Module

What a module actually is

A module takes raw prismatic, cylindrical, or pouch cells — overwhelmingly lithium iron phosphate (LFP) in modern stationary storage, with NMC seen mainly in legacy or space-constrained designs — and combines them into one serviceable assembly.

Inside the housing sit the series/parallel cell group, welded busbars or a cell contacting system, voltage and temperature sense harnesses, and usually a module-level BMS slave board (often called a BMU) that reports every cell voltage and a set of temperatures up to the rack controller. Current utility-scale designs are built around large prismatic LFP cells of typically 280 to 314 Ah, each storing roughly 0.9 to 1.0 kWh.

Electrically, a module is described by its cell configuration in xPyS notation: parallel groups times cells in series. A 1P16S module of 3.2 V LFP cells is a nominal 51.2 V unit; 1P48S or 1P52S designs land around 150 to 170 V, and current 5 MWh-class containers usually run 1P104S — about 330 V per module. The vendor's split between module size and module count per rack balances handling weight, touch-safe voltage during installation and service, and the number of field connections — every extra connector is a resistance, a heat source, and a potential failure point over a 20-year life.

The module is the smallest field-replaceable unit in most utility-scale products. Technicians do not service individual cells; when a fault is isolated, the whole module is swapped. That is why the module boundary — its connectors, voltage class, weight, and diagnostics — is a real engineering interface, not just a packaging convenience. In many designs it is also an electrical protection boundary, with module- or rack-level fusing limiting the energy a fault can draw from the rest of the string.

Why it matters on a real project

The module defines maintainability, warranty, and safety granularity for the asset. Because it is the replaceable unit, its cost, lead time, and availability drive O&M planning and spares strategy over a 15-to-20-year project life. Module-level state-of-health data is what feeds capacity-retention warranties and augmentation decisions, so the module is where commercial performance guarantees are effectively measured and enforced. A rack string is series-limited by its weakest module, which is why one degraded or poorly balanced module can drag the usable energy of an entire rack.

Safety propagation is the other major consequence. Thermal runaway in stationary BESS is analyzed across the cell, module, unit, and installation levels, and the module enclosure is a key barrier slowing cell-to-cell propagation.

The UL 9540A test method drives a single cell into thermal runaway and characterizes fire behavior at progressively larger scales — cell, module, unit, and installation. Those results inform unit spacing, deflagration venting per NFPA 68, and gas detection or ventilation per NFPA 69 — all of which the system integrator must reconcile with the NFPA 855 installation requirements.

Module design also moves the project cost and reliability needle. A modern 20-foot container storing around 5 MWh holds on the order of fifty 1P104S modules, each with its own connectors, sense harness, and communications node — so module count multiplies both factory assembly labor and the population of things that can fail in the field. The hundred-plus module counts often quoted describe the earlier 1P48S–1P52S generation, at roughly 45 to 52 kWh apiece. The industry trend toward larger cells (280 Ah to 314 Ah and beyond) exists partly to cut the number of cells, welds, and modules per MWh.

Typical values and standards

Module voltages range from the tens of volts for 16S-class designs (about 51 V nominal) through roughly 150-170 V for 48S-52S designs, up to about 330 V for the 1P104S modules usual in current 5 MWh-class containers; energy per module runs from a few kWh to roughly 100-105 kWh for a 1P104S module of 314 Ah cells.

Current grid-scale modules commonly weigh about 300 to 700 kg, which is why modules are installed with lifts rather than by hand and why rack-loading ergonomics appear in installation manuals. Modules stack in series to form a rack string of typically 1,200-1,500 V DC to match the PCS input window, with 2,000 V architectures emerging. In a 2-to-4-hour system, modules cycle gently at roughly 0.25C to 0.5C.

On the standards side, keep the roles straight: UL 1973 covers the battery — cell, module and rack — as a stationary-application component, and IEC 62619 is the corresponding international safety standard for industrial lithium cells and batteries. UL 9540 certifies the complete energy storage system, while UL 9540A is the fire and thermal-runaway propagation test method whose data feeds NFPA 855 installation compliance.

UN 38.3 governs transport testing, and lithium modules are commonly shipped at reduced state of charge — about 30 percent or lower. Treat any single voltage, energy, or weight figure as vendor-specific; confirm against the actual datasheet rather than assuming a fixed value.

How it shows up in specs, studies and contracts

On a datasheet, expect the module's configuration in xPyS notation, nominal and operating voltage windows, nameplate energy, weight, and communication interface. Check the reference point of every number: module energy is a BOL DC figure at a stated temperature and C-rate, not deliverable AC energy at the point of interconnection, and nameplate never equals usable or contracted energy.

In supply and long-term service agreements, the module is the unit of warranty administration — capacity tables, spares stocking levels, replacement lead times, and swap labor terms are all written per module, so ambiguity here becomes a commercial dispute later.

Questions a working engineer should ask: is there a module-level UL 9540A test report, and does it match the exact cell and module revision being supplied? What is the documented module swap procedure and time, and can a replacement module built years later remain mechanically, electrically, and firmware-compatible with the original rack?

How must a fresh replacement module be matched — by state of charge and capacity binning — before it joins an aged string? What are the transport (UN 38.3) and shipping-SOC constraints, and who owns module-level data access for warranty claims and state-of-health verification?

Common pitfalls

The word does not mean the same thing across products. Cell-to-pack and so-called module-free architectures integrate cells directly into the rack or pack structure, so one vendor's "module" may be a 14 kWh serviceable brick while another's is a 50 kWh structural sub-pack that is not field-serviceable at all. Terminology imported from the EV world — where pack, module, and brick have their own meanings — adds to the confusion; in stationary BESS, always resolve the term against the vendor's own hierarchy drawing before comparing bids.

Do not sum module nameplates and call it system energy. Series strings are limited by the weakest module, the BMS enforces a usable depth-of-discharge window, and auxiliary and conversion losses sit between DC nameplate and AC energy at the meter — so the delivered figure is always below the arithmetic total, and it degrades from BOL to EOL. Likewise, modules are not interchangeable spare parts across product lines: mixing revisions or vendors within a rack voids the fire-test basis and the warranty, which is exactly why replacement modules are matched, binned, and documented.