Southern California Gas Company recently announced that as part of its ongoing push to achieve net-zero greenhouse gas emissions by 2045, it completed a machine-learning project that helped a manufacturer achieve $150,000 in annual recurring energy savings and reduce nearly 1,000 metric tons of carbon dioxide annually.
Three years ago, SoCal Gas chose The Gill Corporation, a leading manufacturer of high-performance composite materials and products, to take part in a project intended to test ways that machine learning can optimize industrial processes for cost savings and emissions reduction.
In partnership with analytics firm Metron, the project used technology to analyze data and learn where industrial processes resources were needed or where data was lacking and additional sensors were required to fill information gaps. Building appropriate models require a minimum of three months of data or ideally a year or more. According to a SoCalGas news release, the project led to a fully centralized digitization of The Gill Corporation’s operations and identified significant ways to reduce both energy use and carbon emissions. “Machine learning’s capacity to sift through and analyze data sets quickly and accurately makes it a valuable tool,” Alan Leung, SoCalGas research development and demonstration project manager, told American Gas.
He added: “We learned that success on a project like this requires the ongoing involvement of the site team—managers, engineers and operators—and genuine motivation from the company for decarbonization. It’s key to have the right stakeholders involved throughout the process for information sharing and to ensure recommendations and optimizations are implemented that can generate potential energy savings.”—Eric Johnson