Financial simulation

Background

The simulator calculates the financial results of the asset from its simulated generation and the expected market prices. The overall financial performance has four parts:

1. trading simulation
2. calculation of PPA settlements for assets with PPAs
3. calculation of expenses (CAPEX and OPEX)
4. calculation of revenues

Calculation logic

Trading simulation

The simulator models participation in both the day-ahead spot market and the balancing market.

Day-ahead spot market participation

We generate a bid for each day from a perfect forecast: the bid equals the simulated generation. For assets with batteries, this bid will also depend on the results of the battery simulation.

Because the bid matches the generation, the simulation produces no imbalance and does not price imbalance cashflows. To include imbalance costs in the results, add them as an OPEX item under the imbalance cost account (see Expense calculation).

Balancing market participation

For assets with batteries, the simulator also models participation in the offline primary reserve of the balancing market, the frequency control reserve (FCR). This shows the additional revenue a battery can earn from frequency regulation.

Supported assets

Assets with battery storage can participate in the balancing market simulation, regardless of battery capacity, as long as the balancing market is included in the asset's target markets.

info

In practice, individual assets must have a capacity of at least 1 MW to participate in the market. Assets with smaller capacities can participate only when aggregated, provided their total capacity meets or exceeds 1 MW. To simplify the current implementation, all assets with batteries are treated as eligible for participation.

Balancing market participation is excluded:

• during FIT subsidy periods
• during the term of a physical offsite PPA
• outside the operational period of the battery
• before April 2024, the start of the market in the simulation

Trading period

For AC-coupled batteries, balancing market participation is possible in any 30-minute slot without solar generation. DC-coupled batteries sit behind the site inverter and can also participate in slots with solar generation (see battery coupling). A slot is also excluded when a curtailment order leaves the battery with insufficient exchange capacity to fulfill the frequency control commitment.

Bid calculation

For each 30-minute slot, the simulator compares the expected revenue of the possible battery operations, looking ahead over a 12-hour horizon, and selects the most profitable one:

1. Day-ahead spot discharge: discharge the battery to the spot market.
2. Frequency control: commit the battery capacity to the balancing market.
3. Frequency control with grid charging: commit capacity to the balancing market while charging the battery from the grid.
4. Idle: do nothing.

The bid volume for frequency control is the full power capacity of the battery (derated by the site inverter efficiency for DC-coupled batteries). For frequency control with grid charging, the bid volume is the battery power plus the power used to charge from the grid.

When comparing the options, the simulator deducts the battery cycle cost of the expected activations from the revenue of both frequency control options, and the cost of the grid electricity from the grid-charging option.

Operational assumptions

• For each 30-minute slot with a frequency control commitment, the battery is assumed to discharge a net 0.0525 kWh per kW of committed capacity, reflecting expected activations. This energy is recharged in later slots.
• All submitted bids are assumed to clear (be accepted by the market).

Revenue calculation

Balancing market revenues are calculated based on capacity compensation:

• You can input the expected ΔkW price in JPY/kW per 30-minute slot for each grid zone as part of your scenario assumptions. As with other scenario parameters, values are anchored at the start of the Japanese fiscal year and interpolated between the years you specify.
• The simulator calculates compensation for each 30-minute slot as the ΔkW price multiplied by the committed capacity.
• Energy discharged during frequency control operations is netted out of day-ahead sales and does not itself generate revenue.
• Settlements are assumed to occur within the same month as the trades.

The model does not account for any compensation from actual frequency response activations.

Transaction fees

EPRX sets transaction fees annually, in JPY per kW of awarded capacity per 30-minute interval (0.01 JPY/kW in FY2024, 0.03 in FY2025, 0.06 in FY2026). The latest published fee is assumed to remain constant for future years, and total participation fees are recorded as an expense in the simulation results. For the latest fees, see the EPRX fee page.

info

For simplicity, the current implementation covers the spot and balancing markets only. We do not simulate intraday markets or other ancillary services such as secondary and tertiary reserves. These will be added in future versions.

PPA settlement calculation

For assets with a PPA, we also calculate the settlements between the IPP and the offtaker, which affect the total revenue.

Strike price calculation

The strike price depends on the PPA pricing scheme. In the formulas below, $P(t)$ is the reference price: the price of the market and grid zone chosen in the PPA's settlement terms.

• fixed: equal to the fixed PPA price.
• fixed with inflation escalation: equal to the PPA price in the first year of the PPA contract, then increasing every year according to inflation assumptions in the scenario, compounded year on year.
• discount with floor: equal to $P(t) \cdot (1 - d)$, where $d$ is the discount to market as a fraction (0.1 for 10%), and never below the floor price.
• discount with collar: equal to $P(t) \cdot (1 - d)$, kept between the floor price and the cap price.
• collar: equal to $P(t)$ without a discount, kept between the floor price and the cap price.

With flexible pricing, the fixed price, the floor and cap prices, and the discount can follow a schedule that changes by contract year.

Settlement amount calculation

There are two main types of PPAs: physical and virtual. For physical PPAs (Off-site PPAs), the energy produced by the plant during the PPA term is not traded on the market, so the trading simulation is not performed for that period. The offtaker buys the net energy exported at the meter (export minus import, after curtailment and battery operation) and pays the strike price for it, which represents the main revenues for the IPP. During the PPA term, the battery is optimized against the strike price instead of the market price.

During any PPA, physical or virtual, the simulation books no revenue from non-fossil certificates.

For virtual PPAs, the energy is sold on the market, and a settlement amount calculated afterwards may transfer funds between the IPP and the offtaker, depending on market conditions. The settlement amount $S$ in this case is calculated as:

$S = P_{strike}(t) - P(t)$

where $S$ is the price per kWh that the offtaker pays the IPP (or the IPP pays the offtaker, if the value is negative) and $P(t)$ is the reference price. The settled volume is the net energy exported at the meter.

example

If in a slot $S$ = 4.5 JPY/kWh and the net export is 400 kWh, the offtaker pays the IPP 1,800 JPY for that slot.

Expense calculation

We calculate the capital expenses (CAPEX) and operational expenses (OPEX) of each asset. CAPEX follows the CAPEX payments entered for the asset.

OPEX costs are the items configured in the asset's OPEX settings. None of them is built in: each item defines its own

• Account: the expense category, such as O&M, asset management, land lease, insurance, hardware replacement, decommission reserve, or imbalance cost.
• Timing: one-time or recurring, with start and end dates set as exact dates or relative to asset milestones, and a monthly or yearly frequency.
• Amount basis: an absolute amount, an amount per solar DC kW, site AC kW, battery kWh or kW, or land area, an amount per kWh of battery discharge or meter export, or a percentage of revenue.
• Change over time: a fixed (JPY) or percentage change per year or month.

Each payment is booked to the month in which it falls.

Grid electricity bought to charge the battery is booked as a cost.

The asset simulation does not calculate property tax, depreciation, or corporate tax. Property tax is calculated in the SPV financial statements (see SPV property tax settings).

Revenue calculation

Total revenue is the sum of all components: revenues from the markets (day-ahead trades), balancing market revenues, subsidy cashflows, PPA settlements, and non-fossil certificate revenues. Certificate revenues come either from the JEPX non-FIT certificate market or from a bilateral agreement, and apply to solar energy sent to the grid and the solar share of battery discharge.

The total cashflow is the revenue minus all costs.

IRR calculation

We calculate the project IRR as an XIRR, a date-based internal rate of return that is annual by construction. The cashflow of each month is the sum of revenues, CAPEX, OPEX, and costs in that month. It is pre-tax and unlevered, and excludes consumption tax. The calculation starts from the first month with CAPEX.

For assets with a battery, we also calculate a battery IRR. Its cashflow is the battery CAPEX plus the revenue gained from the battery after its COD: the total revenue minus the revenue the solar generation alone would have earned at the asset's FIT rate during the FIT period, or at the day-ahead price outside it.

Where the simulation produces probability bands (P50, P85, P90, and P99), results and IRRs are calculated for each band.

info

Sometimes the IRR cannot be calculated: the data may be missing or incomplete (for example, no CAPEX information), or the IRR would take an extreme value (above 10,000%), which usually points to a problem with the input data.

Project value

The project value chart shows, for each month, the net present value of the remaining monthly cashflows. You enter the annual discount rate as a target IRR (5% by default, limited to between -30% and 100%), and it is converted to the equivalent monthly rate. The simulation does not calculate LCOE.