Models for body-segment-specific thermal sensation and comfort were put forward in 2010 in a three- part series in this journal. The models predict these subjective responses to the environment from thermophysiological measurements or simulations of skin and core temperatures, and apply to a range of environments: uniform and non-uniform, transient and stable. The models are based on unique experimental data, and formulated in a rational but piecewise structure that simplifies further validation and refinement. The models have received much attention and this experience has pointed out two issues needing improvement at the fundamental level. This paper presents solutions to these issues:

a)  In the local sensation model, the neutral set-points for segment skin temperatures are sensitive to the distribution of clothing insulation provided by different clothing ensembles, and to metabolic rate. A new calculation sequence automatically creates accurate segment set-points for specific clothing and activity levels.

b)  In the whole-body (overall) sensation model, the piecewise model construction produced unrealistic jumps in output at the transitions between pieces. A smoothing technique using the model’s key organizational variables was developed and incorporated into the original model.