Engineering Notes

In short. An engineering note answers one narrow question, such as how efficiently a generator converts diesel to electricity or what a load test actually proves, with its inputs, units and boundary made visible. Its value is a traceable basis. A correct equation applied to the wrong boundary still gives the wrong answer, so every note states what was measured, what was assumed and which decision the result supports.

What makes a note different from an article

A technical article teaches a whole method. A note isolates one relationship and works it through cleanly. Typical starting points are a motor that will not accelerate on the set, an unexpected voltage dip when a lift starts, current readings that do not agree between the controller and a clamp meter, or uncertainty about how long the fuel will last. The note identifies the condition, the equipment, the operating state and the single calculation or check that moves the investigation forward. It does not claim to settle the final design.

Rate versus performance: litres per hour and litres per kWh

The note we are asked about most concerns fuel. Litres per hour is a consumption rate. It tells you how fast the tank empties, and nothing about efficiency, because it ignores how much electricity was produced. Litres per kWh relates consumption to useful output over the same period and is the figure that allows two generators, or two months, to be compared.

SFC (L/kWh) = fuel used (L) / energy delivered (kWh)

Fuel used is the measured diesel consumed in the period, energy delivered is the kWh recorded by the generator’s energy meter or computed as average kW times run hours, and SFC is the specific fuel consumption. Worked example (illustrative): Generator A burns 20 L/h while carrying 80 kW, so it delivers 80 kWh per hour and its SFC is 20 / 80 = 0.25 L/kWh. Generator B burns 15 L/h while carrying 30 kW, so its SFC is 15 / 30 = 0.50 L/kWh. B looks thriftier by the hour and is twice as costly per unit. At a diesel price of, say, 90 rupees per litre, A produces electricity at about 22.5 rupees per kWh and B at about 45. The usual cause of a figure like B’s is light loading: diesel engines are most efficient between roughly 60 and 85 percent of rated load.

Without useful output the cost per kWh cannot be calculated meaningfully. Use measured litres and measured load where possible, keep estimates labelled, and see fuel consumption for typical curves.

Making a calculation traceable

Record the inputs, units, assumptions and the source of each significant value. Keep measured values separate from estimates. Check whether the method applies to single-phase or three-phase operation, balanced or unbalanced load, and steady or changing conditions. A three-phase power formula applied to one phase reading, or a nameplate power factor assumed when the load is mostly UPS, are the usual sources of a confident wrong answer. Retain the note’s version with any project record that uses it, because a changed input can change the conclusion.

What a load test proves, and what it does not

A generator’s real test happens under load. A loaded test checks load acceptance, voltage and frequency stability, thermal behaviour and fuel consumption, none of which a no-load run reveals. There are two ways to do it.

Load-bank test

A controlled resistive or resistive-reactive bank applies known load in steps, independent of the site. It is the cleaner test for capacity, for burning off wet-stacking in lightly loaded sets and for proving an overhaul. Steady-state and step-load tests may both be needed. It is a defined scope item and is not automatically part of every service visit; see load bank testing.

Site-load test

A test on the actual connected load proves the real interaction with motors, UPS and the changeover panel. It cannot easily reach full rating and its result depends on what was running that day. A narrow boundary means a narrow claim: a site-load test at 40 percent proves operation at 40 percent.

Site conditions that masquerade as engine faults

A mechanically healthy generator can overheat because the room or airflow path is wrong: recirculated hot air, a blocked louvre, or a radiator discharge duct that is too small. Diagnosis must therefore include the site, which is covered in generator room ventilation. Air-filter restriction matters in dusty stone, construction and mining sites. Exhaust back-pressure from long runs or tight bends raises exhaust temperature and smoke; exhaust pipes should be self-supported, isolated from the engine by a flexible connector and discharged outdoors away from openings. Do not treat smoke colour alone as a diagnosis. When load is added, voltage dips briefly and the AVR restores it; a slow or oscillating recovery points to the AVR, the excitation or the governor, not to the engine block.

Common mistakes

• Comparing sets by litres per hour; the set that seems economical is lightly loaded and costs more per kWh.
• Estimating kWh from rated kVA rather than measured kW; the SFC comes out flattering and hides wet-stacking.
• Calling a no-load run a test; faults in fuel delivery, cooling and AVR response stay invisible until a real outage.
• Diagnosing overheating as an engine fault without checking the room; a costly radiator job leaves the recirculation problem untouched.

What to send us

For a note-style question about your set, send the nameplate photograph, a few days of run hours, fuel dips or flow-meter readings and kW readings from the controller, the ambient conditions, and a short description of the symptom. Photographs of the room, louvres and exhaust route help with thermal questions. Send your service enquiry, or read the related engine service guidance first.

Questions engineers ask

Technical content reviewed by Srinivas Reddy, Managing Partner, Reddy Generators. Last updated 19 October 2026.