Changelog
Each change description and its corresponding rationale detail the updates made relative to the previous version of the methodology.
Version | Date | Change description | Reasons for change |
v1.0 | 31-August-2026 | Initial version of the revenue modelling methodology |
|
Overview
1. Asset selection criteria
Revenues are modelled for generation units that meet the following criteria:
Operational Status: The unit is either operational or under construction. For under construction assets, the revenue forecasts are solved using the hub price curves in this version of the product.
Configuration: The unit is standalone and not connected to an another energy generation unit within the same power plant aggregate.
Category: Revenues are only modelled for front-of-the-meter (FTM) units.
Asset characteristics, including installed capacity, commissioning date, duration of the battery and other relevant information, can be found in the Wood Mackenzie Lens P&R platform.
2. Battery Asset revenue
The Battery Revenue Forecasting Model estimates the revenue potential of front-of-the-meter battery energy storage systems (BESS) through a revenue-maximizing dispatch framework. The methodology co-optimizes participation across energy, ancillary service, and capacity markets while accounting for battery operating characteristics, market participation rules, and operational constraints.
The model simulates battery dispatch at market settlement intervals and determines the economically optimal charging, discharging, and market commitment strategy under forecast market conditions. The resulting dispatch profile is used to calculate revenues, battery utilization, state-of-charge trajectories, and key operating metrics over the forecast horizon.
3. Revenue components
The total revenue of a battery is calculated using a revenue stacking approach that captures the value streams available to battery storage assets in the markets.
Energy Revenue: Energy revenue is earned through energy arbitrage, whereby the battery charges during lower-priced periods and discharges during higher-priced periods. Revenue is estimated using Wood Mackenzie's day-ahead and/or real-time nodal price forecasts, depending on market structure and product configuration.
The dispatch optimizer identifies economically optimal charging and discharging intervals while accounting for efficiency losses, operational constraints, and competing market opportunities.
Ancillary Revenue: Ancillary service revenue represents compensation for providing grid support services such as regulation, spinning reserves, non-spinning reserves, and other market-specific reserve products. The dispatch optimizer evaluates the opportunity cost of reserving capacity for ancillary services relative to energy market participation and allocates battery capacity to the combination of services that maximizes total revenue. Ancillary service participation is modelled using expected deployment factors and minimum energy availability requirements, ensuring that revenues are evaluated alongside the associated operational impacts on battery utilization and state of charge.
Ancillary throughput revenue: Ancillary throughput revenue represents the value associated with energy injected to or absorbed from the grid during ancillary service deployment events. The model estimates expected throughput using service-specific deployment assumptions and incorporates the resulting revenues, efficiency losses, state-of-charge impacts, and cycle consumption within the optimization framework.
Capacity Revenue: In markets with capacity remuneration mechanisms, batteries may earn capacity payments in exchange for maintaining availability during system stress periods. Capacity revenue is calculated using Wood Mackenzie's Long-Term Outlook (LTO) capacity price forecasts and Effective Load Carrying Capability (ELCC) assumptions, which represent the asset's accredited contribution to system reliability. Market-specific operational requirements, including state-of-charge obligations and performance criteria, are incorporated into the dispatch optimization process.
Capacity revenue is not currently modelled in the ERCOT MVP (see Section 6). The methodology above applies where capacity revenue is enabled for a given market.
Tax Credits and Incentives: Federal, state, and local incentive programs may provide additional value to battery energy storage projects. These incentives, including investment-based tax credits and other policy support mechanisms, are asset-specific and depend on eligibility requirements, project configurations, and prevailing regulations. Such incentives are not currently modelled and are excluded from forecast revenues presented in this product.
4. Price Forecasts
Real-time Energy Price: Wood Mackenzie forecasts sub-hourly, nodal real-time (RT) energy prices using a statistical simulation framework. The methodology combines Wood Mackenzie's long-term power price outlook with historical nodal market behaviour to generate forward-looking price trajectories. This framework is designed to do both simultaneously: preserve the fundamentals-driven trend from Wood Mackenzie's power price outlook while replicating the empirical shape of real-time price distributions.
Key modelling approach
Wood Mackenzie's long-term hub price forecasts are used as the primary market anchor.
Historical hub-to-node price relationships are used to derive location-specific nodal prices.
Historical real-time market data is used to characterize volatility, intraday spreads, scarcity events, and negative pricing behaviour.
A statistical simulation framework generates forward-looking price trajectories that preserve the above mentioned historical market characteristics.
Forecast prices are calibrated to maintain consistency with expected nodal basis relationships and long-term market fundamentals.
Ancillary price: Wood Mackenzie forecasts ancillary service (AS) prices using a two-stage, market-specific framework that separates the forecast of physical reserve requirements from the forecast of price formation. This sequencing is deliberate: AS prices are a function of scarcity, and scarcity is itself a function of how procurement volumes evolve relative to the pool of resources able to supply flexibility. Forecasting requirements first allows structural shifts in system flexibility needs, driven by renewable penetration, resource adequacy standards and storage deployment, to flow through mechanically into price outcomes rather than being inferred from historical price trends alone.
Key modelling approach
Two-stage architecture: (i) reserve procurement requirement forecast (MW), followed by (ii) price forecast conditioned on projected scarcity and Wood Mackenzie's energy market outlook.
Requirement forecasts are driven by system fundamentals, renewable penetration and growth, resource adequacy margins, flexible thermal capacity, and battery storage deployment, rather than simple trend extrapolation.
Regulation Up, Regulation Down, Spinning Reserve and Non-Spinning Reserve are modelled jointly to preserve the historical co-movement between products, avoiding independent forecasts that would understate correlated scarcity events.
Price formation combines statistical time-series methods with machine-learning-based estimators, validated through out-of-sample back-testing against realised settlement prices.
4.1 Revenue Scenarios
Revenue forecasts for battery assets are modelled with several price curve scenarios and assumptions,
Power & Renewables forecast: The primary revenue forecast for battery asset is based on Wood Mackenzie's long-term power price forecasts. For example, in ERCOT market, revenues are currently modelled using the latest available base case forecast, i.e., 2026 H1 SPO Base Case. Additional forecast scenarios will be incorporated as they become available..
Another point to note here is how the price timeseries is constructed for battery assets. For example, the 2026 H1 SPO Base Case forecast begins in January 2026. The storage optimization, however, is solved across the asset's full operating life — from its commissioning date (which may fall before January 2026) through decommissioning. For any year prior to the forecast's start, the model uses realized historical market prices rather than a forecast, and the dispatch optimizer solves for the theoretical optimal charge, discharge, and market participation schedule against those actual prices.
This is a deliberate choice, not a data-availability workaround: using observed prices for elapsed years avoids effectively forecasting the past, and ensures every asset is evaluated with the same optimization methodology across its full life. Note this differs from the Historical Realized (Market Dispatch) scenario below, which reflects the asset's own observed dispatch. This realized-price approach instead estimates what an optimally-dispatched battery could have captured under real historical market conditions — the relevant benchmark for lifetime valuation and IRR, where the question is the asset's underlying value potential rather than how it happened to be bid on a given day.
Historical Realized (Market Dispatch): Where market data availability permits, a historical realized revenue case is provided using observed asset dispatch data and market prices. This scenario is intended to assess actual battery performance in the market. For select ISOs, such as ERCOT, asset-level dispatch information is available through market operator reporting (e.g., SCED reports). Historical dispatch data are combined with realized market prices to estimate market revenues and benchmark observed performance against modelled results. These values are only available in the Lens Direct table.
5. Battery Dispatch Optimization
The battery is modelled as a price-taking asset and is assumed not to influence market prices. The optimization is performed using a perfect foresight framework, whereby future market prices within each optimization horizon are assumed to be known. The dispatch optimizer uses these forecast prices to determine the economically optimal charging, discharging, and market participation decisions subject to battery and market constraints. Actual realized revenues may differ due to forecast uncertainty, bidding strategies, operational outages, market rule changes, and other real-world factors.
6. Model Assumptions and Constraints
Parameter | Methodology | ERCOT Assumption (MVP) |
Project lifetime |
| 20 years |
Dispatch framework | Price-taking, revenue-maximizing optimization | Price-taking optimization |
Optimization approach | Perfect foresight within the optimization horizon | Perfect foresight within 3-day rolling window |
Optimization horizon | Rolling 3-day optimization window | 3 days |
Optimization refresh | Dispatch re-optimized every 2 days | 2 days |
Round-trip efficiency (RTE) | Asset-specific assumption | User-defined (default: 85%) |
Power capacity | Asset-specific MW rating | User-defined |
Energy capacity | Asset-specific MWh rating | User-defined |
State of charge (SOC) limits | Asset-specific minimum and maximum operating SOC | User-defined |
Initial state of charge | Default assumption | 50% |
Cycle limits | Asset-specific daily and/or annual cycling constraints | 2 cycles/day |
Degradation/ Variable operating cost | Asset-specific assumption | $10/MWh physical cost, $70/MWh costless adder |
Energy charging and discharging | Simultaneous charging and discharging are permitted within the same market interval | Applied |
Energy market participation | Day-ahead commitments are enforced within real-time dispatch intervals | RT Energy Purchase and RT Energy Sale enabled; DA Energy not optimized |
Ancillary deployment | Service-specific deployment assumptions applied to estimate energy throughput | Reg Up = 30%, Reg Down = 30%, Spin = 0%, Non-Spin = 0%, Non-Spin Slow = 0%. The percentage value shows the expected deployment of ancillary service capacity. |
Ancillary readiness requirements | Product-specific minimum energy availability and SOC requirements | Reg Up = 1 hr, Spin = 1 hr, Non-Spin = 4 hr, Non-Spin Slow = 2 hr, Reg Down = 1 hr |
Capacity market participation | Market-specific accreditation and SOC obligations incorporated where applicable | Not currently modeled in ERCOT MVP |
Eligible market products | ISO-specific participation rules | DA Reg Up (20%), DA Reg Down (20%), DA Spin (100%), DA Non-Spin (100%), DA Non-Spin Slow (100%), RT Energy Purchase (100%), RT Energy Sale (100%). The percentage value indicates the maximum participation level allowed for each product. |
7. Model Outputs
The optimization model generates results at multiple temporal resolutions (annual, monthly, and hourly/5-minute dispatch data):
Revenue Outputs: Market revenues and costs by service and year (For ERCOT - Energy Revenue, Ancillary Revenues - Reg Up/Down, Spin (RRS), Non-Spin, Contingency Revenue (ECRS), Ancillary Throughput Revenue).
Dispatch Outputs: Annual operational metrics including market awards, charging/discharging activity, energy throughput, and state of charge (SOC).
Ancillary Service Throughput: Time-series of ancillary service deployment used to quantify battery degradation and throughput-related impacts.
Monthly Outputs: Monthly aggregation of revenues, dispatch, and utilization metrics for trend analysis.
Hourly/Sub-hourly Outputs: Chronological operational schedules, market participation, charging/discharging behavior, and SOC trajectories used for detailed operational assessment.