Peaky Flexers
Electricity Market Price Formation — Welfare-Maximising Capacity Investment & Dispatch with Zero Variable Cost Intermittency
Get Flexing — How to Use This Model Documentation & README on GitHub Contact me
Load Duration Curve Blocks
i
Load Duration Curve Blocks
Real electricity demand varies hour by hour. Rather than modelling all 8,760 hours, the model uses a Load Duration Curve that groups hours by demand level into three blocks.
Block Represents Default Winter Peak Coldest, highest-demand hours 300h Shoulder Spring, autumn, milder winter 3,000h Low Demand Summer & overnight 5,460h
Set the hours for each block (should sum to 8,760). Demand levels come from the tier definitions.
These are not calendar months — think of them as buckets grouping hours with similar demand and weather conditions.
Load Duration Curve Blocks
Real electricity demand varies hour by hour. Rather than modelling all 8,760 hours, the model uses a Load Duration Curve that groups hours by demand level into three blocks.
| Block | Represents | Default |
|---|---|---|
| Winter Peak | Coldest, highest-demand hours | 300h |
| Shoulder | Spring, autumn, milder winter | 3,000h |
| Low Demand | Summer & overnight | 5,460h |
Set the hours for each block (should sum to 8,760). Demand levels come from the tier definitions.
Renewables Availability Profile (Sub-Blocks)
i
Renewables Intermittency
Each LDC block is split into two sub-blocks to capture intermittency:
- Low Renewables — hours when renewable output is scarce
- High Renewables — hours when output is plentiful
Parameter Meaning % Hours Fraction of block hours in this sub-block (must sum to 100%) % Availability Fraction of installed renewables capacity available
This creates intra-block price differentials that make storage valuable within a block — charging when renewables are plentiful, discharging when scarce.
Renewables Intermittency
Each LDC block is split into two sub-blocks to capture intermittency:
- Low Renewables — hours when renewable output is scarce
- High Renewables — hours when output is plentiful
| Parameter | Meaning |
|---|---|
| % Hours | Fraction of block hours in this sub-block (must sum to 100%) |
| % Availability | Fraction of installed renewables capacity available |
This creates intra-block price differentials that make storage valuable within a block — charging when renewables are plentiful, discharging when scarce.
Each block is split into Low Renewables and High Renewables sub-blocks to capture intermittency. % Hours must sum to 100 per block.
Demand Tiers & Expandable Demand
i
Demand Tiers
Demand in each block is divided into three tiers by willingness to pay:
- High-value — critical loads (hospitals, data centres). High VoLL, high shift cost.
- Mid-value — important but flexible (commercial heating). Moderate VoLL.
- Low-value — deferrable loads (EV charging, water heating). Low VoLL.
Parameter Meaning Quantity (GW) GW of this tier present during the block VoLL Max price this demand will pay; above this it is curtailed Shift Cost Cost to move demand to a later, cheaper block (downward only)
Expandable demand activates when prices fall below its value — representing new consumption attracted by cheap power.
Demand Tiers
Demand in each block is divided into three tiers by willingness to pay:
- High-value — critical loads (hospitals, data centres). High VoLL, high shift cost.
- Mid-value — important but flexible (commercial heating). Moderate VoLL.
- Low-value — deferrable loads (EV charging, water heating). Low VoLL.
| Parameter | Meaning |
|---|---|
| Quantity (GW) | GW of this tier present during the block |
| VoLL | Max price this demand will pay; above this it is curtailed |
| Shift Cost | Cost to move demand to a later, cheaper block (downward only) |
Expandable demand activates when prices fall below its value — representing new consumption attracted by cheap power.
Supply-Side Technology Costs
i
Supply Technologies
The model determines how much of each technology to build. No capacities are pre-set.
- Renewables — zero fuel cost; output limited by availability profile. Input: annualised capital cost.
- Nuclear — firm baseload capacity with low variable cost and zero carbon emissions. Inputs: annualised capital cost, variable cost.
- Gas Peaker — flexible, burns gas. Inputs: capital cost, variable cost, carbon price, emission factor.
- Storage — specified via CapEx: duration (h), connection cost (£/MW), cell cost (£/MWh), asset life, discount rate (%). Annualised using capital recovery factor. No fuel cost — implied cycling cost is an output.
- T&D — network capacity, sized to peak generation. Input: annualised capacity cost.
Storage conversion: power cost = (connection/1000)×CRF; energy cost = (cell cost/1000)×CRF, amortised over cycles/year. CRF uses discount rate and asset life.
Supply Technologies
The model determines how much of each technology to build. No capacities are pre-set.
- Renewables — zero fuel cost; output limited by availability profile. Input: annualised capital cost.
- Nuclear — firm baseload capacity with low variable cost and zero carbon emissions. Inputs: annualised capital cost, variable cost.
- Gas Peaker — flexible, burns gas. Inputs: capital cost, variable cost, carbon price, emission factor.
- Storage — specified via CapEx: duration (h), connection cost (£/MW), cell cost (£/MWh), asset life, discount rate (%). Annualised using capital recovery factor. No fuel cost — implied cycling cost is an output.
- T&D — network capacity, sized to peak generation. Input: annualised capacity cost.
Storage conversion: power cost = (connection/1000)×CRF; energy cost = (cell cost/1000)×CRF, amortised over cycles/year. CRF uses discount rate and asset life.
The model determines optimal capacities. No fixed capacities are pre-set.
Zero variable cost generation (wind, solar). Output depends on the Renewables Availability Profile above.
Firm baseload with zero carbon emissions. Planned outage reduces availability in all periods. Peak unavailability is an additional derating applied only during winter peak sub-blocks.
CapEx is annualised using the capital recovery factor (CRF) with the discount rate: power cost = (connection/1000)×CRF; energy cost = (cell cost/1000)×CRF, then amortised over cycles.
Sized to peak system generation. The constraint binds at the sub-block with the highest total dispatched power.
Household Bill Optimisation (Optional) ▶
Takes system clearing prices and optimises a single household’s demand, storage, and curtailment.
Market Equilibrium Summary
Optimal Capacity Mix
Input Load Duration Curve
Raw demand blocks as entered, before any optimisation, shifting, or storage.
Generation Duration Curve
Total generation by source in each sub-block after optimisation, sorted by GW.
Mouse-over within blocks for more information.
Price Duration Curve
Clearing prices sorted high to low, with demand served, shifted, and storage activity per sub-block.
Demand & Supply by Sub-Block (GW)
Energy by Sub-Block (GWh)
Each bar shows GW × hours for that sub-block. Storage charging appears as demand; storage discharging appears as supply.
Storage Activity
Detailed Sub-Block Results
Comments & Feature Requests
Comments, bug reports, and ideas for new features are very welcome.
If you'd like to make a public suggestion about the model, please open a GitHub issue. If you'd prefer to contact Tony directly, use the email link below.