Nowadays, building cooling loads can be predicted using specialized software that predicts the behaviour of the energy system according to various parameters and conditions such as weather, time of the year and geographical location. In this study the Design Builder V.6, EnergyPlus, and Element software are employed to design and predict the indoor microclimatic conditions. Design Builder is used to define the characteristics of the space, materials and dimensions, thermal zones, activity, weather and every other requirement for the study. Element is used to create the weather data of the experiment location, and EnergyPlus, a simulation engine in the Design Builder software calculates the thermal and energy performance. The computer model is further validated through field experiments that replicate the computer model.

The thermal comfort, energy consumption and overall sustainability of a building is contingent on the indoor microclimate of the building. The indoor thermal characteristics of a building depends on heat transfer in the building.

Heat transfer between a building and its surrounding environment provides a foundation for thermal modelling and occurs in primarily three ways, conduction, convection and radiation.

Conduction occurs when heat transfers from one part of a body at a higher temperature to another at a lower temperature. [23] Conduction is calculated by Fourier’s law which states that heat flux is proportional to the temperature gradient in the direction of the outward normal:

${\stackrel{\cdot }{Q}}_{cd}=k\text{}A\frac{△T}{D}$ (1)

where are: Q_cd - quantity of heat flow/heat flux due to conduction in W; k - thermal conductivity of the material in W/m/K; A - surface area in m²; Δ T-temperature difference of the two media in K; D-thickness of the material in m.

Radiation is heat transfer mechanism that does not require any medium of propagation. In this study, we consider thermal radiation is as it is required to calculate solar gains and heat

Input into buildings from the sun. According to [24], the radiation exchange between two large parallel plane surfaces (of equal area A) at uniform temperatures T₁ and T₂ respectively, can be written as:

$Q_{12}={\epsilon }_{\text{eff}}A\sigma \left(T_1^4-T_2^4\right);$ (2)

${\epsilon }_{\text{eff}}={\left[\frac{1}{{\epsilon }_1}+\frac{1}{{\epsilon }_2}-1\right]}^{-1},$ (3)

where are: Q₁₂ – net radiative exchange between surfaces in W; σ- Stefan-Boltzmann constant (5.67×10^-8 W/m²K⁴); A- area of surface in m²; T₁ - temperature of surface 1 in K; T₂ - temperature of surface 2 in K; ε₁ and ε₂ - emissivities of surfaces 1 and 2 respectively.

Heat transfer by convection, takes place at the surfaces of walls, floors and roofs. According to Newton’s law, convection can be given as:

$Q_{\text{cv}}=h_{\text{c}}A△T,$ (4)

where are: Q_cv - convective heat transfer (heat flux due to convection) in W and h_c - heat transfer coefficient in W/m²K.

The energy balance in the building is established which is the summation all the heat gains and losses associated with the building through the convection radiation and conduction can be expressed as:

Cooling Load = ∑Internal Gains + ∑Solar gains + ∑Lightening gains - ∑Fabric losses - ∑Ventilation losses

According to [21] in his study on creating a thermal model called GRUBCLIM, explained that heat is transmitted through heat conduction through materials and through radiation and convection from surfaces and mediums. These can be categorized into the following: Q̇_W - energy loss through the walls in W; Q̇_F - energy loss into the in W; Q̇_G - energy loss hrough the glazing (window) in W; Q̇_R - energy loss through the roof in W, and Q̇_D - energy loss through the door in W. Therefore, the energy balance for the urban building can be generalized as:

${\stackrel{.}{Q}}_{\text{T}}={\stackrel{.}{Q}}_{\text{W}}+{\stackrel{.}{Q}}_{\text{F}}+{\stackrel{.}{Q}}_{\text{G}}+{\stackrel{.}{Q}}_{\text{R}}+{\stackrel{.}{Q}}_{\text{D}}$ (5)

For this study, ventilation was restricted and no windows were created in order to establish boundary conditions. Therefore, energy balance for this model can be expressed as:

${\dot{Q}}_{\text{T}}={\dot{Q}}_{\text{W}}+{\dot{Q}}_{\text{F}}+{\dot{Q}}_{\text{R}}$ (6)

The heat gain/loss primarily occurs through conduction as energy is transferred from the more energetic particles of a substance to adjacent particles that are less energetic due to interactions between particles [21]. Conduction is quantified by Fourier’s law that says heat flux is proportional to the temperature gradient in the direction of the outward normal.

Considering the transfer through, the basic equation of heat conduction is given by equation (1).

For this study, EnergyPlus, a simulated engine designed to work seamlessly in the Design Builder performs all required analysis pertaining to the building energy model (BEM) and generates the results within the Design building interface. Design Builder offers a range of functions including energy designs in buildings, ventilation, carbon and comfort measurement and regulation in buildings. It compares alternative building designs by using function and performance-based method of assessment in a speedy and economic style. The design builder interface is divided into nine sections, namely; layout, activity, construction, opening, lightening, HVAC, generation, output. In order to restrict boundary conditions for this computer model only the layout, construction was mostly utilized. No opening was constructed in the geometry of the boxes and the activity within the model was set to zero which implicates that there was no human interaction within the boxes during the simulation process. Lightening, HVAC systems, and power generations was disabled in each individual tab for the test objects.

The outdoor temperature and relative humidity readings of Jomo Kenyatta University of Agriculture and Technology (JKUAT), the location of the experiment were taken between May 7, 2019 to May 28, 2019, and utilizing Elements, a software by Big ladder used for editing and creating weather files, and obtaining secondary data of other weather parameters from the EnergyPlus website, a weather file of the test location was created and stored in the design builder software.

To predict the behaviour of the boxes, design began from the layout interface where the design conditions and zones were set, and the weather data created with Elements was selected for the as the site metrological conditions. The 3D physical geometry and dimensions of the boxes is specified in the layout section using component blocks and zones. In the construction tab of design builder, most widely used construction materials are available for selection, and for this study, the material, characteristics, and thickness, of each component of the boxes is specified also. The designed material characteristics and their corresponding U-values are provided in Table 1.

After specifying the conditions required to run the simulation including time step for simulation, the EnergyPlus simulation engine integrated into the Design builder software was run. In EnergyPlus, the heat balance model considers a scenario where the room as a uniform temperature and the air is distributed evenly. throughout and also operates on the assumptions that room surfaces have a uniform surface temperature and short and long-wave radiation, internal heat conduction, and diffuse radiating and reflecting surfaces [25]. In the Design Builder modelling environment, green roofs are calculated using the simple moisture calculation method. The principle is founded on a constant moisture osmosis into the soil. For every soil property updated, it analyses the more saturated layer and transfers moisture to the less saturated layer and distributes the moisture at a steady rate.

This study focuses mostly on the comfort level data, which includes, the envelope’s internal temperature, the outside dry-bulb temperature, and relative humidity. The comfort data is calculated for blocks using floor area weighted averages of the zones in the block. For example block average air temperature:

$T_a=\left(A_1{T}_{\text{az}\text{,1}}+A_2{T}_{\text{az}\text{,2}}+\cdots +A_{\text{n}}{T}_{\text{az}\text{,n}}\right)/\left(A_1+A_2+\cdots +A_{\text{n}}\right),$ (7)

where are: A_n - floor area of zone nand T_az,n - the air temperature of zone n.

Preference for comfort temperature data was ticked while specifying the computer model parameters.

The predicted result is conferred in the results and discussions section.

To validate the predicted thermal behaviour of the boxes developed in design Builder, two exact replicas of the boxes designed in Design Builder were fabricated. The boxes measured 1m by 1m by 2m in length, width and height respectively.The experiment was carried out in Jomo Kenyatta University of Agriculture and Technology, Juja campus, Kenya, an urban area, characterized by a dense population of staff and students, with paved areas, concrete buildings and high vehicular activities. Juja is located in Kiambu county, Central Kenya, the highlands characterized by altitudes between 1,400 m and 1,800 m above the mean sea level. The county is situated between latitude 0°75' and 1°20' south of the equator and longitudes 36°54' and 36°85' east. It experiences bimodal rainfall from March to May and in October and November with annual rainfall above 1,500 mm. [26]. The test objects are fabricated with mild steel sheets, and are located in an open field away from existing buildings that could cast a shadow on or shade them.

The temperature of the indoor space was recorded using a temperature and relative humidity sensor, the HOBO MX1101 sensors. The Onset’s HOBO MX1101 data logger is powered by two AAA 1.5 V alkaline batteries, user replaceable battery, and measures and transmits temperature and relative humidity data wirelessly to mobile devices via Bluetooth Low Energy (BLE) technology. The self-contained wireless data logger, which works with Onset’s free HOBOmobile app for logger setup and data management, enables you to access data anytime from mobile device for iOS or Android over a 100-foot range, and it requires no dedicated equipment beyond a mobile device for configuring the logger or reading out data. The sensors was suspended within the test objects by a string and magnet attached to the inner surface of the boxes at the top. This was done to ensure the uniform temperature within the test objects were recorded. The outdoor temperature and relative humidity were measured with a temperature and relative humidity sensor, the HOBO MX2302 Affixed to pole 5 meters away from the experiment setup with other weather data. Each test object had a single HOBO MX1101 sensor affixed to its roof by a string, and a single HOBO MX2302 was used to collect external temperature and relative humidity data. The Onset MX2302A HOBO Weatherproof External Temperature and Relative Humidity (RH) Data Logger has external temperature and relative humidity sensors. Leveraging Bluetooth Low Energy, the logger enables easy, fast set up and data download directly from your mobile device, and provides high accuracy measurements in harsh outdoor environments.HOBO sensors have been used in predicting the envelope performance in various studies (Gulma, 2014). All measured data was saved in the logger, and once within range of the configured smart phone, uploaded the data into the smart phone. The sensors implemented measure temperature of between -20 to 70 °C accurately to +/-0.21 °C from 0 to 50 °C. The loggers have a radio power of 1mW, a logging rate of 1 second to 18 hours, and a time accuracy of ±1 minute per month at 25 °C. Measurements were taken every 30 minutes. The experimental set-up is shown in Figure 1 below.

Figure 1

The front view of the field test object A (a), test object B (b) and schematic diagram of experimental setup (c)

Temperatures inside the test objects were recorded throughout May, 2019. The computer model was validated using data measured from the experimental set-up with exact dimensions and specifications as used in the computer model. The validated data was recorded between 8 May and 28 May, 2019.

Variations in predicted indoor temperature and the recorded temperatures in the field experiment are shown in Figure 2. Figure 2 shows a generally favourable correspondence between the recorded and predicted values. The graph has a correlation coefficient of 0.85. The peak temperature recorded was 305.15 K compared to the predicted indoor temperature with its peak temperature at 306.1K. Also, coldest temperature value recorded was 288.11K as compared to a 287.12 K temperature value of the predicted indoor temperature.

A two-sample z-test was calculated to establish the relationship between both set of values. The test solved a p-value of 0.1952 against the significance level of 5%. There is a difference between both set of values as shown in the table 2 below.

Figure 2

Graph showing the predicted air temperature from the Design Builder software, the recorded temperature

The results depicted Figure 3 shows the Comparison of the ambient temperature and indoor air temperature in test object A incorporated with passive cooling and test object B without any modification over the 20 days experiment period. The amount of temperature reduction achieved through the combination of three passive cooling methods showed a 14% reduction in temperature with a mean temperature difference of 277.03 K. The highest difference measured during the day was 290.15 K at 3.00 pm on 21^st May, 2019. The significant difference in indoor temperature readings in test object A (passive cooling) and test object B represents a potential for reducing the operating time of air conditioners for cooling purposes indoors. Less cooling demand means a shorter operating time to bring temperatures down to acceptable room temperature of 297.15 K, and less energy consumed. Other important temperature values are displayed in Table 2.

Figure 3

Comparison of the external ambient indoor air temperatures in Test object A (passive cooling) and Test object B

This result shows passive cooling effectively provides a reduction in thermal gains and less erratic temperature fluctuations. This can be attributed to; reduced heat gain by the green roof and walls coated with cool paints compared to bare roof and walls due to incident radiations hindered by the vegetation, soil mass, and incident radiations reflected outwards by the cool paint due to its low solar absorptance. Reduction in internal temperature can also be associated with the evaporative cooling due to irrigation and the increase in heat resistivity caused by lower thermal conductivity in the plants in comparison to the bare metal substituting for an insulation to solar gains by the roof due to the watering of the plants and heat resistance caused by the low thermal conductivity of the plants in comparison to bare metal acting as heat insulators to the heat gain by the roof. In addition, increase in thermal insulation due to the green roof indirectly reduces the energy consumed in heating as heat loss will be reduced too. Moreover, in the test objects, there was no provision for ventilation systems, thus the internal temperature was more stable and less heat was lost.