The use of computer modelling tools enables designers to evaluate options of building design and plant selections at the press of a button. Modelling software will use weather data as made available to it (such as the CIBSE Test Reference Year1) and will attempt to calculate loads across the operational period.
In many cases the development and input of the underlying data set may be complex, and the incorporation of the air conditioning plant details can be challenging, particularly at early design stages. Alongside the computer model it is often useful to have a simpler method not only to aid understanding of the more complex outputs of the software, but also to provide swift feedback as a check of the overall validity of the model. Through the incorporation of appropriate weather data, a psychrometric chart can provide the tool to undertake such checks. This article will explain how to set this up, and explore some first examples of application.
Frequency-based weather data
As has been illustrated in previous CPDs, the psychrometric chart is a useful tool to examine the properties and energy flows associated with air processes. However, when evaluating the comparative operation of different possible solutions, it is important not only to take into account the values of the operational psychrometry in design and part load conditions, but also to consider how frequently this might occur. The designer (whether aware of it or not) will already be applying frequency-based weather data when undertaking any building load analysis, even when calculating a simple steady state building heat loss.
Ready-made analysis
When calculating room heating loads using winter design conditions from the CIBSE Guide2 the selected external temperature would be based on historic weather data (as well as a knowledge of the building thermal inertia). The underlying data in ‘binned’ format (in ‘bins’ or bands of temperature), as shown in Figure 1, provides a basis on which the external design condition may be selected for heat-loss calculations.
.png)
Fig. 1 : Example of binned average temperatures in occasions per year
So, for example, Figure 1 indicates the mean temperature averaged over 24 hour and 48 hour periods for a particular geographic location collected in 1K ‘bins’. So considering the yellow shaded area, the -3oC bin – this bin is the sum of historic occurrences of 24 hourly averaged outdoor temperature between -3.5oC and -2.5oC. Looking at the height of the yellow column, it occurs on the equivalent of just under one daily period per year. But, more importantly, this data or, if available, the numeric data used to create this graph, may be used to see how many 24-hour periods have a hourly averaged temperature below the bin of -3oC. Adding up the (24-hour) columns to the left of the -3oC column comes to about one occasion (ie one 24-hour period) when the temperature falls below -3oC over the average year. According to the guidance in the CIBSE Guide, this would indicate that -3oC was a reasonable temperature to use as the external design condition for heat losses (for low thermal intertia or ‘lightweight’ buildings) – a temperature where it gets colder on just one occasion in a year. When evaluating ‘heavyweight’ buildings, the design condition is again selected based on this single occasion guidance but this time using the 48-hour data. (This method dates back to an article in the predecessor of this Journal by Jamieson in 1955 – see CIBSE Guide A2 for more detail).
The effect of the time period used for averaging the temperature observations may clearly be seen by noting that the 48-hour data is far less extreme than the 24-hour data. The actual underlying data (as collected by the weather station) would be based on the same sets of hourly averages data; however, the longer averaging period will moderate out the high and low values. This moderating effect becomes more obvious as the averaging period is lengthened (for example, if a monthly or seasonal average temperature is considered).
So the assessment of outdoor design conditions for simple heat losses can be chosen from a knowledge of the outdoor temperature, with no need for the ‘complexity’ of a psychrometric chart. But these data are limited to just temperature: what if a better understanding is needed of the coincident external moisture content, or indeed some enthalpy data are needed to evaluate plant performance? And when considering midseason and ‘summer’ building operation, the decisions that will affect the economic viability, selection and sizing of humidifiers, cooling coils and heat recovery devices will depend on the frequency of both dry and wet bulb temperatures.
Weather data
The historic data are available in tables that include (amongst other things) coincident wet-bulb and dry-bulb temperatures, and are ideally suited to act as a basis for the analysis of design conditions. (To purchase specific historic weather data see the CIBSE Guide J section 11 and the CIBSE web site for sources). CIBSE Guide A2, section 2.4 includes frequency data for summer months (June to September) for eight UK locations.
Figure 2 is indicative of a type of weather data that can be sourced. These were taken from data produced by the Meteorological Office3. Many of the 35 main measuring stations are located in relatively remote areas (such as airfields) so actual conditions required for building and system evaluation will need to take account of the local microclimate, compared with that of the nearest weather station. There may also be some concerns as to the effect of climate change on future, predicted values based on historic data. The CIBSE Technical Memorandum 34, Weather Data with Climate Change Scenarios provides extensive guidance based on a number of predictive models, and includes frequency data for UK summer months that can be compared directly with the current tables in CIBSE Guide A2.
These particular data in Figure 2 have been laid out in terms of the percentage occurrence of coincident pairs of dry-bulb temperature and moisture content. So, to take the yellow highlighted data as an example, over a year it might be expected that the dry bulb temperature of between 10.0oC and 11.9oC (ie a 11.0oC ‘bin’) would happen at the same time as a moisture content of 8.00 to 8.99 kg/kg (a 8.5kg/kg ‘bin’) for 0.78% of the hours in a year (about 68 hours per year).
.png)
Fig. 2 : 24-hour weather data for example UK site
The data can also be used to establish how frequently the selected design values are likely to be exceeded. In the right-hand column the totals of all the dry bulb/moisture content coincident pairs are given. So, for example, if we were selecting an external dry-bulb (bin) temperature of -3oC, the total frequency of temperature below that bin is 0.24 + 0.06 + 0.02 + 0.01 = 0.33% (about 29 hours a year) – these data are highlighted in green.
Similarly the data can be quickly used to determine extreme summer time dry bulb temperatures that may be used, for example, to aid in the selection of an appropriately sized air cooled condenser.
Moving the psychrometric chart into the time domain
However, the application of the data becomes far more accessible, both visually and analytically, when it is superimposed onto a psychrometric chart. The data have been taken from Figure 2 and added to the outline chart as shown in Figure 3. (While doing so, the data were multiplied by 100 to remove the decimal places).
.png)
Fig. 3 : Example of psychrometric 24-hour, hourly weather data – frequency % x 100
A couple of the data points appear to exist in the impossible area to the left of the saturation curve; this is due to the approximated sketch of the psychrometric chart and should not distract from the bigger overall picture. If meteorological frequency data is available in terms of coincident dry bulb temperature and wet bulb temperature, then this could have been as simply added to a chart, but using a grid based on the dry-bulb temperature and wet bulb temperature axes.
Immediately, looking at the chart, the data are transformed into a form that allows quick visual evaluations of conditions without any formal calculations. It can be seen that although there are significant tails on both the ‘winter’ and ‘summer’ conditions, the actual times that these are likely to occur are relatively few.
Inferred frequencies of occurrence of wet bulb temperature and specific enthalpy, together with any other air properties that are included in the base psychrometric chart, are now automatically available for analysis.
Visualisation in the fourth dimension
A bounding envelope has been added to the psychrometric data in Figure 4 to emphasise the most frequently occurring psychrometric conditions. The chart is now a tool that not only allows examination of the properties of air but also clearly includes the dimension of time.
A summer external design point is shown – this could be, for example, the summer design condition as proposed by the load calculation software, or the ‘standard’ as used by a designer for a particular location. The graphical representation provides, at the least, a swift visual check on the value, but also can inform and communicate the effect of alternative values on the consequent plant operation. This will not just be, as a standard psychrometric calculation, in terms of power, but by including the frequency information, and hence time, comparative energy use may be examined.
For example this design point of 28oC dry bulb/20oC wet bulb could be used as the starting point to undertake some sensitivity analysis on the size of a cooling coil in a full fresh air conditioning system (see the April 2010 CPD article for a reminder of the basic system and its psychrometry). The load on the cooling coil is determined by the incoming air enthalpy, and at the plotted condition this may be read off the chart (using the superimposed grey enthalpy lines) as 57 kJ/kg. In this case it can be readily seen that by increasing the design wet bulb by 1K to 21oC (maintaining the dry bulb design at 28oC), the system would operate within its capabilities for (about) an extra 10 data points (above that of the current condition). As each data point is equivalent to 1/100% of the hours in a year, this equates to (10/100) % = 0.1% of the time. And the additional design load on a cooling coil to cope with approximately 3 kJ/kg increase in enthalpy difference is likely to be over 10% (based on a ‘typical’ cooling coil air leaving enthalpy of 30kJ/kg) – to meet the extra need for just 24 x 365 x (0.1/100) = six hours per year.
If the underlying data is available explicitly linked to each of the 8,760 hours in a year, then refined subsets of the data may be simply developed. This has been used to produce CIBSE Guide A, tables A2.7 to A2.15 as data for ‘summer’ conditions (June to September) that can be similarly superimposed over a psychrometric chart.
Conclusion
The application of simple tools described in this article are not advocated as a replacement for computer modelling, but can be used both to illustrate quickly and to undertake ‘ball park’ evaluations of different design options. They are particularly useful as a means of illustrating the impacts of system design decisions on those who have less extensive experience of HVAC&R operation.
.png)
Fig. 4 : Summer external design point
© Tim Dwyer 2010
My thanks to Dr Roger Legg for his publications, which provided underlying material for this article
- CIBSE Guide J Weather, solar and illuminance data Section 8, CIBSE, 2002
- CIBSE Guide A2, Environmental Design, Section 2.3, CIBSE, 2006
- Data after -collingbourne and Legg, The frequency of occurrence of outside air conditions in the UK, IoEE TM 58, South Bank Polytechnic, 1979