A methodological framework and computational tool for adaptive predicted mean vote setpoints in EnergyPlus building simulations

Abstract

The Predicted Mean Vote (PMV) index is a widely recognized tool for designing and regulating indoor environments, but it has notable limitations, particularly in estimating comfort levels in spaces without active air conditioning. To address these challenges, the Adaptive Predicted Mean Vote (aPMV) was developed. Despite its potential, aPMV has not been incorporated into widely used simulation platforms like EnergyPlus. This study presents a novel methodology and a Python-based tool for integrating aPMV into EnergyPlus models, enabling the adjustment of heating and cooling setpoints based on aPMV rather than PMV. The proposed method is applied to a calibrated office building model, demonstrating that with an adaptive coefficient of 0.293, the heating and cooling setpoints can be adjusted to −0.59 and 0.44, respectively. This adjustment results in only a 0.4 % increase in annual HVAC electricity consumption compared to conventional ±0.5 PMV setpoints. The broader parametric analysis corroborates these results, demonstrating minimal variations in energy consumption across adaptive coefficients and confirming the robustness of the approach in diverse climatic conditions. Key advantages of this tool include its high customizability and compatibility with other Python-based energy simulation libraries, making it a versatile addition to building performance analysis. An example is available at: https://accim.readthedocs.io/en/v0.7.6/jupyter_notebooks/example_apmv_setpoints_paper/using_apmv_setpoints.html