Sizing a 48 V battery backup for a remote node that already has an aged string
A remote telecom node runs on 48 V DC and draws a measured 18 A continuous. Backup must sustain that load for 4 hours at end of battery life. The site uses 12 V VRLA blocks in a single 48 V string, and the battery room sits at 35 degrees C year round. Manufacturer and operator data: capacity at the 4-hour rate is 0.85 times the 10-hour rating; temperature correction at 35 degrees C is 0.94; maximum depth of discharge allowed is 80 percent; the end-of-life ageing factor is 0.8; the operator's design margin is 1.25. The installed string is four 100 Ah blocks that have served three years at this temperature and now measure 78 percent of rated capacity on an acceptance test. A new 200 Ah bank is available, as are additional 100 Ah blocks of a different batch and age. The operator can also shed a non-critical service to reduce the load to 12 A.
1. Calculate the required nameplate capacity and determine what the existing string can actually deliver.
Read the worked answer
Energy needed per discharge. 18 A x 4 h = 72 Ah at 48 V.
Required nameplate capacity. Apply the operator's margin, then divide by each derating factor:
Required Ah = 72 x 1.25 / (0.85 rate x 0.94 temperature x 0.80 depth of discharge x 0.80 ageing) = 90 / (0.85 x 0.94 x 0.80 x 0.80) = 90 / 0.5114 = about 176 Ah nameplate.
What the installed string gives. Measured capacity is 78 percent of 100 Ah = 78 Ah at the 10-hour rate. Applying the same factors: 78 x 0.85 x 0.94 x 0.80 = 49.9 Ah usable at the 4-hour rate and 80 percent depth of discharge. Against 72 Ah required, that is a shortfall of about 22 Ah, or roughly 1.2 hours of the specified four. The string must be replaced or augmented; it is already past the point where the 1.25 margin could absorb anything.
Sanity check. A 4-hour discharge at 18 A is C/5.5 on a 100 Ah block, which is a realistic rate for VRLA, and three years at 35 degrees C is consistent with the measured 78 percent - roughly a halving of life for every 10 degrees C above 25 degrees C predicts end of useful life around this age.
2. Evaluate the three options: add 100 Ah blocks to the existing string, install the 200 Ah bank, or shed load to 12 A.
Read the worked answer
Adding new 100 Ah blocks in parallel with the aged string: reject. Paralleling VRLA blocks of different age and batch produces unequal impedance. The newer, lower-impedance string carries more of the discharge and the older string more of the recharge stress, so the old blocks fail earlier and the new ones are worked harder than designed. The aged string's measured 78 percent also means the combined bank's usable capacity cannot be predicted from nameplate arithmetic. This option looks cheapest and is the worst engineering choice here.
Installing the 200 Ah bank: sound. Usable capacity = 200 x 0.85 x 0.94 x 0.80 = 128 Ah, comfortably above the 90 Ah that the margin-adjusted requirement implies, so the string meets 4 hours at end of life with about 42 percent headroom against the design figure of 90 Ah. Verify the charger can recharge the larger bank within the allowed recharge window while still carrying the 18 A load, and that the low-voltage disconnect is set for the new string.
Load shedding to 12 A: helps but is not sufficient on its own. At 12 A the requirement becomes 48 Ah delivered, 48 x 1.25 = 60 Ah after margin, so nameplate needed = 60 / 0.5114 = 117 Ah. The existing aged string can still only give about 49.9 Ah usable, which is below 60 Ah, so shedding alone fails. A new single 100 Ah string would give 100 x 0.85 x 0.94 x 0.80 = 63.9 Ah, marginally above 60 Ah - workable only if the shed load is contractually guaranteed to stay off, which is a business risk, not an engineering margin.
3. Recommend a course of action and specify the acceptance test and monitoring you would insist on.
Read the worked answer
Recommendation. Install the new 200 Ah bank as a single string and retire the aged blocks rather than paralleling them. That meets the 4-hour end-of-life requirement with real headroom and avoids the mixed-age parallel-string trap. If capital is constrained, the defensible fallback is a new 100 Ah string combined with a contractual load shed to 12 A, but this leaves only about 4 Ah of margin (63.9 against 60) and should be flagged as such in the design record. Keep the battery room at or below 25 degrees C if at all possible: every 10 degrees C reduction roughly doubles string life and would make the ageing factor less punishing.
Acceptance test.
- A full 4-hour capacity discharge at 18 A with the rectifier off, logging string voltage, individual block voltages and room temperature every minute.
- Pass criteria: no block falls below 1.75 V per cell (42 V string) before 4 hours, and no block deviates from the string average by more than about 0.05 V per 12 V block.
- Record block impedance or conductance at commissioning as the baseline for later trending.
Monitoring to insist on. Float voltage and current, per-block voltage, room temperature, and a low-voltage alarm ahead of the disconnect threshold. Relying on float voltage alone will not reveal a weak block - it is exactly the measurement that looks normal until the site fails.
Also confirm the rectifier is sized for 18 A load plus recharge current, and that the recharge window is achievable without overdriving the string.