Shelter cooling

The air conditioner protects the batteries. When it fails, the clock starts.

Almost every watt the site uses ends up as heat in the shelter or cabinet. The twin knows the heat load from the rectifier and radios, so it can tell a dirty coil from a hot day, and it knows how many minutes you have when the air conditioner stops.

2 × 3.5 kW split units (lead and lag), DC free-cooling fan, 2.5 kW equipment heat, 12 m² shelter

Start with one cluster of sites See how we prove results

Heat load at 35 °C: 2.93 kW
Unit power at 35 °C: 0.99 kW
Minutes to alarm: 24 min
Free cooling a year: 3,088 h

Anatomy and nameplate

1Inside temperature. Held at 25 °C for the batteries.
2Outside temperature. Sets the expected unit power and whether free cooling can work.
3Unit current, lead and lag. Power drawn against what the heat load and weather call for.
4Free-cooling fan state. Outside air does the work when it is cold enough.
5Door contact. An open door lets the heat in and the cooling out.
6Unit alarms. High pressure, low pressure and compressor trips.

Shelter cooling nameplate
Cooling
Units	2 × 3.5 kW split air conditioners, lead and lag
Lag start	Inside above 28 °C, or the lead unit fails
Set point	25 °C
Efficiency at 35 °C outside	COP 2.95
Free cooling	DC fan, 150 W, 0.6 kg/s of outside air
Shelter and heat
Floor	12 m² (3 × 4 m)
Envelope	51.2 m² of insulated panel, U·A = 25.6 W/K
Equipment heat	2.5 kW (radios and transmission, lights, controls)
Rectifier losses	0.17 kW with all modules on
Heat load at 35 °C outside	2.93 kW
Alarm / shutdown	35 °C / 50 °C

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.

Heat load

Q = P_dc·(1/η_rect − 1) + P_equipment + U·A·(T_out − T_in)

Rectifier losses, equipment heat and the heat that soaks in through the walls.

Unit power

P_ac = Q / COP(T_out)

Hotter outside air means more power for the same heat.

Unit efficiency

COP = η_II · T_evap / (T_cond − T_evap)

A dirty coil raises the condensing temperature and pulls the COP down.

Time to alarm

t ≈ m·c_p·(T_alarm − T_now) / Q

24 minutes from 25 °C to the 35 °C alarm at this site’s heat load.

Free cooling

T_out < T_in − ΔT_min

The fan can carry the load when outside air is about 4 °C colder than inside.

Heat load and the coil

The twin knows the heat coming in, so it knows what the unit should draw and how warm the shelter should be. A dirty coil and a hot day look alike to a thermostat. They do not look alike to the model.

Inside temperature against outside air, lead unit alone

2.67 kW of heat inside the shelter

Model, clean coil
Dirty condenser coil
Unit power, clean (right axis)
Unit power, dirty (right axis)

Unit power at 35 °C, clean: 0.99 kW
Unit power at 35 °C, dirty: 1.29 kW

Same heat, same weather. A dirty coil draws more power and runs out of capacity on a warm afternoon, so the lag unit starts. A hot day alone does not do that.

COP = η_II · T_evap / (T_cond − T_evap) Q = losses + equipment + U·A·(T_out − T_in)

Minutes left, and free hours

When the cooling stops, the heat load sets how fast the shelter warms. When outside air is cold, a fan does the work for a fraction of the power.

Minutes to alarm after the cooling stops

Shelter at 25 °C, 35 °C outside

To 35 °C high-temperature alarm
To 50 °C equipment shutdown

With the cooling off, the heat has nowhere to go. More load means less time. The twin knows the load, so the alarm says how many minutes are left.

t ≈ m·c_p·(T_alarm − T_now) / Q

Hours of free cooling by month

Reference climate, set point 25 °C

Hours the fan can carry the load
kWh avoided (right axis)

Free-cooling hours a year: 3,088 h
Air conditioner energy avoided: 990 kWh

Outside air is free cooling whenever it is a few degrees colder than the shelter. A fan stuck off in winter throws those hours away.

Free cooling when T_out < T_in − Q / (ṁ_max · c_p)

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
Dirty condenser coil	Dirt on the outdoor coil raises the condensing temperature. Capacity and COP both fall.	Unit current above the model for the weather. Lag unit starting on warm afternoons.	Compares unit power and inside temperature with the model. A dirty coil runs out of capacity near 37 °C outside instead of 43 °C.
Refrigerant loss	Less refrigerant means less cooling for the same compressor work.	Long run times. Inside temperature creeping up at the same heat load. Unit current lower than expected.	Sees cooling delivered fall short of the heat load the rectifier and equipment explain.
Lead and lag not swapping	One unit does all the work and wears out. The other sits idle and may not start when needed.	Run hours on one unit only.	Tracks lead and lag run hours and flags an unbalanced pair.
Free cooling stuck off	A failed fan or a closed damper sends the work back to the air conditioner.	Air conditioner running on cold days. Fan current zero.	Counts free-cooling hours used against hours the weather made available.
Door left open	Outside air and sun come straight in.	Door contact open. Inside temperature and unit current up.	Ties the temperature rise to the door event and tells the field team.

How we catch it early

A physics twin and a PINN, working on one shelter cooling.

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

Inside temperature against heat load and outside air
Unit current against expected: 0.99 kW at 35 °C
Lead and lag run-hour balance
Free-cooling hours used against 3,088 available

	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.

Free cooling

990 kWh a year

3,088 hours a year when outside air can carry the load

Battery life kept at 25 °C

3.3 years

10.0 years against 6.7 with no working air conditioner

Warning before shutdown

40 min

Alarm at 35 °C after 24 min, 50 °C after 64 min, on a 35 °C day

Dirty coil at 35 °C

+292 W

1.29 kW against 0.99 kW clean

What we do for you

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

Our reliability engineers run the twin on your shelter cooling, 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.