Why we think the PPD model should not be used in practice nor in thermal comfort standards

 The PMV/PPD model is commonly adopted as a design tool in thermal comfort standard and building design. We wrote about our analysis of the accuracy of PMV model in a previous post. The PPD model was removed from the latest ASHRAE Standard 55 due to its low prediction accuracy. Using the world largest thermal comfort field survey database, we evaluated the accuracy of PPD model by comparing the PMV/PPD relationship with people’s actual responses: the observed thermal sensation (OTS) and the observed percentage of unacceptability (OPU).

We published a recent paper describing how we found that PPD failed to predict the percentage of unacceptable votes if the thermal sensation (in x-axis) was predicted using the PMV model (red line in the graph). However, the PPD prediction accuracy on OPU was improved if the thermal sensation was directly reported by respondents (compared with the purple line, OTS – OPU regression). It suggests that poor prediction performance of the PPD model is mainly ascribed by the error of PMV model. Assuming PMV can perfectly reflect subject’s thermal sensation (i.e. PMV = OTS) or thermal sensation is known from a survey response, then the PPD could in some cases be a good predictor of the percentage of people finding the space thermally unacceptable.

 We also studied the PPD model performance for various ventilation systems, building types and climates. For all ventilation and building types, the PPD model accurately predicted around neutral sensations (~5%) but overestimates the observed percentage of unacceptability it both the hot and cold ends of the scale, being especially inaccurate for classrooms and housing. The PPD model performed better in buildings in temperate climates, but larger discrepancies were observed in tropical, arid and continental climates.

Our findings show that the PPD model is accurate only when the real thermal sensation is known and under particular condition (e.g., an office building with temperate climate) but it does not perform equally well among all ventilation, building and climate types. Given the low accuracy of PMV to predict thermal sensation, we think that the PPD model should not be used in practice nor in thermal comfort standards. 

Reference: Cheung T, Schiavon S, Parkinson T, Li P, Brager G. 2019. Analysis of the accuracy on PMV – PPD model using the ASHRAE Global Thermal Comfort Database II. Building and Environment, https://doi.org/10.1016/j.buildenv.2019.01.055

Acknowledgment: This research was funded by the Republic of Singapore’s National Research Foundation through a grant for the Singapore-Berkeley Building Efficiency and Sustainability in the Tropics (SinBerBEST) Program.