OIES EL36 Electricity Market Design for Decentralized Flexibility (2019)

The electricity market design for decentralized flexibility sources. OIES Paper EL36, Oxford Institute for Energy Studies, July 2019. DOI: 10.26889/9781784671433.

Audit note (2026-09-20): an earlier version of this page contained many claims not in the paper — invented Table 4 rows and bid-size figures, a wrong Table 3 typology, an unsourced FCR rationale, an invented aggregation “recommendation”, and an NC DR mapping table contradicting the paper. The paper’s text names no author (the earlier “Spyros Evangelakakis” and the “EL 17 (Keay 2016)” related-paper reference are not in the raw; only Keay & Robinson 2019 is cited).

Full title: The electricity market design for decentralized flexibility sources (OIES Paper EL36).

Summary

The paper analyses how small, distributed flexibility assets — decentralized flexibility sources (DFS), defined as residential active-consumer demand-side resources (generation is explicitly excluded) — can participate in markets designed for large dispatchable generators. Its core findings: explicit DSM is preferred over implicit DSM; DFS are strategically important for VRE integration; aggregators are central. Preconditions in current markets “sometimes act as discrimination against aggregated resources and DFS”, and the recent trend is barrier removal. (The earlier thesis that the barriers are “unnecessary … serve incumbent interests” is not in the paper.) It also notes there is no regional market beyond pilots, and discusses two aggregator models (independent aggregators, 41, and retailer-aggregators, at least 22), VPP capacity of about 18 GW, and pilots such as GOPACS, an EPEX local platform and Power Potential.

DFS characteristics (Table 3)

The paper characterises DFS by direction, power/energy type, response time (tr, the maximum-power temporal ratio) and availability ratio (ar):

• Residential loads: power type, 5 s < tr < 5 min, response time seconds, ar < 0.1
• Public lighting: unidirectional downward; LED (energy type) or older (power type); ar 0.2–0.5 during peak hours
• Heating and cooling: bidirectional, tr ≈ 15 min, ar 0.4–1
• Batteries: tr from 4 s to 10 h, ar ≈ 1
• EVs: tr from 30 min to 6 h; bidirectional in V2G mode The paper gives response time as “seconds” for all DFS types and treats tr as a separate parameter; it has no “conventional generator” comparison column (earlier version’s column and qualitative descriptions like “slow response (hours)” were unsourced).

VRE integration cost categories (Table 1)

Profile costs (overproduction, full-load-hour reduction, adequacy), balancing costs, and grid-related costs (the paper: “rising costs from the extra care for grid management because of the increasing penetration of VRE”). The paper has no “who bears it” column. It says only that “any means that can help to shift demand or supply would significantly reduce the profile costs” (citing Ueckerdt); the claims that DFS reduce all profile and balancing costs but may raise grid costs if poorly located are not in the paper.

Explicit vs implicit demand-side management

Per the paper, implicit DSM is simpler and cheaper and needs no baseline, but its benefit to the customer is ambiguous, consumers don’t respond effectively, bills are averaged, and only explicit DSM can provide network services and engage short-timescale balancing markets. (Not in the paper: “synchronisation peaks”, “cannot scale without systemic balancing risk”, “no transaction costs”, and the 1–10 MW vs 1–100 kW bid-size figures.) Aggregation is needed because individual DFS are too small.

Aggregation and the BRP-aggregator problem

An aggregator’s activation changes the BRP’s imbalance position. Per the paper, only France and Switzerland have a settlement framework; elsewhere aggregators typically negotiate a bilateral agreement with the BRP, which raised the entry barrier. The Swiss model has the TSO correct the BRP’s position, with compensation based on the quarter-hourly day-ahead price. The transaction-cost discussion covers fixed costs (registration, insurance), acquisition and hardware costs; baseline uncertainty is discussed; cross-market optimisation appears as an innovation aggregators pursue (simultaneous bids), not as a cost driver. The paper says regulation must balance competition, scale and scope, and that BRP/aggregator coordination is best done at EU level, favouring central TSO correction of BRP positions. (The earlier “aggregation must be a licensed, regulated function” recommendation, the “three cost drivers” and the battery/FCR/DSO example are not in the paper.) Wiki context (not the paper): this is the barrier that Network Code on Demand Response and the Swedish BSP/BRP split address.

Market access barriers (Table 4)

Table 4 has three rows:

FCR

The paper explains why FCR has been the first balancing product to become feasible for DFS: it has “the lowest energy component and relatively high capacity remuneration” and is “more acceptable to BRPs because FCR only causes marginal energy imbalance”; without a baseline, capacity-based remuneration is favoured. It is a description, not a formal recommendation. FCR is notified 30 s before real time (French example) and Sonnen’s 1 MW block responded in under 30 s; FCR-N is not mentioned and the paper does not relate the symmetric-bid barrier to FCR.

Nodal pricing at distribution level

Some stakeholders argue only value-reflecting marginal pricing works for DFS; the paper notes nodal pricing is politically hard at distribution level and would “evolve in a decentralized manner”, citing the Cornwall Local Energy Market trial (100+ homes). The “theoretically correct / second-best / recommended interim path” framing and “Sweden’s approach is consistent” are wiki interpretation, not the paper.

Relationship to NC DR (wiki reading, not the paper’s content)

The paper concludes that “customized bilateral contracts and technology tailored prequalification are preferred at the current development stage” and that a common framework “could end up becoming too complex” — the opposite of the earlier “standardize prequalification” recommendation. It does not propose mandatory flexibility registers or cite NC DR articles; any mapping to NC DR is the wiki’s own.

Relevance to wiki

• Demand Response — core explicit vs implicit distinction (paper’s cons of implicit DSM: ambiguous customer benefit, poor response, averaged bills)
• Aggregation — BRP-aggregator settlement conflict; aggregation economics
• Balancing Markets — FCR as preferred DFS entry market; symmetric bidding
• Flexibility Market — market access barriers; nodal pricing as long-run option
• Network Code on Demand Response — almost every recommendation maps to an NC DR article
• Elmarknadshubb — FIS/settlement infrastructure as enabling condition
• SWITCH — Swedish DFS market design follows EL36 framework