What It Actually Takes to Buy and Connect a Battery: Inside a 7 MW, 28 MWh BESS Project

Prismecs delivered owner's engineering on a 7 MW battery energy storage system with 28 MWh of capacity in North America, built to capture excess energy from photovoltaic solar installations.

The scope was the owner's side of the project, not the supply side. Prismecs supervised the creation of the technical, performance and safety guidelines, formulated the RFI and RFP documents, guided the assessment of designs from competing BESS manufacturers, built a PV model to quantify the excess energy being lost, sized the battery against that loss, and drafted the utility interconnection application on the client's behalf.

That sequence is what this article is about, because almost nothing published about battery storage describes it. The supplier pages explain what a battery does. The standards bodies publish the codes. Nobody writes down what the owner has to do.

One correction before the detail. Storage does not solve intermittency. Most grid-scale lithium-ion systems are built for 2 to 4 hours of duration. What storage does is capture energy you are currently throwing away and move it to when you need it, which is a smaller claim and a real one.

What Storage Actually Solves

Battery storage shifts energy in time and stabilises output. It does not add energy, and it does not replace dispatchable capacity.

Four things it does well:

Captures clipped and curtailed energy. On a solar plant whose inverters cannot pass peak DC output, or whose grid connection is constrained, storage captures production that would otherwise be lost entirely.

Firms variable output. Firming means converting a variable production profile into a dependable one the system can rely on. Solar and wind produce when the resource allows; most loads do not follow that shape.

Shaves peaks and manages demand charges. Discharging during the facility's highest demand interval reduces the demand charge, which on many industrial tariffs is 30 to 50 percent of the bill.

Responds in milliseconds. Frequency response, voltage support and ride-through are functions no thermal plant can match on timescale.

What it does not do

It does not produce energy. It does not cover a multi-day outage. A battery is a container, and the question is always how big and how long.

The Duration Limit Nobody States Plainly

Most grid-scale lithium-ion battery systems are built for 2 to 4 hours of duration, and that single constraint explains why storage firms rather than replaces.

Duration is how many hours a system can discharge at its rated power. A 7 MW battery with 28 MWh of capacity has a nominal duration of four hours, because 28 divided by 7 equals 4.

Industry reference designs reflect this. A commonly used benchmark configuration is 100 MW / 400 MWh: lithium iron phosphate chemistry, four-hour duration, around 86 percent round-trip efficiency and roughly 2 percent auxiliary load.

Round-trip efficiency is the proportion of energy put in that comes back out. At 86 percent, roughly one unit in seven is lost to conversion and thermal management.

Why four hours is the practical ceiling

Beyond four hours, the cost of additional energy capacity rises faster than the value of the hours it serves, because the marginal hour is used less often. Longer-duration chemistries exist and are not yet cost-competitive at scale for most applications.

The consequence is that a four-hour battery cannot cover a still, overcast week. For that, you need generation. Storage and thermal capacity are complements, not alternatives, and a procurement decision framed as one or the other is framed wrongly.

So is the transition solved

No, and claiming otherwise damages credibility with the people who have to operate the result. Storage closes a specific gap very well and leaves others open. Naming which is the first thing an owner's engineer does.

For chemistry comparison, cost per kWh and the wider storage picture, see our guide to battery energy storage systems.

DC Nameplate Is Not AC Usable Energy

A nominal DC container does not deliver its nameplate energy at the AC point of interconnection, and the gap between the two is where most BESS procurement disputes start.

A nominal 5 MWh DC container does not automatically provide 5 MWh at the AC point of interconnection, because state-of-charge limits, conversion losses, auxiliary demand, reserve capacity and degradation all reduce usable output.

Five deductions, in order:

State-of-charge limits. Operating between, say, 10 and 90 percent rather than 0 to 100 preserves cycle life and removes capacity from the usable total.

Conversion losses. DC to AC through the power conversion system, plus transformer losses to the point of interconnection.

Auxiliary demand. Thermal management, controls and fire detection consume power continuously, typically around 2 percent.

Reserve capacity. Energy held back for ancillary services, black start or control stability, depending on duty.

Degradation. Capacity falls over the life of the asset, which is why augmentation, adding battery capacity at defined intervals, is planned from the start. Reference designs commonly schedule augmentation at year 5 and year 10.

What to specify instead

State whether the energy figure is nameplate DC or usable AC, and at which measurement boundary. Then state the duration at a named power level, because duration without a power reference is not a specification.

If degradation outpaces the model

That is what the augmentation plan and the capacity warranty exist for. Require a capacity test at commissioning and at defined intervals, measured at the same boundary the guarantee is written against, or the warranty is unenforceable.

Sizing a Battery Against PV Clipping

The commercial reason to install storage on a solar plant is usually to capture energy the plant is already losing, which means the sizing question is answered by modelling the loss first.

Clipping occurs when a PV array's DC output exceeds what the inverters can convert, so the surplus is simply not produced. It is designed in deliberately through the inverter loading ratio, the ratio of DC array capacity to AC inverter capacity, because oversizing the array improves output in weak light at the cost of losing some peak.

Curtailment is different: output reduced on instruction from the grid operator, or because prices have gone negative.

The method used on this project

Prismecs created a photovoltaic model to measure the magnitude of excess energy loss, then used that model to determine the appropriate battery capacity to mitigate the surplus loss from the PV system.

That is the correct sequence and it is frequently reversed. A battery sized before the loss is quantified is sized to a budget or a supplier's standard container, not to the asset. Too small and you leave energy on the table; too large and you have paid for capacity that never fills.

DC-coupled or AC-coupled

DC-coupled connects the battery on the DC side of the inverter, which allows it to capture clipped energy that never reaches the AC side at all. AC-coupled connects after the inverter, which is simpler to retrofit and cannot recover clipped energy.

If the objective is capturing clipping, DC coupling is the architecture that does it. If the objective is grid services or demand management on an existing plant, AC coupling is usually simpler.

For PV system fundamentals, see our guide to solar power plants and how solar panels work.

Writing the Requirements: What the RFP Must Contain

Missing inputs force suppliers to price assumptions, and assumptions change power conversion sizing, cooling, civil works, delivery scope, compliance evidence and lifecycle guarantees.

On this project, Prismecs supervised the creation of the technical, performance and safety guidelines, and then formulated the RFI and RFP documents. That order matters: the requirements exist before the request, or the request is a wish list.

The minimum data set

What a full scope book covers

Utility model scope books for BESS procurement run to dozens of pages and cover, beyond the electrical scope: project controls, cybersecurity, environmental requirements, site fire protection, site security, temporary installation and laydown areas, tools, spare parts and consumables, project utilities and redundancy, and control system and communication requirements.

They also require that calculations be signed by a professional engineer of record licensed in the state or jurisdiction, which is a procurement requirement with a staffing consequence.

If the duty cycle changes later

Degradation, warranty and augmentation are all priced against the stated duty cycle. Changing it after award is a variation, and on a battery it is one that affects the capacity guarantee rather than just the price.

Evaluating Designs That Are Not Comparable

BESS bids arrive in different units, against different boundaries, with different duty assumptions, which means the evaluation work starts with making them comparable.

On this project, Prismecs guided the assessment of designs from different BESS manufacturers. That is a distinct activity from running a tender, and it is where the technical risk is decided.

The four normalisations

Energy basis. Nameplate DC, installed DC, or usable AC. Convert all bids to the same one before comparing anything.

Measurement boundary. At the battery terminals, at the PCS output, or at the point of interconnection. Each strips out more losses than the last.

Duty cycle assumed. A system warranted for one cycle per day and one warranted for two have different economics and different augmentation schedules.

Degradation and augmentation. Whether augmentation is included, excluded, or offered as an option, and whose cost it is at year 5 and year 10.

What guarantees to require

Capacity retention against a stated curve, measured by a defined test at a defined boundary. Round-trip efficiency at a stated state of charge and temperature. Availability, with the exclusions written out. And the remedy if any is missed.

A guarantee without a test method and a boundary is not enforceable, which is the same lesson that applies to gas turbine heat rate guarantees and solar performance ratio thresholds.

Financing

Lenders and tax equity assess capacity warranties, supplier creditworthiness and compliance evidence during diligence. A bid that is cheapest and least bankable costs more, because it reappears as a financing condition.

Safety Evidence: What the AHJ Actually Asks For

Battery storage permitting turns on fire safety evidence, and the gap between the configuration that was tested and the configuration being offered is where projects fail.

On this project, Prismecs supervised the creation of the safety guidelines alongside the technical and performance ones. That work is governed by a specific and well-defined set of standards.

The standards that matter

UL 9540A is the only consensus standard explicitly cited in NFPA 855 for large-scale fire testing, and the only national standard in the US and Canada for BESS fire safety testing methods. It provides the method for evaluating system behaviour when design or installation conditions exceed the limits set by NFPA 855, NFPA 1 or the International Fire Code.

The relationship is straightforward: UL 9540 is the listing, UL 9540A is the test, and NFPA 855 is the installation standard that uses the test results to set setbacks.

UL 9540 is technically a voluntary standard. It is nonetheless required by most state building codes, by utility interconnection agreements and by insurance policies, which makes it mandatory in practice.

How to verify a certificate

This is the detail that separates a real evaluation from a box-tick. For every safety document, ask for the certificate issuer, the certificate number, the exact model, the report revision, the test laboratory, the configuration covered, the installation limitations, and any difference between the tested system and the offered system.

That last item is the one that matters. A UL 9540A report covering a different module arrangement, a different enclosure or a different spacing does not support the design you are buying.

Timelines

Certification is a schedule item, not a formality. UL 9540 listing typically takes 6 to 12 months and UL 9540A testing 3 to 6 months. A product without current listings is a programme risk regardless of its price.

Insurance and permitting

Insurers require the listings and increasingly require the UL 9540A report to assess siting. The AHJ sets setback distances from it. A design that cannot evidence its tested configuration will not be permitted and will not be insured, which makes safety evidence a gating item rather than a documentation task.

For the broader compliance picture including chemistry and cost, see our guide to battery energy storage systems.

The Interconnection Application

Connecting a battery to the grid requires a utility interconnection application, and on this project Prismecs drafted it on the client's behalf.

That is the longest pole on most storage projects and the one owners are least equipped for. The application defines the point of interconnection, the maximum import and export, the protection and control scheme, and the operating modes the utility will permit.

IEEE Std 1547-2018, Standard for Interconnection and Interoperability of Distributed Energy Resources with Associated Electric Power Systems Interfaces, governs the technical requirements. IEC 62116 covers anti-islanding test procedures, the verification that a resource will not continue energising a de-energised network.

Why it takes so long

Interconnection queues have become the binding constraint across most US markets. PJM waits approach seven years for large projects, and ERCOT has been tracking roughly 474 GW of large-load interconnection requests. For the full picture, see our analysis of power generation trends and the interconnection bottleneck.

What determines whether it succeeds

Completeness and consistency. An application whose single-line diagram, protection settings and operating modes do not match each other goes back for revision, and each revision costs queue time.

Realistic operating modes. Requesting export capability you do not need adds study scope. Requesting less than you will need means reapplying.

Early engagement. Pre-application meetings with the utility are available in most jurisdictions and are consistently underused.

If interconnection is refused or constrained

A constrained interconnection is common and not fatal. Export limits, charging restrictions and curfews can all be designed around, provided they are known before the system is sized. Discovering them after procurement means a battery sized for a duty it is not permitted to perform.

Who Does This Work

The scope described above is owner's engineering, and the question for most owners is whether they have it or need to buy it.

An EPC turnkey contract transfers design and delivery to a contractor, which is efficient and leaves the owner without an independent check on specification, safety evidence or bid comparability.

Owner-procured with independent engineering support keeps the owner in control of the requirements and the evaluation, with a specialist doing the work.

The distinction that matters: the party writing the requirements should not be the party bidding against them. On this project Prismecs wrote the guidelines, formulated the RFI and RFP, evaluated the manufacturers' designs and drafted the interconnection application. None of those is a supply activity.

Staffing

The skills are specific: PV and storage modelling, electrical design review, codes and standards literacy, and interconnection process experience. Most industrial owners need them for one project every few years, which is the classic case for buying the capability rather than building it.

For the operating phase that follows commissioning, including performance verification and warranty claim recovery, see our guide to renewable energy asset management.

Microgrids for Industrial Sites

A microgrid is a localised energy system that generates, distributes and stores electricity and can operate either connected to the main grid or independently of it.

For industrial operators the case is resilience rather than development. A microgrid boosts energy resilience by disconnecting from the primary grid during disasters or failures and sustaining vital facilities, which is why hospitals, communication centres, data centres and process plants are the typical hosts.

Islanding is the capability to separate from the grid and continue supplying local load. It is not automatic: most grid-parallel installations cannot island without the transfer equipment, protection and control scheme to do it safely.

Black start is the ability to energise from a completely de-energised state without grid supply. Where a site cannot tolerate an extended outage, it is a specified requirement rather than an assumption.

When a microgrid is justified

Where the cost of an outage exceeds the cost of resilience, where grid supply is unreliable or absent, or where a site needs to continue operating through a grid event. Where the grid is reliable and outages are inexpensive, a microgrid is an expensive answer to a cheap problem.

For microgrid architecture in detail, see our guide to decentralisation and the rise of microgrids. For grid-side integration standards and demand response, see our guide to smart grid and distributed energy resources.

Standards Summary

Nine standards govern a grid-connected battery storage project in North America, and the specification should name them.

UL 9540 for the system listing. UL 9540A for thermal runaway fire propagation testing. UL 1973 for battery packs in stationary applications. UL 1741 for the power conversion system. NFPA 855 for installation of stationary energy storage systems. NFPA 70, the National Electrical Code, with Article 706 for energy storage and Article 690 for solar PV. IEEE Std 1547-2018 for distributed resource interconnection. IEC 62933 for system-level electrochemical storage safety and testing. IEC 62116 for anti-islanding test procedures.

Outside North America

UL and NFPA requirements are North American. IEC 62933 and IEC 62619 apply internationally, European installations work to EN standards and CE marking, and India applies BIS IS 17855. Confirm the destination market before specifying, because a product listed for one market may not be accepted in another.

How Requirements Vary by Application

The engineering is constant. What changes is what the battery is for, and therefore how it is sized.

Solar-plus-storage. Sized against modelled clipping and curtailment. DC coupling where recovering clipped energy is the objective.

Industrial and commercial sites. Sized against the demand charge profile and the peak interval. Demand charges on many industrial tariffs run to 30 to 50 percent of the bill, which is the economic driver.

Utility-scale and merchant. Sized against market products: energy arbitrage, capacity, frequency response. Four-hour duration dominates because capacity market rules commonly require it.

Oil and gas sites. Frequently remote with no grid connection, where storage firms generation rather than supplementing a utility supply, and hazardous area classification constrains equipment selection.

Data centres. Sized for ride-through and transition rather than for energy, with response time and reliability outranking capacity. For redundancy architecture, see our guide to data centre power redundancy.

Remote and island systems. Delivered fuel cost dominates, which raises the value of every captured kilowatt-hour and makes hybrid configurations economic at scales that would not work on-grid.

For technology comparison across capacity, efficiency, start time and cost, see our guide to power generation equipment compared.

What Prismecs Does

Prismecs provides owner-side engineering, procurement and delivery for energy storage and power infrastructure, including the 7 MW, 28 MWh BESS project described in this article.

The delivered scope on that project: supervision of the technical, performance and safety guidelines; formulation of the RFI and RFP documents; assessment of competing BESS manufacturers' designs; a photovoltaic model built to quantify excess energy loss; battery capacity sizing against that modelled loss; vendor evaluation and selection support; and drafting the utility interconnection application on the client's behalf.

Other delivered project scope includes DC-coupled battery energy storage for solar and hybrid applications; eight TM2500 dual-fuel units totalling 260 MW at Birr, Switzerland, online in six months with a new 220 kV interconnection; four TM2500 units totalling 110 MW at Duqm, Oman, kept grid-ready with resident O&M crews, CMMS and parts support; and an LM2500XPRESS plant at Miaoli, Taiwan delivered in ten months.

Capability spans owner's engineering for independent technical representation, technology and consulting for modelling, sizing and options analysis, distributed energy solutions for microgrids and storage integration, renewable energy solutions, EPCM services for delivery, I&C services for controls and commissioning, and O&M services for the operating phase.

Prismecs is OEM-agnostic, which on a competitive design evaluation is the whole point: the party assessing the bids is not bidding.

Apply this article's criteria to any adviser, including us. Ask whether they will model the loss before sizing the battery. Ask how they will normalise bids onto a common energy basis. Ask what they will do with the UL 9540A report. Ask who drafts the interconnection application.

To discuss storage feasibility, sizing, procurement support or interconnection, send your site location, PV or load profile, grid connection details and target in-service date to sales@prismecs.com or call +1 (888) 774-7632.

Frequently Asked Questions

Does battery storage solve renewable intermittency?

No. Most grid-scale lithium-ion systems are built for 2 to 4 hours of duration, so a battery cannot cover a still, overcast multi-day period. What storage does is capture energy that would otherwise be clipped or curtailed, shift it in time, firm variable output and respond in milliseconds. Storage and dispatchable generation are complements rather than alternatives, and procurement framed as one or the other is framed wrongly.

What is duration and how is it calculated?

Duration is how many hours a system can discharge at its rated power, calculated as energy capacity divided by power rating. A 7 MW battery with 28 MWh of capacity has a nominal duration of four hours. Duration quoted without a stated power level and measurement boundary is not a specification, because the same container delivers different durations depending on where the energy is measured.

Why does a DC container not deliver its nameplate energy?

Because five deductions sit between the cells and the point of interconnection. State-of-charge limits preserve cycle life and remove usable capacity. Conversion losses occur through the power conversion system and transformer. Auxiliary demand for thermal management and controls runs around 2 percent continuously. Reserve capacity is held back for control stability. And degradation reduces capacity over the asset's life.

What round-trip efficiency should I expect?

Reference designs for lithium iron phosphate systems commonly assume around 86 percent round-trip efficiency with roughly 2 percent auxiliary load. At 86 percent, approximately one unit of energy in seven is lost to conversion and thermal management. Specify efficiency at a stated state of charge and temperature, because it varies with both, and require it to be demonstrated by a defined test at a defined boundary.

What is augmentation and when does it happen?

Augmentation is the planned addition of battery capacity to offset degradation over the asset's life. Reference designs commonly schedule it at year 5 and year 10. It should be established at procurement: whether it is included in the bid, excluded, or offered as an option, and whose cost it is at each interval. An augmentation plan agreed after award is a variation to a capacity guarantee.

How do I size a battery for a solar plant?

Model the loss first. Build a photovoltaic model to quantify the magnitude of excess energy being clipped or curtailed, then size the battery to capture that surplus. Sizing before the loss is quantified produces a battery matched to a budget or a supplier's standard container rather than to the asset. Too small leaves energy unrecovered; too large pays for capacity that never fills.

What is PV clipping and how does it differ from curtailment?

Clipping occurs when a PV array's DC output exceeds what the inverters can convert, so the surplus is never produced. It is designed in deliberately through the inverter loading ratio, because oversizing the array improves weak-light output at the cost of peak. Curtailment is different: output reduced on instruction from the grid operator, or because wholesale prices have gone negative.

Should the battery be DC-coupled or AC-coupled?

DC-coupled connects on the DC side of the inverter and can capture clipped energy that never reaches the AC side. AC-coupled connects after the inverter, is simpler to retrofit, and cannot recover clipped energy. If the objective is capturing clipping, DC coupling is the architecture that does it. If the objective is grid services or demand charge management on an existing plant, AC coupling is usually simpler.

What data must a BESS RFP contain?

Required MW output, usable MWh, required duration with power level and measurement boundary, duty cycle, point-of-interconnection voltage, maximum import and export, reactive power requirement, site environmental conditions, controls and communications interface, safety evidence required, and guarantee terms. Missing inputs force suppliers to price assumptions, which change power conversion sizing, cooling, civil works, delivery scope, compliance evidence and lifecycle guarantees.

Why are BESS bids hard to compare?

Because suppliers quote against different bases. Energy may be nameplate DC, installed DC or usable AC. The measurement boundary may be the battery terminals, the PCS output or the point of interconnection. Duty cycle assumptions differ, and augmentation may be included, excluded or optional. Normalise all four before comparing price, or you are comparing four different products.

What is the difference between UL 9540 and UL 9540A?

UL 9540 is the system-level listing for energy storage systems and equipment, required for authority having jurisdiction approval. UL 9540A is the test method for evaluating thermal runaway fire propagation, and it is the only consensus standard explicitly cited in NFPA 855 for large-scale fire testing. NFPA 855 then uses the test results to set installation setback distances.

How do I verify a BESS safety certificate?

Ask for the certificate issuer, the certificate number, the exact model, the report revision, the test laboratory, the configuration covered, the installation limitations, and any difference between the tested system and the offered system. That last item is the one that matters: a UL 9540A report covering a different module arrangement, enclosure or spacing does not support the design being purchased.

How long does BESS certification take?

UL 9540 listing typically takes 6 to 12 months and UL 9540A testing 3 to 6 months. That makes certification a schedule item rather than a formality, and a product without current listings is a programme risk regardless of its price. UL 9540 is technically voluntary but is required in practice by most state building codes, utility interconnection agreements and insurance policies.

What governs grid interconnection for storage?

IEEE Std 1547-2018 sets the technical requirements for interconnection and interoperability of distributed energy resources, and IEC 62116 covers anti-islanding test procedures. The application itself defines the point of interconnection, maximum import and export, the protection and control scheme, and permitted operating modes. Interconnection queues are now the binding constraint in many markets, with PJM waits approaching seven years for large projects.

Tags: Battery Energy Storage System (BESS) BESS Procurement NFPA 855 Grid Interconnection Owner's Engineering