Most people know that airconditioning chews up a lot of energy but they might be surprised to know that efforts to optimise Heating, Ventilating and Air Conditioning (HVAC) systems are stuck in the dark ages. But, a Melbourne start-up has developed a machine-learning-enabled solution that can drive major energy and emissions savings.
The opportunity to optimise HVAC in commercial buildings is significant, says Iain Stewart, a senior analyst at independent adviser ClimateWorks.
Conservative estimates from research conducted by Climateworks found that electricity saved from commercial HVAC optimisation could be 8.6TWh (3.5 per cent of Australia’s total generation) and 6.4Mt of potential emissions savings (1.2 per cent of Australia’s total emissions).
Recognising the scale of this opportunity, two years ago Stewart started building an HVAC optimisation software platform under the company name Exergenics.
He told The Fifth Estate the technology is expected to save around 15 per cent of a building’s energy use. It will be trialled in Monash University buildings over the summer.
The start-up is currently a one-man-band but Stewart says it won’t be long before he starts hiring staff, thanks to being selected as part of EnergyLab’s start-up Acceleration Program.
Why is AC not already optimised?
Stewart says the building controls industry relies on an “if it’s not broke, don’t fix it” ethos that is preventing easy energy and emissions savings.
“It’s not that they don’t work, they just don’t work well,” he says.
The management of HVAC systems hasn’t changed much over the past 20 years. An engineer might instruct a staggered pattern to turn on chillers but that’s generally the extent of energy optimisation efforts.
Typically, the building management system is collecting data on the amount of energy these systems are using but Stewart says it’s rarely well utilised, and it is industry standard to delete this information after three months.
Stewart’s software platform builds a digital twin of the plant room – a mathematical representation of the physical system – and runs an algorithm to work out an optimal control strategy.
Using machine learning, the system predicts a building’s cooling load based on a number of factors such as the time of day and ambient weather conditions.
Data driven and open
Stewart says there are other options out there for optimising energy use for HVAC systems but he says the sentiment is that “they cost a lot and don’t do that much”.
His solution is data driven, so it works on any brand of chiller. Some HVAC manufacturers have their own optimisation software but they don’t work across multiple systems.
This solution is also more visible and open than other “black box” operations where “no one knows how it works”.
“It’s more open so that you’ll be able to see energy efficiency as a whole.”