Every week I meet smart engineers who can explain a gas turbine in their sleep but wave vaguely at a battery site and say “the batteries store the energy.” That’s true the way “the plane flies” is true. If you’re going to design, finance, or operate one of these plants, you need to know what’s actually inside the fence.
This is the walkthrough I wish someone had given me: the whole chain from a single cell to the point of interconnection, and the specs that actually matter at each stage.
Before we start, a map helps. This interactive BESS site map lets you click through every component from the cell to the point of interconnection — a useful companion to the walkthrough below.
The building blocks: cell, module, rack
Everything starts with the cell — today, overwhelmingly a lithium iron phosphate (LFP) prismatic cell. The industry spent the last few years standardized around the 314 Ah format, and manufacturers are now shipping much larger cells: CATL’s current storage platform is built on a 587 Ah cell, and 600+ Ah designs have been announced across the supplier base. Bigger cells mean fewer parts, fewer welds, and fewer failure points per MWh — and the economics follow. BNEF’s 2025 Energy Storage System Cost Survey found DC-side systems built on 300 Ah-or-larger cells were about 50% cheaper than systems using smaller cells.
Cells are grouped into modules with voltage and temperature sensing, and modules are stacked into racks. Racks live inside a containerized DC block — and here the scaling story repeats: a modern 20-foot enclosure carries 5–6+ MWh (CATL’s TENER at 6.25 MWh set the benchmark, and competitors have matched or exceeded it). BNEF found DC blocks of 4 MWh and above were about 39% cheaper per kWh than 2–4 MWh designs.
The BMS: the nervous system
The battery management system (BMS) is a hierarchy, not a single box. Cell-level sensing feeds module controllers, which feed rack controllers, which feed a system-level BMS. Its jobs:
- Measure voltage, current, and temperature everywhere
- Estimate state of charge (SoC) and state of health (SoH)
- Balance cells so the weakest cell doesn’t limit the whole string
- Trip contactors and coordinate with rack fusing and DC switchgear to isolate faults before they escalate toward thermal runaway
The BMS is also your evidence locker for warranty compliance — if its logs say you operated outside the allowed envelope, you’ll hear about it from the OEM. Treat BMS data quality as a commercial issue, not just a technical one.
PCS: where DC becomes AC
The power conversion system (PCS) is a bidirectional inverter. It converts the DC battery bus — 1,500 V class is today’s standard, with 2,000 V class designs now emerging to cut balance-of-plant cost — into AC, and back again when charging. The PCS largely defines the power rating of the plant, while the batteries define the energy. That distinction — power vs energy — matters more than any other in this industry, and it drives how you size a system.
Modern PCS units also deliver grid services: frequency response, voltage and reactive power support, and increasingly grid-forming capability, where the inverter creates its own stable voltage waveform instead of following the grid’s. Australia has led grid-forming battery deployments, and specifications elsewhere are catching up.
Transformers and the collection system
The PCS outputs low-voltage AC (typically in the 690 V class), which a medium-voltage transformer steps up to the collection level — commonly 34.5 kV in North America. MV switchgear and collection feeders run to the plant substation, where a main power transformer steps up to transmission voltage at the point of interconnection (POI). A separate station-service supply feeds the plant’s own auxiliary loads. Nothing here is exotic — but main power transformer lead times have been one of the biggest schedule risks in the industry for several years running.
EMS, PPC, SCADA: the brains
Three control layers get confused constantly:
- EMS (energy management system): decides what the plant should do — charge now, discharge into the evening peak, hold headroom for a frequency product. This is where market optimization lives.
- PPC (power plant controller): ensures the plant delivers exactly what was commanded at the POI — coordinating active power, reactive power, voltage, and frequency response across all inverters.
- SCADA: the monitoring, alarming, and data-historian layer underneath everything.
In my experience, more commercial disputes trace back to EMS/PPC integration issues than to the batteries themselves.
Thermal management and auxiliaries
Liquid cooling has become the default for new utility-scale systems, displacing air-cooled HVAC designs. Tighter cell temperature control means slower degradation and higher energy density per container. The price is auxiliary load: pumps, chillers, and controls consume energy around the clock, and that consumption comes straight out of your round-trip efficiency. A well-designed plant delivers roughly 85–92% AC-to-AC round-trip efficiency — but always ask at what boundary, climate, and duty cycle that number was measured. Auxiliary design is also a safety system: cooling, gas detection, and ventilation interact directly with the plant’s fire-safety case.
AC coupling vs DC coupling
For hybrid solar-plus-storage plants there are two architectures. AC-coupled systems give the battery its own PCS and connect at the AC bus — simple, flexible, independently operable, and the default for retrofits. DC-coupled systems share the solar inverter and connect the battery on the DC side — this can capture solar energy that would otherwise be clipped and saves conversion equipment, at the cost of design complexity and reduced operational flexibility.
The specs that matter
When you read a datasheet, these are the numbers to interrogate:
| Spec | What it tells you — and what to ask |
|---|---|
| Power (MW) | How fast you can charge/discharge. At the POI or at the PCS terminals? |
| Energy (MWh) | How much you can store. At beginning of life or end of life? Usable or nameplate? |
| Duration (h) | Energy ÷ power. |
| Round-trip efficiency | AC-AC including auxiliaries? At what temperature and duty cycle? |
| Cycle life / throughput | The warranty envelope you must operate inside. |
| Response time | Milliseconds to seconds — decisive for frequency products. |
| Auxiliary consumption | Continuous parasitic load; climate-dependent. |
FAQ
How long does a grid battery last? Typical design life is 15–20 years, with capacity degrading over time. Most projects plan augmentation — adding capacity mid-life — to maintain contracted energy against degradation.
How big is a typical grid-scale battery project? Utility-scale projects today range from tens of MW to multi-GWh campuses, with the economics of mature markets increasingly favoring 4-hour durations.
What battery chemistry do grid-scale plants use? Overwhelmingly lithium iron phosphate (LFP) today, with sodium-ion entering commercial deployment in 2026.
If you want the full picture — PCS topologies, protection coordination, controls integration, and how all of this shows up in contracts — that’s exactly what my Grid-Scale BESS: Complete Guide covers, lesson by lesson.