Powys council was worried by rises in gas prices, so it decided to make its Stag teaching facility the first office in Britain to reach Passivhaus standards. After horrendous design problems, the team spent 12 months collecting data on how it performed. And, as Thomas Lane reports, the figures came as something of a shock鈥
Ask any green designer whether mechanical cooling uses less energy than natural ventilation and the polite answer will be 鈥渄on鈥檛 ask stupid questions鈥. But ask John Williamson, and the response is 鈥渕aybe鈥. The softly spoken designer and engineer, who heads multidisciplinary firm JPW Construction, talks about building physics with an intensity and understanding that would make any green designer listen. But the real reason for taking on board his views is that he has just spent 12 months monitoring the UK鈥檚 first Passivhaus office - and it is performing better than predicted.
The reason for the 鈥渕aybe鈥 is that Williamson鈥檚 services design uses power-hungry mechanical ventilation to purge heat from the building at night rather than just opening windows. 鈥淚t means you don鈥檛 need complicated and expensive control systems for the windows,鈥 he says. 鈥淎nd if there isn鈥檛 any wind how do you naturally ventilate a building?鈥
Going down the mechanical route is a decision Williamson did not take lightly. The Passivhaus standard stipulates that space heating must not use more than 15kWh/m2/yr, which is a tough nut to crack in a home. In an office full of computers it is less of an issue because of the heat they produce. But Passivhaus requires that total energy use in the building does not exceed 120kWh/m2/yr, which is hard with all those computers, and is made more challenging as getting rid of waste heat in summer increases energy use, unless natural ventilation is used. So how has Williamson designed a building that ticks all these boxes - and how well does it perform?
Desperate design
The office, called Canolfan Hyddgen (the Stag Centre), has a floor area of 400m2. It is used for front-of-house council services such as taking council tax payments in Machynlleth, Wales. It is also used for day and evening classes, including computing, so has three large rooms full of computers. Powys council was prepared to become the first client in the UK to go for a non-residential Passivhaus because it spends 拢20m a year on gas and had been tipped off that gas bills were about to rise 40%. 鈥淭he 85% reduction in space heating really sold it to Powys,鈥 says Williamson.
The team had to deliver the building in double-quick time as Welsh assembly funding had to be spent by the end of April 2008, which gave the team eight months from design to completion. Also, the funding was dependent on the building achieving a BREEAM 鈥渆xcellent鈥 rating, and that standard does not always dovetail with Passivhaus. For example, BREEAM awards points for natural light, so the building has lots of south-facing glazing; this is great for solar gain in the winter, but not so good in the summer.
But the main challenge was keeping the building鈥檚 internal temperature as close as possible to 20潞C while staying within Passivhaus鈥 energy limits. Williamson鈥檚 calculations showed that this could be achieved with an annual primary energy use of 112kWh/m2/yr, which is below the Passivhaus target.
Then, just as the design was being finalised, the council said it wanted to add a server room that would service other buildings as well as Canolfan Hyddgen. It also said it wanted to use Dell computers because of an existing framework agreement, rather than the ultra-low energy Tranquil PCs that Williamson had counted on. 鈥淪uddenly our energy use rocketed because of the server and as we couldn鈥檛 use low-energy computers,鈥 says Williamson, who, understandably, believes it is crucial to sort out the IT strategy at the beginning of the design process.
The changes meant that predicted primary energy use soared to 184kWh/m2/yr, and heat gain from the server room and PCs would make summer cooling even harder. So the designers had to search for every conceivable way of cutting energy demand. The first change was simply to add brises-soleil to the front of the building. Another was equally commonsense: shut everything down at night (a task accomplished by mechanical timers, in keeping with Williamson鈥檚 鈥渒eep it simple鈥 principle).
Then the ventilation plant was examined to see what savings it could offer. The system had a high, medium and low setting, with medium being the default for public areas. The solution depended on the fact that each of the three computer rooms could be ventilated independently of the main system supplying the public areas. This meant that if movement detectors were fitted in each, lighting could be turned off if the room was unoccupied, and ventilation switched to the low setting. The highest setting would only kick in when internal temperatures exceeded 26潞C.
Heat could be purged from the building at night using the mechanical ventilation system, and the building has concrete floors and blockwork internal walls to add thermal mass. As with all Passivhaus designs, a mechanical ventilation system with heat recovery was fitted to retain heat in the winter. At the end of winter the heat exchangers are removed so fresh air can
flow into the building without being warmed by the extracted air. The fans kick in after 11pm if the internal temperature of the building exceeds 24潞C and switch off once it drops to 18潞C.
With these changes, the predicted primary energy use dropped to 144kWh/m2/yr with space heating right on the Passivhaus threshold of 15kWh/m2/yr. Because the
server room supplies other buildings the Passivhaus Institut was prepared to certify the building even though it was 24kWh/m2/yr over the requirements, and because Williamson had made every effort to reduce primary energy use.
What the data told us
But Williamson need not have worried - a year of monitoring shows that the building has trounced the Passivhaus energy requirements, coming in at just 80kWh/m2/yr.
He now wants to improve on these figures and reckons mechanical cooling could be the answer. If he can get a grant to fund it, he will fit a 1.5kW heat pump that would pre-cool the air entering the ventilation system in the summer. The reasoning that is this would eliminate the need to run the fans at night. 鈥淔rom the comfort point of view it has advantages and could be a good option for buildings in noisy locations,鈥 he says. If Williamson is right, it鈥檚 back to school for the industry鈥檚 green designers.
The data presented here spans the period of October 2009 to the end of September 2010 and so covers a particularly long, cold winter. This period is after the building was commissioned and had time to dry out, so is representative of normal use.
Data taken prior to this period shows the effect a damp building and teething problems with commissioning the services had on performance. Heating energy use for the winter of 2008/09 was 16kWh/m2/yr, which exceeds the Passivhaus limit of 15. Williamson attributes this to the additional energy needed to dry the building out.
The building was also too hot during the summer of 2009. An external sensor was damaged by a football and early morning sun was shining directly on another sensor. This shut the mechanical ventilation system down prematurely which meant heat was not purged from the building.
Then, during the monitoring period after the building had dried out and been properly commissioned, the building used 14.8kWh/m2/yr for heating, which is slightly better than the predicted figure of 15. Williamson puts this down to a cold but bright winter, in which solar gains were higher than predicted.
To put this in perspective, the annual gas bill for this 400m2 building was just 拢127. Williamson says this figure could be reduced further as the building is fitted with a 24kW boiler despite the heating load being a mere 4.8kW. 鈥淭he M&E contractor was reluctant to fit a smaller boiler,鈥 says Williamson. The boiler鈥檚 minimum output is 9kW, which is still nearly twice the actual heating load.
The electricity use, meanwhile, comes in at 80kWh/m2/yr, well under the Passivhaus maximum of 120, and much better than the predicted use of 144. This costs 拢2,410 a year. The reason why the energy figure is so much lower is probably because computer use is less than predicted (which goes to show how difficult it is to predict primary energy use).
The building is fitted with a 7kWp array of solar panels that generate 5,987kWh/yr, which is a quarter of the building鈥檚 electricity use. The power produced by these panels is not included in the overall energy figures above. If the panels had not been partially grant-funded, the power produced would have benefited from feed-in tariff income, which would have paid all the energy bills. As it is, PV installations that are grant-funded do not qualify for the feed-in tariff.
The human factor
The maximum internal temperature recorded inside the building was 26潞C during the summer. In the winter, the temperature was between 20潞C and 22潞C depending on whether
occupiers turn the radiators up. In the winter, when the building is fully occupied and all the computers on, heat gains are sufficient to keep the radiators off. But of course, the users have to understand a little about the designer鈥檚 intentions for the building鈥檚 operation.
Alyson Jones is the centre鈥檚 administrator. She says it is 鈥渉ugely better鈥 than the office she worked in. She says the building stays cool on hot days over the summer. Understanding how the building works helps maximise energy efficiency and comfort levels, so Jones explains this to new teachers who, in turn, instruct their students.
Machynlleth is a magnet for alternative lifestyles - for example, the Centre for Alternative Technology is just up the road - so people do not find the Passivhaus concept alien. 鈥淲e鈥檙e living in an area where we鈥檝e had new technology for the past 35 years, so it鈥檚 a good place for a building like this - people don鈥檛 see it as abnormal,鈥 says Jones.
Powys council is also happy with the project, particularly as it cost 拢1,784/m2, including photovoltaic panels, which compares favourably with Faithful + Gould鈥檚 benchmark figure of 拢1,711/m2 for an educational building.
鈥淔rom the financial and carbon point of view it鈥檚 been a success for the authority,鈥 says Nick Hinkley, the project manager for Powys. 鈥淲e didn鈥檛 end up paying a
premium for the building as it was something planned from the beginning. 鈥淲e鈥檙e very pleased with it and would very much like to use the system on another project,鈥 he adds. Indeed, the council wants to build a community campus next to the building if funding from the Welsh assembly comes in 鈥
FACTS AND FIGURES
WHAT鈥橲 THE BUILDING LIKE?
This building is all about well thought-out practical design rather than wow architecture. An enormous porch isn鈥檛 intended as a smokers鈥 haven, but as a space for visitors to take off theirwet coats before entering the building. Visitors enter through a lobby with automatic doors on either side, which stops warm air escaping in winter.
Inside, there are some architectural gestures - a double-height reception, for example, and a corridor that runs the length of the building on the ground and first floors, which is flooded with light from the south-facing facade. The computer rooms are on the other side of this and are glazed on the corridor side to cut down on the lighting needed. Separating the computer rooms from the facade by the corridor limits glare, which can also be controlled using blinds. The more utilitarian spaces such as the plant room are simply laid out and the toilets have local water heaters with short pipe runs to cut down on energy use.Externally the building has a green roof and a 7kWp photovoltaic panel array.
In terms of construction, the concrete flooring and internal walls were clad with a panellised timber system. The panels are filled with 200mm Warmcel insulation and clad with 40mm-thick fibreboard covered with timber and lime render. Impressive levels of airtightness were achieved: O.37m3/m2/hr@50Pa, which is 27 times better than the UK regulatory limit, and less than half of the Passivhaus requirement.
Downloads
Floor plan
Other, Size 1.74 mbFacts and Figures
Other, Size 1.84 mb
No comments yet