A survey study on the development and application of data-driven model predictive control in buildings (ASHRAE RP-1934)

Many studies have developed data-driven Model Predictive Control algorithms and demonstrated their advantages in building control applications; however, their adoption in real buildings remains limited. Although most related studies have been conducted by researchers, deployment decisions depend primarily on industry stakeholders. Thus, it is crucial to survey industrial practitioners to understand the barriers and drivers. Existing efforts have been small-scale surveys and do not provide a comprehensive industry view. This study addresses that gap through a broad survey targeting control vendors, building practitioners and technicians, building owners and facility operators, utility representatives, researchers, and other professionals. The results identify several dominant barriers, including high first costs, fragmented hardware and software ecosystems, limited data accessibility and labeling, and the specialized expertise required to build and maintain models. The analysis also highlights key drivers, including decarbonization targets, renewable energy integration, and advances in artificial intelligence and analytics. Based on these findings, the study proposes actionable pathways for academia and industry. Recommendations include standardized datasets and benchmarks, transparent field demonstrations, interoperable plug-and-play architectures, and workforce training focused on practical deployment. These steps can help transition advanced building controls from bespoke pilot projects to scalable real-world practice.