Diesel generator set

A generator that runs light burns more fuel for every kWh, and wears faster too.

Tower loads are small, so most site generators (the DG set, in tower terms) spend their life at a quarter of their rating or less. Fuel per kWh climbs as load drops, and the engine never gets hot enough to burn clean. The twin puts a price on each hour run light and shows what battery-hybrid cycling would save.

20 kVA / 16 kW prime, 400 V 3-phase, 50 Hz, 1,500 rpm

Built and run to ISO 8528-1 (prime and standby ratings)

Start with one cluster of sites See how we prove results

Fuel per kWh at 25%: 0.54 L/kWh
Fuel per kWh at 75%: 0.34 L/kWh
Hybrid cycling, fuel: 35% less
Hybrid cycling, run hours: 67% fewer

Anatomy and nameplate

1Voltage and current per phase. Real power from the phases, not the controller’s rounded kW.
2Engine speed. 1,500 rpm on a 4-pole alternator is 50 Hz. Drift means the governor is working hard.
3Coolant temperature. Checked against what the load and outside air should give.
4Oil pressure. Read against speed and temperature. A reading with the engine stopped is a sender fault.
5Start battery voltage. 12 V. The dip while cranking shows a weak battery before a failed start.
6Run hours and start count. Service is due on hours, not on the calendar.
7Fuel level. From the tank sender, set against the burn the load explains.
8Controller alarms and mode. Auto, manual or off. A set left in manual will not start in an outage.

Diesel generator set nameplate
Rating
Prime rating	20 kVA / 16 kW at 0.8 power factor
Standby rating	22 kVA / 17.6 kW
Output	400 V, 3-phase, 50 Hz
Current at prime	28.9 A per phase
Speed	1,500 rpm, 4-pole alternator
Engine and fuel
Engine	4-cylinder diesel, water cooled
Starting	12 V electric start
Fuel at full prime load	5.0 L/h (0.314 L/kWh)
Fuel at 75% load	4.06 L/h (0.339 L/kWh)
Fuel at 25% load	2.14 L/h (0.536 L/kWh)
Fuel at no load	1.18 L/h
Service interval	500 run hours
Reference site
Run hours a year	1,081 h over 320 outages
Average load	3.3 kW (20% of prime)
Fuel a year	2,125 L

Telecom site power chain

The grid or the generator feeds the site through the transfer switch. The rectifier makes −48 V DC for the radios and keeps the battery charged. The air conditioner keeps the battery cool.

Air
AC power
−48 V DC
Diesel
Has its own twin page

The physics

The equations behind every curve on this page.

Power

P = √3 · V · I · PF

Real power from the three phases, not the controller’s rounded kW.

Load factor

LF = P / P_prime

Measured against the 16 kW prime rating, not the standby rating.

Fuel (Willans line)

F = a · P_prime + b · P

Burn has a fixed part that does not drop with load: 1.18 L/h at no load.

Fuel per kWh

F / P = a / LF + b

At 25% load it is 1.7 times the full-load figure.

Frequency

f = poles · rpm / 120

4 poles at 1,500 rpm gives 50 Hz.

Hybrid cycling

r = P_L / (P_L + P_charge · η_rt)

Share of the day the engine runs when it charges the battery and then stops.

Fuel per kWh

Every engine burns some fuel just to turn over. At low load that fixed part is spread over few kWh. Tower sites run their generators light almost all the time.

Litres per kWh against load

20 kVA / 16 kW prime generator

Fuel per kWh, Willans line

Full load: 5.0 L/h
No load: 1.18 L/h
25% against full load: 1.71×

An engine burns a fixed amount just to turn over. At a quarter load that fixed part is spread over few kWh, so every kWh costs far more fuel.

F = a·P_prime + b·P F/P = a/LF + b

Run hours by load band, one year

1,081 run hours at the reference site

Hours in each 10% band

Hours under 30% load: 1,081 h (100%)
Average load: 20% of prime

Tower loads are small. Every hour this generator ran was under 30% of its rating, where fuel per kWh is high and the engine runs cool.

LF = P / P_prime, P = √3·V·I·PF

Battery-hybrid cycling

Instead of running light all day, run the engine near its best load to carry the site and charge the battery, then stop it and let the battery carry the site.

One day: running continuously against battery-hybrid cycling

4 kW site, lithium string cycled 20–90%

Output, continuous
Output, cycled
Fuel, continuous (right axis)
Fuel, cycled (right axis)

Run hours a day: 24 h → 7.9 h
Litres a day: 51 L → 33 L
Starts a day: 4.3

Run the engine hard to charge the battery, then stop it. The same day takes 35% less fuel and 67% fewer run hours.

r = P_L / (P_L + P_charge · η_rt) fuel = r · F(0.8·P_prime) against F(P_site)

What cycling saves, by site load

Generator at 80% load while it runs

Run hours saved
Fuel saved, lithium
Fuel saved, VRLA

The lighter the site, the more cycling saves. A battery that loses less energy on the round trip keeps more of the saving.

η_rt = 93% lithium, 82% VRLA

Failure modes

What breaks, what it looks like, what the twin does about it.

Failure mode	The physics	Signals that show it	How the twin catches it
Wet stacking	Below about 30% load the cylinders run cool. Unburned fuel and soot build up in the exhaust.	Hours below 30% load. Low coolant temperature while loaded. Oily residue at the exhaust.	Counts hours by load band and flags the sites that need a load-bank run or hybrid cycling.
Failed start	A weak start battery, fuel starvation or a latched fault stops the engine from starting.	Start battery dip while cranking. Start attempts with no run. Long crank times.	Tracks every start, and flags failed or slow starts and a falling start battery.
Overheating	A fouled radiator, a failed fan belt or low coolant leaves heat in the engine.	Coolant temperature above the model for the load and outside air.	Compares coolant temperature with the expected value at that load and ambient.
Low oil pressure	Wear, low oil or a faulty sender.	Oil pressure against speed and temperature.	Flags readings that are impossible with the engine stopped (a sender fault) apart from real low pressure.
Service overdue	Oil and filters age with run hours.	Run hours since the last service.	Drafts the service work order at 500 run hours, not on the calendar.

What the twin does not do

The twin does not start or stop the engine.
Remote start only through your own controller and policy, with a person approving.

How we catch it early

A physics twin and a PINN, working on one diesel generator set.

The twin is a set of equations for this machine, tuned to its nameplate. Give it the flow, speed and conditions and it returns what every reading should be. It always gives the same answer for the same inputs.

A physics-informed neural network, or PINN, sits on top. It is trained on the plant’s own data and it is penalized whenever it breaks a physics rule. So it can only learn what the equations leave out: wear, fouling, friction. It cannot invent an answer that the machine could never produce.

Step 1
Measure
Flow, pressure, current, temperature, vibration. Whatever the asset already has.
Step 2
Physics twin
The equations of this machine, set to its nameplate. It says what the readings should be right now.
Step 3
Residual
Measured minus expected. Demand, weather and speed are already taken out.
Step 4
PINN
A small neural network that learns the cause of the gap, under the physics rules.
Step 5
Finding
A named cause, a date, and the kWh or hours at stake. An engineer reviews it.

loss = (measured − predicted)² + λ · (physics residual)²

The second term is the rulebook. It is zero when the network obeys the equations.

Why this beats black-box AI

A black-box model learns patterns from history. It needs examples of each failure, and it goes quiet when the pump runs at a point it has not seen. The twin starts from physics, so it works on day one. The PINN adds what physics cannot know about this particular machine.

That is also what makes the savings countable. The baseline is a calculation, not a guess, so the gap between baseline and meter is a number a customer can check.

What the twin checks on this asset

Litres per kWh against the Willans line: 0.60 L/kWh at this site’s 20% load
Hours below 30% load: 1,081 of 1,081 run hours
Coolant temperature against load and outside air
Start success and start-battery dip on every start
Oil pressure plausibility, running and stopped

	Fixed alarm	Black-box AI	Twin + PINN
Needs failure history to learn	No	Yes, lots	No
Says what is wrong	No	Rarely	Yes, by name
Works at a flow it has never seen	No	Often fails	Yes, the equations still hold
Puts a number on the saving	No	No	Yes, in kWh and dollars
Lead time on slow wear	Late	Varies	Weeks to months
Every number can be traced	Yes	No	Yes, to an equation

What the twin puts a number on

Energy and cost, from the equations.

Fuel per kWh, 25% against 75% load

0.54 against 0.34 L/kWh

1.58 times the fuel for each kWh at a quarter load

Hybrid cycling, 4 kW site

35% less fuel

51 L a day down to 33 L. 6,557 L a year where the generator carries the site all year.

Run hours saved

67% fewer

24 run hours a day down to 7.9, with 4.3 starts a day

Cost of running light

$1,119 a year

932 L more than the same kWh at 75% load, at $1.20 per litre

What we do for you

We watch it. We catch it. We prove it.

Our reliability engineers run the twin on your diesel generator set, review every finding, and help your crew fix it. After the repair we compare the meter to the baseline and write the result in the ledger.

Start with one cluster of sites See the 90-day pilot

More assets

More from telecom towers, then the lead asset of each other industry.

Diesel generator set

Anatomy and nameplate

Telecom site power chain

The equations behind every curve on this page.

Fuel per kWh

Litres per kWh against load

Run hours by load band, one year

Battery-hybrid cycling

One day: running continuously against battery-hybrid cycling

What cycling saves, by site load

What breaks, what it looks like, what the twin does about it.

What the twin does not do

A physics twin and a PINN, working on one diesel generator set.

Measure

Physics twin

Residual

PINN

Finding

Why this beats black-box AI

What the twin checks on this asset

Energy and cost, from the equations.

We watch it. We catch it. We prove it.

More assets