Imagine for instance that sensors to detect a building’s temperature have been placed near the entrance of a shopping centre with doors that continually open and close. There will inevitably be a wrong reading of the building’s energy performance.
This means you can’t afford to rely only on technology.
Another example provided by Josh Nicholson senior technology consultant relates to an accident at an industrial facility that involved a truck damaging a sensor.
“That sensor, instead of going ‘fail’, went ‘all air quality is bad’, so we had a mechanical fan running at 100 per cent throughout the weekend because it happened on a Friday afternoon.
“While connecting systems are fantastic, and that’s the path we’re going down, we can’t go down there without understanding the ramifications of letting mechanical and computerised systems and sensors control our assets.” he says.
Which is why it is still necessary to have facility managers.
AI can help and is being overlaid in the data analytics space in integrated building platforms and ISPs (Integrated System Platforms) to provide a higher level of certainty and checks when things go wrong as well as to answer queries such as “how’s my chiller running? Or what should I be expecting for electrical usage today; how does this compare to yesterday?”
But it still takes “someone with a fair amount of understanding and knowledge of what the data means to know if it’s correct and if it’s running as it should be”, Nicholson says.
Traditionally, large quality assets have relied on very senior engineering and facilities teams to properly understand the asset, but as companies now continually “roll through different facility management companies”, the trend is to attempt to move away from depending on key personnel and instead implement tools that make “everybody educated about an asset”.
Utilising AI in BMS
But when it comes to AI, the goalposts continue to move.
“As we understand more and more about what we can achieve with AI, we keep moving our expectations of what it’s going to give back to us,” Nicholson says.
“Inherently, we still need the right understanding of building systems to ask the right questions.
“Everybody in every industry we keep jumping over thinks that this AI layer will perform miracles. And it will if we give it really good quality data, with really good quality systems to work on. But no building is like that – no building is designed that way.”