Energy storage for businesses – applications and sizing

Sizing an energy storage system for a business should not start with the question: “how many kWh?”. First, you need to establish what problem the system is meant to solve: short power peaks, photovoltaic surpluses, limiting grid import, variable energy prices, a need for flexibility or backup power. Only then can you determine the required storage power, energy capacity, control strategy and expected operating profile.

Two plants with the same monthly, or even annual, energy consumption may need completely different storage systems. One may have power peaks lasting only several minutes during equipment start-up, while another may have several hours of surplus from industrial photovoltaics. In the first case, adequate power will be the key parameter; in the second, the amount of energy that can be shifted over time may matter much more.

Auditor’s note: the information that a business consumes, for example, 500 MWh per month says almost nothing about whether the storage system should be rated at 100 kW, 500 kW or 1 MW, or how long it should sustain that power. Proper sizing requires a time-series profile and a clearly defined operating algorithm.

First define the problem the energy storage system is meant to solve

An energy storage system is a controllable asset. Whether it delivers the expected result depends not only on battery technology, but above all on the purpose for which it will be charged and discharged. The same system can technically perform several tasks, but those functions may compete for available power and state of charge.

Before analysing offers, it is worth answering several basic questions:

Which specific parameter should be improved: peak demand, renewable energy self-consumption, energy imports during specific hours, security of supply or availability of flexible capacity?
How long does the event that the storage system needs to respond to usually last?
How often do such events occur, and can they be predicted?
Where will the storage system charge from: the grid, PV, cogeneration or several sources?
Should part of the stored energy remain permanently reserved for emergencies?
Which EMS/BMS will make the charging and discharging decisions?

Only after the function has been defined should you move on to technical parameters. In practice, the two most important values are power [kW or MW], which indicates how quickly the system can absorb or deliver energy, and energy capacity [kWh or MWh], which defines how much energy can be shifted over time. The energy-to-power ratio approximately describes how long the system can operate at a given power level.

Applications of energy storage in a business

At an industrial site, an energy storage system can perform several very different functions. Distinguishing between them is important because each function leads to a different sizing approach.

Peak shaving – limiting short power-demand peaks

Peak shaving means discharging the storage system when the site’s demand approaches a predefined threshold. The objective may be to limit maximum import at the connection point, reduce the risk of exceeding contracted capacity, or relieve internal infrastructure during short peaks.

In this application, the storage system often has to provide relatively high power for a relatively short time. What matters is therefore not monthly energy consumption, but the magnitude, duration and repeatability of the peaks. The controller should respond to current demand and maintain an adequate state of charge before an expected peak.

This example shows the difference between power and energy capacity. If we want to maintain a constant 200 kW reduction in grid demand for a full 15 minutes, we theoretically need to shift 50 kWh of energy (200 kW × 0.25 h). If the same reduction must be maintained for a full 2 hours, 400 kWh is required (200 kW × 2 h). This is only the starting point: the design must also account for usable SoC limits, efficiency, operating reserve and degradation.

Increasing self-consumption of renewable energy

When working with photovoltaics, the storage system charges during periods when generation exceeds the site’s current demand and then supplies that energy later. In this application, the key factors are the amount of surplus energy and the time between the surplus occurring and the subsequent demand.

A storage system cannot be sized reliably for PV using annual generation and annual consumption alone. Their coincidence in time is what matters. A site consuming 10 GWh per year may have very high daytime demand and export almost no PV energy, or it may operate mainly in the evening and at night, creating large surpluses during sunny hours. The annual balance may look similar in both cases, while the storage potential will be completely different.

For the analysis, renewable generation and site demand need to be compared using the same time intervals. The control system can additionally use generation and load forecasts so that the battery is not filled too early or discharged before the period when it will be needed most.

Power management and connection-capacity constraints

An energy storage system can also be part of a broader energy monitoring and management system. It can help keep grid imports below a set value, support operation of a site with limited connection capacity, and coordinate with vehicle chargers, local generation and controllable loads.

This application is broader than conventional peak shaving. The control system does not have to respond only to the highest peak of the month. Over many hours, it can coordinate storage with the production process and other assets to comply with technical or contractual constraints. This requires knowledge not only of the energy profile, but also of the installation topology, transformer ratings, protection systems, distribution agreements and planned load changes.

Price arbitrage

Arbitrage involves charging the storage system when energy is cheaper and using it when energy is more expensive. In a business, whether such control makes sense depends on the actual energy procurement model: the tariff, a market-indexed product, hourly prices or the terms of cooperation with an aggregator.

The price spread alone is not yet a saving. The model must account for round-trip efficiency, charges that depend on how energy is imported and exported, storage availability and the cost of consuming its finite cycling resource. An arbitrage strategy may involve more frequent cycling than a system used only for occasional peak shaving, so it has a major impact on projected throughput and degradation.

Flexibility and service provision

Storage is a controllable resource: it can quickly increase or decrease power import or export. This means it can support the business’s energy flexibility and – once the relevant technical and market conditions have been met – be used to provide services either directly or through an aggregator.

In this application, power, response speed, communication with the supervisory system and maintaining an appropriate SoC are particularly important. Potential revenue from flexibility services should not be entered automatically into the financial model. First, confirm whether the asset can participate in a specific mechanism, what availability requirements apply and whether participation conflicts with the storage system’s primary function at the site.

Backup power and resilience to supply interruptions

If the objective is to maintain critical loads, the sizing approach changes fundamentally. You need to determine the power of the critical loads, the required backup duration and the method of transition to backup operation. Starting currents, the sequence in which loads are connected and the ability to operate in island mode are also important.

Not every storage system operating in parallel with the grid automatically provides power during a grid outage. Backup capability must result from the architecture of the entire system: the inverter, automation, switchgear, protection and method of isolation from the grid. If a reserve must be guaranteed at all times, part of the SoC has to remain unavailable for arbitrage or increasing self-consumption.

In a broader context, storage can be one element of a company’s energy security, but it should not replace a reliability analysis of the entire power supply system.

One installation, several functions? Yes, but a hierarchy of priorities is required

Combining applications, often referred to as value stacking, can improve storage utilisation, but it requires deliberate optimisation. Available energy capacity and power are limited, so every EMS decision affects the ability to perform the next task.

A backup reserve reduces the energy available for arbitrage and self-consumption.
Frequent arbitrage increases the number of cycles and throughput, which can accelerate wear.
Peak shaving requires power to be available exactly when the peak occurs, so earlier discharge for another purpose can deprive the system of this function.
Charging from the grid in preparation for a peak can conflict with the objective of limiting current grid demand.
Participation in flexibility services may require part of the power and SoC to be kept in reserve.

For this reason, the storage specification should describe not only the hardware parameters but also the control priority order. For example, security of supply may be the first priority, the import limit the second, while arbitrage operates only within the remaining operating range.

Power in kW and energy capacity in kWh – why must they be sized separately?

In energy storage offers, capacity is often the most prominently presented parameter, but from the site’s perspective the power rating of the PCS/inverter is equally important. Two storage systems with the same capacity can have completely different capabilities when it comes to handling a short power peak or starting loads.

Application	What primarily determines power?	What primarily determines energy capacity?	What must the control system do?
Peak shaving	Magnitude of the peak to be reduced	Duration and number of consecutive peaks	Respond to demand and preserve energy for the peak
Renewable energy self-consumption	Maximum available surplus and subsequent demand	Surplus energy to be shifted between time periods	Coordinate PV, load, forecasts and SoC
Power management	Difference between forecast demand and the limit	Duration for which the limit must be maintained	Coordinate storage and other resources in real time
Arbitrage	Charging/discharging rate required during price windows	Energy that can be shifted economically	Optimise operation against prices, losses and cycling cost
Flexibility	Required power and response speed	Required service duration	Maintain power availability and SoC in line with service requirements
Backup	Critical-load power and starting requirements	Required backup duration	Maintain reserve and manage island-mode operation

There is therefore no universal “correct” power-to-energy ratio. Short-duration power applications require a different configuration from shifting energy over several hours. Sizing should result from a simulation of the specific profile and function, not from a simple capacity-to-annual-consumption ratio.

What data is needed to size an energy storage system correctly?

The minimum data set depends on the system’s function, but for a business it most often includes:

a grid import and export profile in intervals of at least 15 minutes for as complete and representative a year as possible; higher resolution may be required when analysing very short peaks,
a PV, cogeneration or other generation profile at the same time resolution,
maximum power values, peak durations and information on which equipment causes them,
energy supply contract terms, the pricing method and distribution settlement rules,
contracted capacity, connection capacity and constraints at the connection point,
a single-line diagram of the installation, parameters of transformers, switchgear and main protection devices,
a list of critical loads and the required backup duration if the storage system is to provide backup power,
information on planned new loads and generation sources: production lines, EV chargers, PV, cogeneration or process electrification,
site and environmental conditions, as well as available space for the BESS and associated infrastructure.

Why are monthly energy bills not enough?

A monthly bill allows you to assess the scale of consumption and costs, but it aggregates thousands of events into a single value. It does not show whether 500 MWh was consumed evenly or concentrated in several intensive periods. It also does not show the coincidence of PV generation and demand, the duration of exports to the grid, or whether high power levels occurred for 15 minutes or for several hours.

Auditor’s note: monthly consumption is useful for initial screening, but not for final BESS sizing. If the objective is peak shaving or self-consumption, we need to see the time profile – precisely in the intervals in which the storage system will make decisions.

What should the analysis and simulation of storage operation look like?

A proper analysis is not a one-off calculation that divides consumption by battery capacity. It should recreate storage operation step by step using actual data and test how the system would behave with different parameters.

Define the objective and KPIs. For example: an 800 kW import limit, reduced PV export, 2 hours of backup for critical loads, or operation based on hourly energy prices.
Prepare the time-series data. Check for missing data, outliers, production changes and atypical periods.
Establish the baseline. What happens without storage: when peaks, exports, high prices and constraints occur.
Simulate power and energy capacity. Test several variants rather than one preselected size.
Simulate the EMS algorithm. The storage system should be charged and discharged according to the intended operating logic, taking SoC, power and efficiency constraints into account.
Analyse cycling and degradation. Calculate not only the number of full cycles, but also energy throughput and the operating pattern in each application.
Test scenarios. Vary production, prices, PV capacity, site load, planned investments and the required reserve.
Perform technical verification. Electrical integration, connection, protection, communications, location and safety requirements.

The model should compare the storage case with a no-storage scenario. Otherwise, it is easy to attribute to the BESS benefits that actually result from photovoltaics alone, a change in the production profile or a different energy procurement model.

Efficiency, cycling and degradation – parameters that cannot be ignored

Round-trip efficiency

Every charge and discharge cycle involves losses. Round-trip efficiency (RTE) describes the ratio of energy recovered to the energy used to charge the system after returning to a comparable state. Offers should specify at what system boundary the efficiency is measured – for example, whether it covers the entire AC–AC path including the PCS and auxiliary loads, or only the battery section.

This is particularly important for arbitrage and increasing self-consumption. You cannot assume that 1 MWh delivered to the storage system means that 1 MWh will later be available to the loads.

Number of cycles and throughput

Two storage systems averaging one cycle per day may age differently if they operate at different depths of discharge, temperatures, power levels and SoC ranges. The operating model should therefore also track throughput, meaning the total energy flowing through the system, as well as equivalent full cycles.

The control strategy therefore has a direct impact on asset wear. A storage system used occasionally to protect against peaks may have a completely different degradation profile from a system that performs arbitrage and PV energy shifting every day.

Calendar and cycle degradation

The capacity available in the first year is not constant over the entire operating life. Ageing is affected both by the passage of time and by how the system is operated. Depending on the technology, relevant factors include depth of discharge, temperature, cycling intensity and the length of time spent at high or low SoC.

Correct sizing should therefore also be verified for later years. If the minimum capacity required to perform the function is available only at the start of the project, the system may stop meeting the assumptions after several years even though the battery is still operational. The specification should compare time-based warranty terms, guaranteed throughput/cycles, operating conditions and guaranteed capacity retention.

Should battery technology be selected first?

No. The storage technology should follow from the requirements, not be the starting point. A technology-based classification does not answer the investor’s most important question: can the technology perform the required function for the required period and at an acceptable life-cycle cost?

When selecting a solution, compare, among other things, the power and discharge-duration range, whole-system efficiency, thermal conditions, degradation, safety, warranty, service, siting requirements and the method of integration with the electrical installation. Only then does it make sense to compare specific technologies and manufacturers.

Electrical integration, control and safety

Even a correctly sized energy capacity is not enough if the storage system cannot work properly with the site’s electrical installation. At the design stage, verify:

the BESS connection point and power flows in the individual sections of the installation,
transformer, switchgear and protection parameters,
short-circuit levels and protection selectivity to the extent required by the design,
AC-coupled or DC-coupled architecture where the storage system works with renewable generation,
communication between the meter at the connection point, the BMS, PCS, EMS and other generation sources,
the operating mode during a grid outage where backup operation is required,
installation conditions, ventilation/cooling, service access and fire protection requirements.

In Poland, the formal procedure depends on the installation parameters, connection arrangement and location. Under Article 32(1)(2)(a) of the Polish Energy Law, carrying out business activity involving electricity storage in storage facilities with a total installed electrical capacity greater than 10 MW requires a licence. The electricity system operator maintains a register of storage facilities connected to its network, and the obligation to provide data for the register applies to storage facilities with installed capacity greater than 50 kW in the cases specified in Article 43g of the Polish Energy Law. For capacities above 50 kW and not exceeding 10 MW, the Polish government portal gov.pl describes registration instead of licensing. Additional obligations may arise, among other things, from the location, project area, connection conditions, and building and environmental regulations, so the procedure must be checked for each specific project against the current legal framework.

How to size an energy storage system for a business correctly, step by step

Define the primary function. One sentence should describe the problem, for example: “keep grid import below 900 kW” or “shift PV surpluses from 10–15 to the second shift”.
Collect data at the appropriate resolution. Monthly bills are only the starting point; simulation requires a time-series profile.
Determine the required power. It follows from the maximum difference between the baseline and the target limit, or from the power of the loads that need to be supplied.
Determine the required energy. It follows from the time for which the storage system must maintain the given function and from subsequent events occurring before it can recharge.
Add real-system constraints. SoC, charging and discharging power, efficiency, availability, temperature and reserve.
Design the EMS algorithm. Without control logic, even correctly sized hardware can operate at the wrong time.
Calculate cycling and degradation. Check not only the first year, but also the requirements at the end of the assumed operating period.
Verify electrical and formal integration. Connection, protection, island-mode operation, communications, location and required procedures.
Only then compare the economics of the variants. First, determine which variants can technically perform the task. The economic analysis should be based on current energy procurement and distribution conditions, the actual operating algorithm, losses, degradation, and investment and operating costs.

How to assess storage economics without confusing it with sizing?

The economic assessment should be carried out only for variants that can technically perform the required function. Otherwise, an attractive financial result may be obtained for a configuration that in practice will not shave the required peaks, maintain backup power or provide enough energy capacity to shift renewable energy surpluses.

The economic model should use the current energy and distribution settlement method, the actual prices or pricing formulas applicable to the business, CAPEX and OPEX, round-trip losses, auxiliary energy consumption, degradation and any future capacity augmentation. Revenue from arbitrage or flexibility services should be included only when the technical and market conditions for delivering them have been confirmed.

Audit of energy storage application and sizing – what should the outcome be?

A good audit should not end with the statement “we recommend a 500 kWh storage system”. The outcome should be a documented relationship between the site’s problem and the system parameters. Depending on the objective, the analysis may include:

a map of power peaks and their duration,
a profile of renewable energy surpluses and deficits,
a comparison of several power and energy-capacity variants,
an SoC simulation for a representative year,
assumptions for the charging and discharging algorithm,
the number of cycles, throughput and the effect of degradation on the ability to perform the function,
electrical integration and control requirements,
sensitivity analysis for changes in the production profile, renewable generation and price conditions,
a clear answer as to when storage makes sense and when another element of the energy system should be changed first.

In practice, storage should be analysed together with other investments. If a business plans to expand PV, electrify a process, add new chargers, change its connection capacity or carry out a major energy-system upgrade, the post-investment load profile may be more important than the historical profile.

The most common mistakes when sizing energy storage

sizing capacity as a percentage of annual or monthly consumption,
failing to distinguish power [kW] from energy [kWh],
analysing PV and consumption using annual totals instead of a common time series,
assuming that the storage system will always be available simultaneously for peak shaving, arbitrage and backup,
ignoring round-trip losses and auxiliary energy consumption,
calculating the benefit only for the first year without degradation,
assuming flexibility revenue without confirming market requirements,
assuming that every on-grid system automatically operates during a grid outage,
selecting technology before defining the function and control algorithm.

FAQ – energy storage for businesses

How should energy storage capacity be sized for a business?

There is no universal ratio between storage capacity and annual consumption. First define the function, required power and duration, and then simulate operation using time-series data for consumption, generation and prices. Capacity is the result of that analysis, not its starting point.

Are monthly energy bills enough to size an energy storage system?

Not for final sizing. Bills are useful for screening, but they do not show the magnitude and duration of peaks, the coincidence of PV generation with demand, or charging and discharging windows. Typical business analyses require data at least at 15-minute resolution, and faster phenomena may require higher-resolution data.

Does energy storage always increase photovoltaic self-consumption?

Only when there are actual PV surpluses that can be stored and used later. If the site consumes almost all PV generation as it occurs, the potential for additional self-consumption may be small.

Can one storage system perform peak shaving, arbitrage and backup at the same time?

Technically, often yes, but the functions compete for power and SoC. The EMS needs a hierarchy of priorities, and the analysis should check whether enough operating range remains for the other applications after energy has been reserved for backup.

Does a grid-connected storage system provide power during an outage?

This should not be assumed. Backup power requires the right inverter architecture, protection, automation and island-mode capability. The backup function should be defined at the design stage.

What data should be prepared for an energy storage sizing audit?

The most useful data set includes at least 12 months of 15-minute data from the point of supply, renewable generation profiles at the same resolution, bills and contracts, contracted and connection capacity, the electrical diagram, a list of critical loads and information on planned changes at the site.

Summary

An energy storage system should be sized for the task, not simply for the scale of a business’s energy consumption. Peak shaving, renewable energy self-consumption, power management, arbitrage, flexibility and backup require different power-to-energy-capacity ratios, different reserve levels and different EMS logic. Only time-series analysis shows which variant can actually perform the required function.

If a business is collecting offers, it is worth preparing a 15-minute profile and defining the storage system’s primary function before selecting equipment. Energy Trend analyses BESS applications and sizing for businesses using measurement data and the conditions of the specific site. Contact Energy Trend if you need an independent assessment of the application, power and energy-capacity variants, and the control and BESS integration requirements.