Comparative study of thermal comfort and adaptive actions for modern and traditional multi-storey naturally ventilated hostel buildings during monsoon season in India

Journal of Building Engineering 23 (2019) 90–106

Contents lists available at ScienceDirect

Journal of Building Engineering journal homepage: www.elsevier.com/locate/jobe

Comparative study of thermal comfort and adaptive actions for modern and traditional multi-storey naturally ventilated hostel buildings during monsoon season in India

T

⁎

Sanjay Kumara, , Manoj Kumar Singhb,c,d, Rajeev Kukrejaa, Shailendra Kumar Chaurasiyaa, Varun Kumar Guptaa a

Mechanical Engineering Department, Dr. B R Ambedkar National Institute of Technology, Jalandhar, Punjab 144011, India Joint International Research Laboratory of Green Buildings and Built Environments (Ministry of Education), Chongqing University, Chongqing 400045, China National Centre for International Research of Low-carbon and Green Buildings, Chongqing University, Chongqing 400045, China d Department of Human and Social Systems, Institute of Industrial Science, The University of Tokyo, Tokyo 153-8505, Japan b c

A R T I C LE I N FO

A B S T R A C T

Keywords: Thermal comfort Hostel buildings Thermal adaptation Composite climate India

The prime objective of the hostel buildings is to provide an adequate thermal environment to the students for their good health and learning performance. However, a few studies have been reported so far in such buildings around the world. This paper reports the findings of a questionnaire based thermal comfort study in six naturally ventilated hostel buildings located in Jalandhar city, lies in composite climate of India, during the monsoon season (August-September 2018). The study was carried out in two newly constructed (aged less than 5 years) and four traditional (aged more than 25 years) multi-storey naturally ventilated (NV) hostel buildings. During the study, a total of 945 completely filled questionnaires from 470 occupants were collected. About 80% and 75% of subjects voted within central three categories of ASHRAE thermal sensation scale in traditional and modern NV hostel buildings, respectively. Results from the probit analysis revealed that 80% of subjects voted within ± 1 TSV when indoor operative temperature ranged between 27.2 and 31 °C. The mean indoor comfort temperature as calculated by Griffiths method is 29.9 °C (sd( ± σ) = 2.16). To restore comfort primary adaptive action of occupants was found to be switching on the fans followed by the opening of external doors and windows.

1. Introduction and background The environmental conditions required for thermal comfort are not the same for everyone [1,2]. Adaptation affects a subject's perception and thereby comfort temperature of subject's varies from one building to another and from one climate to another [3]. Studies conducted on people living in free-running buildings in hot and warm climate countries concluded that they are tolerant to higher temperature limits than recommended by existing comfort standards [3–5]. A study conducted by Nicol and Roaf [6] in the offices of Pakistan demonstrated high neutral temperature during the summer monitoring period. Heidari and Sharples [7] study in Iran during hot summer conditions for the long term and short term found neutral temperature 26.7 °C and 28.4 °C respectively. Tablada et al. [8] and Feriadi and Wang [9] studied subjects in the buildings of warm-humid climate reported that they felt neutral at 28.5 °C and 29 °C respectively. Nguyen et al. [10] collected

⁎

subject responses from hot and humid climates of South-East Asia and found the neutral temperature of 27.9 °C in free-running buildings. Yao et al. [11] carried out a study for naturally ventilated buildings in hot summer and cold winter zone of China and showed the discrepancies in subject responses with the heat balance model. Researchers in India conducted thermal comfort studies mostly on adult subjects in different built environments i.e. residential buildings, office buildings, university classrooms and university laboratories in different climatic zones of India. Indraganti [12] carried out thermal comfort study based on transverse and longitudinal surveys in naturally ventilated multi-storey apartments in Hyderabad and observed a neutral temperature of 29.2 °C for the studied group. Indraganti et al. [13,14] conducted another set of field studies for NV and AC office buildings of two cities Hyderabad (composite climate) and Chennai (warm and humid climate) in Southern India. The study proposed an adaptive thermal comfort model in close agreement with the CEN 7730

Corresponding author. E-mail address: [email protected] (S. Kumar).

https://doi.org/10.1016/j.jobe.2019.01.020 Received 5 October 2018; Received in revised form 11 January 2019; Accepted 11 January 2019 Available online 14 January 2019 2352-7102/ © 2019 Elsevier Ltd. All rights reserved.

Journal of Building Engineering 23 (2019) 90–106

S. Kumar et al.

different parts of India with varied thermal comfort perception and history), age difference, clothing variations and activity level of students. However, the literature review suggested that a few field studies have been conducted so far to assess the thermal environment conditions in the naturally ventilated hostel's buildings in India. Scopus database (accessed on 30th October 2018) search returned only one study carried out by Dhaka et al. [27], in multi-storey NV hostel buildings in composite climate of India. The study was conducted in Jaipur city during the summer season and for traditional buildings only. The study further concluded that hostels in composite climate of India are not appropriately designed to meet the thermal comfort requirement of student's as well as national and international standards. Moreover, thermal environment conditions in NV hostel buildings are not similar to the office or residential buildings due to differences in occupancy patterns, age group, behaviour and activities [25,27]. In India, most of the higher educational institutes provide accommodation to the students in hostels inside the campus and these hostels are constructed to operate under naturally ventilated mode. So more studies are required to be carried out in these naturally ventilated buildings to understand better the thermal environment, adaptation behaviour and comfort needs of its occupants. The present study evaluates the existing thermal environment conditions in two modern (constructed in last 5 years) and four traditional (aged more than 25 years) multi‒storey naturally ventilated hostel buildings in premises of National Institute of Technology, Jalandhar, India. Questionnaire based study is being carried out following the principle of Adaptive thermal comfort to collect the data of the thermal environment in the hostels. Statistical analysis of the collected data was carried to analyze the overall acceptability of occupants about existing thermal environment conditions during monsoon season in composite climate of India. The study further explored the different thermal adaptation behaviour exercised by occupants of hostel dormitories to restore thermal comfort.

[15] and ASHRAE standard 55-2013 [1]. Singh et al. [16,17] conducted an adaptive comfort study and performance evaluation of vernacular architecture in three bio-climatic zones of North-Eastern India. The results demonstrated seasonal and regional differences in subjective responses, preferences and in thermal neutrality. The study also demonstrated the presence of various solar passive features in the existing vernacular architecture. Recently authors have reported a field study involving the occupants in the office buildings of North East, India and proposed an adaptive thermal comfort model [18]. Dhaka et al. [19] carried out a thermal comfort study on naturally ventilated buildings in Jaipur city, lies in composite climate of India and reported a neutral temperature of 27.2 °C for all seasons. The study also proposed an adaptive model for thermal comfort. Kumar et al. [20,21] also uses the extended data from same field surveys and defines adaptive thermal comfort zones on psychrometric chart for different air velocity ranges. New boundaries of thermal comfort zones extend up to 35 °C for airspeed of 1.5 m/s. These new comfort boundaries take into consideration the thermal adaptation of occupants of composite climate. Currently, in India, more than 10 million students of different age group reside in the hostel buildings across the country [22,23]. Hostel buildings are the special type of residential buildings where the prime objective of the built environment is to provide better thermal environments to the students for their good health and learning performance. Few studies have been conducted for multi-storey NV hostel buildings in India and abroad in the last decade. Dalhan et al. [24] carried out a field study of thermal comfort in three high rise hostels in the hot and humid climate of Malaysia. The study collected 298 responses from girl students only during rainy season over a period of four weeks. The study found neutrality at 28.8 °C and also assessed the effect of operative temperature on the thermal sensation of occupants during rainy and cloudy days [25]. Lai [26] has explored the user expectation and satisfaction criteria for hostel buildings considering the gap theory and an indicative postoccupancy evaluation approach. Six main performance aspects of the hostel's facilities were identified and analyzed, namely visual comfort, thermal comfort, aural comfort, fire safety, hygiene, and communication via information technology. The study found that users were most satisfied with their visual aspect. The study done by Dhaka et al. [27] assessed the existing thermal environment conditions of traditional hostel buildings and its effect on student's perception during summer season (August–November) in Jaipur city, lies in composite climate of India. The study found the neutral temperature 30.2 °C through linear regression analysis, with comfort bandwidth of 7.9 °C (range of 25.9–33.8 °C). During the field study, the acceptable indoor air speed and relative humidity were found to be 0.41 m/s and 36%, respectively. The study also analyzed the variation of thermal neutrality based on different occupancy and gender. The neutral temperature for female students was slightly lower (29.9 °C) than male students (30.1 °C). In the last few years, Government of India initiative towards higher education enabled the educational institutions to grow at a much rapid rate compared to the past decade. This created a huge demand for infrastructure for hostel buildings to accommodate students with better shelter and facilities. Also, most of the field studies conducted in India is restricted to office buildings, residential buildings or classrooms in different climatic zones of India. So more studies are required to be carried out in these naturally ventilated buildings to understand better the thermal environment, adaptation behaviour and comfort needs of its occupants.

3. Research methodology 3.1. Investigated hostel buildings Four traditional and two modern multi‒storey NV hostel buildings located at premises of National Institute of Technology, Jalandhar (31.32°N, 75.57°E, mean sea level = +228 m) has been selected for thermal comfort field study (Fig. 1). In the present study, selection of hostel buildings (modern and traditional hostels) are based on regionspecific architecture and typical characteristic of Indian hostel buildings, representing suitable samples for conducting thermal comfort study. The external wall of hostel buildings (traditional as well as modern) are built of brick of thickness 0.20–0.23 m (U-value = 2.45 W/m2K) and roofs are constructed of poured concrete slab with a thickness of 0.15 m (U-value=3.1 W/m2K). The thermal characteristic of surveyed hostel buildings is presented in Table 1. Window assemblies in modern hostel buildings were double glazed type with 0.004–0.008 m of thickness with aluminum frames (U-value=4.7 W/ m2K), while in traditional buildings windows are single glazed clear glass of 0.003 m of thickness (U-value = 5.7 W/m2K) with steel frame having six thermal breaks. To estimate the overall heat transfer coefficient (U-value) for external wall and roofs in these buildings the authors have referred the studies conducted by Kumar et al. [28] and Kumar and Suman [29] for composite climate in India. A total of 419 rooms from all floors (~G+3) have been surveyed during the two-month field study. Surveyed rooms were mostly single or double bed accommodation with a maximum occupancy of 1–2 students. Driven by the constraint of accessibility all possible effort was made to conduct the surveys on all floors. During the field study, a total of 945 valid subjective responses (after removal of inconsistencies from samples) from 470 occupants were collected from 419 rooms. Table 2 shows the samples collected and physical characteristics of the

2. Objective of the study In India, most of the hostel buildings are naturally ventilated and there is a huge diversity in student's (Students come to study from 91

Journal of Building Engineering 23 (2019) 90–106

S. Kumar et al.

Fig. 1. Pictures of surveyed buildings (a), (b) Modern hostels and (c), (d) Traditional hostels.

Table 1 Thermal characteristics of surveyed hostel buildings. Buildings

Building envelop

Layer by layer construction

Overall heat transfer coefficient, U-value(W/m2K)

Modern hostel

Wall (External)

Plaster (20 mm) Brick burned (230 mm) Plaster (20 mm) Plaster (20 mm) Brick burned (120 mm) Plaster (20 mm) Ceramic tiles Poured concrete Plaster (20 mm) Poured concrete (150 mm) Plaster (20 mm) Double glazed with aluminum frame Plywood Plaster (20 mm) Brick burned (230 mm) Plaster (20 mm) Plaster (20 mm) Brick burned (110 mm) Plaster (20 mm) Ceramic tiles Poured concrete Plaster (20 mm) Poured concrete (150 mm) Plaster (20 mm) Single clear glass (0.003 m) with steel frames (six thermal breaks) Plywood/steel

2.45

Wall (Internal)

Floor (Ground) Roof (Above)

Traditional hostel

Windows Doors Wall (External)

Wall (Internal)

Floor (Ground) Roof (Above)

Windows Doors

92

2.01

0.65 3.14

4.7 2.02 2.45

1.98

0.65 3.14

5.7 2.02

Journal of Building Engineering 23 (2019) 90–106

S. Kumar et al.

Table 2 Details of physical characteristics and collected samples from surveyed buildings. Building Name

No of floors

Floors surveyed

Data collected (N)

Type of window

Room Area

Occupancy

Total Occupancy

(m )

Percentage of Windows operable (%)

15

0.81

100

384

1/2

15

0.81

100

359

1 1/2 1 1/2

20 20 20 20

1.31 1.31 1.31 1.31

60 60 60 60

133 146 134 186

Ceiling Fans

Window to Wall ratio (%)

1/2

2

(m )

Mega B

6

All floors

377

Operable

11.28

Mega F

5

All floors

265

Operable

11.28

H-3 H-4 H-6 H-7

3 3 3 3

All All All All

floors floors floors floors

43 53 32 175

Operable Operable Operable Operable

8.68 8.68 8.68 8.68

Single/ double bed Single/ double bed Single bed Single bed Single bed Single bed

Window Area 2

1/2 Ceiling fans: 1 or 2 fans is shared by two persons.

Fig. 2. Pictures of various environmental controls adaptively used by occupants during the field study.

subjective informations during the surveys. While subjects were filing to the questionnaire, the surveyor noted the surrounding thermal parameters, personal parameters (clothing and activity level) and environmental controls (as binary data) being used in the vicinity of the subject. The instrument setup was placed on a table close to the subjects at a height of 1.1 m as per the ASHRAE standard 55 Class-II protocols [1] (Fig. 3). During the survey, occupant' were asked not to modify their thermal indoor conditions and use of controls. Data thus collected was processed to remove outliers and inconsistent entries. The filtered data is then used to carry out the subsequent analysis in this study. All subjects residing in the hostels were found to be native of Indian sub-continent. However, students came from different provinces and states across India. In the present study, the occupants of hostel buildings were mostly undergraduate and postgraduate students and residing in the hostel for more than one year. A total of 470 students participated voluntarily in the surveying from all hostel buildings which returned a total of 1024 questionnaires completely filled. The subjects were healthy young adults and were in the age group of 16–30 years (mean= 20.6 years, sd( ± σ) = 1.6). Physical characteristics of the hostel residents such as height (m), weight (kg) along with body surface area as calculated using De Bois formula [27] are presented in Table 3.

surveyed hotels buildings under natural ventilation mode. Investigated rooms were mostly equipped with ceiling fans, operable windows, blinds and curtains to improve indoor thermal conditions (Fig. 2). This study aims to priorities these adaptive controls based on their use in improving the thermal comfort conditions. 3.2. Field survey and sample size description A field study was conducted in six NV hostel buildings in the month of August and September 2018. Each surveyed buildings was visited at least once by the surveyor in a week during the two months monitoring period. The survey was conducted three times a day i.e. morning (9:00–11:00 a.m.), afternoon (1:00–3:00 p.m.) and in the evening (5:00–7:00 p.m.). A fresh transverse questionnaire was used every time a subject was interviewed and a minimum interval of two hours was kept between any two interviews of a single subject. A group of 4 students from master and undergraduate courses assisted in carrying out the surveying. A brief introduction and discussion were provided to occupant's by surveyor regarding the objective of the study. Discussion with the occupant's also included the information about the type of data to be collected and seeking their consent. Though the occupant's were conversant with the English language but the language used by the occupant's in conversation is other than English. The questionnaire was also explained in detail in the conversation language of occupants. This briefing was necessary to minimize the human error in recording the 93

Journal of Building Engineering 23 (2019) 90–106

S. Kumar et al.

Fig. 3. Pictorial view of typical survey environments in hostel buildings and instrument setup used in the field study.

parameters. The study follows ASHRAE standard 55 Class-II protocols [1] and measurements were carried out at height of 1.1 m from the ground level. Subsequently, the responses of subjects were captured on ASHRAE's seven-point sensation and Nicol's [31] five-point preference scale, respectively (Table 4). Overall thermal acceptability for existing environmental conditions was also noted down as binary data, 1: Unacceptable, 0: Acceptable. The second part of the questionnaire includes the checklists of clothing ensemble, activity level, behavioral and environmental controls used by the subject and their preferences. It also recorded the thermal environmental parameters such as temperature, globe temperature, relative humidity, air velocity, and carbon dioxide

3.3. Subjective questionnaire, protocols and thermal scale used Transverse type and “Right Here Right Now” questionnaires were administered to the occupants. The questionnaire developed for the study was adopted from previous studies conducted in educational, residential and office buildings of composite climate in India [12,19–21,30]. The questionnaire had two parts consisting of 15 questions and each question was provided with checkboxes for their response. The first part of the questionnaire consists of personal and buildings location identifiers, responses for thermal sensation, preferences and overall acceptability for different environmental Table 3 Descriptive statistics of subjects investigated during the field study. Buildings

No. of subjects

No of rooms surveyed

All

470

419

Modern

268

222

Traditional

202

197

a

Maximum Minimum Mean sd( ± σ) Range Maximum Minimum Mean sd( ± σ) Range Maximum Minimum Mean sd( ± σ) Range

Age

Height (m)

Body surface Areaa (/m2)

Weight (kg)

Icl,tot (clo)

Activity (Met)

30 16 20.8 1.97 14 30 17 20.8 2.1 13 30 18 20.7 1.65 12

1.88 1.50 1.73 0.06 0.38 1.88 1.52 1.72 0.07 0.36 1.87 1.52 1.72 0.06 0.35

2.20 1.42 1.76 0.14 0.78 2.21 1.45 1.77 0.15 0.76 2.14 1.45 1.76 0.14 0.69

100 45 65.5 9.95 55 100 45 65.9 10.2 55 95 48 64.8 9.35 47

0.78 0.10 0.29 0.09 0.70 0.78 0.19 0.30 0.09 0.59 0.64 0.10 0.28 0.08 0.54

1.4 1 1.1 0.08 0.1 1.4 1 1.1 0.09 0.4 1.4 1 1.1 0.11 0.4

Using de Bois equation [27]; Icl,tot: Total clothing insulation. 94

Journal of Building Engineering 23 (2019) 90–106

S. Kumar et al.

Table 4 Description of questionnaire scale used for sensation and preferences for indoor parameters. Sensation scale

Preference Scale

Scale value

Thermal sensation

Humidity sensation

Air movement sensation

Thermal Preference

−3 −2 −1 0 +1 +2 +3

Cold Cool Slightly cool Neutral Slightly warm Warm Hot

Very dry Moderately dry Slightly dry Neutral Slightly humid Moderately humid Very humid

Very still Moderately still Slightly still Neutral Slightly moving Moderately moving Much moving

– Much Cooler A bit cooler No Change A bit warmer Much warmer –

concentration in ppm. The second part was filled by the surveyor itself to minimize error in recording parameters during the survey.

Humidity Preference

Air movement preference

Much less dry Slightly dry No Change Slightly humid Much more humid

Much less movement Slightly less movement No Change Slightly movement Much More movement

Thermal Acceptability

Acceptable Unacceptable

3.5. Assessing clothing insulation and activity level The clo values were estimated using the information provided in the checklist of Standard CEN 7730 [15] and ASHRAE standard 55‒2013 [1]. Authors added undergarment's insulation i.e. about 0.04 clo for each calculated clo value. For metabolic rates, the standard checklists provided in ASHRAE standard 55–2013 [1] have been used. During the survey, subjects were asked about their present and activities 30 min prior to the survey. Metabolic rate i.e. ‘met’ value (1met = 58.2 W/m2) corresponding to the reported activity is selected from the checklist provided in ASHRAE standard 55‒2013. The mean activity of the subjects was observed to be nearly sedentary, i.e. 1.2 met (sd ( ± σ) = 0.08) while some occupants were reported taking sleep (i.e. 30 min prior to the survey) or walking indoor/ outdoor during the field survey.

3.4. Outdoor and indoor environmental variables Jalandhar city falls under composite climate of India and has four seasons: summer (April-June), monsoon (July-September), autumn or moderate (October and March) and winter (November-February) season. In the present study hourly outdoor data (i.e. dry bulb temperature, relative humidity, and precipitation (mm)), was extracted from an online weather tool viz. weather underground for entire study period [32]. The climate of Jalandhar city is very hot during peak summer season (May, and June) and chilling cold in the winter season (December and January). During the summer season, outdoor maximum temperature rises up to 45 °C while during the winter season, minimum temperature falls down to 0 °C. Around 70% rainfall is received during July‒September months. Outdoor maximum temperature during monsoon season varies between 24–36 °C and relative humidity varies between 75–100%. Such a combination of high temperature and high relative humidity during the monsoon season of this climatic zone may lead to thermal discomfort conditions for the occupants of the naturally ventilated building. During the survey, measurements of indoor thermal variables were done in the vicinity of the subject using high accuracy calibrated digital instruments (Table 5). A thermo-hygrometer (Tr–76Ui) was used to measure indoor air temperature (Ta), relative humidity (RH) and CO2 concentration. A probe thermometer (Tr–52i) was used to measure indoor globe temperature which is further used for calculation of mean radiant temperature (which includes the effects of radiant fluxes from all surrounding surfaces). A multi-directional hot-wire anemometer (Testo−405 V1) was used to measure the indoor air speed. Fig. 3 shows the survey environments and instruments used in the study.

4. Result and analysis 4.1. Hygro-thermal parameters observed Rainy period in this region spans for three months i.e. July, August and September. During the field study, ambient air temperature varied between 26.4–36.8 °C (mean Tout = 32.9 °C, sd( ± σ) = 2.5) with relative humidity varied between 55% and 98% (mean RH = 73%, sd ( ± σ) = 6.4). High humidity along with high temperature in the rainy season may have a role in thermal discomfort. In addition, this region is known for growing paddy in the rainy season and water in the paddy fields further increases the water content of ambient air even during clear sunny days leading to more intense humid conditions. Table 6 shows the profile of observed indoor and outdoor temperature and relative humidity for the entire monitoring period. 4.1.1. Thermal indices used in the study In this study indoor thermal environment was monitored with different thermal indices like globe temperature, air temperature and

Table 5 Details of instruments used during the field study. Description

Make of Instruments

Parameter used

Range

Accuracy

Function

Thermo hygro CO2 meter

TR – 76Ui

Air temperature Relative humidity

0–55 °C 10–95% RH

± 0.5 °C ± 5% RH

To measure indoor air temperature For measuring indoor relative humidity, Co2 level

Globe thermometer Fluke 61 infrared thermometer

Tr− 52i, globe (dia 0.075 m) Fluke 61

Co2 level Globe temperature Surface temperature

0–9999 ppm −60 to 155 °C −18 to 275 °C

± 50 ppm + 5% ± 0.3 °C ± 2 °C

Testo 405 Thermal anemometer

Testo

Hobo (4-channel loggers)

Hobo U-12

Air velocity Air temperature Air temperature Relative humidity

0.01–10.00 m/s −20 to 50 °C −0 to 50 °C 5–95%

0.01 m/s ± 0.1 °C ± 0.35 °C ± 3.5%

Lighting level

0–48 klx

± 5%

To measure indoor globe temperature To measure the wall and roof surface temperature For measuring indoor air velocity To measure indoor air temperature

To measure indoor relative humidity

95

To measure the indoor lighting level (lux)

Journal of Building Engineering 23 (2019) 90–106

S. Kumar et al.

relative humidity was also observed high (mean RH = 70%, sd( ± σ) = 5.9) about 2% more than mean outdoor relative humidity (mean RH = 68%, sd( ± σ)= 11). Fig. 4 shows the percentage distribution of indoor temperature and relative humidity observed in traditional and modern hostel buildings during the field study. Variation in operative temperature was also observed among different floors. The ground floor rooms of modern and traditional hostel buildings were on average 3 °C and 3.5 °C cooler than top floor rooms which have roof exposed to the sky, in both the hostels respectively. Relative humidity between the ground floor and the top floor varied by more than 10% and we noted this difference statistically significant (p ≤ 0.05). In addition to this mean indoor airspeed was observed high (mean = 1.7 m/s, sd( ± σ) = 0.60) in traditional buildings compared to modern hostels (mean = 1.04 m/s, sd( ± σ)= 0.59) for entire monitoring period. Results show that occupant in traditional buildings preferred high airspeed by switching “ON” ceiling fans and applying other environmental controls i.e. opening/closing windows and doors to maintain comfort conditions.

Table 6 Distribution of indoor and outdoor parameters observed during the field study. Variables

Hostel buildings Modern N

Ta(°C) Tg(°C) Top(°C) Tout(°C) Rhi (%) Rho (%) Va(m/s) Icl,tot (clo) Activity (met)

642

Traditional Mean

sd ( ± σ)

31.6 31.1 31.1 32.9 71 70 1.04 0.29 1.1

1.4 1.5 1.5 2.5 6 13 0.59 0.09 0.11

N

303

Mean

sd ( ± σ)

31.3 30.7 30.6 31.6 70 68 1.71 0.28 1.1

0.9 0.9 1.0 2.4 6 10 0.62 0.08 0.09

N: No of samples; Ta(°C): Indoor air temperature; Tg(°C): Indoor globe temperature; Top(°C): Indoor operative temperature; Tout(°C): outdoor air temperature; Rhi(%): Indoor relative humidity; Rho(%): outdoor relative humidity; Va(m/s): Indoor air speed; Icl,tot (clo): Total clothing insulation.

4.2. Quantification of personal parameters: Clothing Insulation and activity level

mean radiant temperature. The mean radiant temperature was calculated using the Eq. (1) and as per the definition provided in ASHRAE standard 55–2013 [1].

Tmrt = ⎡ (Tg +273.15) + ⎢ ⎣

108V 0.6 (Tg − 1.1× εD0.4

In this study clothing insulation for an individual subject was estimated from the values provided in the checklist of ASHRAE standard 55–2013 and other published scientific sources on educational buildings in India [27,30]. Present study is conducted in the province of Punjab state (India) where the Sikh community has a major population. The occupants from this community wear a ‘turban’ as a ritual in their religion (Fig. 5). It was noticed that occupants from this community wrap the turban in two ways with different lengths of clothes and the corresponding weight of the turban was about 0.34 kg and 0.37 kg, respectively. For this study, to calculate the extra “clo” value of ‘turban’ we have used the relation given by Hanada et al. [36]. To accommodate the extra insulation an additional clo values of 0.33clo (~0.34 kg turban) and 0.36 clo (~0.37 kg turban) in total clothing insulation of such subjects was added. Table 8 shows the typical clothing ensemble wear by subjects in hostel buildings during the survey period. The mean clothing level was observed to be 0.29 clo (sd( ± σ) = 0.09) (Fig. 6). In the morning, students wore light clothing ensemble i.e. a combination of t-shirt and shorts with inner garments (mean 0.19, sd( ± σ) = 0.08). While during afternoon students were found engaged in their daily routine work of attending classes and other academic activities. Hence clothing insulation was a little higher (mean = 0.42, sd( ± σ) = 0.11). Field studies in naturally ventilated buildings have shown that the clothing insulation has a strong correlation with outdoor temperature [18,21,37]. In the present study, it was also observed that clothing insulation has a significant correlation (p < 0.05) with both indoor operative temperature and outdoor air temperature (Fig. 7). Further, to quantify the relation of clo value with temperature, clothing values are regressed against indoor operative temperature and outdoor air temperature [37,38]. Following linear regression equations are obtained

0.25

Ta) ⎤ ⎥ ⎦

− 273.15 (1)

where, Tmrt, Tg and Ta is the mean radiant temperature (°C), globe temperature, and air temperature respectively. ε (~0.95 ) and D (~0.075 m) used in Eq. (1) refers to the emissivity of black surface and diameter of globe used in the study. Nicol and Humphreys [31,33] and Rijal et al. [34] have used indoor globe temperature as thermal comfort indices for predicting the comfort temperature in naturally ventilated office buildings of European countries. It has been recommended by some researchers that indoor globe temperature considers the effect of airspeed only up to 0.2 m/s [35]. Present study is conducted in naturally ventilated hostel buildings where subjects used elevated air speed through the operable windows, doors, and ceiling fans where the airspeed is usually more than 0.2 m/s so operative temperature is being used as thermal indices to quantify the thermal environments. In subsequent analysis of thermal conditions in surveyed hostel building operative temperature is used. The indoor operative temperature for the measured data is calculated using Eqs. (2) and (3). Eq. (2) is acceptable for occupants engaged in near sedentary physical activity (with metabolic rates between 1.0 met and 1.3 met) but not exposed to direct sunlight and to airspeed greater than 0.20 m/ s. Eq. (3) is applicable for occupants engaged in near sedentary physical activity (with metabolic rates between 1.0 met and 1.4 met), but not exposed to direct sunlight, but exposed to air velocities greater than 0.20 m/s.

Top =

Top =

(Tmrt + Ta) 2

m ⎛0
(Tmrt +(Ta ×√10v)) 1 + √10v

m ⎛Va>0.20 ⎞ s⎠ ⎝

(2)

Icl,tot = −0.018Top+0.86 (N = 19,

R2 = 0.81, p < 0.001, S. E. =0.06) (4)

(3)

Icl,tot = −0.007To+0.47 (N = 19,

R2

= 0.36, p < 0.001, S. E. =0.08) (5)

An interesting phenomenon is observed here. It was found that clothing insulations are showing a strong correlation with indoor operative temperature rather than outdoor air temperature. This characteristic is supported by the similar findings reported by several researchers [12,21,37]. The regression slope in Eq. (4) indicates that a perturbation of about 6 °C in indoor operative temperature is needed to change the clothing insulation of about 0.1clo. Indraganti et al. [14] noted a variation of 6 °C and10 °C in indoor temperature is required for the addition of 0.1clo for office buildings in Japan and India during

4.1.2. In between buildings effect: Traditional versus modern hostel buildings Table 6 presents the indoor and outdoor thermal parameters collected from all naturally ventilated hostel buildings during the field study. The mean operative temperature (mean Top =31.1 °C, sd( ± σ) = 1.5) recorded in modern buildings was on average 0.5 °C warmer than traditional buildings (mean Top =30.6 °C, sd( ± σ) = 1.0). The strength of the relationship of Tg and Top with Tout observed to be robust for both hostel buildings as shown in Table 7. The mean indoor 96

Journal of Building Engineering 23 (2019) 90–106

S. Kumar et al.

Table 7 Correlation between observed subjective thermal responses and indoor thermal variables. Buildings

N

Pearson correlations factor

Tg:To (°C)

Top:To (°C)

TSV: TP

TA: TSV

Modern Traditional Pooled

642 303 945

r= r= r=

0.74 0.68 0.72

0.69 0.66 0.68

−0.26 −0.36 −0.28

0.42 0.41 0.44

Correlation is significant at the 0.01 level (2-tailed test); TA: Overall thermal acceptability; TSV: Thermal sensation, Tg: Indoor globe temperature(°C), Top: Indoor operative temperature(°C), To: Outdoor air temperature (°C).

4.3. Data accuracy and removal of outliers Removal of inconsistency in collected data is important to increase the accuracy in analysis and minimize error. In this study method suggested by Montazami et al. [40] and Teli [41] was used. In this method, the summation of thermal sensation and thermal preference should not increase the overall sensation of warm or cold i.e. (TSV +TPV) < -2 or (TSV +TPV) > + 2 of the individual vote. This is based on the fact that TSV's within (-3,-2) and (+2, +3) are considered to express the dissatisfaction by subject and subject would not deliberately wish to enhance that feeling [40]. In the present study, there are about 8.5% (N = 89) such inconsistencies or outliers and were removed from the complete database. Interestingly, this figure matches closely with the study conducted by Teli et al. [41] in school classrooms. 4.4. Analysis of subjective sensation and preferences In this section sensation and preference votes of the occupants and the corresponding indoor thermal variables were analyzed. Following sub-sections discuss the same. 4.4.1. Thermal sensation (TSV), preferences (TP) and thermal acceptability (TA) The survey questionnaire collected the subject's sensation and preference votes under “Right Here Right Now” condition. The subjects voted on ASHRAE 55 thermal sensation and preference scale. Fig. 8(a, b) shows the frequency distribution of TSV (thermal sensation vote) and TP (thermal preference vote) for individual buildings. About 80% (N = 243) of the occupants from traditional hostel buildings voted within three central categories of TSV compared to 75% (N = 482) occupants in modern buildings. In traditional buildings, about 3% (N = 9) of the occupant perceived their conditions cool and cold while 17% (N = 48) subjects reported indoor conditions warm and hot. Contrary to this observation, in modern buildings, only 2% (N = 15) of the occupants perceived their conditions cold while 22% (N = 142) subjects perceived indoor conditions warm and hot. When data of modern and traditional hostel buildings put together, 77% (N = 728) of occupants voted within three central categories of TSV and can achieve comfort by using adaptive opportunities. The mean of TSV in the traditional hostel and modern hostel buildings were TSVmean= +0.4, sd( ± σ) = 1.1 and TSVmean= +0.7, sd( ± σ) = 1.2 respectively (Table 9). These values are more towards warmer side of the thermal sensation scale. This shows that occupants in modern hostel buildings feel slightly warmer than those residing in traditional hostel buildings. However, it should be noted from Fig. 3, that during the field study ambient conditions observed in traditional hostel buildings were comparatively less harsh than that of modern NV hostel buildings. The mean value of TSV value for all data combined was found + 0.5, sd( ± σ) = 1.2; during the monsoon season. Regarding occupants thermal preferences, it was noticed that about 68% (N = 642) occupants desired cooler environment and a very few subjects, about 7% (N = 66), preferred hot and warm conditions in traditional as well as modern hostel buildings. About 25% (N = 237) of the subjects voted for no change in their existing thermal environment for combined data. Fig. 9 shows a cross-tabulated summary of TSV and

) Fig. 4. Statistical details of (a) indoor air temperature and (b) indoor relative humidity observed during the field study for different building types (temperature is binned to 1 °C and relative humidity is binned to 2%).

summer season respectively. Carli et al. [39] also reported that 11 °C variation in indoor temperature is required to add or remove 0.1clo for the global field data collected under RP‒884 project. The mean activity of the subjects was observed to be nearly sedentary activity, i.e. 1.2 met (sd ( ± σ) = 0.09). During the study occupants, activity varied between 1.4 met to 1 met. These activities were corresponding to walking indoor/outdoor and sleeping conditions. However, most of the occupants were observed seated passively while some subjects were doing light seated work of reading and writing.

97

Journal of Building Engineering 23 (2019) 90–106

S. Kumar et al.

Fig. 5. Typical clothing ensembles worn by hostel occupants during the field study.

and want ‘no change’ in temperature while 35% (N = 107) occupants in traditional building voted neutral. More than 60% subject's wanted cooler environment though they have voted neutral on TSV scale. About 85% subjects preferred ‘no change’ while voting towards cooler side of TSV scale, and about 15% of subjects preferred the same while voting on the warmer side of TSV scale. The mean of TP in traditional and modern buildings was -0.7 (sd( ± σ) = 0.7) and -0.8 (sd( ± σ) = 0.8) respectively. For combined data, it was -0.7 (sd( ± σ) = 0.7). This pattern is found to be similar to the results reported by different studies conducted for different built environments in composite climate of India during summer and monsoon season [12,14,19]. To analyze the preferences, thermal preference scale is modified to change the votes to binary form following the methodology of Singh et al. [10]. Probit analysis was carried out using the binary data plotted as Fig. 10. From the figure it can be seen that the lines representing the ‘wanting warmer’ and ‘wanting cooler’ preferences intersect towards the cooler side of the thermal sensation scale. It indicates that residents of hostel buildings prefer a cooler environment during monsoon season. Several studies conducted in hot and humid climates, across the world, have reported the identical findings for studies conducted during summer season [4,6,9]. On the direct question of thermal acceptability about the existing thermal environmental conditions, about 70% (N = 662) occupants voted ‘acceptable’ while 30% (N = 283) felt unacceptable on combined data (Fig. 8(c)). A small difference in “overall acceptability” of thermal environment in traditional and new hostel buildings was also noticed. About 70% (N = 449) subjects voted towards “overall thermal acceptance” of their existing environment in modern hostel buildings compared to 68% (N = 206) subjects in traditional hostel buildings. Also, the strength of the relationship was week between TSV with thermal preference (TP) (N = 945, r = -0.28) while it was moderate between TSV and TA (N = 945, r = 0.44) as shown in Table 7. This shows that occupants were able to correlate their immediate thermal sensation with overall acceptability rather well compared to that of thermal preferences. Peeters et al. [42] also noted that occupants always preferred cooler environments during a warmer period, irrespective of building type or acceptability level (at 95% confidence level). Further,

Table 8 Typical clothing insulation observed during the field study. Sr. No.

Clothing

Total clo Value

1. 2. 3. 4.

Short shorts + Men's briefs T-Shirt (~Baniyan) + Short shorts + Men's briefs T-Shirt + Straight trousers (thin) + Men's briefs Shirt (Half Sleeve) + Straight trousers (thick) + Men's briefs Shirt(Full Sleeve) + Straight trousers (thick) + Men's briefs

0.10 0.18 0.27 0.46

5.

0.61

Fig. 6. Percentage distribution of clothing value for all data.

TP votes in modern and traditional hostel buildings, respectively. It can be seen from the figure that in modern hostels about 33% (N = 212) subjects voted the thermal environment ‘neutral’ at the time of voting 98

Journal of Building Engineering 23 (2019) 90–106

S. Kumar et al.

Fig. 7. Relationship of clothing insulation with (a) indoor operative temperature and (b) outdoor air temperature, respectively (Temperature scale is binned at 1 °C).

in the present study, no significant differences (p < 0.05) were noted between the relationship of TSV with TP and TA considering different building type. 4.4.2. Humidity sensation (HSV) and preferences (HP) Present study was conducted during the monsoon season and the relative humidity was observed comparatively high during the monitoring period. During the field study period, the indoor and outdoor relative humidity varied between 65% and 90% and 55% and 98%, respectively. Looking at the prevailing high humidity authors thought that it will be interesting to analyze the characteristics of the occupant's perception towards prevailing humidity conditions inside the hostel rooms. Seven-point scale (from ‘-3’ to ‘+3’) and five-point scale (from ‘-2’ to ‘+2’) was used to record the sensation and preference of subjects respectively. Fig. 11 shows the distribution of HSV (humidity sensation votes) recorded for different building types. About 26% (N = 246) subjects reported ‘neutral’ on existing humidity conditions in both traditional and modern hostel buildings. Further, the survey results showed that about 19% (N = 180) of subjects reported existing humidity condition as dry while 54% (N = 510) of the subjects reported the conditions humid and very humid. These results are on the expected lines as the field study was conducted during rainy season and indoor relative humidity is comparatively high in this period. Mean sensation of humidity was observed to be + 0.6 (sd( ± σ) = 1.3) i.e. toward the humid side of the humidity sensation scale. Mean relative humidity of 70.2% was found for the combined data.

Fig. 8. Frequency responses of (a) thermal sensation, (b) thermal preferences, and (c) thermal acceptability votes for modern and traditional hostel buildings, respectively.

Subsequently, we noted that mean humidity preference of the subjects was towards the slightly dry side of preference scale (mean=-0.3, sd( ± σ) = 0.8) (Table 9). Interestingly, about 65% (N = 197) and 75% (N = 482) of the subject's accepted the existing humidity conditions (when the subject's voted HSV=0) in traditional and modern hostel 99

Journal of Building Engineering 23 (2019) 90–106

S. Kumar et al.

buildings, respectively and still prefers ‘no change’ (Fig. 12). About 48% subjects preferred ‘no change’ while voting on the humid side of HSV scale in traditional and modern hostel buildings. Similarly, a total of 53% and 65% occupants voted no change while accepting the humidity conditions as dry on HP scale in traditional and modern hostel buildings, respectively. Nicol [3] and Rijal et al. [43] has shown that relative humidity has a negligible effect on the thermal comfort of occupants, particularly in naturally ventilated buildings. The possible reason shall be various adaptive opportunities available to the occupant to nullify the effect of high relative humidity. For example, high airspeeds (by means of operation of fans) are more effective for increasing the heat loss and evaporation from the skin. ASHRAE standard 55 [1] suggests, at higher airspeed upper relative humidity limit for comfort can be reached up to 95%.

Table 9 Statistical details of subjective responses observed for different indoor variables during the study. Variables

Hostel buildings

N=

Modern 642

TSV TPV TA HSV HP AVS AVP PMV PPD

Traditional 303

Pooled 945

Mean

sd ( ± σ)

Mean

sd( ± σ)

Mean

sd( ± σ)

0.7 −0.8 0.4 0.6 −0.3 −0.2 0.8 1.6 56

1.2 0.8 0.3 1.4 0.8 1.4 0.9 0.7 30

0.4 −0.7 0.3 0.6 −0.3 −0.2 0.6 1.0 34

1.1 0.7 0.4 1.2 0.7 1.2 1.0 0.7 23

0.5 −0.7 0.3 0.6 −0.3 −0.2 0.7 1.4 49

1.2 0.7 0.5 1.3 0.8 1.3 0.9 0.8 30

4.4.3. Air speed sensation (AVS) and preferences (AVP) During the field study, all surveyed hostel rooms were found under ceiling fan “ON” condition. Data analysis shows a significant variation in airspeed in different buildings and on different floors during the field study. For example, the mean indoor airspeed was observed high (mean =1.70 m/s, sd ( ± σ)= 0.60) in traditional buildings compared to modern hostels (mean=1.04 m/s, sd ( ± σ)= 0.59) for whole monitoring period. Also, the mean indoor airspeed was observed low in ground floor rooms (mean=0.84 m/s, sd( ± σ) = 0.54) as compared to top rooms(with roof expose to the sky) (mean=1.6 sd ( ± σ) = 0.61). The data collected through questionnaires were further analyzed to assess the impact of different airspeeds on sensation and preferences of occupants. Fig. 13 shows the distribution of airspeed sensation for traditional and modern surveyed hostel buildings. About 26% (N = 246) subjects feel neutral at the existing indoor air speed on seven-point AVS (air velocity sensation) scale. While more than 35% (N = 331) subjects voted for more air movement and 37% (N = 350) felt the air movement slightly less in these investigated NV hostel buildings. Mean AVS and AVP (air velocity preference) for hostel buildings were found to be -0.2 (sd ( ± σ) = 1.4) and + 0.7(sd ( ± σ) = 0.9), respectively (Table 9). This indicates that occupants perceived airspeed is slightly less than the desired airspeed and preferred more airspeed during monsoon season. However, no significant differences (p < 0.05) was noted between the air movement preferences of occupants of traditional and modern buildings during the study. Fig. 14 shows the cross-tabulation of airspeed sensation and airspeed preferences for traditional and modern hostel buildings. It can be seen from the figure that about 27% (N = 82) and 33% (N = 212) of subjects from traditional and modern buildings accepted the air movement and preferred no change. it was also observed that about 67% (N = 203) and 58% (N = 372) of subjects from traditional and modern NV hostel buildings felt air movement ‘neutral’ at the time of voting but desired more airspeed while expressing their preference votes. A very few votes about 22% (N = 67) and 8% (N = 51) subjects from traditional and modern NV hostel buildings preferred less air movement while voting ‘neutral’ on AVS scale.

N: Sample size, TSV: Thermal sensation vote, TPV: Thermal Preference vote; TA: Thermal acceptability; HSV: Humidity sensation vote; HP: Humidity preference; AVS: Air speed sensation, AVP: Air speed preference; PMV: Predicted mean vote; PPD: Percentage people dissatisfied.

Fig. 9. Cross-tabulation of thermal sensation (TSV) and thermal preferences (TPV).

4.5. Thermal comfort zone for surveyed subjects In thermal comfort research community, it is widely accepted that if the subjects/occupants in naturally ventilated buildings voting within the central three categories of the thermal sensation scale ( ± 1TSV) then subjects can make themselves comfortable by going through preferred or immediately available adaptation process. This corresponds to the thermal comfort zone of the surveyed population with predicted mean vote (PMV) range of ± 0.85 and percentage people dissatisfied (PPD) < =20%. For estimating the thermal comfort zone of surveyed subjects, we have used the Probit analysis method. Rijal et al. [44] discussed this method to predict the proportion of votes for each category of the ASHRAE seven-point scale. To carry out this analysis, data for both buildings type were combined. This includes the ordinal

Fig. 10. Relationship between thermal sensation and thermal preference of subjects preferring warmer and cooler environments.

100

Journal of Building Engineering 23 (2019) 90–106

S. Kumar et al.

Fig. 13. Frequency distribution of air speed sensation (AVS) for surveyed buildings.

Fig. 11. Frequency distribution of humidity sensation (HSV) for surveyed buildings.

Fig. 12. Cross-tabulation of humidity sensation (HSV) and humidity preferences (HPV) for surveyed buildings.

regression analysis with Probit (Z) as the link function and the operative temperature as the covariate (Table 10). Mean temperatures for each equation were estimated by dividing the constant for each equation with the estimated Probit regression coefficient. Meanwhile, the standard deviation of cumulative normal distribution was the inverse of the probit regression coefficient. All equations presented in Table 10 were plotted into sigmoid curves (Fig. 15a) using the function

Probabilityof voting = CDF. NORMAL(quant,mean, sd)

Fig. 14. Cross-tabulation of airspeed sensation (AVS) and airspeed preferences (AVP) for surveyed buildings.

and modern buildings using the procedure defined in standards ISO CEN-7730 [15] and ASHRAE standard 55–2013 [1] and then compared the PMV model with the actual thermal sensation votes (TSV). Fanger's heat balance model [45] of thermal comfort prescribed the comfort limit within PMV bandwidth of ± 0.50 (~90% acceptability) and ± 0.85 (~80% acceptability). The mean PMV was found to be + 1.46 (sd ( ± σ) = 0.8) and the corresponding PPD was 49% for the combined data. Interestingly, the mean PMV was observed lower in the traditional building (mean PMV=+1.0, sd ( ± σ) = 1.4) compared to modern hostel buildings (mean PMV=+1.6, sd( ± σ) = 1.4). However, on combined data, PMV significantly overestimated the thermal sensation of occupants (TSV). Fig. 16 shows that there is a significant discrepancy between calculated PMV with observed TSV for both traditional and modern hostel buildings. Various researchers [10,11,31,37] through field experiments explained the reasons for such deviation in PMV and TSV. They argued that the discrepancy is mainly due to the methods of

(6)

where CDF.NORMAL is the cumulative distribution function for normal distribution and ‘quant’ is the operative temperature (°C). In the present study, neutrality of 30.5 °C has been observed using a straight line corresponding to “neutral TSV = 0” at the probability of 0.50. Further to the results, more than 80% of subjects were comfortable i.e. voted within central three categories ( ± 1TSV) when operative temperatures varied between 27.2 and 31 °C (Fig. 15b). 4.5.1. Comparison of predicted mean vote (PMV) and thermal sensation (TSV) The study calculated the mean PMV and PPD for both traditional 101

Journal of Building Engineering 23 (2019) 90–106

S. Kumar et al.

Table 10 Description of Probit analysis for thermal sensation and indoor operative temperature. TSV

Probit regression lines

Mean temperature (°C)

sd( ± σ)

N

SE (estimated)

Nagelkerke R2

< < < < < < <

0.336Top + 7.46 0.336Top + 8.24 0.336Top+ 9.40 0.336Top + 10.37 0.336Top + 11.32 0.336Top + 12.21 0.336Top + 13.15

22.2 24.5 27.9 30.9 33.7 36.3 39.1

2.97

945

0.027

0.17

=− 3 =− 2 =− 1 =0 =1 =2 =3

TSV: Thermal sensation; sd( ± σ): Standard deviation; N: No. of samples; SE: Standard error of estimate.

psychological, physiological and behavioral adaptation of the occupants in the field data which is lowering the regression coefficients. 4.6. Predicting comfort temperature The comfort temperature can be defined as ‘the indoor air temperature’ at which, a healthy subject will vote ‘neutral’ on the thermal sensation scale. In this study, two methods are used to estimate the comfort temperature. They as follows a) Linear regression method b) Griffiths Method. 4.6.1. Linear regression method In the linear regression method, the neutral temperature is determined by finding the intersection point of the regression line the horizontal line corresponding to sensation “neutral”. Fig. 17 shows the scatter plot of the subject's thermal sensation votes and corresponding room operative temperature. The relationship between thermal sensation and room operative temperature is shown by the regression Eq. (8). By substituting the TSV = 0 in Eq. (8), a neutral temperature of 29.7 °C has been calculated for combined data. A comfort bandwidth of 26.7–32.6 °C corresponding to the TSV ± 1 is obtained through the results of the regression analysis (Fig. 12). The slope of the Eq. (4) is 0.34/°C indicating that if there is a 3 °C change in indoor operative temperature, thermal sensation vote would shift by a unit. The slopes of regression equations can, therefore, be viewed as respondents’ sensitivity with respect to indoor operative temperature, Top. The results are comparable to the other studies conducted in hostel dormitories. For example, Dhaka et al. [27] and Dalhan et al. [25] obtained a slope coefficient of 0.302(~ Neutral temperature 30.2 °C) and 0.42(~ Neutral temperature 29.8 °C) for their field study in hostel buildings. Interestingly, the regression slope obtained from the present study resembles with a slope of Mishra and Ramgopal [46] i.e. 0.36. Also, the study finds a close similarity to the gradient observed by Indraganti [12] for field study in multi-storey residential buildings of Hyderabad city. Fig. 15. Probit analysis (Z) results showing the probability of voting on (a) thermal sensation and (b) proportion of comfortable votes with indoor operative temperature for the field data.

4.6.2. Griffiths method Griffiths emphasized that whenever the indoor temperature range obtained from the field surveys is narrow, the regression method could not provide a reasonable estimate of comfort temperature. It was emphasized by Humphreys and Nicol [31,33] that presence of thermal adaptation behaviour of occupants (i.e. opening of windows, use of elevated speed by means of fans or clothing adjustments) in field data will tend to lower the regression coefficients (as observed in Eq. (8)) and may deviate from actual thermal neutrality. So, Griffiths method was used for the estimation of individual comfort temperature for each TSV. Griffiths comfort temperature was calculated by using Eq. (9).

assessment of clothing and metabolic rates in field studies, particularly in hot and arid climates. TSV and PMV were further regressed against indoor operative temperature which resulted in following Eqs. (7) and (8). The relations were significant at p < 0.001.

PMV = 0.50Top − 14.0 (R2 = 0.79, N = 945, S. E. =0.26

, p < 0.001) (7)

TSV = 0.34Top − 10.1 (R2 = 0.16, N = 945, S. E. =0.11

, p < 0.001)

TC = Top +

(8) 2

The low regression coefficient (R value) observed between TSV and the indoor operative temperature is supporting the presence of

0 − TSV G

(9)

Where, Tc is the comfort temperature (°C), Top is the indoor operative temperature (°C), TSV is the thermal sensation vote, and G is the 102

Journal of Building Engineering 23 (2019) 90–106

S. Kumar et al.

Fig. 16. Scatter plot of the TSV and PMV versus the operative temperature for different buildings (with lines of 95% confidence interval). Table 12 Descriptive statistics for mean operative temperature in regard to neutral TSV and change of TP in surveyed hostel buildings. Hostel buildings

Variables

Ta(°C) Tg(°C) Top(°C) Ta(°C) Tg(°C) Top(°C) Ta(°C) Tg(°C) Top(°C)

Modern

Traditional

Pooled

TSV = 0 N 182

109

291

Mean ( ± σ) 31.2 30.8 30.7 31.5 30.5 30.4 31.2 30.7 30.6

(1.3) (1.4) (1.4) (0.7) (0.8) (0.7) (1.1) (1.2) (1.2)

TP = 0 N 164

72

236

Mean ( ± σ) 31.6 31.3 31.1 31.0 30.5 30.4 31.1 30.6 30.5

(1.2) (1.3) (1.4) (0.8) (0.8) (0.9) (1.3) (1.4) (1.4)

Ta(°C): Indoor air temperature; Tg(°C): Indoor globe temperature; Top(°C): Indoor operative temperature; N: No. of samples; ( ± σ): Standard deviation, TP: Thermal preference.

Griffiths constant. The scale value for neutral sensation is represented by ‘0’ in the equation. Griffiths comfort temperature for all votes using a scale of 0.25, 0.33 and 0.50 as Griffiths slope. Analysis of mean indoor operative temperature for neutrality (‘0’ TSV scale) on the TSV scale has shown close agreement with comfort temperature obtained with 0.50 as

Fig. 17. Scatter plot of TSV with indoor operative temperature and linear regression equation with 95% confidence interval.

Table 11 Descriptive statistics for comfort temperature calculated by the Griffiths method in surveyed buildings. Hostel buildings

Modern

Traditional

GC (/°C)

0.25 0.33 0.50 Voting Neutral 0.25 0.33 0.50 Voting Neutral

TaC (°C)

TgC (°C)

TopC (°C)

N

Mean

sd( ± σ)

Mean

sd( ± σ)

Mean

sd( ± σ)

642

29.2 29.8 30.4 31.2 29.5 29.9 30.4 31.5

4.3 3.2 2.2 1.3 4.4 3.3 2.2 0.7

28.8 29.4 30 30.8 28.9 29.4 29.9 30.5

4.2 3.2 2.1 1.4 4.4 3.3 2.1 0.8

28.8 29.4 29.9 30.7 28.9 29.3 29.8 30.4

4.2 3.2 2.1 1.4 4.3 3.3 2.2 0.7

182 303

109

N: Sample size; GC: Griffiths constant; Tac: Comfort air temperature (°C); Tgcglobe temperature (°C); Topc: Comfort operative temperature (°C); sd ( ± σ) : Standard deviation. 103

104

25.9 29.8 26.5 Indraganti et al. [14] Singh et al. [30] Mishra et al. [46]

NV: Naturally ventilated; Top: Indoor operative temperature; Tg: Indoor globe temperature.

NV NV NV

Composite Composite Composite

Summer and Monsoon Summer Summer

352 900 116

0.215Tg − 5.682 0.19 Tg − 5.04 0.36Top − 10.73

21.4–30.5 23.5–31.5 19.4–33.7

0.29 0.41 0.47 N.A. 29.9 30.2 30.6 29.2 0.34Top + 10.4 0.30Top+ 9.43 0.18Top − 5.04 0.31Tg − 9.06 682 429 1220 3962 Monsoon Summer Summer Summer Composite Composite Composite Composite NV NV NV NV

Hostel buildings Traditional Hostels Residential and offices Multi-storey Residential Apartments Offices University classrooms Laboratory Present study Dhaka et al. [27] Kumar et al. [21] Indraganti [12]

Conditioning type

Climate

Time of study (Season)

Sample size

Neutral temperature/ Comfort temperature (°C)

28.7–31 25.2–33.1 25.2–30.6 26.0–32.4

Mean clo value

Nicol and Humphreys [3,31] laid down the basic principle of adaptive thermal comfort approach. They revealed that in adaptive approach the occupant in built environment acts as active agents who respond to the changing indoor environments through various thermal adaptations and actions which tend to restore thermal comfort. In the present study, we noted the different personal adaptations taken by subjects to restore their comfort. The most commonly used adaptive actions come across were a change of clothing insulation, use of controls like windows and doors (for the natural flow of air) and use of

Regression Equation

Comfort range (°C)

4.7. Occupant's thermal adaptation behaviour analysis

Building type

Table 13 Comparison of comfort temperatures among different studies conducted for different building types in composite climate of India.

Griffiths coefficient (with least standard deviation) as shown in Table 11. The mean operative comfort temperature was also matched closely to the mean temperature for which TP is equal to ‘0: No change’ on five-point scale (Table 12). The mean indoor operative comfort temperature as calculated by Griffiths method is 29.9 °C (sd( ± σ) = 2.16) on combined data (Fig. 18). The mean indoor operative comfort temperature as calculated by Griffiths method is 29.9 °C (sd( ± σ) = 2.16) on combined data. This may be due to various adaptive opportunities available to hostel occupants like the use of high indoor air speed by means of opening the windows or switching on ceiling fans (Fig. 3). Higher airspeeds are more effective for increasing the heat loss and evaporation from the skin, even at higher operative temperatures. Kumar et al. [20] also observed that occupants of NV buildings were feeling comfortable up to 35 °C at airspeed of 1.5 m/s (using ceiling fans) in composite climate of India. Interestingly, this means comfort temperature is very close to the thermal neutrality (~29.7 °C) obtained using a linear regression analysis for the present data. No significant difference (p < 0.05) in mean comfort operative temperature was observed considering inter-building effects i.e. traditional versus modern NV hostels (Table 11). We further compared the mean comfort temperature obtained in the present study with other studies conducted in India during the summer season in India (Table 13). The comfort temperature obtained in this study is comparable to the study conducted in Hyderabad [12] and Jaipur for residential buildings [21]. However, comfort temperature was observed higher than that of office buildings under similar climatic conditions [13,14]. Such deviation can be attributed to the fact that different thermal adaptation opportunity available to the occupants of office and hostel buildings. The hostel occupants enjoy the choices of clothing adjustment, use of environmental controls and its accessibility, and activity level whereas in office spaces these are very limited.

Study

Fig. 18. Frequency distribution of Griffiths comfort temperature for combined data using Griffiths coefficient of 0.5.

0.71 0.36 0.50

Journal of Building Engineering 23 (2019) 90–106

S. Kumar et al.

Journal of Building Engineering 23 (2019) 90–106

S. Kumar et al.

Subjects participated voluntarily after a briefing of the objective and purpose of the study. However, this study has the following limitations: (1) the survey was conducted in selected multi-storey buildings considering the availability and accessibility for surveying. (2) The study used the checklist provided in ASHRAE standard 55–2013 for calculation of total clothing value and other traditional clothing commonly used in India considering published sources from India [12,21]. However, the study further observed that some occupants were wearing ‘turban’ as their religious belief required. So, this study is making the first attempt to add extra clothing insulation value for ‘turban’ in traditional clothing using the equation of Hanada [36]. This includes an extra value of about 0.33clo (~0.34 kg turban) and 0.36clo (~0.37 kg turban) to the total clothing value. (3) The surveying was conducted on all floors and subjects participated voluntarily based on their availability and interest. 5. Conclusions A questionnaire-based field study of thermal comfort following ASHRAE standard 55 class-II protocols were conducted in 419 rooms of six multi-storey NV hostel buildings during monsoon season (Aug-Sept, 2018) in the composite climate of India. The study collected about 945 valid subjective thermal responses (out of 1024 samples). Indoor thermal parameters were measured using high precision and digital instruments. Statistical analysis was carried out using SPSS version 21. The study found the comfort temperature of the studied group using Griffiths method and possible thermal comfort zone using Probit analysis. The study also analyzed the effect of personal adaptations i.e. clothing change and use of controls like opening and closing of windows, doors and switching “ON” ceiling fans to restore thermal comfort. Following are the main conclusion derived from the results and analysis:

Fig. 19. Percentage breakdowns of environmental controls in use from the field surveys.

elevated speed through the switching “ON” ceiling fans. In the subsequent sections, we analyzed and prioritize some of these personal adaptations. 4.7.1. Use of environmental controls During the field study, we also recorded the availability and use of environmental controls as binary data in the vicinity of the occupant. The subjects operated the windows (Open/close), doors (Open/close), and fans (switch “ON/OFF”) to restore comfortable conditions indoors. Subsequently, data was analyzed to prioritize these controls as a behavioral adaptation of occupants. Fig. 19 shows the proportion use of different environmental controls during the field study for Individual buildings. It was observed that switching “ON” fan i.e. almost 100% was primary adaptive action preferred by occupants than followed with the opening of external doors i.e. more than 80%. Since, fans offer a significant adaptive opportunity for users at high indoor temperatures and high relative humidity, especially during summer and rainy season. During the survey, the mean air speed observed during fan operation was more than 1 m/s and reached a maximum of 3 m/s. The next mostpreferred adaptive action was the opening of external doors about 80%, while 50% of the occupants adopted opening and closing of windows as their thermal adaptation measure. Interestingly, the findings are showing a close resemblance to the adaptive behaviour of classroom student's in composite climate of India [47]. This shows that environmental conditions encountered by the students, early in the morning in the hostels develop a psychological behaviour which gets reflected throughout the day (during classes). A less proportion of fan use was observed early in the morning compared to the afternoon and the evening during the entire daytime. Also, the opening of windows was observed maximum about 55%, during the evening time. The occupants were taking advantage of night ventilation technique to flush out the stored heat during the day through the opening of windows in the evening time. Windows opening behaviour was almost similar for both traditional and modern hostel.

1. The operative temperature recorded in traditional NV hostel buildings (mean Top=30.6 °C, sd ( ± σ)= 1.0) was on average 0.5 °C cooler than modern NV hostel buildings (mean Top=31.1 °C, sd ( ± σ) = 1.6). In addition to this, the significant difference (p < 0.05) was observed in indoor operative temperature (~3.5 °C between the ground floor and top floor) on different floors of the same building. 2. The mean of TSV in the traditional hostel and modern hostel buildings were TSVmean= +0.4, sd( ± σ) = 1.1 and TSVmean= +0.7, sd( ± σ) = 1.2 respectively (Table 9). These values are more towards warmer side of the thermal sensation scale. The subjects preferred cooler environment (mean =-0.7, sd( ± σ) = 0.7) throughout the survey period. 3. About 26% subjects reported ‘neutral’ for prevailing humidity conditions in both traditional and modern NV hostel buildings. Mean humidity preference of the subjects was towards the dry side (mean=-0.3, sd ( ± σ) = 0.8) on five-point humidity preference scale. 4. Mean airspeed sensation (AVS) and preference (AVP) were found to be -0.2 (sd ( ± σ) = 1.4) and + 0.7 (sd ( ± σ)= 0.9), respectively. This indicates that occupants perceived the air movement slightly less than desired airspeed and preferred more air movement during the monsoon season. 5. Probit analysis revealed that 80% of subjects voted within central three categories of TSV scale i.e. ± 1TSV, when indoor operative temperatures range between 27.2 and 31 °C. 6. The mean indoor comfort temperature as predicted by Griffiths method is 29.9 °C (sd( ± σ)= 2.15). However, no significant differences (p < 0.05) was noted between traditional and modern hostel buildings. 7. We observed that clothing insulation correlated strongly and significantly (p < 0.05) with indoor operative temperature. The study also ascertains that a perturbation of about 6 °C in indoor operative

4.8. Study limitations The present study is conducted in six multi-storied hostel buildings, constructed between 5 and 25 years, in premises of National Institute of Technology, Jalandhar city under composite climate of India. The study follows “Right here and Right Now” questionnaires and ASHRAE standard 55 and follows Class-II protocols for data collection [1]. 105

Journal of Building Engineering 23 (2019) 90–106

S. Kumar et al.

temperature is needed to make a 0.1clo change in clothing insulation value. 8. Primary behavioral adaptive action preferred by occupants was switching “ON” the ceiling fan (~100%) followed by opening of doors (~80%) and windows (~50%) to restore thermal comfort conditions.

[19] S. Dhaka, J. Mathur, G. Brager, A. Honnekari, Assessment of thermal environmental conditions and quantification of thermal adaptation in naturally ventilated buildings in composite climate of India, Build. Environ. 86 (2015) 17–28. [20] S. Kumar, J. Mathur, S. Mathur, M.K. Singh, V. Loftness, An adaptive approach to define thermal comfort zones on psychrometric chart for naturally ventilated buildings in composite climate of India, Build. Environ. 109 (2016) 135–153. [21] S. Kumar, M.K. Singh, V. Loftness, J. Mathur, S. Mathur, Thermal comfort assessment and characteristics of occupant's behaviour in naturally ventilated buildings in composite climate of India, energy, Sustain. Dev. 33 (C) (2016) 108–121. [22] Ministry of Human Resources Development, Government of India. Annual report. Department of School Education & Literacy and Department of Higher Education, 2016–2017. [23] 〈https://www.dreducation.com/2013/08/data-statistics-india-student-college. html〉 (Accessed 6 July 2018). [24] N.D. Dahlan, P.J. Jone, D.K. Alexander, Operative temperature and thermal sensation assessments in non-air-conditioned multi-storey hostels in Malaysia, Build. Environ. 46 (2011) 457–467. [25] N.D. Dahlan, P.J. Jone, D.K. Alexander, E. Salleh, D. Dixon, Field measurement and subjects’ votes assessment on thermal comfort in high-rise hostels in Malaysia, Indoor Built Environ 17 (2008) 334–345. [26] J.H.K. Lai, Gap theory based analysis of user expectation and satisfaction: the case of a hostel building, Build. Environ. 69 (2013) 183–193. [27] S. Dhaka, J. Mathur, A. Wagner, G. Agarwal, V. Garg, Evaluation of thermal environmental conditions and thermal perception at naturally ventilated hostels of undergraduate students in composite climate, Build. Environ. 66 (2013) 42–53. [28] S. Kumar, M.K. Singh, J. Mathur, A. Mathur, Thermal performance and comfort potential estimation in a high thermal mass naturally ventilated office building: an experimental study, J. Build. Eng. 20 (2018) 569–584. [29] A. Kumar, B.M. Suman, Experimental evaluation of insulation materials for walls and roofs and their impact on indoor thermal comfort under composite climate, Build. Environ. 59 (2013) 635–643. [30] M.K. Singh, S. Kumar, R. Ooka, H.B. Rijal, G. Gupta, A. Kumar, Status of thermal comfort in naturally ventilated classrooms in composite climate of India, Build. Environ. 128 (2017) 294–304. [31] F. Nicol, M. Humphreys, S. Roaf, Adaptive Thermal Comfort Principles and Practice, Routledge, London, 2012. [32] Wonderground in: www.wonderground.com. Available at: 〈https://www. wunderground.com/weather/in/jalandhar〉 (Accessed 15 September 2018). [33] M.A. Humphreys, H.B. Rijal, F.J. Nicol, Updating the adaptive relation between climate and comfort indoors; new insights and an extended database, Build. Environ. 63 (2013) 40–55. [34] H.B. Rijal, H. Yoshida, N. Umemiya, Seasonal and regional differences in neutral temperatures in Nepalese traditional vernacular houses, Build. Environ. 45 (12) (2010) 2743–2753. [35] A. Auliciem, V.Z. Szokolay, A note on thermal comfort, published in PLEA: Passive and Low Architecture International, 2007. [36] K. Hanada, K. Mihira, Y. Sato, Studies on the thermal resistance of men's underwear, J. Jpn. Res. Assoc. Text. End.-Users 24 (8) (1983) 363–369. [37] M. Indarganti, D. Bousaa, Comfort temperature and occupant adaptive behaviour in offices in Qatar during summer, Energy Build. 150 (2017) 23–36. [38] M.S. Mustapa, S.A. Zaki, H.B. Rijal, A. Hagishima, M.S.M. Ali, Thermal comfort and occupant adaptive behaviour in Japanese university buildings with free running and cooling mode offices during summer, Build. Environ. 105 (2016) 332–342. [39] M. De Carli, B.W. Olesen, A. Zarrella, R. Zecchin, People's clothing behaviour according to external weather and indoor environment, Build. Environ. 42 (2007) 3965–3973. [40] A. Montazami, M. Gaterell, M. Nicol, M. Lumley, C. Thoua, Developing an algorithm to illustrate the likelihood of the dissatisfaction rate with relation to the indoor temperature in naturally ventilated classrooms, Build. Environ. 111 (2017) 61–71. [41] D. Teli, M.F. Jentsch, P.A. James, Naturally ventilated classrooms: an assessment of existing comfort models for predicting the thermal sensation and preference of primary school children, Energy Build. 53 (2012) 166–182. [42] L.F. Peeters, R. Dear de, J.L.M. Hensen, W. D'Haeseleer, Thermal comfort in residential buildings: comfort values and scales for building energy simulation, Appl. Energy 86 (5) (2009) 772–780. [43] H.B. Rijal, M.A. Humphreys, F. Nicol, Adaptive thermal comfort in Japanese houses during the summer season: behavioral adaptation and the effect of humidity, Buildings 5 (3) (2015) 1037–1054. [44] H.B. Rijal, M.A. Humphreys, J.F. Nicol, Towards an adaptive model for thermal comfort in Japanese offices, Build. Resour. Inf. 47 (2017) 717–729. [45] P.O. Fanger, Thermal Comfort, Analysis And Applications In Environmental Engineering, McGraw-Hill, New York, 1972. [46] A. Mishra, M. Ramgopal, Thermal comfort field study in undergraduate laboratories–an analysis of occupant perceptions, Build. Environ. 76 (2014) 62–72. [47] S. Kumar, M.K. Singh, A. Mathur, J. Mathur, S. Mathur, Evaluation of comfort preferences and insights into behavioural adaptation of students in naturally ventilated classrooms in a tropical country, India, Build. Environ. 143 (C) (2018) 532–547.

The composite climate of India is seasonally and geographically diverse. The results presented in this work are limited to multi-storey hostel buildings under natural ventilation operation mode in monsoon season of the composite climate of India. So, the results from the present study can be helpful to the building designers and architects to better understand the thermal comfort requirement of occupants in these naturally ventilated hostel buildings. In addition, the effective use of adaptive controls available in these hostel buildings can enhance the overall acceptability and thermal environmental conditions of occupants during thermal discomfort conditions. Acknowledgment We would like to express our sincere thanks to all the respondents residing in hostel buildings for participating and answering the questions during the field study despite their busy working schedules. We also acknowledge the hostel staff and administrative authorities for helping the authors and surveyors throughout the field study. References [1] Thermal environmental conditions for human occupancy, American Society of heating, Refrigerating and air-conditioning Engineers Inc, ASHRAE (2013) 55–2013. [2] R.J. de Dear, G.S. Brager, Developing an adaptive model of thermal comfort and preference, ASHRAE Trans. 104 (1) (1998) 1–18. [3] M.K. Singh, S. Mahapatra, S.K. Atreya, Adaptive thermal comfort model for different climatic zones of North-East India, Appl. Energy 88 (7) (2011) 2420–2428. [4] J.F. Nicol, Adaptive thermal comfort standards in the hot and humid tropics, Energy Build. 36 (2004) 628–637. [5] A. Sharafat, M.R. Sharma, Tropical summer index – a study of thermal comfort of Indian subjects, Build. Environ. 21 (1) (1986) 948–960. [6] J.F. Nicol, S. Roaf, Pioneering new indoor temperature standards: the Pakistan project, Energy Build. 23 (1996) 169–174. [7] S. Heidari, S. Sharples, A comparative analysis of short-term and long-term thermal comfort surveys in Iran, Energy Build. 34 (2002) 607–614. [8] A. Tablada, A.M. De la Pena, F.D. Troyer, Thermal comfort of naturally ventilated buildings in warm-humid climates: field survey, in: Proceedings of the 22nd Conference on Passive and Low Energy Architecture, Beirut, Lebanon, pp. 1–6. [9] H. Feriadi, N.H. Wong, Thermal comfort for naturally ventilated houses in Indonesia, Energy Build. 36 (2004) 614–626. [10] A.T. Nguyen, M.K. Singh, S. Reiter, An adaptive thermal comfort model for hot humid South-East Asia, Build. Environ. 56 (2012) 291–300. [11] R. Yao, J. Liu, B. Li, Occupant's adaptive responses and perception of thermal environment in naturally conditioned university classrooms, Appl. Therm. Eng. 87 (3) (2010) 1015–1022. [12] M. Indraganti, Using the adaptive model of thermal comfort for obtaining indoor neutral temperature: findings from a field study in Hyderabad, India, Build. Environ. 45 (2010) 519–536. [13] M. Indraganti, R. Ooka, H.B. Rijal, G.S. Brager, Adaptive model of thermal comfort for offices in hot and humid climates of India, Build. Environ. 74 (2014) 39–53. [14] M. Indraganti, R. Ooka, H.B. Rijal, Field investigation of comfort temperature in Indian office buildings: a case of Chennai and Hyderabad, Build. Environ. 65 (2013) 195–214. [15] ISO 7730, Moderate thermal environment – determination of PMV and PPD indices and specifications of the conditions for thermal comfort, International Organization for Standardization, Geneva, Switzerland, 2005. [16] M.K. Singh, S. Mahapatra, S.K. Atreya, Thermal performance study and evaluation of comfort temperatures in vernacular buildings of North-East India, Build. Environ. 45 (2010) 320–329. [17] M.K. Singh, S. Mahapatra, J. Teller, Development of thermal comfort models for various climatic zones of North-East India, Sustain. Cities Soc. 14 (2015) 133–145. [18] M.K. Singh, R. Ooka, H.B. Rijal, M. Takashu, Adaptive thermal comfort in the offices of North-East India in autumn season, Build. Environ. 124 (2017) 14–30.

106