Battery Swapping Stations: Where Should Cities Build Them? A New Data-Driven Answer

Electric vehicles are supposed to be the future of urban transport, but in fast-growing cities across the developing world, a stubborn problem keeps slowing them down: drivers cannot reliably charge or swap their batteries. A new study from researchers at VIT Business School in Chennai, India, offers a data-driven answer to one of the most practical questions in the electric mobility transition — where exactly should a city build its battery swapping stations? The research, published in Frontiers in Sustainability, applied a hybrid multi-criteria decision-making (MCDM) framework to the Chennai Metropolitan Area (CMA), one of India’s most densely populated urban regions, and identified Tambaram, a peripheral node of the metropolis, as the single most suitable site for a new battery swapping station (BSS).

The stakes are considerable. India’s road transport sector contributes roughly 12 percent of the nation’s energy-related carbon dioxide emissions and is a major driver of urban air pollution, according to figures cited by the International Energy Agency. The Indian government has backed the global EV30@30 campaign, which aims for at least 30 percent of new vehicle sales to be electric by 2030. Yet adoption remains hampered by three familiar barriers: range anxiety, insufficient charging infrastructure, and long waiting times at chargers. Battery swapping offers a way around all three. Instead of plugging in and waiting, a driver simply exchanges a depleted battery for a fully charged one — a process that can be automated or done manually in minutes, much like exchanging a domestic gas cylinder.

The research team, Siva Priya V and Asi Vasudeva Reddy, focused on the Chennai Metropolitan Area, which spans Chennai city and parts of the Kancheepuram, Tiruvallur, and Chengalpattu districts. The region has seen EV adoption rise at a rate of 20.21 percent over the five years to 2024, according to the Parivahan vehicle registration database. Tamil Nadu has installed 1,524 public charging stations over the last five years, but Chennai accounts for only 218 of them, and the battery swapping infrastructure in the metropolitan area remains insufficient to support the growing EV user base. The mismatch between demand and infrastructure is precisely what the study set out to address.

What distinguishes this study from previous work on charging infrastructure siting is its explicit treatment of EV adoption rate as a quantifiable decision criterion rather than an assumption. The researchers defined five criteria: EV adoption rate, proximity to residential areas, land cost, proximity to emergency services such as fire stations, and proximity to public facilities including government hospitals, colleges, and transport terminals. The adoption rate was calculated from five years of vehicle registration data from the Tamil Nadu Regional Transport Office, categorised by fuel type, while land costs were drawn from government-guided street-level valuation records. Proximity measures were computed using real-world road distances from Google Maps via a custom Google Apps Script function, providing an automated and realistic alternative to straight-line geographic measures.

Methodologically, the framework combines three techniques. The CRITIC method — Criteria Importance Through Intercriteria Correlation — derives objective weights for each criterion directly from the data, measuring both the contrast intensity of each criterion and the degree of conflict between criteria. This approach deliberately avoids the subjective expert judgments that dominate many MCDM studies, reducing bias in the weighting process. The weights were then fed into two parallel ranking engines: TOPSIS (Technique for Order of Preference by Similarity to Ideal Solution), which measures each alternative’s geometric distance from an ideal and an anti-ideal solution, and COPRAS (COmplex PRoportional ASsessment), which evaluates the proportional contribution of beneficial and non-beneficial attributes to each alternative’s overall utility.

Six candidate locations were evaluated: Ambattur (L1), Tiruvallur (L2), Poonamalle (L3), Tambaram (L4), Sriperumbudur (L5), and Chengalpattu (L6). The CRITIC weighting produced a clear hierarchy. EV adoption rate emerged as the dominant criterion at 31.97 percent of the total weight, followed by land cost at 26.43 percent. Proximity to emergency services (14.42 percent), proximity to public facilities (14.10 percent), and proximity to residential areas (13.08 percent) trailed behind. Notably, although land cost showed the strongest variability across the alternatives — its standard deviation of 0.401 indicated the greatest discriminating power — the EV adoption rate achieved the highest overall CRITIC score because of its negative correlation with the other criteria, meaning it carried more unique, non-redundant information.

The rankings were strikingly consistent. Under TOPSIS, Tambaram scored a performance value of 73 percent, well ahead of Poonamalle at 62 percent, with Chengalpattu and Tiruvallur tied at 48 percent, Sriperumbudur at 44 percent, and Ambattur at just 16 percent. The COPRAS method produced the same winner, awarding Tambaram a utility index of 100, with Poonamalle second at 83. Spearman’s rank correlation coefficient between the two methods was 0.8286, indicating strong agreement. Only minor differences appeared in the middle of the ranking, where Tiruvallur, Sriperumbudur, and Chengalpattu shuffled positions slightly, while the top and bottom ranks remained fixed.

To test whether these results were an artefact of the particular weights used, the researchers ran a Monte Carlo sensitivity analysis with 10,000 iterations, randomly perturbing each criterion weight within ±20 percent of its original value and renormalising after each draw. Tambaram retained first place in 100 percent of the simulations, and Poonamalle consistently held second. The Spearman rank correlation between the deterministic ranking and the simulated mean ranking was a perfect 1.00. A separate cross-validation using subjective weights derived from expert judgment via the Analytic Hierarchy Process (AHP) produced rankings consistent with the objective approach, confirming that the optimal alternative is unchanged under reasonable uncertainty.

The spatial findings carry a message that may surprise planners accustomed to concentrating infrastructure in city cores. The peripheral nodes of the Chennai Metropolitan Area outperformed other candidate locations because of higher EV density and greater public awareness. Tambaram’s advantages include a strong EV adoption rate, significant transportation hubs, and proximity to employment centres such as the Madras Export Processing Zone, which draws large numbers of commuting IT and industrial workers. The authors argue that developing peripheral nodes like Poonamalle, Tiruvallur, and Chengalpattu supports Sustainable Development Goal 11 — specifically targets on reducing the environmental impact of cities and strengthening urban–rural linkages — by decentralising infrastructure and reducing charging congestion in the metropolitan core.

The study is not without limitations, which the authors acknowledge candidly. It relies on secondary data and a deliberately restricted set of five criteria; parking availability was excluded because site-level spatial survey data were unavailable, and renewable energy integration would require energy mix modelling at a spatial resolution beyond the scope of the work. The findings may also differ in regions with different socio-economic conditions, grid reliability, or traffic patterns. Still, the framework offers something increasingly rare in infrastructure planning: a transparent, computationally efficient, and reproducible decision-support tool that grounds demand in measured behaviour rather than assumption. As cities worldwide race to decarbonise transport, the lesson from Chennai is clear — the best place to build a battery swapping station is where the data says drivers are already switching.

Subject of Research: Data-driven site selection for electric vehicle battery swapping stations using hybrid multi-criteria decision-making in the Chennai Metropolitan Area

Article Title: A hybrid MCDM approach for site selection of electric vehicle battery swapping station

Article References: V, S. P., & Reddy, A. V. (2026). A hybrid MCDM approach for site selection of electric vehicle battery swapping station. Frontiers in Sustainability, 7, Article 1867240. https://doi.org/10.3389/frsus.2026.1867240

Image Credits: AI Generated

DOI: 10.3389/frsus.2026.1867240

Keywords: electric vehicles, battery swapping stations, multi-criteria decision-making, CRITIC, TOPSIS, COPRAS, Chennai Metropolitan Area, EV infrastructure, Monte Carlo simulation, sustainable mobility, site selection, urban planning