Weak Grids and Frequent Outages: Setting Depth of Discharge for African Base Stations

Article

August 22, 2017

When grid availability falls from 99% to 60%, a battery stops being float standby and starts cycling daily. Batteries specified on UPS assumptions fail within eighteen months across Nigeria, South Africa, Ghana and Tanzania — the problem is the duty assumption, not the product. This article quantifies depth of discharge against cycle life and gives a capacity factor table by grid availability.

One Battery, Two Completely Different Duties

Commercial districts of Lagos, provinces across South Africa during load shedding, Ghana through its 'dumsor' load management period, the urban fringes of Tanzania and Uganda, and towns in western Kenya all face the same grid reality: power does not fail occasionally, it fails several times a day for several hours at a time. That looks like a question of frequency, but it replaces the battery's entire operating mode. Where availability is 99%, a battery discharges fewer than ten times a year and sits on float; life is governed by grid corrosion and a twelve-year design life really is twelve years. Where availability is 60%, the battery completes one or two full cycles every day — over 500 a year — and life is governed by cycle count. The same product can differ in service life by more than five times between the two duties. Yet procurement specifications usually state only 'four hours of backup' with no mention of cycles, and that omission is exactly why batteries die at eighteen months.

The Quantitative Relationship Between DOD and Life

The effect of depth of discharge on cycle life is strongly non-linear, and not simply a matter of using less to last longer. For conventional AGM, 30% DOD may give around 1,200 cycles, 50% about 600, 80% only around 300, and 100% fewer than 200. In other words, raising DOD from 50% to 80% increases energy per discharge by 60% but halves cycle count, so total lifetime throughput — cycles times energy per cycle — actually falls. That curve explains why 'use it more deeply' is an illusion: it just buys replacements sooner. The correct approach is to find the DOD that maximises lifetime throughput, usually between 50% and 70%, subject to an important constraint: recharge time after deep discharge. Where the grid is present only a few hours a day, an 80% discharge needs more than ten hours to fully recharge, the next day may not complete it, and the battery sits chronically undercharged with accelerating sulphation — which is the real killer.

Why 'Four Hours of Backup' Fails in Africa

Four hours of backup is a capacity figure, not a life figure, and it fails in weak-grid duty along three paths. First, cycle count is ignored entirely: four hours of backup implies roughly 50% to 60% DOD, so at one cycle a day and 365 a year, a conventional AGM rated 600 cycles is exhausted in about twenty months — almost exactly the eighteen months seen in the field. Second, the charging window is insufficient: a string sized for four hours of load typically charges at only 0.1C, so going from 50% to 100% takes over ten hours, while the grid is present only four to six hours a day; the battery never reaches full charge and lives permanently in PSoC. Third, heat compounds it: many African sites run at 35°C to 45°C, and cycle life is discounted again at those temperatures. Stack the three and you get eighteen months. A specification must therefore state three things together: target cycle count, available charging window, and site ambient temperature.

Capacity Factors by Grid Availability

Here is a table you can use directly. The principle is that capacity is sized to the daily energy deficit, not to backup hours. Above 95% availability the duty is float standby: size to the longest single outage times load, hold DOD within 50%, and the standard front-terminal series is sufficient. Between 80% and 95%, with roughly one two-to-four-hour outage a day, apply a cycling factor of 1.2 on top of backup hours and prefer a deep-cycle series (FTC, 100–170Ah, rated 1,500 cycles or more). Between 50% and 80%, the battery is the daily cycling prime mover: apply a factor of 1.5 to 1.8 and move to lead-carbon (LLC, 300–1,000Ah, four to five times cycle life in PSoC) or LiFePO4 (SDA10, continuous 1C discharge). Below 50%, the site is effectively off-grid: size for one full cycle a day plus three days of autonomy through overcast weather, which means going straight to a solar hybrid design — adding battery alone is not economic.

A Selection Decision Tree

Condense the analysis into four steps a site engineer can follow. Step one, grid availability: above 95%, take the standard 12V front-terminal series (6-FMX, 50–200Ah, twelve-year design life) at lowest cost; 80% to 95%, take deep-cycle (FTC) or gel (6-XFMJ, fifteen-year design life, tolerant of heat with low water loss); 50% to 80%, take lead-carbon (LLC, twenty-year design life, 30% better charge acceptance so the limited charging window is actually used); below 50%, move to solar hybrid and size the battery for daily cycling. Step two, charging window: where grid presence is under six hours a day, charge acceptance dominates, and lead-carbon or lithium beat conventional AGM. Step three, ambient temperature: above 45°C persistently, prefer lead-carbon or a high-temperature series; where temperature is controlled, lithium's volume advantage can be used. Step four, remaining life: a site with under five years of lease left does not need a twenty-year technology — take the cheapest compliant option for the remaining term. After four steps there is little room for argument.

Five Mandatory Items in the Specification

In weak-grid duty the specification must state the duty, not just the capacity. Five items are mandatory, and omitting any one returns you to the eighteen-month failure cycle. First, grid availability as a percentage, with the reporting period stated — monthly or annual, since the two produce very different figures — and this item alone decides whether the battery floats or cycles. Second, longest single outage duration and average outages per day, stated separately: the first determines capacity, the second determines cycle life, and most specifications capture only the first. Third, available charging window per day, which sets the charge acceptance requirement; below six hours, fast-charging technology is mandatory or the battery never reaches full charge. Fourth, site temperature range, stated as cabinet interior rather than outdoor, with the summer peak, because cabinet interiors typically run more than 10°C above ambient. Fifth, target service life and remaining site lease, so a site with three years of lease left is not fitted with a twenty-year battery — that is not prudence, it is waste. With these five stated, bidders can no longer hide behind a single sentence about four hours of backup, and evaluation has a comparable basis. Many projects fail not because the wrong product was chosen but because the duty was never specified, so suppliers quoted against the most favourable assumption while operators ran the site under the least favourable reality. View Telecom Backup Power Selection