Adopting RL-based energy optimization for a building management system on AWS - AWS Prescriptive Guidance

Adopting RL-based energy optimization for a building management system on AWS

Ivan Cui, Gauhar Bains, Jake Chen, and Jack Tanny, Amazon Web Services (AWS)

August 2023 (document history)

Global temperatures are on the rise with greenhouse gas (GHG) emissions as the primary contributor. Industrial facilities are among the top contributors to GHG. The Paris Agreement stipulates that facilities need to be 30 percent more energy efficient and net carbon neutral by 2050. Many companies have set new targets to reduce their emissions in recent years. For example, Amazon’s mission is to be net neutral by 2040, and its 2022 sustainability report has touched on how the company is using innovative design to build sustainability into physical Amazon campuses.

The energy optimization of facilities must be a key component of your organization’s plan to operate more sustainably. This strategy provides information about how companies can operate and maintain their existing buildings more efficiently by using reinforcement learning (RL) to optimize energy consumption of heating, ventilation, and air conditioning (HVAC) equipment. This guidance can also be extended to other energy consumption systems such as grain mills and plant chillers, as noted in the case studies in the Resources section.

This strategy is for industrial facility managers, sustainability officers, building engineering managers, CIOs, and CTOs who are tasked with reducing energy consumption in their industrial facilities. Although the motivation for this effort is often to reduce GHG, you should also expect a reduction in energy costs. Predictive maintenance could further reduce GHGs while simultaneously reducing the operating costs of facilities.