Engineers are often criticised for not revisiting projects to learn from experience. There are a lot of pressures that stop it happening, not least a lack of time and resources. And they are not alone. When was the last time a client invoked the feedback clause, Part M, in the RIBA Plan of Work?
Which is why the PROBE project has been successful in plugging the gap between design expectation and performance. Since 1995 building physicist and social scientists have been dissecting well-known exemplar buildings to find out what makes them tick.
Have the energy targets been reached? Do the lighting controls work? How much more energy does air conditioning really consume over natural ventilation, and which system delivers the best comfort conditions for occupants?
The PROBE team has studied 18 buildings in the last six years, and studies carried out in the last 12 months include buildings which can be regarded as representative of the latest practice in comfort control. These are typified by advanced natural ventilation – normally with bems-controlled windows, night cooling, automatic lighting controls and high levels of insulation – or buildings which use a mix of air conditioning, mechanical ventilation and openable windows – the so-called mixed-mode approach. Here are some of the key messages.
Ventilation
The PROBE buildings cover a range of ventilation types, from conventional or advanced naturally ventilated buildings to fully air conditioned prestige head offices. Most of the offices were air conditioned, while many of the naturally ventilated buildings were in the education sector.
By using the method described in CIBSE TM22: Energy Analysis Reporting Method1, it was possible to accurately assess the gas and electricity requirements of the PROBE buildings by ventilation type. Figure 1 shows how the ventilation systems affected electricity consumption for building services, and figure 2 the gas consumption.
From figure 1 it can be seen that the air conditioned buildings in the dataset consumed up to 10 times the energy required by naturally ventilated and mixed mode buildings: typically around 200 kWh/m2/y as against 25-50 kWh/m2/y.
Looking at figure 2, the results show a mixed distribution, with the three best performing buildings being those which were super-insulated with low energy a key design priority. The highest performing building used a mixed-mode system: natural and mechanical ventilation with high efficiency heat recovery. Note that this beats the performance of the two next-best buildings, a doctors' surgery and a simple warehouse.
The buildings in the centre of the distribution are a mix of ventilation types, falling within the range 60-160 kWh/m2/y. At the high consumption end of the graph are two prestige offices and a naturally ventilated training centre. The common denominator for all three was that they are very leaky buildings, with leakage through the structure or through ill-fitting (often motorised) windows. This is clearly reflected in the ventilation heat load.
Lighting
Figure 3 shows how the PROBE buildings performed on lighting energy consumption. The typical and good practice benchmarks are those in the latest edition of ECON 192.
Again, the difference between the best and worst performing buildings varies by a factor of 10 to one. Most buildings fell between the good practice and typical levels, and four buildings were significantly worse than typical.
None of the PROBE buildings equipped with automatic lighting controls (65%) performed well. Only 10% of the buildings had lighting controls that the occupants understood. Most of the systems operated the lighting wastefully, some of them irritated the occupants by turning lights on and off at inappropriate times, and most of the attempts at daylight-linking were unsuccessful.
This widespread failure of lighting controls is due to a number of reasons. Many systems were never properly commissioned. Lights were often operated wastefully, particularly in communal areas such as corridors and reception areas where no individual person felt they had "ownership" of the controls. Daylight linking often failed when blinds were drawn to combat glare. Once the blinds were down, they tended to stay down.
The message for designers is twofold. First, lighting controls have to be very simple for occupants to use. They are also rarely used by trained facilities staff. Switching generally falls to administrative staff, but they have more work-related tasks to perform.
Second, lighting controls are rarely demand-responsive. They tended to come on under presence detection (whether needed or not), and remain on for a pre-set period long after people had vacated the space.
Not one of the PROBE buildings had the more energy efficient alternative of absence detection, where lights are switched on manually and turn off automatically. This type of system tends to prevent lighting from coming on in rooms that are briefly occupied or adequately daylit, and when cleaners or security staff are walking through the building in the evening.
While lighting control systems can be programmed to prevent all the lights coming on, this assumes the system has been properly commissioned and is easy to control by the facilities manager. When either of these two are lacking, lighting will default to on – irrespective of how advanced lighting controls might be.
Delegates wanted to know which method of construction delivered the most air tight building
Overall energy consumption
Generally, those buildings which are air conditioned (with humidification) tend to use more energy than the simpler mechanically or naturally ventilated buildings, with a factor of 10 between the highest and lowest consumers of energy. When the energy consumption data for all the PROBE buildings is normalised into CO2 emissions, this translates to about 26 kg CO2/m2/y for the best performing naturally ventilated building to 176 kg CO2/m2/y for the worst performing office building.
Bear in mind that the ECON 19 Best Practice Type 1 value (naturally ventilated, cellular) has recently been updated to 23 kg CO2/m2/y, while the typical CO2 emissions value for prestige offices (Type 4) is now 127 kg CO2/m2/y.
The mixed-mode buildings tend to fall between the two extremes, which is probably what could be expected for buildings that can cycle between mechanical ventilation and openable windows.
The PROBE buildings only come out as average performers, and as they were considered exemplars of their kind they might reasonably be expected to be closer to good practice. This is a worrying finding, particularly for the UK average.
A conventional solution to controlling energy consumption is via a building energy management system. Around 80% of the buildings in the PROBE dataset had a bems. Unfortunately only 20% of the buildings possessed an in-house or contracted facilities manager who had a good understanding of hvac controls. This might explain the minority of PROBE buildings with controls that operated plant in reasonable accord with the design intent. Consequently, 80% of buildings demonstrated shortcomings in the operation of plant, or controls which inhibited economical operation. In about half the buildings studied, small loads like computer rooms caused main plant to operate wastefully.
Once again, bear in mind that in this context it is salutary to note that these buildings are exemplars of their kind, where the designers had placed particular emphasis on the quality of the controls specification.
Comfort issues
PROBE buildings are all subjected to a rigorous occupancy survey, based on the Building Use Studies (BUS) questionnaire. By polling 100 occupants per building (or 100% of occupants in the smaller properties) it is possible to build a picture of their feelings about comfort conditions, noise, lighting and how these factors influenced their perceived productivity.
The advanced naturally ventilated buildings performed poorly against the BUS benchmark database, with mixed mode buildings tending towards the productive end of the scale. The best building was naturally ventilated, but the next best an air conditioned office, with attentive facilities management.
Why did advanced natural ventilation perform relatively poorly? Manual control of the systems (such as openable windows) was found to be psychologically important. In naturally ventilated buildings where motorised windows are opened automatically, it was important to provide manual override as the controls were not always tuned in accordance with the desire for ventilation.
This was true for many buildings that had a poor quality motorised components. Some were noisy and intrusive, opened and closed inappropriately, or did not close the windows properly which led to draughts and excessive heat loss. That said, the PROBE studies found many conventional windows with poor operability.
Conversely, it is risky to rely on manual control for night venting – security staff are too eager to close them, and the use by occupants is unreliable. But where night venting is automatic, it must be properly commissioned to enable the building to discharge its daytime heat gains.
Discussion
The results of the probe building studies were discussed by the CIBSE seminar delegates. One delegate queried whether the perceived productivity ratings were more a function of the respondents' views of a building's management and less a comment on a building's engineering systems.
Most questions in the PROBE questionnaire cover specific issues such as comfort and temperature, and answers to those questions should not be unduly influenced by the occupants' worries about their job or the management. The anecdotal responses often reveal a distinction, but when asking about productivity it is difficult to separate the quality of the building from that of the management.
The poor performance of lighting controls vexed the delegates. Many said that while lighting controls are expected in the brief, the money to do the job is often cut by the cost consultant. The PROBE response to this problem is to argue for lighting controls, not on the basis of energy efficiency, but on how much happier the occupants will be if the lights are working sensibly. If the clients are made to appreciate that lighting controls exist to make occupants productive, then perhaps quality will not be wrung out of the specification.
The dissatisfaction with advanced naturally ventilated buildings stems from inadequate engineering of the important interfaces, such as windows. Mandatory pressure testing is likely to make everyone aware of this problem.
Delegates were interested to know which method of construction delivered the most airtight building. The buildings which were easier to get airtight were those of masonry construction with a wet plaster finish. This is possibly because masonry requires fewer junctions between different components and the use of formless materials, like plaster, helps to seal the building. But airtightness has more to do with the quality of detailing (such as bay windows, eaves, trickle vents and motorised windows) and construction.
Delegates wanted to know whether high mass buildings are more comfortable than lightweight buildings. The answer in brief is yes: high mass is good in principle, but if you don't flush the heat then discomfort during the day is inevitable.
Downloads
Source
Building Sustainable Design
Reference
References
1 TM22: Energy Assessment and Reporting Methodology: Office Assessment Method, CIBSE 1999.
2 Energy Consumption Guide 19: Energy Use in Offices, BRECSU January 2000.