Lithium-Ion Battery Pack Design Simulation for Telecom Towers and UPS Backup

The Standby Paradox

A backup battery has a strange job. For 99% of its life it does nothing: it sits at full charge in a cabinet at the foot of a tower, or next to a rack in a server room. Then the grid fails, and it must deliver its full rated runtime, on the first attempt, with nobody on site.

That “idle, then perfect” duty changes what good design means. For vehicle packs, engineers fight peak power, fast charging and vibration. For telecom and UPS packs, the enemies are slower and quieter: heat soaking into the cabinet for years, high state of charge accelerating ageing, a BMS fault nobody notices, and a single weak cell that spoils the runtime on the one night it matters.

Simulation lets teams find these problems on screen. The IST Li-ion Battery Simulator is a browser-based workbench that takes a pack from application targets to a finished design report, covering sizing, cell modelling, thermal behaviour, BMS logic, ageing and compliance checks, with nothing to install.

Where These Packs Are Used

• Telecom tower shelters and outdoor cabinets: Typically 48 V-class DC systems backing up radio equipment, often in hot climates and remote locations.
• Data centre and enterprise UPS: Rack or cabinet batteries bridging the gap until generators start.
• Commercial and industrial UPS: Protecting production lines, banks, hospitals and control rooms.
• Hybrid sites: Towers and microgrids that add solar and cycle the battery more often.

Each has different voltage, runtime and environment targets, which is why the design starts with the application rather than the cell.

Five Questions Every Backup Pack Design Must Answer

1. Does it deliver the required runtime?
Runtime depends on usable energy, the load and how the pack behaves at the end of its life, not just on day one. The simulator sizes the pack from your voltage window and energy target and flags every target the chosen cell and BMS cannot meet.

2. Will it survive its environment?
Cabinet temperature drives ageing more than discharge heat does. The guide’s example application uses ambient up to 50 °C, which is a realistic stress case for shelters in hot regions.

3. How long will it actually last?
Standby packs often fail on calendar ageing rather than cycle count. The ageing model captures SEI growth, which is faster at high state of charge and high temperature, exactly the conditions of a float-charged battery in a warm cabinet.

4. Will it protect itself, and report problems?
Unattended sites need a BMS that trips correctly, balances reliably and raises clear fault codes over CAN or RS485.

5. What happens if a cell fails?
In an enclosed cabinet or rack, one cell going into runaway must not take down the whole installation.

The Workflow, Step by Step

1. Define the application

You enter the application type, voltage window, energy capacity, power, ambient range, IP target, cooling method, weight budget and the standards you need (such as UN 38.3 and IEC 62619). These targets drive every later check, so the guide advises filling them in carefully.

2. Select the cell and BMS from a database

The simulator’s libraries cover cells, BMS boards (including ESS and UPS reference designs), thermal parts, busbars, fuses, contactors and more. For stationary use the main trade-offs are:

• LFP vs NMC: The guide describes LFP as the safest and longest-lived of the common chemistries, which is why it is popular for backup.
• Large prismatic vs small cylindrical: Fewer, larger cells mean fewer welds and a simpler BMS, while smaller cells give finer granularity.
• Charge temperature limits: Cold sites may need a heater, and hot sites need margin below the cell’s discharge limit.

The cell-to-BMS check compares series range, OVP and UVP against the cell window, over-current ratings, over-temperature trips and communication protocol.

3. Size the pack

Enter Ns, Np and module count, or let the simulator suggest them. Formulas are transparent: pack voltage is Ns times cell voltage, capacity is Np times cell capacity, and resistance is Ns times DCIR divided by Np, plus a 10% allowance for busbars, welds and contactors.

Illustrative example (using the 150 Ah, 3.2 V, 0.8 mΩ LFP cell from the guide’s worked example, with an assumed load):

That 20 W is the key insight: during a normal backup event the pack barely warms itself, so passive cooling is often adequate. The real thermal risk is the cabinet’s ambient temperature over years. Substitute your own cell and load figures.

4. Check thermal behaviour

Two separate questions need answers:

• During discharge: The 3-D cooling simulation handles natural convection, forced air through cell gaps or a liquid cold plate, reporting maximum temperature, cell-to-cell temperature difference and energy balance. The guide suggests targets of Tmax at or below 45 °C and ΔT at or below 5 K.
• During abuse: An auto-generated ARC-style runaway analysis shows onset temperature and the margin to the BMS over-temperature trip, and a propagation module compares barriers (none, air gap, mica, aerogel or ceramic fibre) to find the cheapest one that stops spread or buys time.

One honest caution: the 3-D cooling tool is not a full CFD solver, so detailed cabinet airflow design still calls for full CFD or a prototype test.

5. Verify BMS logic

• Protection and balancing state machine simulation
• SOC and SOH estimation, useful for remote runtime reporting
• A full-day duty-cycle run with the BMS in closed loop (the closest built-in stationary profile is the ESS day of rest, solar charge and evening load, so adjust the settings for outage-style use)
• A fault-injection matrix covering over-charge, over-discharge, over-current, short circuit, cooling failure, cold charging and open sensors, with detection latency and diagnostic codes
• C code generation with software-in-the-loop tests and a CAN database file, which is template-based and not certified, with hardware short-circuit protection still required

6. Predict life

Set the years, cycles per day (often near zero for backup), depth of discharge, rest SOC, seasonal ambient swing and end-of-life threshold. The model returns end of life in years and equivalent full cycles, SOH and resistance curves, and a calendar-only reference. That comparison shows how much extra life you gain from a cooler cabinet or a lower resting SOC. Warranty prediction adds expected claim rate and the longest warranty that stays under your target.

7. Size the electrical circuit and run safety checks

The tool derives design current, derated fuse, busbar cross-section, contactor ratings and minimum isolation, and checks fuse breaking capacity against short-circuit current. Packs above 60 V are treated as class B with contactor and pre-charge protection, which matters for higher-voltage UPS strings. The safety cards cover capacity, resistance, isolation, overcharge, short circuit, drop, shock, vibration and protection timing, then roll up into a release decision.

8. Generate the report

One click runs every stage and produces an A4 design report with the schematic, results, standards checklist and bill of quantities, a useful starting point for RFQ responses.

What You Get Out of It

• Targets your cell and BMS cannot meet are flagged before you order anything
• A realistic view of cabinet temperature, runtime and life under your site conditions
• Evidence that the BMS protects the pack and reports faults correctly
• A barrier choice backed by a propagation comparison
• One shared pack definition feeding every analysis, so changes ripple through consistently

Limits Worth Stating Plainly

The simulator is a design and screening tool. Pack arithmetic is exact for your inputs, but thermal, abuse, ageing and system models are reduced-order. It does not replace full CFD for cabinet airflow, or the physical tests needed for approval. The built-in vehicle standards (AIS-156 and AIS-038) apply to vehicles, so for stationary packs rely on the ticked standards such as IEC 62619 and UN 38.3, and confirm operator-specific requirements with your customer and test lab.

A sensible validation path is to replace synthetic cell data with measured HPPC, EIS and ARC data, test a prototype module, run SIL then HIL on the BMS, complete certification tests with an accredited lab and start calendar-ageing tests early, so the life model can be refitted before you commit to a warranty.

Conclusion

Telecom and UPS packs earn their keep on the worst night of the year. Simulation lets design teams stress that night in advance: the hot cabinet, the aged cells, the missed fault, the failing cell. The result is a pack designed around reliability rather than just a datasheet.

Explore pack design at Academy of EV Technology and follow the step-by-step guide from mission targets to a full design report.