Carnegie Mellon cuts energy usage with cloud machine learning solution

Sean Dudley
Sean Dudley
By Sean Dudley on 24 September 2014
Carnegie Mellon cuts energy usage with cloud machine learning solution

Carnegie Mellon University (CMU) in Pittsburgh, US, has significantly cut its energy use with a cloud machine learning solution. Thanks to Microsoft Azure and Microsoft Global ISV partner OSIsoft’s PI System, the university has been able to reduce building maintenance, energy costs and energy use by as much as 20%.

CMU faces big data challenges in fields such as astrophysics and building management, the latter of which includes the operational efficiency of the university’s own buildings.

Traditionally, the monitoring of operational efficiency in buildings centres around the collection and analysis of disparate data which is captured by sensors and actuators that control features, such as heating, cooling, lights, ventilation, plug load and security systems. However these systems cannot predict failures or reduce energy use, problems which can result in expensive system failures and wasted energy.

CMU adopted the PI System as its infrastructure in 2011 to connect sensors, data and people, helping deliver real-time insights into facility performance.

In 2013, CMU extended the solution with the self-service business intelligence capabilities of Microsoft Power BI for Office 365. The university migrated the PI system to a hybrid on-premises/cloud configuration using Microsoft Azure infrastructure-as-a-service (IaaS).

CMU wanted to add real-time predictive analytics to the solution in order to help building managers act proactively and carry out tasks such as repairing or replacing worn components before failure. The university also wanted automated building systems to act more cost-effectively and precisely by anticipating when and by how much thermostats can be adjusted to anticipate heating and cooling requirements. The predictive analysis requirements hinged on speed, ease of implementation, and accessibility for non-technical personnel on a daily basis.

To achieve this, the university extended its solution with Azure Machine Learning, a Microsoft Azure platform-as-a-service (PaaS) offering that uses a highly visual interface with prebuilt models and templates to help reduce the time, cost, and complexity of creating and training predictive models for use with applications such as the PI System.

CMU used Azure Machine learning alongside historical data from the PI system to meet the challenges of fault detection and diagnosis for environmental control-system components. Researchers at the university used the temperature of the water in the building, which was released by a valve, as a proxy measure of its performance. They then used the PI System and Azure Machine Learning to compare predicted and actual water temperatures, highlighting deviations between these to identify potential failures.

The solution uses four key components in the extended PI System and Microsoft Azure solution. It begins with an on-premises PI Server that collects sensor data from across the campus and forwards it via Azure-based PI Cloud Services to a PI Server running in Microsoft Azure IaaS. A research tool from OSIsoft then cleanses, aggregates, shapes and transmits the data to a working repository of Microsoft Azure table storage, where it is analysed by Azure Machine Learning. Predictive insights can be accessed through Power BI, and these predictions are stored in the PI Server for use by the building systems applications.

Based on experimental results, CMU researchers estimate the solution could cut energy costs by 20%, and plans are being discussed to implement it campus-wide.

“The savings come both from reducing energy use and from being able to shift some energy use to hours of lower demand and cost,” said Bertrand Lasternas, a researcher at the Center for Building Performance and Diagnostics at Carnegie Mellon University. With the new solution, creation and test times for machine learning models have been significantly reduced.

“We immediately began using Azure Machine Learning without having to prepare on-premises software; everything’s ready-to-use in the cloud,” said Lasternas. “It’s significantly easier to use than other tools we’ve tried, and it fit seamlessly with the PI System and Microsoft cloud solution we already had.”

“We see Azure Machine Learning and the PI System ushering in an era of self-service predictive analytics for the masses,” he added. “We can only imagine the possibilities.”

Number of views (8853)/Comments (-)

Comments are only visible to subscribers.