A large university ran dozens of buildings like a city with the lights left on — air-conditioning at full power in empty lecture halls, and expensive lab equipment failing without warning. We gave the campus eyes.
The campus — dozens of buildings, halls and labs — ran blind. Air-conditioning and lighting in huge halls sat at full power even when the rooms were completely empty.
Meanwhile, sensitive and expensive research equipment — lab fridges, servers, ventilation — failed without warning, wiping out valuable research samples and stalling academic projects. And the old room-scheduling system was illogical: it scattered small lectures across giant halls, or forced a whole building to be heated, cleaned and secured for two or three evening classes. Waste at every turn, because nothing connected the timetable to the building.
Sensors and schedules in one loop — so the campus powers down what nobody's using and fixes what's about to fail.
We joined the academic timetable with operational sensors across the air-conditioning, electrical and lab systems campus-wide — one connected view of what's booked, what's occupied and what's drawing power.
Logic that syncs the hours: if a wing has no lectures from 16:00, the system automatically drops its energy use and closes it down operationally for cleaning and maintenance — no more heating an empty building.
The model watches temperature and power-draw fluctuations on lab equipment and servers, and opens a maintenance ticket to replace a part before it collapses and takes research down with it.
Small lectures stop landing in giant halls. Matching room size to the actual class cuts the footprint that has to be lit, cooled, cleaned and secured — especially in the quiet evening hours.
By catching the early signs of equipment drift, the system stopped losing the samples, experiments and projects that a surprise failure used to wipe out.
It runs against the real buildings and real timetable in production, adapting as the schedule changes — the campus stays optimised week to week, not just on the day it was set up.
The waste came out, and the failures stopped being surprises.
Campus energy, electricity and maintenance costs fell by around 18% — a large, recurring annual saving — and moving to predictive maintenance meant zero incidents of lost research materials or critical lab shutdowns.
By matching rooms to real demand, the operations team cut the number of halls kept active in the quiet hours by 30% — taking load off the cleaning and security staff and shrinking the footprint to run.
It lasts because the loop is closed: the schedule tells the building what to do, the sensors tell the model what's happening, and the model acts — on energy and on failing equipment — before a person would have noticed. Live operational data wired to automatic action is the same discipline behind our facilities-management work, applied to a campus.
Reef TRH