environmental impacts. By adopting the best modelling practices, IPD models can enrich the understanding of the benefits and challenges of increasing energy storage systems in the grid or microgrids.
Acknowledgment: Special thanks to my friends, Babatunde Odetayo and Wale Akanni who reviewed my notes.
References
[1] Trahey, L., Brushett, F.R., Balsara, N.P., Ceder, G., Cheng, L., Chiang, Y.M., Hahn, N.T., Ingram, B.J., Minteer, S.D., Moore, J.S. and Mueller, K.T., 2020. Energy storage emerging: A perspective from the Joint Center for Energy Storage Research. Proceedings of the National Academy of Sciences, 117(23), pp.12550–12557.
[2] Doluweera, G., Husseini, H., and Umeozor, E., 2018. Economic and Environmental Impacts of Transitioning to a Cleaner Electricity Grid in Western Canada. Canadian Energy Research Institute Study Report 174, Calgary, Canada.
[3] Rancilio, G., Lucas, A., Kotsakis, E., Fulli, G., Merlo, M., Delfanti, M. and Masera, M., 2019. Modeling a large-scale battery energy storage system for power grid application analysis. Energies, 12(17), p.3312.
[4] DiOrio, N., Dobos, A., Janzou, S., Nelson, A. and Lundstrom, B., 2015. Technoeconomic modeling of battery energy storage in SAM (No. NREL/TP-6A20–64641). National Renewable Energy Lab. (NREL), Golden, CO (United States).
[5] Eto, J.H., Undrill, J., Mackin, P., Daschmans, R., Williams, B., Haney, B., Hunt, R., Ellis, J., Illian, H., Martinez, C. and O’Malley, M., 2010. Use of frequency response metrics to assess the planning and operating requirements for reliable integration of variable renewable generation (No. LBNL-4142E). Lawrence Berkeley National Lab (LBNL), Berkeley, CA (United States).
