Techno-Economic Analysis of Shared Mobile Storage for Demand Charge Reduction

arXiv:2606.20163 · eess.SY, cs.SY · Submitted 2026-06-18 · Read on arXiv

Listen

Transcript

Introduction to the show: ident: Robotics Radio. Generated commentary on the latest robotics and control papers.

Rosa: Today's paper: "Techno-Economic Analysis of Shared Mobile Storage for Demand Charge Reduction".

Dev: This paper investigates how shared electric vehicle (EV) fleets can be economically viable for demand charge reduction by developing a high-fidelity management framework that accounts for complex operational realities.

Rosa: First, who's behind it and why it matters.

Paper summary: Rosa: Hey Dev, so we're looking at this paper titled "Techno-Economic Analysis of Shared Mobile Storage for Demand Charge Reduction," and it seems the main idea is that shared electric vehicle fleets can actually be economically viable for reducing demand charges by using them as mobile storage.

Dev: Right? The abstract claims they developed a high-fidelity management framework because previous models often ignored crucial things like transit overheads, labor costs for the drivers, and battery degradation.

Rosa: Exactly! It moves beyond those idealized scenarios by formulating the dispatch problem as a mixed-integer linear program that tries to minimize both demand charges and the total cost of ownership simultaneously.

Dev: That's the core mechanism they're proposing to link these complex operational realities together in one optimization structure.

Taro: I’m really interested in how they handled those practical logistical constraints; it sounds like a big step beyond just modeling energy flow through a stationary battery.

Rosa: It is, Taro, because they explicitly account for the spatio-temporal coupling of energy consumption, labor costs for EV drivers, and battery degradation when formulating that MILP (<ref:2606.20163#pg0>).

Dev: And then they tackle the computational headache of deciding which EV goes where and when using a marginal-value-based heuristic algorithm to keep things from getting bogged down with complexity (<ref:2606.20163#pg1>).

Taro: I think that heuristic approach is interesting because it aims for near-optimal performance while keeping the computation time manageable for real fleet operations, which is a practical concern.

Rosa: Speaking of real operations, they use data from San Francisco businesses served by PG andE to test this framework (<ref:2606.20163#pg1>).

Dev: And what did that real-world testing reveal about the drivers? They found that labor costs ended up being the dominant factor affecting the revenue in their analysis (<ref:2606.20163#pg1>).

Rosa: That finding about labor costs being dominant really hits home for me, Dev, because it suggests that even if we optimize energy perfectly, the human element in operating these fleets is a huge part of the economic equation.

Dev: It certainly seems so; when you look at their case study results, they also validated a tiered charger deployment strategy—mixing DC fast and AC Level-two chargers across different users—which they showed yielded superior net savings compared to just having uniform infrastructure (<ref:2606.20163#pg1>).

Paper summary: Taro: That’s interesting because it implies that the physical placement of the charging infrastructure matters as much as the vehicle itself for maximizing savings.

Rosa: I agree, Taro, and looking at their findings regarding seasonality, they noted that profitability was particularly in summer months compared to winter (<ref:2606.20163#pg1>).

Dev: And when we talk about the fleet size needed for those strategies, they found that for the tiered setup during winter, three EVs achieved monthly net savings of approximately thirty thousand dollars (<ref:2606.20163#pg1>).

Taro: That's a concrete number showing how scalable this concept could be if implemented correctly.

Dev: And for the summer months with that same tiered setup, they saw savings approaching one hundred thousand dollars with six EVs (<ref:2606.20163#pg1>).

Rosa: That jump from thirty thousand to nearly a hundred thousand dollars highlights how sensitive the economics are to when you deploy the assets, especially given the tariff structures they were analyzing under Schedule B-ten or B-nineteen tariffs (<ref:2606.20163#pg0>).

Taro: The sensitivity analysis they did regarding labor cost showed that net savings decrease as rates go up, identifying a critical labor cost around one hundred twenty/hr in winter and two hundred ten/hr in summer (<ref:2606.20163#pg1>).

Dev: It also flagged that uncertainty matters, showing that net savings vanish beyond moderate forecast uncertainty levels, like five percent in winter or ten percent in summer (<ref:2606.20163#pg1>).

Rosa: So, to wrap up this paper "Techno-Economic Analysis of Shared Mobile Storage for Demand Charge Reduction," it really shows that the shared mobile storage business model has economic appeal, especially during the summer (<ref:2606.20163#pg0>).

Dev: The authors successfully quantified the true economic value of these assets by integrating operational factors like transit energy and battery wear into that unified MILP structure (<ref:2606.20163#pg1>).

Taro: I think the implication here is that this isn't just a theoretical exercise; it provides a realistic baseline for commercial viability because it grounds the assessment in actual operating conditions rather than some idealized model (<ref:2606.20163#pg0>).

Rosa: Thinking about what this means for the wider world, Taro, if we can effectively deploy these mobile storage solutions across commercial and industrial facilities, it could fundamentally alter how we manage peak energy loads in urban centers (<ref:2606.20163#pg1>).

Dev: The framework’s focus on spatio-temporal coupling suggests that smart energy management systems need to be deeply integrated with dynamic fleet dispatch capabilities to be truly effective (<ref:2606.20163#pg2>).

Paper summary: Taro: And what about when the world misbehaves, like during unexpected load spikes or infrastructure failures? The paper doesn't really cover that scenario in detail; it focuses heavily on optimized operation under steady conditions (<ref:2606.20163#pg1>).

Rosa: That’s a fair point, Taro; the paper does state its limitation in that it doesn't extend to incorporating uncertainties in load profiles or EV availability, which means its practical application might require more advanced forecasting systems (<ref:2606.20163#pg5>).

Dev: So, what’s the next step for this research? The authors mention future work will extend this by incorporating those very uncertainties you brought up earlier (<ref:2606.20163#pg5>).

Taro: I think moving into that uncertainty analysis is the most crucial next piece because real-world energy demands are rarely perfectly predictable (<ref:2606.20163#pg5>).

Rosa: It sounds like this paper really lays a solid foundation for understanding the economic trade-offs involved in deploying mobile energy storage for demand charge reduction (<ref:2606.20163#pg5>).

Dev: We should keep an eye on how their proposed tiered charger deployment strategy performs when we move beyond San Francisco data to other regions with different tariff structures (<ref:2606.20163#pg1>).

Taro: It’s a really interesting piece of work that connects high-level optimization theory with concrete operational realities, Rosa; it gives us a much clearer picture of the viability of shared EV storage (<ref:2606.20163#pg5>).

Rosa: Exactly, Taro, and the fact that they’ve validated a tiered deployment strategy suggests that infrastructure planning should prioritize flexibility in deployment over just maximizing raw energy capacity (<ref:2606.20163#pg1>).

Dev: It seems like the main implication is that for C andI users dealing with high demand charges, the operational cost of labor and infrastructure configuration are as important as the direct energy savings themselves (<ref:2606.20163#pg1>).

Taro: That’s a big picture point; it shows that solving this kind of problem requires looking at the entire system—energy, labor, and physical deployment—together (<ref:2606.20163#pg5>).

Rosa: Well, to wrap up our discussion on "Techno-Economic Analysis of Shared Mobile Storage for Demand Charge Reduction," this paper gives us a strong techno-economic argument for using shared EV fleets as mobile storage assets (<ref:2606.20163#pg5>).

Dev: The central message is that the model works, provided you account for the complex operational factors like battery degradation and labor costs in your planning (<ref:2606.20163#pg1>).

Taro: It’s a valuable resource because it sets a high bar for how we should approach modeling shared mobile resources in real-world scenarios (<ref:2606.20163#pg5>).

Conclusion: Rosa: So we’ve been digging into how shared electric vehicle fleets can actually cut down on those big electricity bills using them for mobile storage, and now we’re at the conclusion of this study by the authors who put it all together. Dev, what are your thoughts on the overall title and who came up as the main players here?

Dev: Yeah, I think that title really captures what they did—it's not just about EVs; it's about a techno-economic analysis where they tie operational costs directly to energy savings. The authors seem very focused on building a high-fidelity model that actually accounts for how things run in the real world, which is important for me because we need to know if this loop rate and latency stuff is realistic.

Taro: I agree with Dev; the authors clearly spent a lot of time making sure their mathematical formulation reflected practical realities like battery degradation and labor costs, which pushes us toward thinking about how this system performs outside of a perfect lab setting. The implication here is that the real test will be how robust this framework is when things get messy.

Rosa: It seems the paper suggests that the main contribution isn't just proving EVs can store energy, but showing exactly *how much* saving you can realistically expect while keeping your total cost in check, and those results are quite compelling. What do you see as the biggest implication for companies looking to adopt this idea?

Dev: I think the biggest implication is that they prove profitability isn't just about finding cheap batteries; it’s about the entire operational structure—the charging strategy, the dispatch algorithm, and crucially, managing those labor expenses effectively. That tells us that infrastructure design is a massive variable we need to control for smooth operation.

Taro: From an autonomy standpoint, the implication is that as these fleets scale up across multiple facilities, the management framework needs to become incredibly smart at handling dynamic demand spikes without failing service continuity or incurring excessive operational costs due to inefficient dispatching. That's where I want to focus next.

Rosa: It really highlights that this isn't just a theoretical exercise in energy storage; it’s a blueprint for how urban centers can manage peak load demands through flexible, distributed assets. We need to keep looking at these models as we try to deploy them in actual city environments.

Elmore Family School of Electrical and Computer Engineering, Purdue University · School of Advanced Engineering, Great Bay University

eess.SY, cs.SY

Submitted: 2026-06-18

Updated: 2026-10-05

Comments: 22 pages, 26 figures, journal

License: http://creativecommons.org/licenses/by/4.0/

Importance score: 75/100

The gist: This paper investigates how shared electric vehicle (EV) fleets can be economically viable for demand charge reduction by developing a high-fidelity management framework that accounts for complex

Key concepts

High-Fidelity Management Framework: This is a detailed system designed to manage a group of EVs. It goes beyond simple scheduling by including real-world operational details like how much energy the car uses while driving, the cost of paying the driver, and how much the battery wears down over time. This makes the economic prediction much more accurate.
Mixed-Integer Linear Program (MILP): This is a mathematical tool used to solve complex decision problems in this study. It helps find the best way to dispatch EVs to users while minimizing two main things: the utility's demand charges and the total cost of owning and operating the vehicles.
Tiered Charger Deployment Strategy: This strategy suggests mixing different types of charging stations (like fast DC chargers and slower AC chargers) across various user locations. The study found this setup provides better net savings than using only one type of charger everywhere, making it a key recommendation for maximizing profit.

Terminology

Fidelity — 10× in this paper · explained in 2 episodes
Mixed-Integer Linear Program (MILP) — 4× in this paper · explained in 2 episodes
Core — 3× in this paper · explained in 2 episodes
Puget Sound Energy — 3× in this paper
Aggregation — 1× in this paper · explained in 1 episode
Approximation Bound — 1× in this paper · explained in 1 episode
Benchmarking — 1× in this paper · explained in 3 episodes
Modularity — 1× in this paper · explained in 9 episodes
Objective Function — 1× in this paper · explained in 2 episodes
Physical Constraints — 1× in this paper · explained in 7 episodes
Unified Framework — 1× in this paper · explained in 5 episodes

Summary

This paper investigates how shared electric vehicle (EV) fleets can be economically viable for demand charge reduction by developing a high-fidelity management framework that accounts for complex operational realities. The core finding is that a modest number of EVs can achieve significant savings sufficient to recover ownership and operational expenses, with profitability being highly sensitive to tariff structures, fleet size, and cost components.

The Gist

A high-fidelity fleet management framework explicitly accounts for the spatio-temporal coupling of energy consumption, labor costs for EV drivers, and battery degradation during the formulation of a mixed-integer linear program (MILP) that jointly minimizes demand charges and total cost of ownership.

Framework Development and Model Formulation

The research develops a centralized fleet management framework to evaluate profitability by coordinating a fleet of EVs to provide peak-shaving services to multiple commercial and industrial (C&I) facilities. The objective is to minimize the users’ demand charges and the EVs’ operational cost, which includes transit energy consumption, travel-dependent dispatch coupling, labor cost, charging infrastructure cost, and battery degradation costs represented through established depreciation models.

The optimization problem is formulated using decision variables such as dispatch matrices (Mj), where an entry Mj,t,i = 1 represents EV j’s dispatch to user i during interval t. The model explicitly captures the spatio-temporal coupling of mobile storage by integrating transit energy consumption and labor costs into a unified objective function. A key technical challenge addressed is the physical constraint that at most one EV can actively provide service during any interval t, which is handled using a bilinear constraint (1d) to permit necessary transit overlap for uninterrupted service continuity.

Heuristic Algorithm for Efficiency

To tackle the computational complexity arising from user-EV dispatch decisions, the authors propose a marginal-value-based heuristic algorithm. This algorithm iteratively dispatches the least used EVs to the most-valued services based on a calculated priority value, Vi,s = DCi − DCRi(s) / s. The process involves several steps:

Identifying the most valued user i⋆ and corresponding number of services s⋆ using optimization problem (10a).
Selecting the least used EV j⋆ from Ja, defined as the vehicle with the lowest total energy expenditure across discharging and transit.
Determining if dispatch is profitable by comparing the marginal demand charge reduction, s⋆Vi⋆,s⋆, against the increase in charging cost δCC.

Case Study and Key Findings

The analysis utilizes real-world data from San Francisco businesses served by PG&E, focusing on users with average monthly peak demands above 75 kW under Schedule B-10 or B-19 tariffs. The case study reveals several critical insights:

labor costs as the dominant factor affecting the revenue.

we propose and validate a tiered charger deployment strategy (mixing DC fast and AC Level-2 chargers across users), demonstrating that it yields superior net savings compared to uniform infrastructure configurations.

The results show that profitability is particularly in summer months compared to winter, with the Tiered setup yielding the highest savings across both seasons. The optimal fleet size for the Tiered setup is three EVs during winter, achieving monthly net savings of approximately 30,000 dollars, and six EVs during summer, approaching 100,000 dollars.

Sensitivity Analysis and Infrastructure Impact

The study evaluates the impact of infrastructure configuration (All-AC vs. All-DC vs. Tiered) and operational costs (labor cost sensitivity). The Tiered setup yields the highest savings across both seasons. While All-DC setups show higher demand charge reductions for B-19 users, their higher infrastructure cost makes them less favorable than the Tiered configuration under the PG&E tariffs. Sensitivity to labor cost shows that net savings decrease with increasing rates, with a critical labor cost identified around 120/hr in winter and 210/hr in summer. Furthermore, uncertainty analysis indicates that net savings vanish beyond moderate forecast uncertainty levels (5% in winter and 10% in summer).

Conclusion

The paper concludes that the shared mobile storage business model is economically attractive, particularly during the summer months. The framework successfully quantifies the true economic value of these assets by explicitly integrating operational factors like transit energy and battery wear into a unified MILP structure, providing a realistic baseline for commercial viability. Future work will extend this to incorporate uncertainties in load profiles and EV availability.

How it works

The research develops a high-fidelity fleet management framework to quantify the true economic value of shared mobile energy storage for demand charge reduction by accounting for operational realities such as transit energy consumption, labor costs, and battery degradation within a unified MILP formulation. This ensures that the economic assessment is grounded in practical operating conditions rather than idealized models.

Improvements for AI systems

Here are the specific improvements and capabilities that an AI system, informed by this research, could achieve:

) Improved Demand Response (DR) Optimization for Smart Buildings:

The AI system can move beyond simple peak shaving by implementing a demand response orchestration module. It can predict user load profiles with high accuracy (using historical data like the San Francisco study) and proactively dispatch EV fleets to not just reduce the maximum demand, but to shape the entire load profile across multiple Time-of-Use (TOU) periods, thereby minimizing the total weighted demand charge across all periods.

) Spatio-Temporal Logistics Engine:

The system can incorporate a high-fidelity spatio-temporal coupling model. This allows it to optimize fleet dispatch decisions not just based on immediate financial gain, but by planning routes and charging schedules that account for:

Transit energy consumption (based on real road networks).
Battery degradation costs over the long term.
Labor costs associated with driver availability and scheduling constraints in specific geographic zones (e.g., optimizing drivers based on localized peak-event heatmaps).

) Dynamic Infrastructure Deployment Strategy:

Instead of assuming a fixed infrastructure, the AI can dynamically recommend charger deployment strategies based on real-time tariff structures (like the PG&E vs. PSE examples). It can determine the optimal mix of AC Level-2 and DC Fast chargers per user type to maximize net savings under specific economic conditions, as demonstrated by the Tiered charger deployment strategy finding.

) Multi-Objective Fleet Sizing and Configuration Advisor:

The system can serve as a decision support tool for fleet operators, providing prescriptive recommendations on optimal fleet size based on seasonal demand patterns. It can explicitly model the trade-off between increasing service coverage (which increases demand reduction) and increasing operational costs (labor, charging, depreciation), yielding specific optimal fleet sizes (e.g., 2 EVs in winter for Tiered setup; 6 EVs in summer).

) Robustness and Uncertainty Management Module:

The system can be trained to handle forecast uncertainty. Using methods derived from the sensitivity analysis (Figure 20), it can adjust dispatch decisions based on probabilistic load forecasts, ensuring that the demand charge reduction remains profitable even when actual loads deviate from predictions by a certain percentage (e.g., maintaining profitability up to a 10% error margin).

) Cost-Aware Service Interval Scheduling:

The AI can determine the optimal service intervals for each user dynamically. It can decide whether to provide service in the current interval or wait for a future interval where the marginal demand charge reduction outweighs the cost of waiting (waiting at location vs. immediate dispatch), effectively optimizing the service schedule variable from Problem (2a) within a continuous learning loop.

) Economic Feasibility Screening:

For any proposed operational plan, the AI can instantly screen its economic viability against various tariff structures (PG&E vs. PSE). This allows for rapid scenario planning to assess profitability under different utility rate regimes before physical deployment, directly addressing the Remark 1 sensitivity findings.

Abstract

This paper investigates the techno-economic viability of shared electric vehicle (EV) fleets for demand charge reduction under practical logistical and operational constraints. Unlike idealized models that overlook transit overheads, we propose a high-fidelity fleet management framework that explicitly accounts for the spatio-temporal coupling of energy consumption, labor costs for EV drivers, and battery degradation. We formulate the dispatch problem as a mixed-integer linear program (MILP) that jointly minimizes demand charges and total cost of ownership. To address the computational complexity arising from path-dependent constraints, we develop a marginal-value-based heuristic algorithm that achieves near-optimal performance with high computational efficiency. Using real-world data from San Francisco, our analysis reveals that a modest number of EVs can achieve significant demand charge savings, sufficient to recover the ownership and operational expenses. Our results also show how tariff structures, fleet size, and cost components influence overall profitability.

Sources

Mobile Energy Storage in Power Network: Marginal Value and Optimal Operation