Cold silence

Forget Egan. Barclays Bank is way ahead. Back in 1996 when British Airports Authority was lauding its controversial Framework Agreements, Barclays was putting the finishing touches to a spectacularly innovative headquarters office building in Northampton.

Hailed at the time as a successful fusion of "green engineering" with otherwise conventional open-plan space, the building is delivering everything the client hoped for at a very low energy penalty1. Indeed the Property Services arm of the Barclays Group (BGPS) was so pleased with what it got that the entire design and construction team was awarded the contract to design a similar building in Essex.

No competition, no Framework Agreement, "but a clear commitment to construct a similar building to that at Northampton for a clearly defined level of cost, and to work together in a similar manner," is the way Barclays Gordon Harrison put it.

The result is just coming out of the ground on-site in Basildon. When completed late this year, the 9000 m2 (net) building will be home to the bank's financial services arm BarclaysLife.

While the Northampton building bristled rather courageously with innovative environmental engineering – ammonia refrigeration, chilled beams, mixed-mode ventilation, exposed mass, lake cooling and so on – BarclaysLife is slightly less adventurous. The high costs of decentralised ammonia refrigeration couldn't be justified, and the client wanted something greener than bespoke chilled beams.

That said, Barclays has moved on from the institutional norms that hamstrung aspects of the design at Northampton. The heavily-ribbed, exposed concrete coffers on a 1·5 m module – driven by institutional norms – has been relaxed at BarclaysLife to more pragmatic 3 m centres. This has allowed the architect Fitzroy Robinson to design a very elegant flat-domed coffer.

Services engineer Troup Bywaters & Anders (TBA) wanted to rely on the exposed structure for passive cooling, but the shortcomings of the heat transfer properties of concrete would not provide enough capacity to cope with the estimated 30 W/m2 internal cooling load. Some way of boosting the thermal capacity and response time of the slab was obviously needed, and this led TBA to propose a radical idea of mechanically chilling the concrete coffers.

With main contractor Tilbury Douglas appointed at the outset, the designers entered into a series of design iterations. Ultimately this led to the design and fabrication of 3 m-wide precast concrete coffers, embedded with plastics piping to carry chilled water. Warmafloor Structural Conditioning was appointed to work with the fabricator Tilbury Douglas, which precast the coffers in a local fabrication yard in Lenwade, Norwich.

Warmafloor laced the 20 mm polybutylene pipes to each coffer's steel reinforcement cage, which was then upended, fitted to one of four moulds, and filled with concrete. Once cured, the coffers were craned to site and simply dropped into position.

On-site, there is little to inform the visitor that the building's 380 barrel-vault coffers are anything other than conventional precast elements. The only clue is the pipework termination on the top of each coffer, where the stub ends lie protected in a small recess. Secondary chilled water pipework will then connect these to header manifolds on each floor, just like an underfloor heating circuit.

Concept development

The simplicity and elegance of the solution belies the considerable research that was carried out on the viability of the chilled concrete coffer.

In early 1998 the BSRIA was commissioned to determine the heat transfer characteristics of a prototype chilled concrete slab. The tests confirmed that the idea was sound in principle, and also helped to optimise the spacing of the pipes, the depth of the pipes within the concrete, and the diameter of the pipework.

The BSRIA carried out a series of tests on the performance of the slabs during April 1998. The basis of the test was the construction of an insulated calorimeter box around the suspended slab, whereby the heat absorbed by the slab could be balanced by 100 W strip heaters in the spaces above and below. 27 embedded thermocouples enabled the heat flows to be measured and recorded. The slab was fed with temperature controlled water generated from a chiller. A three-port valve of the cws primary circuit was operated under computer control to maintain the slab inlet temperature. In most cases the tests were carried out over a 24 h period.

Temperatures were recorded for the surface of the slab using an infrared spot temperature meter. Wall temperatures were also recorded using a surface probe and a Squirrel data logger.

The BSRIA's tests showed that the test slab was capable of providing 6·1 W/m2K of cooling in the lower chamber. For the operating conditions of 13°C flow, 16°C return and a 25°C air temperature, this corresponds to 64 W/m2. The BSRIA concluded that the radiant cooling effect would "significantly" reduce the dry resultant temperature under these conditions.

Surface area was the most important criterion, as a result the pipes follow the curvature of the coffer to obtain maximum exposure to the space below the curve. The concrete to pipework depth was set at 45 mm to maximise the downward radiant component, with 150 mm of concrete on top to minimise the upward component. These dimensions were shown by the BSRIA tests to obviate the need for any floor slab insulation.

Tests also showed that the maximum thermal transfer could be achieved with just 100 m of pipework (at 150 mm centres with flow and return pipes laid alternately) per 3 x 7·5 m precast coffer, half that initially anticipated by the design team.

There was no evidence of condensation on the exposed plastics pipework in any of the tests. In one test, the flow temperature of 13°C and an air temperature of 25°C yielded a lower surface temperature of 16·8°C. Under these operating conditions, condensation could be expected above a humidity of 12 g moisture per kg dry air (equivalent to a relative humidity of 60% at 25°C). This level of humidity is rare in the UK – around 4% of the hours during June, July and August.

The table shows that the radiant cooling from the slab is quite significant, with the dry resultant temperature being some 1·9°C lower than the air temperature. Approximately 90-95% of the cooling effect of the slab was directed through the lower surface, this, in addition to a 300 mm floor void above the slabs, meant that top surface insulation was not required.

Installation issues

The 380 precast coffers are now in position ready to be connected to the secondary chilled water circuit. Flexibility in operation has been allowed for by splitting each floor into two zones and supplying chilled water from six 200 kW chillers. Each multi-stage chiller will supply a maximum of 100 kW into its nominated slabs, with the other 100 kW for cooling the ventilation air via the floor void and dehumidification in extreme weather conditions.

Without the need for a multifunctional service beam like that at Barclaycard Northampton, the ceiling contractor SAS has designed a winged luminaire which will be suspended from the centre of each flattened coffer. Being fabricated in factory conditions has endowed the concrete with a high quality finish, and the fully finished test bay on site is certainly elegant in only a single coat of white paint.

So are there any downsides? In mainland Europe embedding cooling coils in a building's internal fabric is relatively popular, the Ka Ro system of capillary coils in precast gypsum plasterboard being one example2. Not so in the UK, primarily because clients prefer tried and tested solutions and dislike being guinea-pigs, and second, because sequential terms of appointment still tend to militate against such holistic solutions.

Without the close working relationship between the architect, consultant and main contractor on this job, a function of the client's enthusiasm for long-term partnering arrangements, there would not have been time nor enough trust between professionals to develop up the concept. The early appointment of main contractor Tilbury Douglas helped considerably, as did the early appointment of the pipework manufacturer Warmafloor, who demonstrated a willingness to work with Kent Concrete in the fabrication yard.

There is a small question-mark over the reliability of the 110 m runs of pipework inside the coffers. While the multi-layered plastics pipework used by the underfloor heating industry comes with a 25 year warranty, it would be foolhardy to claim 100% reliability. Unlike underfloor heating systems there will be no quick means of getting access to repair leaks (although solutions do exist), so the sequence of installation, commissioning and water treatment has to be undertaken assiduously.

At least the pipework has been encased in concrete in factory conditions, which is one step better than a site-based process. While the terminations are protected by the recess in the top slab, they need to be carefully capped to prevent debris getting into the pipes – no small matter when one considers that the terminations have to survive 90% of the construction period.

Building Services Journal intends to return in a later article to report on whether or not the chilled slab system is as robust in operation as its designers promise.