Smart energy utilization up and down the yard lines
Machine learning and artificial intelligence deliver insights and predictions that allow us to be more intentional in our actions, including the steps we take towards better energy usage to achieve greater efficiency. Logic20/20’s smart energy solutions continue to make progress on data-driven utilization, mitigating the challenges we’re experiencing from increased demand and more extreme weather patterns.
For the chief engineer overseeing operations at a nearly 2 million square foot sports stadium, miscalculations on when to engage the cooling systems equated to excess energy demand that meant hundreds to thousands of dollars in surplus costs annually.
As part of an overarching quest to support smart sustainability measures, our team had the opportunity to assess the stadium’s approach towards energy waste management, leveraging machine learning-driven tools to decrease the carbon footprint.
How we made the 40-yard pass
Driving smart energy usage across a football field is ripe with opportunities to fumble. We worked closely with business stakeholders to identify and define the requirements necessary to meet their goals for greater energy efficiency. We charted a roadmap to make sustainable updates to current data; created dashboards that informed business operations, designed waste management interfaces and architecture to scale the solution, and–perhaps most importantly, provided it all in a mobile-friendly format.
Our team went to work on how they could best predict optimal HVAC runtime. We knew if we could accurately narrow the runtime window, we could achieve the desired stadium climate without energy waste. To get started, we analyzed available data sources and made enhancements; this included information on the current and forecasted weather, stats on solar radiation, current power demand (kW), and data from the chiller units themselves.
We leveraged Application Programming Interfaces (APIs) and other techniques to collect weather data and data from the chiller units. The weather data was intricate and came from highly responsive sensors that were located throughout the stadium itself. The sensors included precise details about the stadium’s environment, such as wind speed, humidity, pressure, and elevational temperature variances.
To visualize the data in ways that could help the operations crew make timely decisions, we built a sustainability dashboard in Power BI that included the following user interfaces:
Energy: this view shows the utilization of energy, including peak demands (annual, monthly, and daily), current demand, and monthly consumption. We sourced data directly from the stadium’s three power generators to gain insight, offering the operations crew a macro view of the stadium’s overall usage.
Chiller: this report page shares data from the chiller units on a more granular level, including what temperature goes in and what temperature goes out right at that point when the HVAC system is turned on.
Weather & Forecasting: this report describes the outdoor temperature and solar radiation at different intervals throughout the day (i.e., 12am – 12pm). It also indicates the forecasted temperature over the next 24, 48, and 72 hours in relation to the HVAC system times. By applying machine learning techniques to both the weather and energy datasets, we are able to then predict chiller lead time over 24, 48, and 72 hours, determining the ideal time to turn the units on and off. With this prediction, the operations crew would know the exact amount of runtime they would need before an event to reach optimum temperature in the stadium’s “bowl” – for example, turning the units on precisely 1.5 hours before showtime.
Combination: captures all the information from the energy, weather, and chiller reporting into one mobile-friendly screen. This report offers a holistic view of the current state as well as the determining factors to achieving desired state.
Having all this data available in one place saves the operations crew valuable time and tedious steps – they no longer need to log into each HVAC and energy system separately. The Logic20/20 solution is a one-stop-shop where they can see all the dependent factors both within the stadium itself, and externally – including forecasted weather and solar data.
Prior to delivering this solution, the operations crew was, by standard operating procedure, turning on the chillers 72 hours prior to an event. Using the new predictive models, they can now pinpoint the optimal start time for the chillers, which is typically 1-2 hours before an event, saving days of extraneous runtime.
With operations professionals that are frequently away from their desks, a key requirement for the solution was a mobile-ready view of the data, with a clean layout and clear insights on the operations systems. Our platform agnostic dashboard delivers the right information on mobile, tablet, and desktop.
The stadium operations manager can now access the dashboard right from his phone, making critical decisions on how to manage the system with just one screen view. In doing so, he is able to protect the stadium’s infrastructure and maintain a comfortable climate in the bowl.
Originally, our client was solely anticipating an analysis of their energy and weather data. But because our team had a solid understanding of the client’s vision, we were able to add the chiller data, which truly upped the game in directing operational actions that could optimize consumption, reduce CO2 emissions, and reduce costs.
Making it to the championships
Looking forward, our team is focused on hardening the existing solution using Azure best practices and integrating continuous improvements and delivery. We’re driving enhancements by leveraging Azure, machine learning, and our expertise in advanced analytics. Improving the tool’s reliability will involve boosting predictive qualities as well as evolving internal processes, the tech stack, and inter-application communication.
Hardening the existing solution will reduce overall risk, secure continuous operations, and chart the course for long term operability. In future phases, our team will evaluate additional data collection functions, build a roadmap of target architecture improvements, and establish greater disaster recovery by improving source control. Lastly, we plan to strengthen sustainability by continuing to improve operational efficiency, including new failure notifications and scheduled backups to prevent data loss. End goal? The operations team can continue to create a fantastic fan experience while hitting their long-term sustainability targets.