A Thermal Sensation Index for Real-Time Tuning and Energy-Optimal Control of HVAC Systems

Abstract

In this paper we discuss and develop a thermal comfort index that addresses the limitations of applying thermal comfort indices to control applications. The derivation closely follows the derivation of PMV, but certain changes and simplifications make the index an explicit, linearly parameterized function of environmental variables. We show that the differences between the derivation of this index and the derivation of PMV do not reduce the accuracy of the index in comparison to PMV. Since this index is linearly parameterized, the parameters can be quickly and efficiently tuned in real time to reflect the thermal sensation of the specific occupant. Parameter tuning makes it possible to accurately predict the thermal sensation of the occupant without exact knowledge of the activity level or clothing insulation of the occupant when these two quantities are known to be constant. Additionally, the tuning process makes the thermal sensation prediction relatively insensitive to sensor location because biases and scaling errors are absorbed by the estimated parameters. Real-time parameter tuning is demonstrated experimentally for a seated, stationary occupant. The feasibility of using variable air flow and variable heat flow to regulate the thermal sensation index in a way that minimizes power consumption is investigated. The simplified index provides a quantitative means for determining the most energy efficient comfortable conditions. The analysis demonstrates that for low to moderate outdoor relative humidity there is an energy optimal combination of air flow and heat flow.