Analysis of Energy Efficient Features in Traditional Hill Architecture of Himachal Pradesh, India for its Indoor Performance in Sub-Tropical Climate

. Vernacular architecture of any region highlights the architecture and construction methods developed by the locals, in need to provide indoor thermal comfort amidst outdoor climatic conditions which is done using the locally available materials and construction techniques in response to the climatic conditions and geographic location. At present in India, standardized building materials and construction methods are adopted for modern buildings due to the easy availability of materials and also serving well economically. This paper presents the results of a field study and simulations to investigate the thermal and energy performance of selected case study, two traditional houses using adobe blocks, rammed earth and stone construction are compared to a modern house using a masonry envelope, the construction material is altered by adding insulation to the building envelope to study the differences in the performance of original construction. The case study is located in Dharamshala town in the upper reaches of Kangra valley, Himachal Pradesh, India having a sub-tropical humid climatic condition. Hourly indoor temperatures of the houses were monitored and a field survey was done to collect the data, and simulations were carried using dynamic thermal simulation software DesignBuilder, the monitored data was then used to match the parameters in simulation models. The results show that indoor temperatures of traditional mud house provide better thermal comfort when compared to modern house, however, slightly high energy demand in traditional house is observed due to increase in discomfort hours.


Introduction
The climatic conditions play a major role in the design and construction techniques used in vernacular architecture, but it's also the culture and historic background of people that shape this architectural style, same is the case in the settlement of Dharamshala town located in the middle Himalayan range of Himachal Pradesh, which uses indigenous materials like mud, stone, and wood in its traditional architecture.
In a modern context the need to reduce the energy consumption of buildings through sustainable development and energy efficiency has become a significant requirement, vernacular architecture can provide solutions to some of the problems faced by a modern conventional architectural style. According to the Government of India, statistics show residential and commercial spaces account for 32% of the country's total electricity consumption and overall building energy use consumes 37% of India's total annual primary energy consumption [1] . With the present standardized construction materials and methods in the Indian market, there will be a further increase in the use of energy in built spaces, thus there is a need to find solutions to improve the thermal performance of building envelope which is a primary source of energy loss. In this sense energy-efficient features of vernacular architecture have been highlighted by many authors in their studies [2], [3] , and concluded, to adopt certain traditional architecture and passive solar techniques to improve energy efficiency and sustainability in modern buildings. Many research tasks were done to study building envelopes, in this direction especially the earth construction and passive design features with different climactic conditions. Cheikhi et al. [4] led a study comparing the thermal performances of buildings using rammed earth and concrete masonry, the results show the differences in heating requirements and heat loss through the building envelope.
Amitava Sarkar et al. [5] conducted a comparative study in Mandi town of Himachal Pradesh, comparing the architecture and construction methods of traditional and modern houses, results show that traditional houses provide more thermal comfort. The winter indoor environment was comfortable without the use of any mechanical means when compared to the modern house. Heat loss from each construction, summer, and winter loads, and the thermal performance of rammed earth construction, stone construction, and concrete masonry in a sub-tropical humid climate are studied in this paper. Another comparative study conducted by Pragya Gupta et al. [6] compared rammed earth with assemblies of different thickness and insulation in hot desert climatic, the results show a significant difference in heating and cooling loads while indoor temperature difference of 5°C -6°C was observed when compared to uninsulated rammed earth, the composite walls show excellent thermal properties in hot climate zones but the influence of adding insulation to rammed earth walls and whether it is suited for the subtropical humid climate is not explored. This paper tries to stick to this research track, a comparative study between the thermal and energy performances of uninsulated rammed earth and insulated rammed earth construction, stone construction, and concrete construction in Dharamshala town, Himachal Pradesh is presented. The study is done in two phases, the first phase collects data in form of the indoor monitored temperature, relative humidity, and conducting field survey. Then the data is calibrated to carry out simulations, using thermal dynamic simulation of both variants, led by the software Design Builder [7] .

Study area
The state of Himachal Pradesh is located in the western Himalayas of India from 30° 22' 40" N to 33° 12' 40" N lat. and 75° 45' 55" E to 79° 04' 20" E long. covering a geographical area (2011) of 55,673 sq. km. The state is divided into 12 districts and is bordered by Jammu & Kashmir on the North, Punjab on the West and South-West, Haryana on the South, Uttarakhand on the South-East and China on the East [8] .
The study area, Dharamshala town is located in the Kangra district of Himachal Pradesh at an average altitude of 1,457m in the upper reaches of the valley (32.2°N lat., 76.32°E long.). The climate of the district varies from subtropical in low hills and valleys to sub-humid in the mid hills, and gets temperate in high hills and receives an average annual rainfall of about 205 cm that goes up from about 100 cm in southern parts to about 250 cm in northeastern areas [9] .

Methodology
The main aim of this study is to investigate the thermal performance and energy impacts of earth, stone, and masonry construction, affecting the indoor comfort conditions in sub-tropical climate zone. To achieve this a comparative study is done using an energy simulation model in DesignBuilder, a graphical user interface for Energy Plus. The study includes the following steps: 1. Three existing houses were selected, with similarities in orientation, shape, and passive design features. The first house is made of adobe brick wall, rammed earth construction and openings facing south, the second house uses traditional stone construction techniques, this house also has its openings facing south, and the third house uses masonry construction and benefits from its 'solar house plan'.
2. Typical Meteorological Year weather files of Dharamshala were used for simulations [11], [12] and outdoor climatic data was collected from the state meteorological department. Hourly indoor temperatures and relative humidity were recorded in the living room of these houses using handheld equipment see 'figure 1' in winter and summer months for a period of 7 days (January 2022 and July 2022), living area was chosen to record the temperature, based on occupants' information on the occupancy of the building. A survey questionnaire was also prepared to gather information on the occupants' background, activities, clothing, and activity rate for thermal comfort and to calibrate the simulation model 3. The results obtained from the simulation are then compared, indoor temperatures are examined and two houses are chosen which perform the best. Insulation is added to the walls and roof and their performance is then compared to the uninsulated construction material. Also, to quantify the energy consumption of the chosen two variants, thermal balance was added by selecting an ideal profile for HVAC. However, it must be noted that all three actual houses did not use mechanical heating and cooling therefore a free-run (non-heating and cooling) mode for simulation was used to calculate predicted thermal comfort in terms of annual discomfort hours and Kansas TSV (thermal sensation vote) in correlation with indoor temperatures.

A Case study of vernacular architecture of Dharamshala, Himachal Pradesh
The vernacular houses of Himachal Pradesh vary from region to region depending on the climatic conditions, which use locally available resources for construction and incorporate a number of passive design features. Most of the houses follow a 'solar house plan' to capture maximum sunlight in winter by using spaces like 'verandah' and placing bedrooms, living rooms, and stores according to the direction of the sun. Some features like small window size in the north to protect against cold winds and wider windows in the south for solar gains, and walls with thickness ranging from 300mm -600mm to provide indoor comfort are seen throughout the vernacular architecture of the state.
Dharamshala, see 'figure 2' lies in the middle Himalayan range and experiences a sub-tropical climate that is affected by the monsoon. The mean minimum temperature in winter, which is December to February, is 6.7°C while the mean maximum is 17°C, which can sometimes go below 0°C. In summer (April to August) the mean minimum temperature is 18.6°C and the mean maximum is 28.5°C [13] . Relative humidity in the region is around 66% although it varies from around 42% during Summer (May) to 89% during the Monsoon (August) [14] . Mean solar radiation is 12.3 Mj/m² in winter and 22 Mj/m² in summer [15] . ' Figure 3', shows global and diffuse solar   The houses are compact one to two storeyed high with rectangular or square form, with gable roofs of slates, walls of mud, dry stone, and wood. The architectural features and materials used have thermal insulation properties which improve the energy efficiency, the identified features are shown in Table 2.

House details
The first house is 80m², two-storeyed high, both external and internal walls are constructed of adobe blocks and have a thickness of 500mm. The floor is made of rammed earth mixed with mud-phuska and coated with cow dung, the roof construction is a sloping gable roof with slate shingles and an attic that acts as an insulation chamber, 'figure 4' shows the ground floor plan of the mud house. The house is occupied by one person, and the occupant mostly spent the time in the living room, this space was used to monitor the indoor temperature where the handheld equipment was placed. The living room is on the ground floor, facing south, the south wall has two windows with a glazing window-to-wall ratio of 21%. The second house is 87m², three storeyed high, both external and internal walls are constructed of dry-stone blocks with mud plaster inside having 500mm thickness, unlike the first house wall construction is slightly different, the wall uses braced timber framing filled with layers of stone joined together with an iron hook. These houses are compact and have less window-to-wall ratio due to the structure of the building, it's rare to find such houses anymore as most of it have been demolished in the region or abandoned and are now are usually found in the upper colder regions of the state. The floor is made of rammed earth mixed with mud-phuska and is supported by bamboo beams. The roof of this house has thicker slates, 'figure 5', shows the first-floor plan of the house. The house has one occupant, the living area (monitored area) has three windows on the east side which have a glazing windowto-wall ratio of 13%. Occasional gas heating was used in winter and during monitoring week the occupant was requested to avoid the use of heater. The third house is 122m², three storeyed high and is square shaped, the wall construction is a typical brick masonry having 250mm thick external and internal walls.
The house follows the 'solar house plan' and has large window openings on each side resulting in high solar gains from external windows. The floor slab is tiled, and the roof is flat which uses typical RCC construction. The house is occupied by three occupants and occasionally four. Similar to the first two houses living room in the north was used for monitoring and occasionally the portable heater was used in the bedroom, however, the bedroom facing south was also monitored because of its window-to-wall ratio of 80% resulting in overheating of the spaces in summer when compared to the space in the north. ' Figure 6', shows the ground floor plan of house three.

House modelling and parameters
The building geometry was modelled using Revit for BIM modelling, thermal properties of the materials was included within the BIM and each space in the buildings was zoned into room volume. The BIM model  was then exported to DesignBuilder using gbXML file format which contains all the geometry information, construction materials, and their thermal properties. Published data was used to calculate the thermal properties of the materials used in house modelling [16], [17], [18] see Table 3. Existing overhangs in the buildings were added for simulation modelling as this was a sensitive input that affected the accuracy of simulation results. It must be noted that the kitchen zone used in the simulation model for house one and the height of the first floor in house two is slightly altered as the original layout from Revit did not produce the correct model resulting in the failure of the simulation, therefore modification of kitchen wall in house one and adjustment in the height of the house two to was carried out. ' Figure 7', show the simulation model used in DesignBuilder. Natural ventilation was only applied in summer to all three houses as the windows were shut in winter, the ventilation rate was estimated based on the window openings observed during summer design week monitoring. A simplified method was used according to ASHRAE 1997b [19] with the following equation.
where Q is the volume flow rate of air (m³/s), Cv is the ventilation coefficient assumed to be 0.2 from the opening, A is the area of the opening for ventilation (m²) and V is the wind velocity outside the building (m/s). The ventilation had minimal impact on the simulation results when compared to scheduled based ventilation as the ventilation rate was based on the zone which is defined in the activity template. The Model infiltration in all three houses was assumed to be 0.7 ac/h, 24/7. The simulation results analyzed are based on the indoor temperature in free-run mode and HVAC ideal load template is added to quantify the energy consumption. The parameters detailed in Table 4 were kept constant, while the wall material and free run mode were altered.

Simulation results
Hourly temperatures of summer and winter design weeks were simulated in all three houses, living room data was then compared to the monitored data to ensure the calibration of obtained results. 'Figures 8', 'figure 9', and 'figure 10' show the comparison of the simulated and measured temperatures for the summer design week simulation from July 03 to July 09, the living room indoor temperatures of the first house (earth construction) show close agreement when compared to the recorded temperatures, the correlation coefficient (R²) obtained was 0.832. Similarly in the living rooms of house two (stone construction) and house three (brick construction) acceptable correlation coefficient was obtained, 0.737 and 0.634 respectively. However, it was noted that in house three, there was a slight difference in the monitored data and the obtained simulation temperatures during the summer period, which showed the simulated temperatures were slightly higher than the measured temperatures because natural ventilation was affected, due to the use of window openings which was assumed to be open during calculation but were in fact kept closed, mostly in the afternoon as the temperatures during this time peaked highest. House three shows the highest indoor temperatures in summer week, firstly due to its envelope with less thermal mass and lastly, the solar gain from exterior windows, while the maximum indoor temperature of the living room in the north reached 30.5°C, the bedroom in the south overheated with a maximum temperature reaching 31.8°C.    Winter design week simulations were carried out from January 18 to January 24 as shown in 'figure 11', 'figure 12', and ' figure 13' the simulated and measured indoor temperatures of the living room were compared and a correlation coefficient of 0.954 obtained was obtained for house one, 0.9085 for house two and 0.9752 for house three. Since the windows were shut during winter no ventilation rate was calculated for this period.

Discussion on the Performance of the houses
The summer design week simulation models for non-airconditioned houses were used for further analysis as the results obtained were acceptable.  Table 6 for minimum acceptable indoor temperatures in summer and winter. It was noted that for house three the rooms in the south had high indoor temperatures when compared to rooms in the north with a temperature difference of 2°C, this was due to high solar gains from exterior windows in the south which have a glazing window-to-wall ratio of 80%. The winter design week simulations were carried out and the obtained simulated results were in close proximity to the measured data. Simulated temperatures show that house two and three were comparatively cooler in winter than house one with a temperature difference of 3°C -4°C. The minimum temperatures in the living room of house one, house two, and three were 10.5°C, 8°C and 7.8°C respectively while the minimum outdoor temperature was 0°C. The average indoor temperatures in house one, house two, and house three were 11.9°C, 9.3°C and 9.7°C, while the outside average temperature was 4.1°C. Due to high solar gain in the south zone of house three, the indoor temperature differences in the north and south rooms were 13°C during the day where the temperature in the south zone got as high as 26°C. ( Figure 14. Heat loss (kW) in house one (a), two (b) and three (c).
Thermal loss in the building envelope was also simulated, which shows heat loss (kW) through various elements of the houses. ' Figure 14' (a), (b) and (c) show heat loss in all three houses, in all three cases highest heat loss is from the walls, roof, glazing, and ceilings. House three glazing has significant heat loss which explains the indoor temperature differences in the north and south zone.

Altering material
The walls and roof of the house one and house three were insulated by adding 50mm of expanded polystyrene (EPS) on the inside, the adobe wall thickness was changed to 450mm, and 50mm of EPS was added, for brick masonry wall thickness was changed to 220m and 50mm of EPS was added, Table 7 shows the thermal properties of the added insulation material. For better results, glazing in the new simulation model was unchanged as high heat loss only occurred in house three, and the free-run mode was continued in these simulations too. The results of house one and three with insulated envelope in 'figure 15', and 'figure16' show that, in summer design week both the houses maintained comfortable indoor temperatures, the difference between indoor temperatures of insulated and uninsulated brick masonry was 5°C. In house one indoor temperature difference of 1°C -2°C was noted, showing similar performance in summer when insulation is added. While in winter 'figure 17', and 'figure 18' show that insulated earth construction has warmer indoor temperatures, with a temperature difference of 3.5°C when compared to uninsulated earth construction, and in insulated brick construction, the indoor temperature     difference was 4°C, when compared to uninsulated brick construction, while the temperature difference between uninsulated rammed earth and insulated brick masonry was 2°C. Thermal loss in the added insulation simulation model, ' figure 19' (a) and (b) show that heat loss in house one is quite similar to the uninsulated model with a slight improvement in heat loss from the roof, however, in house three the insulated simulation model showed a significant reduction in heat loss from the walls, roof which also resulted in reduced heat loss through external windows.
(a) (b) Figure 19. Heat loss (insulated walls and roof) in earth construction (a) masonry construction (b).

6.2.1Kansas TSV
Kansas TSV (thermal sensation vote) is a comfort-related output used for thermal comfort analysis in DesignBuilder that predicts thermal sensation (TSV). TSV is calculated using KSU's two-node thermal comfort level which is the change in thermal conductance between the core and the skin temperatures, these changes are based on skin wetness in a warm and cold environment. The parameters used in calculating this output are the indoor environmental conditions (temperature, humidity, air speed), the number of occupants, and their comfort levels.
To obtain close results, occupants' actual activity level, and clothing reference were used for winter months calculation, and simulated humidity and temperatures were compared with the monitored data used in winter design week to calibrate the simulation model for winter months simulation. House one model performed the best showing better indoor comfort conditions and therefore was used in this simulation, also note that summer months were not calculated for predicting comfort, as indoor comfort conditions were better in the house when compared to winter months.
Where TSV is the thermal sensation vote, and T op is the indoor operative temperature. The simulated indoor comfort optimum temperature for winter months is 15.9°C ('0' TSV) according to the linear regression analysis, the thermal comfort range is 14°C -17°C ('±0.5' TSV) and the minimum acceptable operative temperature is 12°C ('-1' TSV). As per ASHRAE Standard 55 [21] thermally acceptable limit is -1 < TSV < +1, the results show a difference of 3°C when compared to the minimum indoor comfort temperature mentioned in NBC, 2016.

6.2.2Discomfort degree hours
Predicting discomfort conditions based on hours of discomfort (summer and winter), by using adobe brick walls and insulated adobe brick walls, that is indoor conditions requiring additional heating when it is cold, reduced requirement results in better performance. Using the relationship proposed by Humphreys [22] , the optimum comfort temperature in a non-ventilated building was calculated using the equation: T c = 11.9 + 0.534 T o ; 10°C ≤ T0 ≤ 33.5°C Where T c is the optimum comfort temperature and T o is the monthly mean outdoor temperature. This equation was used to calculate the optimum comfort temperature for winter months (December to February) with mean outdoor temperature of 6.7°C, and the calculated optimum comfort temperature of 15.5°C was obtained. The result is in close agreement with the linear regression analysis (2) where the simulated optimum indoor comfort optimum temperature for winter months is 15.9°C corresponding to '0' TSV. It must be noted that the calculated optimum temperature of 15.5°C is not the actual comfort temperature, the occupant of house one did not use any additional heater even when the temperature went below 12°C, whereas the occupants of house three occasionally used portable heaters when the temperature went below 16°C. ' Figure 21', shows the comparison of simulated indoor temperatures of all three houses during winter months with minimum optimum indoor temperatures according to NBC, 2016. The results show December and January are the coldest months where indoor temperatures are below 15°C, earth construction performs the best, while in February month earth and stone construction, indoor temperatures are similar and in masonry construction temperatures spike during daytime reaching a maximum of 21°C. Original wall construction and insulated wall construction model of house one and three were used to compare predicted annual discomfort degree hours. The results show 2790 annual discomfort hours in house one, whereas the insulated envelope showed 1916 discomfort hours, resulting in a 31% reduction. In house three brick masonry construction, discomfort hours were 2479 while the insulated envelope had 1797 discomfort hours, resulting in a 21.5% reduction.

6.3.1Heating and cooling demand
The wall and roof construction of house one and three were altered by adding 50mm of EPS insulation, to quantify the energy consumption in the original house construction and modified construction, mechanical heating and cooling system, using ideal load template was added to the simulation model. Table 8 shows annual heating demand of 74.974 kWh/m², this high demand is due to reduced solar gain in house one in winter, and cooling demand of 6.728 kWh/m² in summer. The insulated wall and roof construction of house one shows a heating demand of 23.156 kWh/m² and a cooling demand of 1.048 kWh/m², showing a significant 70% reduction in total energy demand. For house three with original construction annual heating and cooling demands were 36.09 kWh/m² and 7.22 kWh/m² respectively, in the insulated construction annual heating and cooling demands were 22.96 kWh/m² and 10.94 kWh/m² as seen in ' figure 22' (a) and (b), rise in cooling demand was observed due to poor performance of window glazing resulting in increased solar gain, annual energy demand for the insulated house was reduced by 21.7%.  ' Figure 23' (a) and (b) summarizes the total energy demand in house one and three.
(a) (b) Figure 23. Annual energy demands of house one (a) and house three (b).

6.3.2Heat Loss (kW)
Original house construction and insulated construction of house one and three were compared using a simulation model to get heat loss (kW) from various elements of the building see ' figure 24' (a) and (b), the results show walls, roof, and glazing to be the major source of heat loss from the building envelope. House one has an insignificant change in heat loss when insulation is added while house three has a significant improvement in heat loss when insulation is added.
(a) (b) Figure 24. Maximum heat loss in house one (a) and house three (b).

6.3.3Thermal comfort
It is noted that the number of hours of discomfort (summer and winter) in house one is slightly higher than house two see 'figure 25', this suggests additional heating, resulting in higher energy use in comparison to house three. Using an insulated adobe wall the discomfort hours reduced significantly and resulted in similar discomfort to the insulated masonry.

Conclusion
The comparative study of the traditional and modern houses is presented to analyse the energy-efficient features, affecting the indoor performance, in the traditional hill settlement of Dharmshala, Himachal Pradesh. The main features identified that affect the indoor performance in traditional houses are the building material that is earth and stone construction, roof materials, openings, and the orientation of the building. The high thermal mass of earth and stone construction helps in maintaining indoor comfort conditions in winter without additional heating, and in summer optimum indoor temperatures are maintained when compared to modern brick masonry house. The predicted simulation results show that in an insulated mud house and uninsulated mud house, the temperatures in summer are comparable, however, in winter the temperature in an insulated mud house is high with a temperature difference of 3.5°C. During winter, house one temperatures are comparable to that of the insulated house three where the temperature difference is 2°C, while in house two indoor temperatures in winter were as low as 8°C (simulated) and 7.6°C measured.
The predicted thermal comfort using Kansas TSV and discomfort hours show that, according to ASHRAE -1 < TSV < +1 thermally acceptable range, mud house shows the minimum acceptable indoor optimum temperature of 12°C, the simulated temperatures in winter design week shows a similar range of 10.5°C -12.8°C which is in close agreement with the monitored data, however, this minimum acceptable temperature is below the thermal comfort temperature mentioned in NBC, 2016. The predicted annual discomfort hours show a 31% decrease when the mud house is insulated, it must be noted that discomfort hours in house one is 11% more than house three indicating more heat is required, this is also represented in annual heating and cooling demand.
Since no additional heating devices was used in house one, the utility records (electricity) show annual energy use of 321 kWh per person per year, while in house three occasional additional devices were used in winter and summer resulting in annual energy use of 4120 kWh. This energy use is not a direct result of wall material, rather it shows the perspective of the occupants in terms of thermal comfort. The results obtained from the comparative study of the thermal and energy performance of the traditional and modern house conclude the following: 1) Earth construction is one of the main energyefficient features of vernacular architecture, the same is seen in the region of Dharamshala. Due to its thermal properties and availability, it is used widely in traditional houses and provides better indoor temperatures when compared to brick masonry, however, to be used in modern architecture it's durability and compressive strength need to be improved. 2) The orientation of the building should be the in direction of the sun, in this case south, with open space in the front 'verandah', which captures the sunlight keeping indoor temperatures warm in winter and cooler in summer by using overhangs. The roof in mud houses gets heated up in summer but the attic acts as an insulation chamber keeping the ground floor cooler when compared to a flat RCC roof which overheats the floor.
3) Window openings in the traditional house affected the energy demand, high heating demand was noted due to reduced solar gain from external windows in the south as the glazing window-to-wall ratio was 21%. Brick masonry with 80% glazing window-to-wall ratio shows better indoor temperatures in the south in winter months while overheating in summer.

4)
Insulation of mud house show better indoor temperatures in winter and summer with reduced energy demand, this is due to the decrease in heat loss (kW) from walls, roof, and windows. When compared to uninsulated mud house the only comparable difference was in reduced energy demand of 70%. 5) Insulated mud house show a reduced annual energy demand of 28% when compared to insulated brick masonry. Adding insulation to brick masonry house show high efficiency in terms of heat loss (kW), energy demand, and discomfort hours which were all reduced significantly with better indoor temperatures in summer and winter months when compared to original brick masonry.
This study shows that, in the sub-tropical humid region of Dharamshala, the indoor performance of the traditional house is better when compared to modern brick masonry in terms of its energy-efficient features, confirming the results of other studies, however, adding insulation works best for increasing the energy efficiency of brick masonry houses which increases the time lag of heat transfer, for better indoor temperatures but insignificant changes are seen in a mud house, with significant changes in energy demand and a slight decrease in annual discomfort hours. Heating demand in uninsulated mud houses can be decreased by improving the area of the opening, not more than 40% of the wall area in the south, and using better shading devices and overhangs. Adding insulation in a mud house in winter months does improve comfort but a minimum optimum comfort temperature of 15°C is not maintained, hence further studies can be conducted by using different insulation materials and altering the adobe block thickness to analyse the effect on the thermal performance of an envelope in a sub-tropical climate. The Himachal Pradesh government has introduced laws to make it mandatory to include passive design features in modern buildings, however using concrete and brick masonry construction in the settlement of Dharamshala is not sustainable and does not provide a thermally acceptable indoor environment.