|
|
||||||||||||
|
Abstract Buildings account for a substantial fraction of globalfinal energy consumption, with HVAC systems representing their dominant load. The progressive deployment of dynamic electricity pricing creates a strong economic incentive to shift thermal loads toward low-price periods; however, implementing optimization frameworks on low-cost embedded hardware remains largely unvalidated experimentally. This paper presents the design and 24-hour experimental validation of a two-layer hierarchical control architecture for HVAC systems targeted at resource-constrained embedded microcontrollers. The upper layer consists of a rolling-horizon dynamic programming optimizer computing a global binary on/off schedule over a 24-hour planning horizon, while the lower layer implements a real-time safety controller enforcing hard thermal comfort constraints at every control cycle. The architecture is validated through a hardware-in-the-loop platform, in which the microcontroller exchanges signals with a first-order RC building thermal model on a real-time emulation target. Results demonstrate a total electricity cost reduction of 19.71% relative to a conventional hysteresis thermostat, with indoor temperatures maintained within the prescribed comfort band for 98.8% of the trial duration, confirming that cost-optimal thermal management is achievable on severely resource-constrained hardware. Key words: Hardware-in-the-Loop, HVAC Control, Dynamic Programming, Embedded Microcontroller.
References
|
||||||||||||
![]() |
||||||||||||
![]() |
||||||||||||
|
||||||||||||