Different experiments or surveys will be concerned with different things. It may be that you are interested in the thermal performance of the building in general and not just with the room which the subject inhabits. Roaf (1993 or thesis) found that different rooms in buildings on the Persian Plateau had quite different thermal characteristics, and the strategy of the occupants of the building in moving from room to room at different times of day was to a large extent decided by the thermal profile of each of the spaces concerned. In such a building the thermal performance of the different rooms will be of particular interest.
Another experiment might be trying to relate the temperature in the occupied space to computer predictions and relating these in turn to the thermal response of the occupants. You will need to be clear about the exact meaning of `room temperature' used by the particular thermal model you are using. The temperature towhich humans respond in still air is about half way between radiant and air temperature, rooms respond differently, and the `characteristic room temperature' may reflect this. The problem is discussed by Humphreys (1974) in relation to the `environmental temperature' used for calculations using the admittance method. Also be sure about the definitions used by the programme for radiant, or surface temperature - are walls assumed to be all at the same temperature, is any particular surface assumed to be isothermal? All these factors need consideration when deciding on what measurements you need to take.
Measuring thermal conditions in a room is a problem on which a large body of literature has been written. Since we are not primarily concerned with overall room conditions but with those appropriate to the subjects, we will concentrate on pointing out the way to avoid the more major errors which can occur.
The first problem is to identify a `characteristic room climate'. The personality of the room climate will depend to some extent on the nature of the space and the materials and construction of the building. Thus a closed room will generally have roughly equal air and radiant temperature, some buildings such as the Malayan house (fig 7.1) rely on air movement and buildings in hot dry climates such as the Baghdadi house shown in fig 1.1 may rely on the relative cool of the room surfaces to remain habitable. We can attempt to measure the character of the room in two ways: by measuring the climate at a point in the centre of the room using a clusterof instruments, and also measuring the temperature at a number of points distributed around the room.
The air temperature in a room can vary substantially from place to place. In particular there can be marked vertical layering of air temperature in a room. Humphreys found that the temperature measured near ceiling level in a school classroom with warm air convector heating can be as muchas 10K higher than the temperature experienced by the children sitting on the floor. Clearly if we are interested in the conditions experienced by the occupants of a room, the vertical height at which the sensor is placed should be representative of the occupants' experience. It is also important to avoid placing your sensor in up- or down-draughts which can occur near heated or cooled surfaces (unless the occupant is also subject to them). Your qualitative investigation of air patterns in the room should help, but in general the air temperature should not be measured less than half a metre from any wall.
The air temperature should be measured at a number of places on a horizontal plane at a vertical height of about 0.6m above the floor - representative of sitting height (different occupations may mean a different value is appropriate). One of these points should be at the centre of the room. The exact choice of horizontal distribution would be left to the judgement of the experimental team in the light of room geometry and layout etc.
The problems of finding the radiant temperature have been discussed already. Unless the there is a particular reason for knowing the radiant temperature, it is best to use the globe temperature to represent the room temperature. Again it is important to use some judgement in positioning the instrument so that it is representative of the sort of place that occupants might use. If the room is large or the conditions within it vary from place to place then more than one instrument may be needed. If you are interested in the radiant temperature then the globe thermometer will need to be accompanied by a air thermometer and an anemometer and the measurements taken will need to be accurate. Measures such as radiant asymmetry are beyond the scope of this handbook. Those interested should refer to McIntyre (1980).
The water vapour pressure varies little from place to place in most rooms. A kitchen or a very damp cellar with trickle ventilation might be an exception but in most reasonably well ventilated rooms a single measurement will cover the whole space.
As we said above, start by doing a visualisation of the normal air currents in the room using a smoke puffer or other visualisation technique. This will give you clues to the useful places to put an anemometer - remembering that it is the air movement which the occupants will encounter that matters. The number and distribution of your anemometers will depend on how important this measure is to your study.
We are interested in relating thermal preferences to the climate the subjects live in, and not just to the indoor conditions they happen to meet. In one sense the climate they live in is defined by meteorological data collected at the local weather station. It is, after all, only these readings which we will have at our disposal in deciding what comfort temperatures to recommend. Nevertheless it is useful to investigate the microclimate round the site of our experiment, if only to compare it to the readings at the local weather station and decide how typical our site is of the official climatic data. Measurements of outdoor air temperature and wind speed and direction are the most important together where possible with measurements of solar radiation onto the horizontal. Solar intensity onto the walls of the building will clearly be useful where thermal modelling of the building is envisaged. The exact method for taking meteorological measurements and siting the instruments is beyond the scope of this book, but it is possible to obtain automatic data logging equipment to do the job.
One problem that a lot of experimenters find with comfort surveys is that the temperature often varies very little in some rooms over a period of a week or a fortnight. This can be due to badluck with the weather or to the design of the building you are using. This is particularly common in winter-time surveys in heated buildings - the heating system is, after all, designed to reduce temperature variations. But even in free-running buildings temperature variations can be small. Nicol found in the monumentally heavy-weight headquarters of the Prudential Assurance in Holborn, London, that temperature variations could hardly be measured - even in summer.
The problem with getting a small temperature range is that it becomes difficult to carry out the analysisof the data using the techniques which assume variation (see section 8 below). We end up with an average subjective response to a single set of conditions. The temptation, of course is to `tweak' the controls to give some temperature variation. The experimenter is tempted to turn up the heating or throw open a window just to get some variation. But this is against the spirit of the model and can produce just the effects of unexpected temperature variation which we criticise in climate chamber experiments. Indeed supporters of the climate chamber approach cite the problem of the lack of control over temperature variation as a major criticism of field surveys.
The problem of temperature range is another good reason for extending your survey to home as well as work conditions. As well as allowing you to follow the whole thermal experience of the subject, the information contained in your analysis will be improved by the increased range of conditions. Sharma and Ali (1986) did not set outto make any particular statement about the adaptive model so there was no particular reason to follow the full range of their subjects' experience but they did specifically mention the need for a wide temperature range as one reason to extend their study outside work hours.