Executive summary
3
live sites retrofitted without taking any of them offline: a campus, a hospital and an airport
5+ yrs
the battery life the PestBusters sensor retrofit was built around
Key conclusions
- Retrofits fail first in the physical layer: power supply, equipment certification and signal integrity on cabling installed before sensors were planned.
- On a live facility, the asset owner must sign off before any of this work starts. Skipping that sign-off is the quickest way to lose access to the site.
- Before approving a retrofit, name the recurring manual task it removes and the new one it creates, and state how often each occurs.
A retrofit puts a sensor on a machine, a panel or a building system that was never designed to report data. This is a different job from specifying new equipment, and it fails in different ways. The failures occur in the physical layer: power, certification and wiring installed years before anyone needed sensors.
Teams on these projects tend to start with the platform: what the sensor connects to and which dashboard displays its data. On a live facility, that is the second question. The first question is whether you can fit a sensor to the asset at all without shutting it down, voiding its certification or creating a new fault.
Retrofits fail at four points, in a predictable order
On a real site, the problems appear in this order:
- Power. Older equipment rarely has a spare terminal or a 24V rail sized for an extra load. Connecting the sensor's power is a small electrical project in its own right, and it must be finished before any data moves.
- Certification. Pressure vessels, fire systems and medical enclosures carry certifications that a drilled hole or an added tap can invalidate. The retrofit has to be routed around the certified boundary.
- Signal integrity over the existing cable run. Cable laid decades ago for another purpose was not specified for the electrical noise that a new sensor adds alongside it.
- Sign-off from the asset owner, obtained before any of the above. A retrofit carried out without it is the quickest way to lose access to the site.
We retrofitted three live sites without taking any of them offline
We ran an IoT and building-management sensor deployment across three sites that could not be taken offline: SUTD, Ng Teng Fong General Hospital and Changi Airport. From 2018, the sensors fed real-time environmental data into ENGIE's own systems. None of the three sites was built with this in mind. The main constraint on the project was fitting the installation around a campus, a working hospital and an airport that all had to keep operating throughout.
A retrofit must not replace one manual task with another
The second failure mode is harder to spot: a project solves one manual-checking problem by creating another. At PestBusters' sites, staff used to check every bait box and trap on a fixed schedule, whether or not there was any activity to find. The sensors we deployed there are designed to run for five years or more on a single battery. If a technician had to visit every unit each year to change its battery, the manual round would still exist, only for a different purpose.
Apply this test to any retrofit proposal before you sign it off. Name the recurring manual task it removes and the recurring manual task it creates, and state how often each occurs. If a vendor cannot answer the second part, the retrofit has only moved the labour from checking the asset to maintaining the sensor.
