Common-path evaluation of electric boiler and thermal storage dispatch in Finnish district heating under electricity-price uncertainty

The value of uncertainty-aware district-heating dispatch depends on when prices become known and how long heat remains in storage. This study develops a price-taking mixed-integer linear model for a Finnish system with biomass CHP, a 50 MW th electric boiler (EB), a natural-gas peak boiler, and thermal energy storage (TES). Daily 72 h problems implement the cleared first 24 h and propagate each policy's realised TES state. Deterministic, stochastic, and budgeted robust policies are compared on one 2023 hourly path and on 800 held-out price trajectories; their economic differences are assessed against TES value and parameter sensitivity. With 200 MWh th TES, annual costs are EUR 4.902, 4.920, and 4.939 million, respectively, compared with EUR 4.783 million for a rolling perfect-foresight benchmark. Stochastic dispatch has the lowest held-out mean cost and regret, while deterministic dispatch has slightly lower P 95 and CVaR 95 . TES saves EUR 1.974 million a − 1 relative to matched no-storage dispatch; the stochastic and robust annual premiums are only EUR 0.037 and 0.075/MWh th . Storage mainly shifts heat within one day. A 200-sample annual rolling sensitivity analysis identifies CHP power-to-heat ratio, EB capacity, fuel price, and TES duration as the dominant cost drivers. Biases of ±5% and ± 10% in look-ahead heat demand preserve the annual policy ranking. The EB dispatch-weighted carbon intensity is approximately 26 g CO2/kWhel, below the national consumption-weighted mean of 37.69 g CO2/kWhel.