ABSTRACT

Amidst the recent development in the usage of curtain walls for office buildings, high utilization of energy and poor thermal comfort issues have become paramount. This paper assesses thermal comfort in multi storey (naturally and mechanically ventilated) office buildings in Accra, the Capital city of Ghana using Fanger’s Predicted Mean Votes (PMV) and Predicted Percentage of Dissatisfied persons (PPD) model. The model relates to the American Society of Heating, Refrigerating and Air-Conditioning Engineers (ASHRAE) Standard 55 (which gives an acceptable temperature range of 23˚C - 26˚C) and the International Standards Organization (ISO) 7730. Indoor environmental parameters (temperature and relative humidity) of 4 multi storey office buildings were recorded over a 10 month period. The environmental parameters were analyzed using PMVcalc_v2 software which resulted in the generation of PMV-PPD values. The findings reveal high PMV-PPD values in the Naturally Ventilated Building (NVB) whiles the Mechanically Ventilated Buildings (MVB) fall within the comfort zone. Meanwhile, the Actual Mean Votes (AMV) by the occupants suggest all four buildings are relatively comfortable with the mechanically ventilated offices being more comfortable. Additionally, it is recommended that buildings are orientated with their longer sides facing north-south, with enough shading in order to improve the thermal comfort conditions of work spaces.

Keywords:Thermal Comfort; ASHRAE Standard 55; Office Buildings; Natural; Mechanical

1. Introduction

Man has over the years been successful in creating environment that is conducive for all kinds of activities. One of the basic requirements of this environment is to maintain thermal conditions comfortable for occupants as these have a direct effect on health, productivity and morale [1] . What is more, [1] ascertains that thermal comfort is attained when a thermal balance is achieved: a situation in which no heat storage occurs in the body. Although this can be achieved over a wide range of thermal environmental conditions, thermal-comfort is associated with conditions to which the body can readily adjust. Reference [2] defines thermal comfort as the “state of the mind that expresses satisfaction with the surrounding environment”. Standards such as ASHRAE Standard 55 [2] and ISO Standard 7730 [3] have been the basis in the determination of thermal comfort by most practitioners and researchers [4] . Nonetheless, people in different climatic zones feel comfortable at different indoor air temperatures: a situation which can vary considerably from that of the world standards.

The factors affecting thermal comfort depend on four environmental (i.e. dry bulb temperature, mean radiant temperature, relative humidity and air velocity) and two personal (clothing-insulation and physical activity) parameters [5] . The thermal comfort standard prescribed by ISO 7730 was the first to have been used on a worldwide basis [6] . It’s based on Fanger’s work in climate chamber experiments using young Danish students and eventually the PMV model.

Architecturally, the hot and humid regions are the hardest climates to improve through design: due to high outdoor thermal conditions (humidity and temperature) that result in soaring indoor temperatures exceeding the ASHRAE summertime comfort upper limit of 26˚C. A temperature limit should at least get 90% of occupants feeling thermally comfortable for most part of the year [7] .

1.1. Problem Statement

In recent years, most multi storey office buildings in Ghana have become what is termed the “glass box” exposing all or most part of its facades to the high day time temperatures. The effect of which is the contribution of thermal discomfort within the indoor spaces. In an attempt to ameliorate the effect of this exposure, multi storey office buildings have resorted to the use of complex HVAC systems which comes with its problems such as high usage of energy and greater emission of Carbon dioxide (CO₂) into the atmosphere.

1.2. Objectives of the Paper

The objectives of the research include the followings:

• To measure indoor and outdoor thermal comfort parameters of the selected buildings;

• To compare the predicted mean votes (PMV) within the office with the actual mean votes (AMV) given by the occupants.

2. Literature Review

The PMV Model

The Fanger-based approach, which is alternatively known as the heat-balance approach, is built on experiments in climate chambers and eventually theorized into the PMV-PPD model. It was fist propounded by Fanger around the 1970s with about 1296 participants dressed in standardized clothing and completed standardized activities constituting a metabolism rate of 1.2 Mets or 70 W/m². The predicted mean vote (PMV) equation has been used in international standards to predict thermal sensation of occupants. It has been widely used by designers to assess indoor thermal environments [8] .

The PMV equation uses the four environmental parameters; air temperature (˚C), mean radiant temperature (˚C), air velocity (m/s), relative humidity (%) and two personal variables (clothing insulation (Clo) and metabolic rate (Met)). The aforementioned when calculated, forms an index that can be used to predict the thermal comfort levels within a space. The index provides a score that corresponds to the ASHRAE thermal sensation scale and represents the average thermal sensation felt by a group of people in a space [9] [10] . Table 1 shows the seven point ASHRAE thermal sensation scale.

Overall, thermal comfort field studies [8] [11] have suggested that the PMV model does not always accurately predict the actual thermal sensation of occupants. Researchers have cited measurement error and contextual assumptions as the two contributing factors to the discrepancies [12] . Measuring of the physical variables with reliable instrument, clothing insulation, activity levels, individual differences, building differences, outdoor climate, behavioural and psychological adaptation are all factors that have been mentioned by thermal comfort

researchers as factors contributing to the bias nature of the PMV model [12] -[14] . The PMV model prediction has been found by a majority of researchers as a better predictor in mechanically ventilated buildings [15] and [16] . In the case of naturally ventilated buildings, differences have been observed between the thermal sensations reported by the occupants and that of the PMV model [17] . Studies have shown that the PMV either underestimates or over-estimates the thermal sensation of occupants. Reference [18] and [19] for instance observed the underestimation of thermal sensation by PMV model in their respective studies.

Air temperature is often taken as the main design parameter for thermal comfort. Hence it is essential for occupants’ well-being, productivity and efficiency [20] . Reference [21] suggested that air temperature alone is a good indicator of thermal comfort. The effects of other environmental parameters on thermal comfort have also been studied by [22] as cited in [5] .

Reference [23] in his study asserted that, heat gain through the exterior window accounts for 25% - 28% of the total heat gain. Moreover, the thermal capacities of various materials respond differently on incident solar radiation [24] .

Orientation and spatial organization also affect the ability of a building to ventilate and receive solar radiation [25] . To minimize solar heat gain and maximize ventilation, buildings should be orientated with their longer sides intercepting prevailing winds and the shorter sides facing the direction of the strongest solar radiation. The positive effects of shading on thermal comfort have been stated by [26] and [27] .

Reference [28] studied thermal comfort and workplace occupant satisfaction in low energy office buildings in Germany. They reported that votes of “just right” and “slightly warm” covered operative temperatures of above 27˚C. Again, they noted that PMV was very wide spread on the sensation scale and only changed slightly, dependent on the class of the subjective votes (“just right”, “slightly warm” or “very warm”).

Similar studies conducted by [17] in UK summer conditions found a mean temperature value of 23.9˚C in office buildings with a range of 21.6˚C to 26˚C. Additionally, there was an average PMV value of −0.25 with a range of −1.6 to 0.5. Reference [16] also found lower mean temperatures of 23.8˚C in offices he studied in London.

Very few studies have been conducted in the area of thermal comfort in Ghana [29] [30] . There is therefore a need to build a growing literature of the subject. Thermal comfort and indoor thermal environment can have an important impact on productivity [31] . Accordingly the aim of this study is to evaluate thermal comfort levels in naturally and mechanically ventilated offices using the PMV-PPD. Eventually, all the stakeholders in the building industry will be informed on best practices that can improve upon thermal comfort in office buildings.

3. Research Methodology

Thermal comfort of the offices was studied by means of quantitative survey along with questionnaire for building occupants. The thermal comfort index (PMV) which formed the basis of the study is considered to be applicable within an environment for the following reasons; Fanger’s investigation was based upon a wide range of environmental conditions, including conditions when air temperature >30˚C, and relative humidity up to 70%. In addition, a moderate thermal environment is deemed to have been achieved when PMV values range from −3 as cold to +3 as hot. Figure 1 below illustrates the functional relationship between PMV and PPD. At PMV neutral (0), 5% of the occupants are still dissatisfied.

Four buildings (Table 2) situated within the central business district of Accra were selected for the studies. The selected buildings are representative of current design trends in Ghanaian high-tech office construction due to their exposed glazing.

3.1. Participants in the Study

In all, 195 occupants in all 4 buildings participated in the study. All participants were workers on the floors monitored and had stayed in their current offices from between a year to ten years. The respondents in total consisted of 80 females and 115 males. Greater majority of the respondents (131) were between the ages of 25 - 35 years, while 8 were under 25 years, 29 participants were between 36 - 45 years and 37 were above 45 years.

3.2. Experimental Design

A field study was carried out from the 14^th April, 2012 to 31^st January, 2013, representing both the raining and the dry seasons in Ghana. Subjective measurements (in form of a thermal comfort questionnaire) and objective data were recorded during the study period. The objective data made use of loggers (onset hobo sensors for environmental parameters) which were installed at various points in the offices. Indoor temperature and relative humidity levels were measured at 10 minutes intervals for the study period. Altogether, a total of 714,286 data set points were measured and used for the study.

3.2.1. Subjective Measurements

Participants were asked to rate their thermal sensation on the ASHRAE seven-point scale. This was to gather information on how the occupants feel within their office spaces, thus their Actual Mean Votes (AMV). These were then compared to the Predicted Mean Votes (PMV). As part of the questionnaire, occupants were also asked to put down what they normally wore to work. This was used to find the corresponding clothing value from a clothing checklist [17] . Thermal insulation of the chairs of participants (0.1) was added to the clothing insulation for each respondent. Metabolic rate of 1.2 mets which is proposed in ISO 7730 (2005) and asserted in [31] was used as occupants in the offices engaged in sedentary activities.

3.2.2. Objective Data

Air temperature (T) and relative humidity (R.H) values were monitored with onset hobo sensors, whilst the hot wire anemometer was used to measure the air velocity (A.V). Measurements were taken at a height of 1.1 m above the floor, which represent the height of an occupant when seated. The accuracy of the measuring instruments is shown below (Table 3).