How reserve procurement works for DSOs

At the DSO level, reserve procurement receives less attention than its better-known TSO counterpart, yet distribution operators face their own obligations and inefficiencies. This plain-language guide explains how the process works, which counterparties are involved, and where costs escape.

The procurement chain at a glance

A regional DSO must ensure that generation supplying its network can meet expected demand throughout the next operating period. For day-ahead planning, it commits enough generation to cover forecast demand and an operating reserve margin before the procurement gate closes.

Several parties form this chain. The DSO forecasts demand across its coverage zone, then sends its reserve requirement to the upstream connection point, which may be a transmission system operator (TSO) or a regional balancing entity, depending on the network architecture. Reserve generation may be procured through a capacity contract with named units, a bilateral agreement with a generation company, or participation in the day-ahead balancing market, subject to local market rules.

In Portugal, REN oversees the balancing market and works with DSOs on reserve requirements. DSOs within REN's zone are not direct price-setters in the MIBEL wholesale market. Instead, they operate through contracted capacity and regulated procurement mechanisms administered by REN.

Day-ahead and intraday procurement

Day-ahead procurement costs less than intraday procurement. Reserve committed 24-hour ahead receives the day-ahead rate because the generator can plan unit commitment and fuel logistics. Capacity committed 4 hours ahead requires a short-notice response to activation, creating a higher opportunity cost and price.

DSOs therefore seek to place as much reserve as possible day-ahead and leave as little as possible for intraday procurement. Each MW shifted from intraday to day-ahead is a saving. Day-ahead load forecast accuracy chiefly determines how much reserve can be confidently sized and committed in advance.

A pessimistic forecast can make a DSO over-commit day-ahead, leaving it paying for unused capacity. An optimistic forecast can cause under-commitment, leaving a residual to cover intraday at a higher price. The target is a day-ahead position that follows actual demand, with a reserve margin reflecting the real probability distribution of demand uncertainty.

Where costs escape: over-procurement

Distribution operators commonly procure too much day-ahead reserve. This is understandable: shortfalls carry a more visible immediate risk than surplus capacity, security-of-supply rules require protection, and uncertain forecasts can push cautious operators toward larger reserve margins.

This pattern stands out when demand falls well below forecast. Contracted generation then runs during part of the dispatch period but is not dispatched later because demand is lower than expected. Its capacity payment remains due. The operator has paid for unneeded generation, yet avoiding a shortage on a high-demand day seemed worth more than the unused reserve.

The financial impact depends on standby rates and how often over-procurement occurs. An operator paying EUR 35 per MW per day under standby contracts, while procuring an average 12 MW too much across 200 operating days annually, spends approximately EUR 84,000 per year on unused reserve. The total increases with feeder count and the severity of the bias.

Forecast timing constraints create urgency

In the Iberian market system, the day-ahead gate typically closes in the prior day's early afternoon. The DSO therefore needs the next-day load forecast and must commit reserve from it by roughly 12:00 to 14:00 on the preceding day.

Once the gate closes, the committed position can be changed only through limited intraday balancing sessions held during the day and evening before delivery. Each session has its own gate closure, while available volumes and prices generally deteriorate as delivery nears.

This schedule imposes a firm forecast deadline. A superior forecast arriving at 20:00 the night before cannot alter the day-ahead commitment. The useful forecast is available by 11:00 to 13:00 on the day before delivery. If improved weather information, such as a meteorological update that materially changes the next day's temperature forecast, arrives after gate closure, its value can be captured only through costlier intraday changes.

That is why forecast refresh frequency matters. An intraday forecast updated every 15 minutes has value, but the biggest cost-reduction opportunity is improving the 12:00 to 14:00 day-ahead forecast, since that decision fixes the largest share of next-day procurement cost.

What better forecasting changes, and what it does not

A more accurate day-ahead load forecast lets the DSO set its reserve margin more precisely, directly reducing over-procurement. A probabilistic forecast using P10/P50/P90 also shows the commercial team how uncertain the forecast is, allowing an informed reserve choice instead of a fixed rule of thumb.

Better forecasting does not alter the market structure, contracted unit costs, or gate closure schedule. A DSO with an expensive generation contract, high standby rates, or inflexible minimum commitment volumes cannot match the procurement efficiency of one with better terms, even with excellent forecasts. Forecast accuracy and contract quality complement each other; neither replaces the other.