BESS Sizing: Why Power and Energy Must Be Evaluated Separately

1. Power and energy are separate design parameters

Power describes the rate of energy transfer, while energy describes the quantity transferred over time. For constant power, E = P × t. A 500 kW load sustained for two hours requires 1,000 kWh. This does not imply that a battery with 1,000 kWh nominal DC capacity can deliver 1,000 kWh at the AC connection point. State-of-charge restrictions, conversion losses, ageing and equipment limits must be considered separately. A battery can have sufficient stored energy but insufficient discharge power, or adequate power but insufficient duration.

2. Define the measurement boundary

Battery energy may be specified as nominal DC capacity at the battery terminals, usable AC energy at the PCS output, or guaranteed delivery at the point of connection (PoC). These are different quantities. Battery racks, the power conversion system (PCS), transformers, cables and auxiliary systems all affect net delivery.

For example, a 10 MW PCS rating does not by itself establish 10 MW export capability. The transformer, connection agreement, conversion losses and auxiliary loads may impose tighter limits. Import and export limits can also differ. The measurement boundary needs to be written into the technical specification and the performance test, rather than inferred from a project headline.

Power capability is not necessarily constant across the full SoC range. Low or high SoC, temperature and BMS restrictions may reduce the available charge or discharge power. A credible design checks the guaranteed operating envelope, not only the nameplate rating.

3. Usable energy and ageing

For an initial energy estimate, E_AC = E_DC,nom × f_SoC × f_retention × η_discharge. Here the factors represent the permitted usable fraction of nominal battery energy, capacity retention at the evaluated age, and aggregate discharge-path efficiency to the defined AC boundary.

This is a simplified energy budget, not a manufacturer performance guarantee. It assumes that the battery starts the event at the required SoC, that the stated operating window remains available and that the efficiency estimate is suitable for the discharge profile. Round-trip efficiency includes charging losses and should not be substituted for discharge-path efficiency.

Battery ageing is influenced by both time and use. Temperature, SoC, cycling depth, throughput and cell chemistry can all affect the degradation trajectory. A fixed annual percentage may be adequate for an early sensitivity study, but it is not a substitute for manufacturer-specific ageing data. Projects with long-term delivery commitments may require initial capacity margin, augmentation or a contractual capacity guarantee.

4. Industrial BESS: peak shaving

Consider a facility with a 1,000 kW import target. Its half-hour average load rises above the target during a production event. The battery must cover the exceedance: P_BESS(t) = max(0, P_load(t) − P_limit). The highest interval requires 500 kW, but the total AC energy over the event is only 750 kWh. The result depends on the time resolution: shorter peaks may require higher instantaneous power than half-hour averages reveal.

5. Industrial capacity calculation

Assume a 90% usable SoC window, 94% aggregate discharge-path efficiency and 80% capacity retention at the evaluated end-of-life point. The preliminary beginning-of-life nominal DC energy requirement is 750 / (0.90 × 0.94 × 0.80) ≈ 1,108 kWh.

This estimate addresses the energy required for the illustrated event, not the complete design. It assumes that the battery is sufficiently charged when the peak begins. A nominal 500 kW / 1.5 MWh configuration provides additional energy relative to this event, but that does not establish it as economically optimal or confirm 500 kW AC delivery at the PoC.

Real facilities rarely repeat an identical load profile every day. Consecutive peaks may arrive before the battery can recharge. Production schedules, weekend operation and grid import restrictions may affect recovery. An engineering study should therefore simulate representative days and seasons, with the actual billing interval and a finer time step where transient power limits matter. Auxiliary loads, control margins and the manufacturer's permitted discharge-power curve also need checking.

The values used here are illustrative assumptions, not project or manufacturer data.

6. Grid-connected BESS: 10 MW / 20 MWh

A project labelled 10 MW / 20 MWh has a nominal power-to-energy ratio of 0.5 h⁻¹, commonly described as a two-hour system. Whether it can export 10 MW for two hours depends on where the rating is defined. A guarantee of 20 MWh net AC delivery at the PoC is not equivalent to 20 MWh nominal DC battery capacity.

Grid-connected systems often have several operating objectives: day-ahead or intraday energy trading, frequency-related reserves and other flexibility services. These uses place different demands on power headroom, usable energy and response time. A dispatch schedule that looks feasible on an hourly energy basis can fail when reserve activation, SoC limits or charging restrictions are included.

The design therefore needs to define both the contractual output boundary and the conditions under which rated power must be available, including the required duration and end-of-life performance.

7. Competing trading and reserve commitments

Suppose a BESS is scheduled to export 8 MW and an upward reserve activation would require an additional 4 MW relative to that schedule. Simultaneous delivery would require 12 MW, exceeding a 10 MW net discharge limit. This is an illustrative operating scenario, not a universal market rule. Actual feasibility depends on the reserve product, baseline, activation and scheduling rules. Energy headroom matters as well as power headroom; downward services can require charging power and room below the upper SoC limit.

8. Reserve qualification and controls

Nominal power and energy capability do not automatically establish eligibility for reserve markets. Fingrid specifies different technical requirements for FFR, FCR, aFRR and mFRR. FFR, FCR and aFRR resources undergo product-specific prequalification tests; mFRR follows a different qualification process rather than the same test procedure. The applicable requirements should be checked against the current product documentation.

Control performance, measurement, communication and availability obligations can influence the equipment and operating strategy. Grid-code compliance and reserve-market qualification are distinct. A BESS may satisfy its connection requirements yet still need additional verification before offering a particular reserve product.

9. Operating strategy and grid connection

The sequence of operating events often matters as much as the largest individual event. A battery may discharge during a high-price period, reserve upward power for activation, then need to recover SoC before the next commitment. If the grid import limit restricts charging, the nominal PCS rating does not describe how quickly the system can recover.

A representative dispatch model should track SoC, charge and discharge limits, discharge-path and charging losses, auxiliary consumption, grid import/export limits, market commitments and degradation costs. It should also prevent incompatible commitments from being counted as simultaneously available.

Simulation resolution must match the application. Hourly models can be useful for preliminary energy trading analysis, but they cannot on their own demonstrate compliance with fast-response reserve obligations. More detailed modelling and equipment-specific testing may be necessary.

10. Choosing the capacity

Additional battery energy can provide flexibility, reduce some cycling stresses and help maintain performance as the battery ages. It also increases capital expenditure and may increase auxiliary or maintenance costs.

For industrial peak shaving, extra capacity may remain unused if exceedances are brief and recharging is readily available. For grid-connected applications, more energy can support longer trading positions or reduce competition between services. Neither benefit should be assumed without modelling the dispatch profile and commercial constraints.

Lifecycle evaluation should consider investment and operating costs, degradation, expected market revenues, availability and any augmentation strategy. The engineering target is a system that satisfies defined power and energy obligations at the relevant boundary, under realistic conditions, over the required service life.