TL;DR #
In charge/discharge testing at 35 °C ambient, a 3.35 MWh liquid-cooled battery enclosure held cell-to-cell temperature spread to under 5 K at end of discharge — comfortably inside the 6 K maximum mandated by GB/T 26276—2023. For buyers specifying containerized BESS above 2 MWh, this confirms that a correctly sized non-contact liquid cooling system is the minimum viable thermal architecture — wind cooling alone cannot reliably meet that 6 K delta at rated load. Before issuing any RFQ, audit the chiller cooling capacity calculation: a 1.1 safety factor applied to net heat load (gross cell heat minus stored temperature-rise energy) is the standard engineering basis.
Overview #
Most procurement teams evaluate liquid cooling by asking the wrong question — “does the system have liquid cooling?” — rather than “is the chiller sized to the actual net heat load, and has the flow balance been physically validated?” The distinction matters because undersized chillers and unbalanced manifolds are the two failure modes that dominate field returns in high-density BESS enclosures.
The analysis presented here draws on engineering validation work conducted at a commercial energy storage equipment manufacturer — a team responsible for integrating and type-testing 3.35 MWh containerized battery enclosures. Testing was performed at 0.5× rated power (0.5 Pe), with full thermocouple arrays on individual cells across a nine-cluster battery stack. The dataset is unusually granular for published industrial work: cell-level temperature readings at discharge start and discharge end, compared against calculated heat load and chiller selection output.
The subject enclosure uses prismatic LFP cells (3.2 V / 280 Ah), organized in a 1P52S module → 1P416S cluster → 9P416S stack hierarchy. Understanding the thermal architecture at this scale is directly relevant for any buyer specifying a 20-foot containerized BESS, since the energy density of that form factor is now migrating toward 6–7 MWh — making the thermal engineering proportionally more critical with each generation.
Liquid-Cooled BESS Thermal Architecture: System Design and Component Selection #
The thermal architecture used in this enclosure is non-contact liquid cooling — cells are bonded to liquid-cooled baseplates via thermally conductive structural adhesive, and the chiller circulates coolant (ethylene glycol/water mixture) through those baseplates. Heat moves from cell surface → adhesive → baseplate → coolant. The chiller then extracts heat from the coolant via a plate heat exchanger on the refrigerant side.
This is worth distinguishing from immersion cooling, which buyers sometimes ask about as a perceived upgrade. Non-contact liquid cooling is lower risk for several reasons: standard coolant (no dielectric fluid compatibility concerns), easier IP sealing, higher energy density per enclosure, and a simpler flow control architecture. For 280 Ah prismatic LFP cells in a utility-scale enclosure, non-contact is the right call — immersion cooling introduces more variables than it solves at this cell chemistry and form factor.
Chiller architecture — key components:
The air-cooled chiller unit contains a variable-frequency compressor (inverter-driven, output modulates to match heat load), an aluminum-fin copper-tube condenser with EC axial fans for stepless speed control, a plate heat exchanger at the refrigerant/coolant interface, a PTC electric heater for cold-climate preheating, and a circulation pump. Control logic runs three states: cooling (compressor on, VFD-modulated), heating (PTC heater on), and standby (compressor off, pump speed reduced).
Chiller sizing — the calculation you must verify:
The cooling capacity calculation has three steps, and buyers should ask suppliers to show all three, not just the final kW number.
Step 1 — Total cell heat generation at 0.5 Pe:
With 3,744 cells each generating 13 W at 0.5 Pe, gross heat load P₁ = 3,744 × 13 W = 48.7 kW.
Step 2 — Subtract heat absorbed by cell thermal mass:
Cell specific heat capacity C = 965 J/(kg·K), cell mass m = 5.4 kg, allowable temperature rise ΔT = 10 K, n = 3,744 cells.
Stored heat Q = 3,744 × 965 × 5.4 × 10 = 195.7 MJ. Over one full 0.5 Pe cycle (4 hours = 14,400 s), Q/t = 13.6 kW.
Step 3 — Apply safety factor:
Net cooling load P₃ = 1.1 × (48.7 − 13.6) = 38.6 kW.
The selected chiller is a 40 kW unit. That 1.4 kW margin above the calculated 38.6 kW is intentional but thin — which is why the 1.1 safety factor (k) cannot be reduced without revalidation. Honestly, some procurement teams get quoted a 35 kW chiller and accept it because the headline system cost is lower. That is a mistake that shows up as chronic overtemperature within the first summer operating season.
Heating load for cold-climate operation:
At −30 °C ambient, cells must be preheated to the minimum charge-permitted temperature before any charging begins. With ΔT₁ = 50 K (from −30 °C to +20 °C target), over a 20-hour standby window (72,000 s), and the same 1.1 safety factor:
P₄ = 1.1 × (3,744 × 965 × 5.4 × 50) / 72,000 = 14.9 kW.
The 40 kW chiller is configured with a 16 kW PTC heater — adequate for this preheating load with margin.
| Parameter | Calculated Value | Design Margin |
|---|---|---|
| Gross cell heat at 0.5 Pe | 48.7 kW | Baseline |
| Heat stored in cell thermal mass | 13.6 kW (subtracted) | Reduces net load |
| Net cooling load (with k=1.1) | 38.6 kW | — |
| Selected chiller cooling capacity | 40.0 kW | +3.6% above calculated |
| Cold-climate heating load | 14.9 kW | — |
| Installed PTC heater capacity | 16.0 kW | +7.4% above calculated |
Can’t find a supplier meeting these specs? Submit your requirements and we’ll match you within 48 hours.
Flow Balance and Manifold Design: Where Thermal Consistency Is Won or Lost #
The chiller output is necessary but not sufficient. A perfectly sized chiller connected to a poorly balanced manifold still delivers uneven cooling — cells in high-flow modules run cool while cells in low-flow modules overheat. This is the failure mode that most RFQ documents don’t even ask about.
The piping architecture in this enclosure uses a four-tier parallel design:
- Tier 1 & 2 (main runs): Stainless steel pipe with clamp fittings — rigid, corrosion-resistant, pressure-rated.
- Tier 3 & 4 (cluster-level branches): Nylon tubing with automotive-grade push-fit connectors — flexible for installation, rated for the operating pressure range.
- Insulation: All external pipe surfaces are wrapped with thermal insulation to prevent condensation, which is critical when coolant temperature runs below ambient dew point.
Simulation of the four-tier manifold shows that inter-module flow variance is kept within 10% — which is the design threshold for acceptable thermal uniformity across the battery stack. That 10% figure matters: if a supplier cannot tell you their simulated or measured flow variance, they haven’t done the analysis.
In supplier qualification, evaluation of enclosure designs from multiple manufacturers has revealed that three out of six samples tested failed to demonstrate flow balance data at all — they could show chiller capacity but had no manifold simulation or measured flow data. Those suppliers were disqualified from the shortlist immediately. This is a more common gap than buyers expect, especially from manufacturers who assemble enclosures using outsourced liquid cooling subsystems.
The IEC 62619 safety standard for stationary battery systems addresses thermal management requirements at the system level — but flow balance uniformity is an engineering design criterion that goes beyond what any single standard mandates. Buyers need to specify it explicitly in their technical requirements document.
Charge/Discharge Test Validation Results #
The thermal management scheme was validated under 0.5 Pe charge/discharge cycling at 35 °C ambient temperature — a representative worst-case for outdoor BESS installations in moderate climates.
Key measured results:
- Discharge start — minimum cell temperature: 24.9 °C
- Discharge end — maximum cell temperature: 33.7 °C
- Temperature rise across discharge: < 10 K (within the 10 K design allowance)
- Discharge end — minimum cell temperature: 28.9 °C
- Cell-to-cell temperature spread at end of discharge: < 5 K
That <5 K end-of-discharge spread is the number that matters for GB/T 26276—2023 compliance, which sets a 6 K maximum for both charge and discharge end conditions. The 1 K margin may look comfortable on paper, but at the next generation of enclosure density (6–7 MWh in the same 20-foot footprint), that margin compresses. Buyers specifying next-generation systems should require the supplier to rerun the thermal validation at the higher energy density, not just accept prior test data.
The charge-temperature operating window for these 280 Ah LFP cells is 0 °C to 60 °C; discharge is permitted from −20 °C to 60 °C. The heating system is specifically designed to bridge the gap in cold climates where ambient temperatures can reach −30 °C — a condition where charging would be prohibited without active preheating.
Most procurement teams don’t realize that GB/T 26276 was updated and now carries an explicit 6 K end-of-charge and end-of-discharge temperature uniformity requirement — the older version of this standard did not include this quantitative threshold. Suppliers quoting compliance with “the national standard” may be referencing an earlier revision. Ask for the specific standard number with year suffix, and verify the test conditions under which compliance was demonstrated.
Practical Guidance for Buyers #
If you are procuring a containerized BESS above 2 MWh, the thermal management system deserves at least as much scrutiny as the cell chemistry. Here is what that scrutiny should look like in practice.
First, require the supplier to show their chiller sizing calculation — not just the selected chiller model. The calculation should include: total cell heat at the specified C-rate, cell thermal mass subtraction, safety factor (minimum 1.1), and the resulting net cooling load. If the supplier gives you a chiller spec sheet without this calculation, they are guessing.
Second, ask for flow balance data — either CFD simulation output or measured differential pressure/flow data across modules. The threshold to require is ≤10% inter-module flow variance. This is not a pass/fail on a certificate; it is a calculation result or a measurement — and suppliers who cannot produce it have not designed the manifold properly.
Third, confirm the test validation conditions. Accept only test data taken at ≥35 °C ambient under rated (or near-rated) charge/discharge conditions. Test data taken at 25 °C ambient is not representative of summer field conditions and should not be used for compliance assessment.
At CompactBESS, our sourcing team works with verified Chinese manufacturers of containerized BESS and liquid-cooled energy storage enclosures — connecting overseas OEM buyers and system integrators with suppliers who can demonstrate engineering-level thermal validation, not just certification paperwork. If you need to identify qualified suppliers for liquid-cooled battery enclosures meeting GB/T 26276—2023 thermal consistency requirements, reach out directly.
Need help identifying qualified suppliers for liquid-cooled BESS enclosures? Talk to our sourcing team →
Supplier Qualification Questions #
- What is the calculated net cooling load (in kW) for the chiller, and can you show the three-step calculation — gross cell heat at 0.5 Pe, thermal mass subtraction using cell specific heat capacity and allowable ΔT, and the applied safety factor of 1.1 or greater?
- What is the simulated or measured inter-module coolant flow variance across the parallel manifold, and is it confirmed to be within 10%? Can you provide the CFD simulation report or measured flow data?
- What cell-to-cell temperature spread (in K) was recorded at end-of-discharge during your charge/discharge validation test, and at what ambient temperature was that test conducted? (Require ≤6 K at ≥35 °C ambient to satisfy GB/T 26276—2023.)
- What is the installed PTC heater capacity (in kW), and can you demonstrate that it meets the cold-climate preheating requirement — specifically, heating all cells from −30 °C to ≥0 °C within the standby window at your specified heating load calculation?
- For the liquid cooling circuit, what pipe materials and connector types are used at each tier of the manifold, and are the connectors at cluster level automotive-grade rated push-fit fittings? What insulation specification is applied to prevent condensation on cold-side piping?
Sourcing Checklist #
- [ ] Chiller cooling capacity is ≥38.6 kW (or supplier-calculated equivalent with k ≥ 1.1 safety factor) for a 3.35 MWh / 3,744-cell enclosure at 0.5 Pe operation
- [ ] Cell-to-cell temperature spread at end of discharge is confirmed ≤6 K under GB/T 26276—2023, validated at ≥35 °C ambient
- [ ] Inter-module coolant flow variance is ≤10%, supported by CFD simulation output or measured differential flow data
- [ ] PTC electric heater capacity is ≥14.9 kW (or calculated equivalent) to support preheating from −30 °C ambient to +20 °C cell temperature within the standby window
- [ ] Coolant piping uses stainless steel for primary/secondary runs and automotive-grade push-fit connectors at tertiary/quaternary cluster-level branches
- [ ] All cold-side pipe external surfaces are insulated to prevent condensation in sub-ambient coolant operation
- [ ] Chiller control logic supports three operating modes: active cooling (VFD compressor), active heating (PTC), and standby (reduced pump speed), with BMS integration for remote setpoint control
- [ ] LFP cell charge temperature window confirmed as 0 °C to 60 °C and discharge window as −20 °C to 60 °C, with documentation per IEC 62619
Key Specifications Table #
| Parameter | Recommended Value | Verification Method |
|---|---|---|
| Cell-to-cell temperature spread (end of discharge) | ≤5 K (target); ≤6 K (GB/T 26276—2023 maximum) | Thermocouple array on individual cells during 0.5 Pe discharge at ≥35 °C ambient |
| Chiller net cooling capacity | ≥38.6 kW for 3,744-cell / 3.35 MWh enclosure | Engineering calculation: P₃ = k × (P₁ − Q/t), k ≥ 1.1; verify against chiller rated output |
| Inter-module coolant flow variance | ≤10% | CFD simulation report or measured flow differential across parallel manifold branches |
| Cell specific heat capacity (LFP prismatic) | 965 J/(kg·K) | Used in thermal mass calculation; verify supplier uses this value or provides test-measured equivalent |
| PTC heater capacity (cold climate) | ≥14.9 kW | Calculation: P₄ = k × (nCmΔT₁/t₁); verify against installed heater rating |
| Cell maximum temperature rise during discharge | ≤10 K | Thermocouple measurement from discharge start to discharge end |
| Chiller safety factor (k) | ≥1.1 | Review chiller sizing calculation document |
| Cell charge temperature range | 0 °C to 60 °C | Datasheet + BMS protection threshold confirmation |
Can’t find a supplier meeting these specs? Submit your requirements and we’ll match you within 48 hours.
Frequently Asked Questions #
Why is the 6 K temperature spread limit specifically important for LFP cells in grid storage applications?
Uneven temperature distribution across a cell cluster drives divergent state-of-charge during cycling — hotter cells age faster, which accelerates capacity fade in the highest-temperature cells and forces the BMS to curtail the entire cluster to protect them. Over a 10-year operating life, a chronic 8–10 K spread versus a controlled 4–5 K spread can translate to a measurable reduction in usable cycle count. The 6 K threshold in GB/T 26276—2023 is an engineering compromise between what is achievable with a well-designed liquid cooling system and what is necessary to protect long-term calendar life — it is not an arbitrary regulatory number.
Can a wind-cooled (air-cooled) BESS enclosure meet the GB/T 26276—2023 6 K uniformity requirement at full load?
At low energy densities and modest C-rates, some air-cooled designs can pass the 6 K test under controlled laboratory conditions. In field operation at high ambient temperatures and at the energy densities now standard for 20-foot containerized systems (3–7 MWh), consistent compliance is extremely difficult. The fundamental issue is that air has roughly 1/800th the volumetric heat capacity of water-based coolant — flow rate requirements to achieve equivalent cooling uniformity become impractical at this cell count and pack density. Liquid cooling is the standard for any system above approximately 2 MWh.
What coolant mixture is used, and does it require special handling or sourcing?
Standard ethylene glycol/water mixture — the same antifreeze used in automotive and HVAC applications. No dielectric fluid, no proprietary formulation. This is one of the practical advantages of non-contact liquid cooling over immersion cooling systems. Concentration ratio should be specified by the supplier to match the project’s minimum ambient temperature requirement.
How does the chiller interact with the BMS during standby (non-cycling) periods?
During extended standby, the enclosure’s thermal load drops significantly — cells are not generating heat, so continuous active cooling wastes energy. The control architecture supports BMS-driven chiller setpoint adjustment: the BMS communicates operating state (standby vs. active cycling) to the chiller controller, which then relaxes the target coolant temperature and reduces pump speed. This is an energy efficiency feature, not just a convenience — at grid scale, parasitic thermal management load across a multi-enclosure installation adds up.
What should I look for in a factory acceptance test (FAT) protocol for liquid-cooled BESS thermal management?
Require FAT conditions to include: (1) 0.5 Pe charge/discharge cycling at ≥35 °C ambient, (2) full thermocouple logging at cell level — not just module-level or cluster-level averages, (3) confirmation of ≤6 K end-of-discharge cell spread, (4) coolant flow balance verification across all parallel branches, and (5) heating function test at simulated cold-climate inlet temperature. Test data presented at 25 °C ambient with no cell-level temperature logging does not constitute adequate FAT documentation.
For deeper technical context on cell format selection and thermal design integration, see our resources on cell formats and form factors and BMS protection circuit design.
Published by compactbess.com Technical Team | Request a sourcing quote
Data source: Thermal Management Design and Validation of Liquid-Cooled Energy Storage Battery Enclosures for Grid-Scale Applications, H. Zhang et al., Journal of the Electrochemical Society, 2024
