Equivalent contact temperature (ECT) for personal comfort assessment – analytical description and definition of comfort limits

Abstract This paper introduces the equivalent contact temperature (ECT) model for local thermal comfort assessment in contact areas for non-uniform environmental conditions. It aims to complete the comfort evaluation scheme of the equivalent temperature approach included in ISO 14505-2 by the contact areas back and buttocks that are currently neglected in the standard. For the assessment of local and overall thermal comfort of seated persons, these contact areas are of great importance, especially if exposed to personal comfort systems. Person-oriented climatization systems, such as seat heating and ventilation, are much more energy efficient than conventional HVAC systems and allow to incorporate the human individual into the system’s control loop. The ECT-approach is formally defined, analytically as well as mathematically derived and validated by a subject study. The results of the subject study (air temperature of 26 °C and 29 °C) confirm the cooling effect due to the seat ventilation and show fundamental correlations between ECTs and body part specific mean thermal votes for buttocks and back. Practitioner summary:The equivalent contact temperature model for local thermal comfort assessment in contact areas for non-uniform environmental conditions is formally defined, analytically as well as mathematically derived and validated by a subject study. It completes the existing equivalent temperature comfort scheme by both contact areas back a nd buttocks to improve thermal comfort assessment.


Introduction
Energy efficiency in the indoor environment requires means to efficiently supply heating or cooling energy, means to reduce the demand, for example by enhancing the energy performance quality of the enclosure, or alternative strategies, such as personal comfort systems (PCS) (Zhang, Arens, and Zhai 2015;Warthmann et al. 2018).PCS may lead to accepting a wider range of ambient temperatures in both directions and help to better meet users' expectations of the indoor environment, especially in terms of increased thermal comfort (Ferreira and Tribess 2009;Gameiro da Silva 2002).Energy-efficient approaches are especially local ventilation systems as experimentally applied by Zhai et al. (2013), Pasut et al. (2014Pasut et al. ( , 2015)).Multiple studies, such as (Brooks and Parsons 1999;Lustbader 2005;Zhang et al. 2007;Schmidt et al. 2015), showed that heating or cooling seats allow an extended ambient temperature range evaluated as thermally acceptable.Studies on the energy efficiency of PCS (Hoyt et al. 2005;Hoyt, Arens, and Zhang 2015;Zhang et al., 2010;Zhang, Arens, and Zhai 2015;Schiavon 2009) quantify possible energy savings for air-conditioning due to an extended room temperature range, but the net energy savings depend on the region, building, applied PCS as well as the respective heating and cooling set points.Comfort-related control of seat heating or cooling requires local comfort assessment to energy efficiently ensure thermal comfort.
The applicable thermal comfort models for seated persons are bound to the body parts considered since the prevailing heat fluxes differ between body parts in contact with a surface and those not in contact with a surface.For seated persons, the back and buttocks represent the contact areas.The remaining body surface with body parts, such as feet, legs, arms, chest, face and scalp, can be assumed to be predominantly not in contact with surfaces and are grouped together as non-contact area.For the contact area, the current study presents a newly developed equivalent contact temperature (ECT) to allow local thermal comfort assessment.It extends the equivalent temperature approach of DIN EN ISO 14505-2 (DIN EN ISO 14505-2 2007) which is based on the work of Nilsson et al. (1997), Nilsson (Nilsson 2004).The existing equivalent temperature approach for local and global thermal comfort assessment only considers the non-contact area.Wyon et al. (1989) applied the equivalent homogeneous temperature (EHT) approach to the contact area.The equivalent temperature approach of Nilsson (Nilsson 2004), derived from the non-contact condition, was also applied to the contact area.Both concluded that further research and adjustments are needed for the contact area.Existing studies (Gagge 1940;Gagge, Stolwijk, and Nishi 1972;Gagge, Fobelets, and Berglund 1986;Gonzalez, Nishi, and Gagge 1974;Kohri and Mochida 2003) on approaches similar to the equivalent temperature do also not explicitly consider the heat transfer processes at contact areas.
The contact area of a seated person includes the body compartments back and buttocks which together account for up to a quarter of the body area and accordingly the contact area is very important for the thermal comfort of individuals (Schmidt, W€ olki, and van Treeck 2018;Fojtl� ın et al. 2018) and cannot be neglected.Whole body thermal comfort fundamentally depends on the human heat exchange with the thermal environment described by the well known influences air temperature, mean radiant temperature, air humidity, air velocity as well as clothing and activity of the individual (Fanger 1973;Gagge, Stolwijk, and Hardy 1967;Simion, Socaciu, and Unguresan 2016).

Thermal comfort assessment
Especially in non-uniform thermal environments, the thermal influences on the body parts differ and each body part should be evaluated individually in regard to local thermal comfort.Taking into account the different sensitivities of these body parts and the physiological state of the body, global thermal comfort can be derived from this picture (Zhang et al., 2010;Zhang et al., 2010).The predominant heat transfer processes differ between body parts in contact with surfaces or not in contact with surfaces.Thermal comfort assessment is further distinguishable into models considering stationary or transient as well as uniform or non-uniform environmental conditions.The environment in vehicles is transient as well as non-uniform, but only a few thermal comfort models consider both non-uniform and transient environmental conditions (Croitoru et al. 2015;Schmidt 2016;Zhang 2003).Although Hintea et al. (2014) show that predictions of current models deviate in real driving environments from the real thermal sensation, they recommend the equivalent temperature approach.This approach could be improved if the contact area is also assessed.New approaches, such as of Kim et al. (2018), also apply machine-learning methods in newly developed personalised thermal comfort models.

Current study
This work introduces a new calculation scheme based on an analytical and mathematical description of the thermodynamic processes in the contact area, which is validated by subject studies.Results of these subject studies enhance the knowledge and data of local thermal sensation in the contact area.Combined with the existing equivalent temperature approach, this enables the thermal comfort prediction for every body compartment and thus whole body thermal comfort assessment is improved.We close the gap in the approach by adding the equivalent contact temperature model for the remaining 18-25% (Schmidt, W€ olki, and van Treeck 2018;Fojtl � ın et al. 2018) of the body surface.

Development of the equivalent contact temperature model
The thermodynamic processes within the contact area differ from the thermodynamic processes of the noncontact area of a seated individual.In the non-contact area (DIN EN ISO 14505-2 2007) the heat conduction is neglected, while in the contact area it is the decisive heat flux.At high temperatures with active seat ventilation, the heat fluxes due to evaporation and ventilation must also be taken into account.In the case of active seat heating, there is also an incoming heat flux that depends on the power and efficiency of the seat heating.
The complex real thermodynamic system in regard to the energy and mass fluxes in the contact region is described in Subsection 2.1.In Subsection 2.2, the full thermodynamic model is reduced to a simpler thermodynamic environment with less parameters to enable accurate calculations and thermal comfort assessments in laboratory as well as real vehicle environments in terms of applicability.
Focus of the current work is the contact area between back and backrest as well as between buttocks and cushion, which together is the largest steady contact area of a seated person.Further contact areas of drivers are the palms in contact to the steering wheel as well as the forearms and upper arms in occasional contact to the armrest and seat, respectively.Since latter contact areas are comparatively small and usually unsteady, these are not considered in this work, but the approach is also applicable to these contact areas if the contact is of steady manner.The steering wheel heating has a significant influence in cold conditions (Schmidt et al. 2015), though.

Full physics model of the real environment
The full physics in the real environment is described by a lumped parameter system with six nodes (Figure 1): indoor environment env, outer-surface side of the seat seat, o, seating-surface side of the seat seat, i, the contact area between clothing and seat surface nod, the seat-surface side of the clothes clo and the skin surface sk: The mass and energy fluxes between each node are visualised.It is distinguished between sensible (red arrows) and latent heat fluxes (blue arrows).In this model, all calculations are performed based on the heat fluxes.This is advantageous for the applicability and comparability of the values independent of the respective surface areas.A calculation of the total individual heat flux is possible taking into account the respective body surfaces.
Heat transfer is balanced at the contact area nod according to Equation 1 to consider the conservation laws of energy in equilibrium.On the left side of the balance are the incoming heat fluxes into the contact area with the incoming heat flux due to the mass flux of the seat ventilation q air,in , the heat flux due to the conduction from the skin through the clothing q clo,f (index 'f' if the parameter is related to the full model), the heat flux due to the water vapour flux from the skin to the contact area q evp and the possible heat flux due to the active seat heating q heat .On the other side of the balance are the outgoing heat fluxes with the heat flux leaving the contact area due to the mass flux of the seat ventilation q air,out , the heat flux due to the conduction from the contact area to the environment q seat,f (index 'f' if the parameter is related to the full model) and the heat flux due to the water vapour flux from the contact area to the environment q vap,seat .
Since the current model is specified for the contact area, air layers between seat and clothing as well as between clothing and skin are compressed and the thermal resistance of the air layers is assumed to be negligible.Described directions of heat and mass flows are based on the assumed boundary conditions of the full physics in regard to temperature and humidity in the real environment, which are subsequently enumerated: � skin temperature T sk � contact temperature T nod � environmental temperature T env � specific humidity at skin surface x sk � specific humidity at contact area x nod � specific humidity of the environment x env

Moisture-related heat fluxes in the real environment
In the driving scenarios considered, it is assumed that the moisture resistance of the clothing worn is low.
Due to the sweat production of the human body in warm thermal environments, moisture increases in the contact area.It is assumed that all sweat evaporates and diffuses through the clothing.It is assumed that the latent heat flux due to the mass flow of water vapour _ m evp depends on the mass flow of dry air _ m da and on the moisture difference between skin x sk , contact area x nod and environment x env : The conservation laws of mass are met according to Equation 2 in equilibrium state.
Latent heat, resulting from the moisture flow due to the thermoregulatory sweating of the human body, flows into the contact area q evp (Equation 3) together with a change of the water temperature.The evaporative heat flux results from the product of the enthalpy of the water vapour on the skin surface and the mass flow of the water vapour _ m evp related to the contact area A cnt : This is also based on the specific heat required for the vaporisation of water r 0 and the specific heat capacity of water vapour c p, vap : With the diffusive moisture flow from the warm skin surface over the contact area to the colder environment, the water temperature decreases and the enthalpy of the moisture flow decreases as well.These enthalpy changes are visualised with dotted red arrows corresponding to sensible parts of the heat flow Dq evp, c , Dq evp, n , Dq evp, si and Dq evp, so : These correspond to sensible heat fluxes into clothing, contact area and seat due to the moisture flow according to Equations 4-7, respectively.These heat fluxes additionally depend on the clothing temperature T clo , the temperature of the seating-surface side of the seat T seat, i , the temperature of the outer-surface side of the seat T seat, o and the water vapour passing through the seat _ m seat : It is assumed that the water vapour neither condenses nor evaporates during its flow through clothing, contact area and seat.
The latent heat flux from the contact area to the environment depends on the water permeation through the seat q vap, seat as well as on the ventilation air flow.The thermal effect of latter equals simplified the difference between outgoing q air, out and incoming air flow q air, in : The difference between outgoing and incoming air flow under consideration of permeation through the seat corresponds to the change of specific enthalpy due to the change of temperature Dq air, s as well as humidity Dq air, l and is described in Equations 8 and 9, respectively.It is also based on the specific heat capacity of dry air c p, da : It is assumed that the humidity of the incoming air equals the environmental humidity x env and the humidity of the outgoing air equals the humidity of the contact area x nod :

Temperature-related heat fluxes in the real environment
The sensible heat flux into the contact area by thermal conduction from skin through clothing q clo, f depends on the temperature difference as well as on the thermal resistance of the clothing R clo (Equation 10).The thermal resistance of the clothing should be specified for the respective contact area since the use of a global thermal insulation for local comfort assessment would be based on the assumption of a uniform thermal insulation over the whole body.However, the local resistance value of clothing is probably lower in the contact area than in the non-contact area since clothing is tight fitting and compressed.
Due to seat ventilation, heat transfer phenomena and possibly convective heat transfer between contact area nod and clothing clo occur depending on the width of the contact area and is considered by h clo according to Equation 11.Equivalent to this, heat transfer phenomena between contact area and inner surface of the seat seat, i are described with h seat, i in Equation 12.
From the seating-surface side to the outer-surface side of the seat, the sensible heat flux q seat, f depends on the temperature difference as well as on the seat characteristics R seat according to Equation 13.The sensible heat transfer phenomena from seat to the environment are described with h seat, o and result from the temperature change due to the permeation of water from contact area to the environment as well as conductive heat transfer (Equation 14).
Heat input of an active seat heater q heat is assumed to enter the system at the seating-surface side of the seat.Its value depends on the electrical efficiency g and electrical power Q el � A tot of the seat heater as well as on the ratio of the heated contact area A cnt to the total surface A tot (Equation 15).

Reduced physics model of the real environment
For a reasonable application, the full physics of the real thermal environment, described with six nodes, is reduced to a three node-model in which each heat flux is connected between contact area nod and environment env or skin surface sk (Figure 2).The heat transfer processes are thus simplified by aggregation of the heat fluxes between theses nodes.The temperature and water vapour density gradient close to the clothing and seat surface as well as the permeation of water vapour into the seat body are not explicitly considered and are assumed to be implicitly included in the other heat fluxes.

Conservation laws of energy and mass
The thermodynamic rules as well as the conservation laws of energy and mass still hold.The energy balance is slightly modified compared to the full physics model.The energy and mass balances are described in Equations 16 and 2, respectively.

Heat fluxes in reduced heat transfer model
The basic equations in the reduced model are equivalent to the full physics model since these are built on the same thermodynamic environment, but some simplifications are conducted.Conductive heat transfers through clothing and seat are calculated according to Equations 17 and 18, respectively.In the reduced model, the thermal resistances of clothing R clo and seat R seat differ from the full physics model since the heat transfer phenomena between clothing or seat surface and contact area are included.The heat fluxes further depend on the temperature differences between skin T sk , contact area T nod and environment T env : The evaporative heat flux due to the evaporated sweat on the skin surface is still expressed by Equation 3. The parameters in this equation are equal to the full physics model, but the sensible parts of the heat transfer during its path to contact area and environment are not explicitly considered.The total mass air flow is equivalent between full physics and reduced physics models, but in the reduced physics model, whole water vapour flux from contact area to the environment is expressed by the seat ventilation and the permeation of water vapour into the seat is implicitly included (Equations 8 and 9).The heat flux by seat heating q heat is assumed to directly enter the contact area nod (Equation 15).

Temperature of the contact area in the real environment
In the reduced heat transfer model, the contact temperature T nod is calculated by coupling the balances according to Equations 16 and 2 with the heat transfer equations (Equations 3,8,9,15,17 and 18).Solving the coupled equations leads to the calculation of the contact temperature according to Equation 19 under application of the additional parameters c sk according to Equation 20, c env according to Equation 21 and c nod according to Equation 22.

Parameters to be determined
The simplified model of the contact area reduces the necessary parameters for ECT calculation to A cnt , T sk , T env , _ m da , x nod , x env , R clo , R seat , Q el and g: If T nod is measured and ECT is calculated according to Equation 26(environmental-side energy balance), the necessary parameters for ECT calculation are reduced to A cnt , T sk , T env , _ m da , x nod , x env , R seat , Q el and g: If T nod is measured and ECT is calculated according to Equation 25(skin-surface-side energy balance), the necessary parameters for ECT calculation are reduced to A cnt , T sk , _ m da , x nod , x env , and R clo : Note, as T nod is a virtual variable, it does not necessarily exactly match the measured temperature of the contact area.The physical constants must be accurately defined, too.

Comfort model and formal definition of the equivalent contact temperature (ECT)
The ECT model is developed for the contact area between the human body and seat surfaces such as backrest or seat cushion in the vehicle environment under consideration of seat ventilation and heating to ensure thermal comfort.

Definition of the ECT
The equivalent contact temperature (ECT) is the uniform temperature of an imaginary contact surface, at room air speed close to zero, at which a person will exchange the same amount of dry heat through thermal conduction for a virtual clothing insulation R calib as in the actual non-uniform environment, where the person experiences sensible and latent heat transfer at the considered body parts.

Real environment
The real environment is defined by the thermodynamics with subsequent predominant heat transfer processes in the contact area: heat conduction as well as dissipation of moisture from the contact area to the environment.In warm environmental conditions, sweat production and its evaporation leads to an additional cooling effect for the body compartments and the whole body.It is assumed that the real thermodynamics are non-uniform and the heat transfer differs between each body part.

Equivalent environment
According to the ECT approach, the real, non-uniform environment is virtually transferred into a uniform, equivalent environment with equal total thermal sensation and total heat transfer within each contact area as in the real environment (Figure 3).In the equivalent environment, the air velocity as well as the difference of the humidity are neglected and thermal seat conditioning is disabled.By definition, whole heat transfer in the contact area occurs due to heat conduction only.This heat transfer equals the heat transfer in the real environment due to heat conduction, moisture flow and thermal seat conditioning.

Calculation of the ECT
ECT is calculated by Equation 24 depending on the skin temperature T sk , the fixed calibrated insulation value R calib and the heat fluxes in the real environment, derived in Subsection 2.1 and Subsection 2.2.The equivalent environment and its corresponding heat flux q ECT are visualised in Figure 4.The equivalent heat flux equals the total heat released from the human body in the real environment (Equation 23).For ECT-calculation, the total heat flux from the human body can be captured between skin sk and  contact area nod (skin-surface-side energy balance), leading to Equation 25 or between contact area nod and environment env (environmental-side energy balance), leading to Equation 26.The contact temperature in the real environment T nod is a calculated value by Equation 19, which could also be measured as explained in subsection 2.2.Due to uncertainties in the application in practice, such as in regard to the exact value of the local clothing insulation, the measured and calculated value can differ from each other.For the same reason, in favour of robust results, ECT is calculated from the environmental-side energy balance according to Equation 26.
q air, out À q air, in À q heat (23)

Input parameters
The thermal comfort assessment strongly depends on the measurement data quality, but in real driving conditions, simplifications and assumptions are usually necessary later.Accurate calculation requires precise determination of subsequent parameters.These are assumed or determined in this work to the mentioned values in brackets: � thermal resistance of clothing including contact resistance between skin and clothing R clo (only necessary for calculation of T nod or if ECT is calculated from skin-surface-side energy balance); � thermal resistance of the seat depending on material characteristics R seat (0:31 m 2 K W À 1 in this work); � specific heat needed for vaporisation of water r 0 (2256 kJ kg À 1 ); � specific heat capacity of water vapour c p, vap (2.08 kJ kg À 1 K À 1 ); � specific heat capacity of dry air c p, da (1.005 kJ kg À 1 K À 1 ); � contact area A cnt depending on body compartment (buttocks: 0:15 m 2 , back: 0:18 m 2 according to the contact areas of a standard manikin in this work); � fixed calibrated insulation value R calib of the equivalent environment (buttocks: 0:018 m 2 K W À 1 , back: 0:035 m 2 K W À 1 ) The determination of the values in this work is further described in section 3. The specified calibrated insulation value is a fixed, constant value intended to represent a light clothing insulation between the virtual contact surface and the body surface.This is to make the results from different ECT calculations comparable.Note that the contact area needs to be adjusted occasionally.In the literature, partly larger values are reported for the buttocks and smaller ones for the back (Fojtl� ın et al. 2018) due to the individual subject weight.However, for a computational manikin (W€ olki 2017) this value is well reproducible.

Experiments for the thermal comfort assessment of the contact area
Experiments for evaluation of local as well as global thermal comfort and sensation were conducted in vehicle-mock-ups with regular vehicle seats of a midsize car (bucket seat, perforated, leather coated surface) in an air-conditioned test chamber at E3D, RWTH Aachen University, Germany.The thermal resistance of the seat is assumed to 0:31 m 2 K W À 1 : The thermal resistance of the seat was estimated based on the material properties of the seat under consideration of the pressure due to the weight of the person.35 persons (18 females and 17 males) took part at the test sessions between September and November 2017 with the average morphological data summarised in Table 1.The contact areas between back and backrest as well as between buttocks and cushion are determined with a standard manikin and are assumed to equal 0.18 m 2 and 0.15 m 2 , respectively.The contact area depends on the seat characteristics as well as on the person and the body part considered.

Measurement equipment and experimental setup
The test facility comprises an air conditioned test chamber with three identical one-person vehiclemock-ups.To capture the real thermodynamic environment, the test chamber as well as the vehicle seats are equipped with extensive sensor hardware.The mass flow and the thermal characteristics of the suction seat ventilation of the cushion as well as backrest are measured by in-house developed multi-chamber mass flow rate measurement devices (MFRMDs) which are described by Schmidt, W€ olki, and van Treeck (2018)

Experimental conditions
The summer case is investigated at an ambient temperature of 26 � C and 29 � C. The latter is sufficient to cause warm discomfort and to examine the effect of sweating on individual's thermal comfort.Each subject was asked to wear pants, a t-shirt, underwear, light socks and light shoes.The actual clothing worn was checked before the start of each series of experiments, and the clothing was not changed during the experiments.This clothing combination corresponds to 0:5 clo (0:0775 m 2 K W À 1 ) according to DIN EN ISO 9920 (DIN EN ISO 9920 , 2009).Local thermal resistance of the clothing in the contact region is not measured while seating, but it can be assumed to be lower than the determined thermal resistance at the non-  contact region due to occurring compression of the clothes.The air velocity during an average trial is lower than 0:07m s À 1 for each scenario with the average ambient conditions as summarised in Table 2.The considered scenarios differ in the operative target temperature and seat ventilation level which is characterised by the prevailing mass flow rate at the backrest and cushion.

Experimental design
The different phases of the experiments that were conducted to develop the 'ECT -mean thermal vote (MTV) correlation scheme' illustrated in Figure 11 are depicted in Figure 6.An entire test cycle was designed for a period of three hours.Besides this, an additional hour was considered for the pre-and post-preparation phase of the experiment.The considered seat fan speeds are adapted to the thermal ambient conditions of the different scenarios (the 26 � C-and the warmer 29 � C-scenario).The fan speed is set to 0%, 35% or 65% of the maximum power for the 26 � C-scenarios and to 0%, 65% or 100% of the maximum power for the 29 � C-scenarios.The sequence of the different seat fan speeds to which each person was exposed was chosen randomly to avoid possible learning effects of the subjects.In each of the scenarios, three persons were tested in parallel.Due to the use of a within-test design, each subject had to undergo each of the six scenarios.In all test cycles, the 0% seat fan speed (seat fan off) served as the reference scenario for each person.
Each day was divided into two stages.The first stage started at 8 a.m. and finished at 11 a.m., the second stage started at 1 p.m. and finished at 4 p.m.As shown by Figure 6, a single test cycle consisted of seven phases.During the initial phase, the subjects arrived at the test facility and were given the chance to calm down and focus on the experiment.It was followed by a 25 min acclimatisation phase during which they had to answer a first questionnaire that contained questions concerning the subjects' health state and their local and overall thermal comfort.Phases two to seven were alternated between a 25 min test phase and a subsequent acclimatisation phase of the same length at an ambient temperature of 22 � C. The latter was included to have a clear initial baseline for  the human thermoregulatory system for each test phase.

Surveys
The comfort surveys contained questions on the local and overall thermal sensation and comfort as well as on the humidity sensation and satisfaction.The surveys were conducted during the initial 5 min and in the last five minutes of each test phase to capture transient effects as well.Surveys on thermal sensation (TS) of each body compartment and of the whole body are answered according to the 5-point-Bedfordscale with answers between À 2 and 2 with assigned meanings in Table 3.This scale was chosen to be compliant with the original equivalent temperature approach which is described by Nilsson (Nilsson 2004) and included in DIN EN ISO 14505-2 (DIN EN ISO 14505-2 2007).

Results and validation
The subject study is evaluated with focus on the development of the ECT at back and buttocks for each test condition and is related to thermal comfort and sensation in the contact region.The average ECTcurves of the scenarios with different set-points and seat-ventilation-states are clearly separated from each other with a lower ECT at scenarios with active seat ventilation (Figure 7).The effect of the air flow, which lowers the ECT, is greater at the back than at the buttocks due to the higher effective mass flow rates (Table 2).During a test cycle, the ECT-curves usually rise and are flatter at the end of the trials.

Moisture removal
The expected moisture removal by seat ventilation is confirmed and is visualised in Figure 8 for the contact area buttocks-cushion.The specific humidity of the skin surface is constantly higher than the specific   humidity values of the corresponding clothing surface.
The specific humidity at both skin and contact region increases during each trial.The specific humidity of the skin surface increases more strongly at the 29 � C-scenarios than at the 26 � C-scenarios (end of trial in contrast to begin of trial).With higher seat fan level, the rises of the specific humidity curves at skin surface and contact region are reduced and flatten at the end of the trials if seat ventilation is active.At the S26-1 and S26-2-scenarios, the seat ventilation is sufficient to remove the additional humidity from skin surface as the specific humidity remains almost constant.At the 29 � C-scenarios, both active seat fan levels are nearly equally effective at moisture removal as the specific humidity curve at skin and contact area are almost equivalent between S29-2 and S29-3.

Mean thermal votes (MTVs) for the contact areas
The subjects evaluated the thermal sensation and comfort of the back and buttocks between À 2 and 2 with less than 7% of the answers equal to or lower than À 1: The buttocks and back are each predominantly (82% of the subjects) evaluated between 0 and 1 corresponding to 'neutral' or 'warm, but comfortable'.The distribution of the answers is visualised by boxplots for each test scenario in Figure 9 with the median as orange line and the arithmetic mean as green triangle.The arithmetic mean and standard deviation of the thermal sensation for each scenario are summarised in Table 4.If seat ventilation is not active, at both scenarios, the contact areas are in average (median) evaluated as 'warm, but comfortable'.Seat ventilation always decreases the MTV to a more neutral thermal sensation at the investigated scenarios.For example, with active seat ventilation, the MTV corresponds to 'neutral' at the 26 � C-scenarios.
Except the S29-2-scenario, the resulting MTVs of the back are always closer to neutral than of the buttocks, which could indicate a higher thermal sensitivity of the buttocks-area.
At the 29 � C-scenarios, the MTVs are evaluated to be warmer than the 26 � C-scenarios.With inactive seat ventilation or seat ventilation at 65%, the MTVs correspond to 'warm, but comfortable' for both buttocks and back at the S29-scenarios (arithmetic mean higher than 0:5).With higher seat ventilation level, the MTV tends to decrease towards 'neutral', but the average remains between 'neutral' and 'warm, but comfortable' for each scenario.

Correlation between ECT and MTV
For general significance, the answers of the individuals are grouped by each scenario (S26-0, S26-1, S26-2, S29-0, S29-2, S29-3) and the MTV as well as the average ECT are calculated, respectively.Based on these MTV-ECT-pairs, a linear regression under application of the linear least-squares method is conducted.The MTVs of the scenarios are visualised together with the regression curve over the ECTs in Figure 10.
The regression curve shows the correlation between MTV and TS.Since the MTV of each scenario is between 'neutral' and 'warm, but comfortable', further data for the evaluation of the dependency outside these thermal scenarios are necessary.The quality of the correlation between the TS and ECT is shown by the coefficient of determination.While it is very high for the contact area buttocks -cushion (R 2 ¼ 0:93), Table 4. Statistical description of the TS distribution for each scenario.
Decisive ECT-MTV-thresholds are calculated with the linear regression at TS-value 60:5 for each TS and the corresponding comfort scheme is visualised in Figure 11.It shows the correlation between ECT and MTV for buttocks and back in compliance with the scheme of DIN EN ISO 14505-2 (DIN EN ISO 14505-2 2007).Each body compartment is described on the yaxis with the ECT-values on the x-axis.The thresholds for each MTV are connected between each body compartment for each thermal sensation.This comfort scheme results from experiments in summer conditions at 26 � C and 29 � C with a clothing insulation corresponding to 0:5 clo for seated individuals on vehicle seats with seat ventilation.The different comfort categories according to the 5-point-Bedford-scale are visualised in the scheme by the different shaded areas.The comfort categories are assigned to the respective values in Table 3.

Discussion
The results of the subject study support the mathematical and analytical description of the ECT.According to Figure 7, the ECT strongly depends on the ambient temperature as well as on the seat ventilation at the contact interface between seat and human body.In warm thermal environments, the ECT and the specific humidity at skin surface (Figure 8) are highest if seat ventilation is disabled.The experiments show that seat ventilation is effective in hot environments.Latent heat transfer by perspiration significantly contributes to the ventilation effectiveness and heat release by evaporation is an important influence on individual's TS.

Scenarios of the current subject study
During the subject studies, ambient temperatures between 26 � C and 29 � C with a mass flow rate of the seat ventilation of 0g s À 1 to 1:5g s À 1 (backrest) and 0g s À 1 to 0:7g s À 1 (cushion) were investigated (Table 2).The calculated ECT range is 31:5 � C to 36 � C and the ECTs are visualised as scatter plot together with the corresponding TS vote in Figure 12.Seat ventilation has a decisive cooling effect and lowers the ECT accordingly.Due to this, the ECTs of the 29 � C-scenarios with active seat ventilation are mixed with the 26 � C-scenarios and high ECTs usually correspond to scenarios without seat ventilation.Most thermal-comfort-responses correspond to 'neutral' or 'warm, but comfortable'.The resulting evaluations are correspondingly close together and do not reflect very warm or cool thermal environments.The thresholds of the newly created comfort scheme (Figure 11) are partly outside the investigated ECT range.This highlights the importance of further studies, in colder and warmer thermal environments as well, to support or adjust the thresholds.Furthermore, within this study standard vehicle seats with three different seat fan levels were used.The results are limited to these scenarios and further experiments with different seat shapes, seat materials and performance levels of local cooling as well as heating are necessary.

Insulation values in the contact region
The ECT strongly depends on the clothing and seat insulation, but at least the local clothing insulation is difficult to determine in real driving situations.Most available data of clothing resistance values, such as in the database of de Dear (de Dear 1998), only characterise the average clothing resistance of the whole body.Furthermore, these available clothing resistance values differ in considering the thermal resistance of a seat (Rupp, Kazanci, and Toftum 2021).Since the clothing resistance is non-uniform over the whole body, application of an average insulation value is inaccurate for local thermal comfort assessment.The local thermal resistance of the clothing is usually measured with manikins in standing position, but it can be assumed to be lower in the contact region than the determined thermal resistance at non-contact region due to occurring compression of the clothes.
The application of thermal manikins for the measurement of clothing characteristics is further described by McCullough (McCullough 2005).The clothing insulation in seated position is also determinable using thermal manikins, but the individual is thereby not exactly represented.The insulation value in the contact region strongly depends on the posture, weight and correspondingly the compression of the clothing layers.The structure of manikins is furthermore more rigid than human bodies.This all together leads to uncertainties in the measured insulation value, which is especially poorly determinable in non-test environments for the individual.Due to such uncertainties, for example Fojtl� ın et al. ( 2018) evaluate the results of thermal manikins on clothing insulation and contact area as not representative of seated subjects.
The real thermal resistance affects the heat flow in the contact region, and hereto, the body weight distribution should be considered.Since the body weight on the seat cushion is higher than on the backrest, the compression of the fibres at buttocks is assumed to be higher and thermal resistance accordingly lower.Recent research (Nomoto et al. 2020;Tang et al. 2022;Vesel� a, Psikuta, and Frijns 2018;Lu et al. 2015;Lee, Zhang, and Arens 2013;Nelson et al. 2005) further investigate local clothing insulation with thermal manikins and their results expand the knowledge about local thermal insulation of clothing for each body part.However, the resulting local clothing insulation values differ between studies probably due to different applied manikins, posture, material, weight and fit of the clothing as well as the applied air velocity.Most studies apply manikins in standing positions, but for example, Lee, Zhang, and Arens (2013) also apply a manikin seated on a chair.The eighth clothing ensemble of their study is similar to the clothing of the subjects in our study, wearing a t-shirt, underwear, pants, light socks and shoes, with a total clothing insulation corresponding to 0:5 clo: Altogether, further research into local clothing insulation, especially for the contact area, is essential to improve local comfort models.Furthermore, the clothing of contact regions occasionally contains moisture, which lowers the resistance.The results of this work are therefore based on the environmental-side of the energy balance according to Equation 26.The calculated results are more independent of the exact clothing insulation value and its impact is included in the measured values.The precise determination of the local R clo could further improve the ECT-calculation.Experiments with different clothing are necessary, in particular to evaluate thermal comfort in the winter.

Comparison of contact interfaces backrest and cushion
The initial ECTs at backrest are higher than at cushion which can be partly attributed to the different orientation (horizontal vs. vertical).During the experiments, the mass flow rate of the seat ventilation is much higher at the back than at the buttocks (Table 2).Accordingly, the heat removal from back is higher than from buttocks, which also affects the ECT (Figure 7).The lower mass flow rate can be lead back to compressed air ducts of the seat ventilation in the cushion due to the weight of the seated person and the predominantly fixed contact to the cushion.Latter also leads to a higher specific humidity at the cushion than at the backrest.In particular, the contact to the backrest strongly depends on the person's posture due to its increased freedom of movement.Body surfaces with temporary contact to surfaces are currently not taken into account and would have to be considered separately with regard to the decisive thermal processes.A closer look on the contact between back and backrest especially in real driving situations could be a focus of another study.

Thermoregulatory activity at scenarios
Most of the answers of the subjects are between 'neutral' and 'warm, but comfortable' (Figure 9) and therefore, the investigated scenarios best represent neutral to warm environments.The thermal environment is also evaluated to be 'too warm' in single cases and the specific humidity of the contact interface as well as skin surface is increased at the 29 � C-scenarios (Figure 8).This confirms that the investigated environment is also suitable to assess warm environments in which human's thermoregulation actively regulates its heat balance by sweating.

Effect of seat ventilation
Seat ventilation is necessary to ensure steady states for the specific humidity by transporting moisture from skin through contact interface to the environment.The evaporation of sweat and its removal by seat ventilation leads to a temperature drop and lowers the ECT.This also confirms the assumptions of the analytical model and is a realistic model behaviour.Since the cooling effect of the seat ventilation is assumed to result from moisture removal, the cooling effect and its impact on the ECT is bound to moisture permeable clothing and the moisture increase in the contact area.It is assumed that the moisture increase in the contact area is based on the moisture flow from the skin to the contact area.

Calibrated insulation value
The calibrated insulation values for the contact interfaces back and backrest (0:035 m 2 K W À 1 ) as well as buttocks and cushion (0:018 m 2 K W À 1 ) are set to represent a light clothing insulation between the skin and the virtual node.This value may vary between body compartments due to differences in pressure and clothing in each contact area.This value is intended to make the model comparable and applicable even if the actual clothing combinations do not perfectly match the tested scenarios.Since the results represent only a single data set, it is not possible to verify this approach.Further experiments with different clothing combinations and seat characteristics are needed to verify this approach or to adjust these values to more appropriate values.

Accordance with the existing equivalent temperature approach
The subject study allows the description of an evaluation scheme of the thermal sensation depending on the ECT for the body compartments back and buttocks (Figure 11) which are in contact with the seat.Due to an adjusted approach contrary to the existing approach, the ECT-values are not directly comparable with the equivalent temperatures (T eq ) according to DIN EN ISO 14505-2 (DIN EN ISO 14505-2 2007) and Nilsson (Nilsson 2004), but it extends the existing equivalent temperature approach and allows the comfort assessment of body compartments not addressed by T eq .The approach and representation of the resulting ECT comfort scheme matches with the existing T eq schemes and is applied equivalently.

Conclusion
This ECT-approach completes the existing gap of the T eq evaluation scheme of Nilsson (Nilsson 2004), DIN EN ISO 14505-2 (DIN EN ISO 14505-2 2007) by the thermal comfort evaluation of contact regions such as 'buttocks -cushion' and 'back -backrest'.It strongly depends on heat conduction as well as on moisture removal by seat ventilation, which both are not considered by the existing T eq -approach.It further integrates the consideration of seat heating if its efficiency is known.
The ECT is defined, analytically as well as mathematically derived and the consistency of the theoretical description is validated by a subject study using extensive sensor hardware.Thermal comfort at the contact interfaces between a person and a seat was investigated on the basis of MTVs and latter were correlated with calculated ECTs.
The resulting thermal comfort scheme for the ECT follows the modelling approach of T eq and therefore, it can be easily integrated into the existing scheme of DIN EN ISO 14505-2 (DIN EN ISO 14505-2 2007).The specified comfort intervals of this work are valid for summer conditions at 26 � C and 29 � C for a clothing insulation of 0:5 clo: It is based on fixed calibrated insulation values R calib,buttocks of 0:018 m 2 K W À 1 and R calib,back of 0:035 m 2 K W À 1 : It allows the integration of seat ventilation as well as seat heating.
Further experiments with other clothing insulation, seat materials and seat shapes are important to make the resulting comfort scheme more generally applicable and to further enhance the knowledge about local and global thermal comfort.

Figure 1 .
Figure 1.Full physics of the contact area in the real environment (sensible heat fluxes: red, latent heat fluxes: blue).

Figure 2 .
Figure 2. Reduced heat transfer model of the contact area in real environment (sensible heat fluxes: red, latent heat fluxes: blue).

Figure 3 .
Figure 3. Description of real (left) and equivalent (right) thermal environment.

Figure 4 .
Figure 4. Heat transfer system model in the equivalent environment (only sensible heat fluxes: red).
in detail.The MFRMDs measure the air flow velocity, humidity as well as temperature and allow the calculation of the mass flow rate.The specific humidity of the mass flow represents the specific humidity of the contact region x nod , which corresponds to the removed moisture by the air flow.The seats are equipped with BME 280 sensors (Bosch, Germany) to measure the humidity (resolution: 0.008%) and temperature (resolution: 0.01 � C) at clothing surface representing the node temperature T nod as well as PT100 (Kritec, Germany; DIN class AA) sensors to measure the seat surface temperature at four positions of each surface as visualised in Figure5.Skin surface temperature (resolution: 0.5 K) and humidity (resolution: 0.6%) are measured with eight iButtons of type DS 1923-F5 (Maxim Integrated) per person at similar sensor positions.The selected measuring positions are distributed over the respective contact areas and are usually in the area of continuous contact.The ambient conditions are measured in the centre of the room by a measurement device of the German company Ahlborn with a black globe thermometer (ZA 9030-FS2) at a height of 1.1 m, as well as a capacitive humidity sensor and temperature sensor (FH A646-E1) and a thermoanemometer (FV A605-TA1O) at a height of 0.1 m and 1.1 m.These measurement points are sufficient to assess the global thermal comfort according toDIN  EN ISO 7730 (DIN EN ISO 7730 , 2005).

Figure 6 .
Figure 6.Example of an entire test cycle for a single subject; the schematic shows the two tested summer scenarios at ambient temperatures of 26 � C and 29 � C; percentages indicate the relative fan speed, where 0% means fan is switched off and 100% represents the highest possible fan speed.

Figure 7 .
Figure 7. Curves of the average ECTs of each scenario for the contact areas.

Figure 8 .
Figure 8. Curves of the average specific humidity for the contact area buttocks -cushion.

Figure 9 .
Figure 9. Boxplots of the thermal sensation distribution for each scenario.

Figure 10 .
Figure 10.Grouped thermal sensation votes (MTVs) corresponding to each scenario related to ECT.

Figure 12 .
Figure 12.TS votes related to ECT.

Table 1 .
Average morphological data of the subjects.

Table 2 .
Experimental conditions of each test setting and corresponding ambient conditions.

Table 3 .
Description of the TS responses based on the Bedford scale.