Your standby set has one job. Most Ugandan plants find out whether it can do it during the outage.
This is instrumentation for factory-scale standby plant — the set sitting behind an automatic transfer switch on an industrial site, sized in hundreds of kVA, feeding a production line rather than an office. It answers the questions a generator cannot answer about itself: does it start every time, what does each hour actually cost, where does the diesel go, and how long is the gap between the grid dropping and the line running again.
The set runs more often than it used to, and nobody re-costed it.
In August 2026 UEDCL's acting managing director told Parliament's Public Accounts Committee that since the utility took over distribution from UMEME, outages had "almost doubled" and restoration time had gone from about 12 hours to nearly 20. Standby plant that was specified as insurance against occasional interruptions is, on many sites, now running as a routine part of the operation.
That changes the economics without anyone noticing, because the change arrives as fuel invoices rather than as a decision. A set running twenty hours a month and a set running eighty consume service intervals four times as fast, burn four times the fuel, and offer four times as many opportunities for a start failure. Very few sites have re-examined the assumptions behind their standby arrangement since the handover.
The comparison worth having. ERA's approved grid tariff for July–September 2026 is UGX 308.1/kWh for large industrial and 207.7 for extra-large — low by regional standards. Your generator's cost per kWh is a different number entirely, made of diesel, service, and the depreciation of a machine measured in running hours. Most plants can quote the first figure and can only guess at the second. Monitoring makes the second measurable, and the ratio between them is what decides how much resilience is worth buying.
Six questions a monitored set answers.
- Does it start? Start attempts, successes and failures logged from routine exercise runs — so a battery, a fuel solenoid or a controller fault is found on a Tuesday morning rather than at 14:40 during a production run.
- How long is the gap? Time from grid loss to the set accepting load. This is the number that determines whether your process rides through or has to be restarted, and on most sites nobody has ever measured it.
- What does an hour cost? Fuel consumed against kWh generated gives litres per kWh and therefore shilling per kWh — the only basis on which standby generation can be compared with grid supply, solar or storage.
- Where does the diesel go? Tank level trended continuously against run hours. A tank that falls when nothing is running is the clearest signal there is, and on many sites fuel accountability alone justifies the installation.
- How hard is it working? Load factor over time. Sets run far below their rating wet-stack and foul; sets run near their limit have no headroom for the next line you add. Both are visible only in trended load data.
- What is the grid actually doing? Interruptions timestamped at the intake with voltage and frequency behaviour around them — because unstable voltage was among the leading complaints in the EPRC survey and it damages plant without ever fully interrupting supply.
Beyond the set itself, energy sub-metering per line separates grid kWh from generator kWh. They are very different costs wearing the same unit, and a plant that cannot tell them apart cannot cost a production hour properly.
The maintenance schedule the manual actually specifies.
Generator service intervals are written in running hours. Most Ugandan sites service on the calendar, because running hours are read off a counter when somebody remembers to look. With outages up and restoration times close to a full day, the two have diverged: a set doing 80 hours a month reaches a 250-hour interval in three months, not annually.
Continuous run-hour logging turns that back into what the manufacturer intended — service triggered by use, with the interval visible weeks ahead so parts can be ordered rather than air-freighted. It also produces the maintenance record itself, which matters if the set is under warranty or on a service contract where the provider's obligations depend on documented servicing.
Where a site runs several sets, the same data shows which one is carrying the load and which one has quietly become the least reliable — a comparison that is nearly impossible to make from paper logs kept by different shifts.
Uganda's power arithmetic on one table
| Fact (Q3 2026) | Figure | What the gateway logs |
|---|---|---|
| Large industrial tariff (ERA) | 308.1 UGX/kWh; extra-large 207.7 | Grid kWh at tariff vs genset kWh at fuel cost |
| Reliable capacity vs peak | ~1,550 MW vs 1,337 MW | The power exists; the interruptions are the cost |
| Restoration time after the handover | ~12 h → ~20 h (UEDCL evidence to Parliament) | Outage duration per event |
| Businesses reporting harm from the handover | ~42% of 1,152 surveyed (EPRC, Q2 2025) | Voltage quality, outage count, restart time |
On the plant you already have.
- From the controller where possible. Modern generator controllers — DSE, ComAp, Woodward and similar — expose run state, hours, load, fuel and alarms over Modbus. Where yours does, we read it and add nothing.
- Sensors where it doesn't. Older sets get run detection, a fuel level sensor appropriate to the tank, and current measurement on the outgoing feed.
- At the intake. Grid presence and quality logging on the incoming supply, and changeover timing across the transfer switch.
- Battery-buffered nodes. The record has to survive the event it is recording, so nodes store locally and back-fill when the link returns.
- Installed by your electricians or a local contractor under live remote guidance from our engineers in East Africa Time, with hardware pre-configured in Johannesburg before it ships.
For the wider picture of what interruptions cost beyond fuel — lost production, restart ramps, scrap and equipment life — see what power interruptions cost a Ugandan factory.
Generator & power monitoring in Uganda — common questions
What size of installation is this for?
Industrial standby plant — sets behind an automatic transfer switch feeding a factory, processing plant, cold store, hospital or telecom site, typically from around 100 kVA upwards. The economics rest on fuel volume, production value protected and service intervals consumed, which is why it suits factory-scale sets rather than small portable ones.
Can you read our existing generator controller?
Usually, yes. Controllers from DSE, ComAp, Woodward and equivalents expose run state, run hours, load, fuel and alarm status over Modbus, and where that is available we take the data from there rather than adding instruments. Older sets without a communicating controller get run detection, fuel level and current sensing instead. Which of the two applies is the single largest factor in what a site costs.
How does fuel monitoring detect theft?
By comparing tank level against run hours continuously. Fuel consumed while the set is running has an expected profile; a level that drops when nothing is running, or faster than the load explains, does not. The value is less in catching an individual incident than in the fact that the comparison is running all the time — which is usually enough to change behaviour on its own.
Why measure changeover time?
Because it decides whether your process rides through an interruption or has to be restarted, and the two have very different costs. A gap of a few seconds is invisible to a warehouse and fatal to a moulding machine. Knowing your actual measured gap — not the specified one — is what tells you whether the answer is a faster transfer, a UPS on the control system, or nothing at all.
Now that outages are worse, should we add more generation?
Measure before deciding. Uganda has surplus generation nationally and low industrial tariffs, so the question is not whether grid power is expensive but how often and for how long you lose it, and what that costs you. Once you have interruption frequency, duration, your generator's real cost per kWh and the production value at risk, the choice between more standby capacity, a dedicated feeder, solar or storage becomes arithmetic. Before that it is an argument.
What does generator monitoring cost in Uganda?
Set by how many sets and intake points, whether the controllers can be read over Modbus, tank type and size for level sensing, and whether you want production data linked to the power events. We quote a written cost band after a remote review — photographs of the controller, the tank and the nameplate are usually enough to size it. On sites with meaningful fuel volumes, fuel accountability alone typically carries the payback.
Know what every generator hour buys you.
Send an engineer a photo of your controller and tank. You'll get a straight answer on what can be read as it stands and what it would cost to log it continuously.