Practical challenges of model predictive control (MPC) for grid interactive small and medium commercial buildings
Publication Type
Date Published
Authors
Abstract
To the urgent call for mitigating climate change, substantial initiatives have been
undertaken to deploy grid-interactive heating, ventilation, and air-conditioning (HVAC) controls,
such as model predictive control (MPC) for buildings. These efforts typically aim to curtail peak
energy demand, shift load and enhance overall energy efficiency. With the recent development
of low-cost MPC technologies that don’t require extensive instrumentation or manual modeling,
small and medium commercial buildings (SMCBs), which rarely utilize advanced HVAC control
systems, have become candidates for grid-interactive efficient buildings (GEBs). However,
despite the potential benefits and maturity of the technology itself, several practical challenges
remain in real-world implementation. In this paper, we share the practical challenges that we
have encountered in implementing and testing three types of MPC solutions (ON/OFF unit, dualfuel,
and VRF systems) on multiple SMCB sites. We describe the MPC deployment process and
discuss the lessons learned. The site selection, eligibility, and retrofit availability (e.g., utility
price structure, thermostat communications, etc.) are the main discussion points at the beginning
of the project. Also, the modeling automation and the best practices for interacting with endusers
and handling erroneous situations are presented for successful operations.