Conserve It's General Manager Chirayu Shah, and R&D Engineering Manager Michael Berger presented an artificial intelligence-supported model predictive controller (AI-MPC) framework at the 2020 ARBS Exhibitions Ltd Seminar Series Online. AI-MPC uses a high-level model to generate predictions of system load and resources to minimise operation costs. AI methods will be applied to continuously learn and update the system models based on feedback from measurements and to generate predictions and define optimal trajectories.
The AI-MPC focuses on addressing the following two challenges existing in the framework:
1. Reliable and efficient estimation of future status is required for multiple factors in the system, e.g., temperature and energy demand.
2. Joint optimization of the whole system. Building HVAC energy usage consists of various subsystems that, from an energy delivery perspective, can be classified as air side and water side. These subsystems share commonalities and dynamically interact with each other however the most efficient performance of each subsystem may not necessarily lead to optimum performance of the overall system.
Watch the video recording here.