Disturbance-Adaptive Data-Driven Predictive Control: Trading Comfort Violations for Savings in Building Climate Control

Model Predictive Control (MPC) has demonstrated significant potential in improving energy efficiency in building climate control, outperforming traditional controllers commonly used in modern building management systems. Among MPC variants, Data-driven Predictive Control (DPC) offers the advantage of modeling building dynamics directly from data, thereby substantially reducing commissioning efforts. However, inevitable model uncertainties and measurement noise can result in comfort violations, even with dedicated MPC setups. This paper introduces a Disturbance-Adaptive DPC (DAD-DPC) framework that ensures asymptotic satisfaction of predefined violation bounds without knowing the uncertainty and noise distributions. The framework employs a data-driven pipeline based on Willems’ Fundamental Lemma and conformal prediction for application in building climate control. The proposed DAD-DPC framework was validated through four building cases using the high-fidelity BOPTEST simulation platform and an occupied campus building, Polydome. DAD-DPC successfully regulated the average comfort violations to meet pre-defined bounds. Notably, the 5%-violation DAD-DPC setup achieved 30.1%/11.2%/27.1%/20.5% energy savings compared to default controllers across four cases. These results demonstrate the framework’s effectiveness in balancing energy consumption and comfort violations, offering a practical solution for building climate control applications.

💡 Research Summary

The paper introduces a novel control framework called Disturbance‑Adaptive Data‑Driven Predictive Control (DAD‑DPC) for building heating, ventilation, and air‑conditioning (HVAC) systems. While Model Predictive Control (MPC) can reduce energy consumption, its practical deployment is hampered by the need for accurate physics‑based models. Data‑driven Predictive Control (DPC) alleviates the modeling burden but suffers from model uncertainties and measurement noise that may cause comfort violations. To address this, the authors propose to regulate the average comfort‑violation ratio, denoted by α, rather than enforcing hard constraints at every instant.

The core idea is to adaptively tighten or relax the disturbance bound used by the predictive controller based on observed violations. The disturbance bound estimator D(σ) is built using Willems’ Fundamental Lemma, which reconstructs a linear predictor from past input‑output data without explicit system identification, and conformal prediction, which provides a non‑parametric confidence set for unknown disturbances (weather, solar gains, occupancy). The scalar σ∈(0,1] controls the size of the bound: larger σ yields a more conservative set.

Algorithm 1 runs online: at each sampling instant the current indoor temperature yₜ is checked against the time‑varying comfort set Yₜ. A binary indicator vₜ (1 if a violation occurs, 0 otherwise) updates an internal variable αₜ via a simple integrator with gain η, then αₜ is clipped to the interval