Thermal insulation of multi-layer clothing ensembles measured on a thermal manikin and estimated by six individuals using the summation method in ISO 9920

The thermal insulation of a clothing ensemble can be measured with a thermal manikin or estimated according to ISO 9920. In ISO 9920 there are insulation tables for a number of clothing ensembles with specified garments and other tables for a variety of single garments. The insulation of single garments can be summated to estimate the insulation for a clothing ensemble. However, when a clothing ensemble consists of multiple layers, the summation of the insulation of single garments may overestimate the ensemble insulation since the air volume between the layers and between the fibres is probably reduced.
The purpose of this study was to investigate the error of the estimate of the insulation of cold-protective clothing ensembles comprising two or more clothing layers and the inter-rater reliability of the estimate.
Two coveralls, sweaters, thermal pants, t-shirt, socks, gloves, hood and helmet were combined into seven clothing ensembles. The thermal insulation of the ensembles was measured on a standing thermal manikin. The insulation of all combinations were estimated by six individuals ("raters") instructed to use tables in ISO 9920.
Five out of six raters underestimated the basic insulation (Icl) of most clothing ensembles. The maximal deviation of the estimated basic thermal insulation (IclEST) from the measured basic thermal insulation (IclMEAS) was 67 % (underestimation). IclMEAS of four ensembles was underestimated by on average 15 %, one ensemble was overestimated by on average 10 % and the average IclEST of two ensembles were similar to IclMEAS. One rater systematically overestimated IclMEAS. Contrary to the other raters, this individual used the fabric of the garment as a determinant for the insulation estimate. The insulation of the ensemble with only one layer on the upper and lower body beneath a heavy insulated coverall seemed to be most difficult to estimate correctly. The smaller garments (helmet, gloves, socks and hood) gave the highest variability among the raters, coefficient of variation 23-73 %. The raters commented that the tables were badly organised and that many types of garments were lacking in the tables, e g heavy insulated coverall and boots.
In conclusion, the estimations of insulation were more accurate for the ensembles with lower insulation and with only one layer beneath the outer shell than for ensembles with higher insulation and multiple layers. The tables were considered to be difficult to use by the raters. Experience tended to improve the estimates of thermal insulation.