Short answer
Bulk milk coolers and village chilling centres sit outside the plant fence on unreliable grid power with no telemetry, so a compressor failure, a power outage or a lid left open is usually discovered when a tanker is rejected at reception. Treating each centre as a site in one portfolio, with compressor power, temperature profile and lid events reported over cellular or LoRaWAN, turns a quality incident into a maintenance alert hours earlier.
The blind half of a dairy business
A dairy group can instrument its main plant beautifully and still be blind across most of its operation. India's cooperative sector alone has an installed processing capacity of around 43.3 million litres per day and actually processes an average of 33.5 million litres per day, drawn through a collection network that reaches into tens of thousands of villages. Programmes such as NPDD have created roughly 168 lakh litres per day of milk chilling capacity.
Every one of those chilling points is a small refrigeration plant with an electricity supply, a temperature requirement and no one watching it in real time.
What goes wrong at a chilling centre
01
Power that comes and goes
Rural supply is intermittent. A cooler that loses power for three hours in the afternoon may still show an acceptable temperature at the next manual reading, having spent part of the day well outside spec.
What it costsThe quality damage is done before any log sheet records a problem.
02
Compressor degradation with nobody to notice
The same silent efficiency drift that affects plant refrigeration affects a bulk milk cooler, except here there is no engineer on site. The unit runs longer and longer to reach setpoint until one day it does not reach it at all.
What it costsRun hours climb for months before a failure that could have been scheduled.
03
Lids, doors and the human factor
A lid left open, a cooler filled beyond capacity, a unit switched off to save electricity. All are ordinary human decisions taken with no visibility of their consequence.
What it costsEach one is a quality risk that only surfaces at reception testing.
04
No comparability across the network
With hundreds of centres, the question that matters is not how one is doing, it is which ten are worst. Without telemetry there is no ranking, so maintenance attention is distributed by complaint rather than by need.
What it costsField engineering time goes to the loudest centre, not the failing one.
What to measure at a remote chilling point
| Signal | What it catches |
|---|---|
| Electrical power and run hours | Compressor health, short cycling, unit switched off, and the true energy cost per litre chilled |
| Milk temperature profile, continuous | Excursions between manual readings, which is where most quality loss actually happens |
| Ambient temperature | Separates a hot day from a failing unit |
| Lid or door events | Distinguishes a process problem from a behaviour problem, which need different responses |
| Supply availability | Outage duration and frequency per centre, which is the case for a generator or a battery backup |
| Volume received | Energy per litre chilled, the only fair way to compare centres of different sizes |
Why it belongs in the same platform as the plant
It is tempting to treat cold chain monitoring as a separate problem with a separate product. That produces two dashboards, two alerting regimes and two versions of the truth about energy per litre.
In aQ Core, a chilling centre is simply another site in the portfolio, ranked alongside the plants by total consumption, intensity and per litre performance. Missing data flags make an un instrumented centre obvious on day one rather than a gap nobody noticed. And the same anomaly engine that watches a plant compressor watches the cooler.
The economics
The case for cold chain telemetry rarely rests on energy savings alone, though compressor efficiency at scale is real money. It rests on rejection avoidance. A rejected tanker is a full day of collection from multiple villages lost, plus all the energy already spent chilling it, plus the relationship cost with the producers whose milk was in it.
One avoided rejection per centre per year across a large network is usually the entire business case, and everything after that is upside.
Can bulk milk coolers be monitored without reliable internet?
Yes. Remote chilling points are connected over cellular or LoRaWAN rather than fixed connectivity, with local buffering so that data is held and forwarded when the link returns. A wireless survey during the site assessment confirms coverage before any commitment is made.
What should be monitored at a village chilling centre?
At minimum: continuous milk temperature, compressor electrical power and run hours, ambient temperature, lid or door events, and power supply availability. Adding volume received allows energy per litre chilled, which is the only fair basis for comparing centres of different sizes.
How does this reduce milk rejections?
By converting a discovery at reception into an alert at the centre. A temperature excursion, an extended power outage or a compressor that is failing to reach setpoint raises an alert while the milk is still recoverable and while a field engineer can still be dispatched.
How many sites can one portfolio hold?
Aqueduct runs portfolio analytics across estates spanning many dozens of sites in existing deployments, ranking and benchmarking by usage, intensity and cost, and the architecture is designed for portfolio scale rather than single site monitoring. A chilling centre network is treated as sites in the same portfolio as the plants.
