Dr Troy Wilson joins CIM Enviro as Chief Data Scientist

Dr Troy Wilson joined CIM Enviro as Chief Data Scientist in May of 2018.

“I want to help build a scalable system, which can be rolled out to customers around the world, bringing efficiencies, reducing energy usage and emissions and driving a more intelligent and sustainable built environment.”

Troy is responsible for the research, development and implementation of the latest advances in Artificial Intelligence (AI) towards the optimisation of the built environment. His focus is also on monetising the changes in energy consumption within the built environment and its effects on the wider energy and emissions markets for clients.

Troy envisions CIM Enviro continuing their global expansion, because the same issues currently being solved in Australia are present around the world.

“Buildings consume 40% of the world’s energy and produce one third of the world’s carbon emissions. The deployment of scalable technologies to improve the energy efficiency of the built environment, whilst driven by saving money for our customers, has the potential to have a significant global impact."

Troy has a background in Robotics and Machine Learning acquired at the Australian Centre of Field Robotics at the University of Sydney. This is also where he earned his PhD and worked as a researcher developing and applying algorithms for real-time autonomous operation of marine and terrestrial robotic platforms.

“Through my work applying AI and Machine Learning to robotic platforms in the field, I have dealt with large volumes of noisy data in which algorithms are deployed with limited computational resources and tight time constraints.”

Consequently, this led to Troy focusing his research on algorithms which are “efficient and robust to the vagaries of real world data and deployment.”

Troy also has thirteen years of experience in energy and emissions markets and was a Managing Director at Goldman Sachs where he ran the electricity, gas, coal and carbon emissions structured products desk across the UK and Europe. He also set up and ran the Australian energy trading desk for Goldman Sachs.

Today, Troy is using his passion and experience for AI and Machine Learning and applying it to CIM Enviro’s building optimisation technology. He believes AI and Machine Learning in buildings are going to have a large impact on society over the coming decades.

“It’s only by being involved at the forefront that we can separate the hype from reality, and steer the course of development in a way that is beneficial to all.”

When Troy first met CIM Enviro’s CEO, David Walsh, he says he was “excited by the large data sets available in the built environment and the opportunities to apply the latest algorithms in AI and Machine Learning.”

“Recent advances in AI have been enabled by two factors. Large data sets generated by the internet and smart phones, and increases in computing power, recently driven by the utilisation of massive parallel processing power developed initially for rendering of computer games in Graphical Processing Units (GPU's)."

On what CIM Enviro is doing in building optimisation;

“I was amazed by the amount of data buildings generate that is not being fully utilised.”

Facebook, Amazon, Netflix and Google (FANG) along with Baidu, Alibaba, Tencent and Xiaomi (BATX) have generated immense value from consumer data. The Industrial Internet of Things (IIoT) is generally not connected to the internet.

“CIM Enviro has developed the technology to access and standardise this data from the built environment, which presents an incredible resource to drive the next level efficient optimisation."

Troy has developed a deep understanding of the value of flexibility in his experience trading, making and managing risk in energy and emissions markets across the world.

“Though there is a small amount of storage in electricity markets through pumped storage hydroelectric facilities and batteries, the majority of imbalances in supply and demand are met by changes in production. Thermal energy can be stored in the mass of buildings and in some Heating, Ventilation and Air Conditioning (HVAC) systems.”

When combined with flexibility in desired thermal conditions, Troy believes there is an ability to provide demand-side flexibility to electricity markets through timing of consumption.

Optimisation of building operation has historically occurred in a silo. Large amounts of data are collected by Building Management Systems (BMS) and utilised for real-time optimisation against a small set of local objectives, such as current internal temperature.

“This narrow focus can be expanded,” he says. “Internally, savings can be achieved by including information on time-varying energy tariff structures, predicted future conditions and requirements and life cycle impacts on equipment from operating regimes.”

“Externally, changes to energy consumption levels and profiles can have positive externalities on energy and emissions markets, which has large value for producers, retailers, market operators, network service providers and even other consumers in the market.”

He advises that demand-side flexibility provided to the market can also “enhance network resilience and enable increased volumes of non-dispatchable renewable energy sources such as wind and solar to be Incorporated into the energy supply mix."

“Information on the operational performance of building plant and equipment aggregated across sites, can enable informed purchasing and operating decision for owners based on real world analysis of life cycle performance of plant and equipment in varying operational and environmental conditions."

Outside of work, Troy enjoys spending time with his family, cycling, playing soccer, reading “and if there is any time left, robotics projects.”

Troy is so passionate about data that he tracks it even in his personal life. He uses a cycling computer to collect information from a GPS, a heart rate monitor and a power meter on his rides. Last year he rode 14,061km in 500 hours and 49 minutes, and since 2012 has ridden over 75,000 km.

--