Advanced control of heating and cooling systems can substantially reduce energy costs and pollution. However, real-world adoption of popular algorithms among researchers, such as model predictive control (MPC) and reinforcement learning (RL), remains limited due in part to their high deployment and commissioning costs. Here, we develop two nearly commissioning-free controllers tailored to objectives that depend linearly on the controlled thermal load, such as energy costs and pollution. The controllers require at most two thermal parameters. In representative heating simulations, controller performance is robust to large parameter specification errors, suggesting potential for deployment with no tuning. The controllers maintain good occupant comfort while achieving 43 to 98% (depending on the electricity pricing and controller variant) of the performance improvement achieved by an omniscient policy with perfect model information and forecasts. These results suggest that simple, structure-exploiting controllers may capture most of the attainable value of advanced control while avoiding the data, modeling, tuning, and computational burdens that can arise with conventional MPC or RL.
A systematic assessment framework is presented that compares four prominent DRL controllers with a classical control baseline across a diverse set of applied control problems, including non-minimum phase dynamics, flexible mechanical systems, nonlinear marine control, and aerial robotics, and clarifies the trade-offs between learning-based and conventional control.
Klinsmann Agyei, Pouria Sarhadi, Daniel Polani· 0 citations
This paper addresses the problem of energy-efficient thermal management in Battery Electric Vehicles (BEVs), with the goal of extending driving range while ensuring passenger comfort and battery thermal safety. A hierarchical Nonlinear Model Predictive Control (NMPC) framework is proposed, in which the supervisory layer explicitly balances energy consumption and cabin temperature reference tracking over a long horizon and translates this trade-off into optimal actuator command sequences for the integrated heating, ventilation, and air conditioning (HVAC) system and the battery thermal management (BTM) system. These command sequences are then enforced by a lower-layer NMPC, which introduces local corrections to compensate disturbances and modeling uncertainty while tracking an externally specified cabin temperature reference trajectory. To enable predictive control of the highly nonlinear and strongly coupled thermal dynamics, data-driven neural network models are employed as prediction models within both control layers. The proposed approach is developed and validated using a high-fidelity BEV thermal simulator developed at Politecnico di Torino in collaboration with Stellantis N.V., and benchmarked against an industrial rule-based (RB) control strategy over the WLTC driving cycle. Simulation results demonstrate a significant reduction in energy consumption compared to the baseline strategy, while satisfying comfort and battery temperature constraints.
Francesco Ripa, D. Regruto, P. Ceres et al.· 0 citations
Abstract. The need to achieve sustainable production has become an urgent necessity in the conditions of stricter environmental requirements and the rise in the cost of energy worldwide. The classical proportionalintegralderivative controllers and linear Model Predictive Controllers are conventional model-based control strategies that by nature rely on precise process models and fixed optimization horizons and hence are not well suited to the dynamic, non-linear, and complex nature of the modern production environment. The current paper suggests a new adaptive control system based on Reinforcement Learning (RL), where the overall production system is optimized and controlled to achieve Sustainable Production Systems, with the multi-objective rewarding function, which aims to minimize the Overall Equipment Effectiveness (OEE), defect rate, specific energy consumption, and carbon dioxide emissions, and a physics-informed Digital Twin safety filter that stops unsafe policy execution in training and deployment. The proposed framework is evaluated on a multi-machine flexible manufacturing cell benchmark, which includes CNC milling, turning, and robotic assembly, and yields an OEE of 93.6, a defect rate of 0.9, a decrease in specific energy consumption of 27.8, and a decrease in CO 2 emissions of 23.4 compared to PID baseline controllers. Experiments with ablation prove all the above-mentioned in the necessity of each of the architectural constituents, and the Pareto frontier analysis proves that the proposed SAC agent is the best in the OEE-versus-energy trade-off space compared to all of the competing methods.
Apoorva Verma· Materials Research Proceedin...· 0 citations
Flexible loads, such as Thermostatically Controlled Loads, are emerging as key elements for power balancing in smart grids, particularly under increasing renewable energy penetration. However, exploiting their flexibility requires knowledge of their aggregated dynamics, which are difficult to derive analytically. This challenge naturally leads to a dual control problem, where the controller must learn the system dynamics while simultaneously regulating the power consumption, to provide flexibility service to the grid. This paper presents a Set-Membership Dual Model Predictive Control (MPC) framework for the aggregated control of flexible loads that performs simultaneous system identification and optimal control of a set of loads, using Autoregressive with Exogenous Inputs models. The proposed approach actively reduces model uncertainty by minimising the volume of the Feasible Parameter Set, providing an uncertainty-aware dual control strategy. The performance of the method is evaluated through a comparative study against a passive adaptive MPC scheme in a demand response scenario. Simulation results show that the proposed approach reduces the Integral Square Error by 70.5% and the Integral Absolute Error by 57.9%, relative to the benchmark, demonstrating the benefits of incorporating active learning into the control design.
Alessandro Del Duca, Erick Meza, Andres Cordoba-Pacheco et al.· International Conference on...· 0 citations
A reinforcement learning–based temperature control framework tailored for the evaporation process and successfully closes the gap between RL-based control and conventional PID for nonlinear high-temperature processes.
B. Park, Narim Jeong, Hyukjun Yang et al.· IEEE Access· 0 citations
Model Predictive Control (MPC) is widely applied to multivariable systems due to its capability to systematically handle constraints and interactions. However, centralized MPC (CMPC) can become computationally intensive for strongly coupled systems, limiting its real-time applicability. Distributed Model Predictive Control (DMPC) addresses this limitation by decomposing the overall system into smaller subsystems, each controlled by a local MPC. This paper presents the design and experimental evaluation of a non-cooperative DMPC framework for the Temperature Control Laboratory (TCL) system, a thermally coupled two-input two-output process with nonlinear dynamics. In the proposed approach, each subsystem independently solves its local optimization problem while incorporating interaction effects through measured or predicted variables. Both sequential and iterative coordination strategies are implemented to examine the trade-off between computational efficiency and control performance. Experimental results under regulatory and servo conditions demonstrate that DMPC achieves effective disturbance rejection and accurate setpoint tracking with smooth control action. Sequential DMPC shows improved computational efficiency with comparable performance, whereas iterative DMPC enhances coordination at the expense of increased computational effort. These results highlight the practical potential of DMPC as a scalable alternative to centralized control for real-time applications in interconnected multivariable systems.
Priti R. Sukhadeve, S. Jogwar· International Conference on...· 0 citations
We use cookies to run the site and, with your consent, for analytics and to show ads.
See our Cookie Policy.
