Ramasan et al To V2G or Not Residential (2026)

Title: To V2G or Not? Assessing EV Flexibility Across Electricity Markets and User Behaviour in Residential Energy Systems. Chalmers University of Technology, Department of Electrical Engineering. SSRN preprint 7403061, not peer reviewed (every page is marked “Preprint not peer reviewed”); the authors’ README says a journal version is forthcoming.

Funding and project: Vinnova Ref. No. 2023-00785, Implementation of Vehicle-to-Grid Services in Sweden, with Polestar Performance AB, Vattenfall, Göteborg Energi, Svenska kraftnät, Easee and Chalmers. This is the same grant and partner list as the PAVE pilot and the Malakhatka service blueprint: this paper is the project’s economics strand.

Dataset: the authors publish their model results (not the input price and frequency series) as Excel files in a public GitHub repository under an MIT licence. The raw copies here were fetched with curl on 2026-09-29 at commit 045cb03 (see _FETCHED.md). One file, Results_summary_2025_LEC.xlsx, holds results for two households modelled together as a local energy community, which the paper only lists as future work; those results are not described in the paper and are not summarised here.

Summary

A mixed-integer linear programming (MILP) model of one detached house in Gothenburg (SE3) with a 65 kWh EV on an 11 kW bidirectional home charger, run at 15-minute resolution over the full years 2022 and 2025. The household pays Göteborg Energi’s 2025 retail and network tariffs, including a peak charge of 61.55 SEK/kW per month, energy tax and VAT. Inputs are historical SE3 spot prices (ENTSO-E), FCR-N and FCR-D prices (Svk’s Mimer), mFRR prices as the FCR-N activation price, and 0.1-second Nordic frequency data (Fingrid) converted into activation through the FCR droop curves.

Scenarios combine six charging and market strategies (direct charging, smart charging, spot, spot + FCR-N, spot + FCR-D, spot + FCR-N + FCR-D), three work patterns (work from office, hybrid, work from home), optional rooftop PV (10 kWp) and home battery (30 kWh), district heating or a heat pump, and EV battery ageing either inside or outside the cost function. The objective is to minimise the household’s annual electricity cost.

Key findings

Cost by strategy (2025, work from office, district heating, no PV or battery; figures confirmed against the repository’s Results_optimisation KPI_V1.xlsx):

• FCR-N carries most of the value. FCR-N pays for both capacity and activated energy and had higher average prices than FCR-D, so spot + FCR-N beats spot + FCR-D by a wide margin. In the results data, FCR-N revenue is about 11,700 SEK/yr for the office pattern and about 21,000 SEK/yr for work from home; adding FCR-D on top contributes about 1,700–2,900 SEK/yr.
• Availability matters more than the market mix. Working from home turns the all-markets case into net income: −18,250 SEK/yr. Moving from office to home saves 7,528 SEK with direct charging and 6,005 SEK with smart charging, mostly by spreading the same charging over more hours.
• Home PV and battery lower cost further; PV + battery is always cheapest. With spot + FCR-N, adding a home battery gives −28,273 SEK/yr, about 15,000 SEK better than PV alone. With spot + FCR-D a home battery raises the peak (10.0 kW) and the peak charge, so PV alone does better.
• Market conditions dominate. FCR revenue was at least 50 % higher in 2022 than in 2025 for comparable scenarios, driven by higher reserve prices rather than more reserved capacity; spot + FCR-D cost 6,066 SEK in 2022 against 12,943 SEK in 2025. Without reserve markets, 2022 was about 62 % more expensive on average because of high spot prices. In the sensitivity analysis, halving FCR prices gives the highest cost, and doubling spot-price volatility at an unchanged average price gives the lowest cost (about −76 %) but the highest ageing.

Battery ageing. Direct charging ages the EV battery most, about 4.1 % in the year, because the car sits at high state of charge (calendar ageing). Smart charging cuts about 1 percentage point, and spot trading gives the lowest ageing, 2.7 % (1.4 points below direct charging): more cycling adds less wear than the lower average state of charge removes. Reserve markets age slightly more than spot because capacity has to be held in reserve. Putting an ageing cost in the objective lowers ageing further, most in spot + FCR-N (2.9 % → 2.4 %). The authors’ trade-off for the office pattern: all three markets cost about 14,000 SEK/yr less than spot alone, for about 0.2 percentage points more ageing.

Caveats

The authors state the main limitations themselves, and together they make the results an upper bound on what a household could earn:

• Minimum bid ignored. Svk requires 0.1 MW to take part in the reserve markets; a single 11 kW charger is below that, and the model lets the household bid directly. In practice it would need an aggregator, which takes a share.
• Perfect foresight. Prices, frequency, demand and driving are all known in advance, and every reserve bid is assumed accepted; activation-time requirements are not modelled. The authors expect real revenues to be lower.
• Tax assumption. The energy-tax refund is assumed to apply to all exported electricity, since the rules for homes with several flexible resources are unclear.
• No investment costs. The EV, the bidirectional charger, PV and home battery are excluded, and district-heating costs are not counted.
• Simplified ageing model. Calendar ageing is evaluated on the battery’s state at year end, and the model is linearised; ageing from driving is not included.
• A labelling error in the paper. Section 3.1 says its scenarios use the work-from-home pattern, but its figures (32,378 SEK down to 2,342 SEK) are the work-from-office rows in the results data and match section 3.2’s office figures.

Relevance to the wiki

Data gaps

• The results assume a household can reach FCR directly; no figure yet shows what the same household would net through an aggregator after the aggregator’s share and the minimum-bid pooling.