The “District Heating Wall”: A Synergistic Approach to Achieve Affordable Carbon Emission Reductions in Old Terraced Houses ()
1. Introduction
Heat for domestic dwellings accounts for 26% of total UK energy demand and over 21% of UK carbon emissions [1]. 91% of dwellings have central heating, with the majority (87% in 2006) gas fired and less than 1% of heat is from renewable sources [2].
There are approximately 2.5 million pre-1919 terraced homes in England, with the majority having solid wall construction [3]. As part of the UK strategy to reduce residential sector emissions by 12% by 2020, the Committee on Climate Change proposed 2 million such homes should be insulated by 2020 [4]. BRE analysis indicates that installing Solid Wall Insulation (SWI) offers the greatest potential savings of all the insulation measures possible for the overall housing stock [5]. SWI can either be applied to the exterior of the property (external wall insulation or EWI) or to the interior of the property (internal wall insulation or IWI). In 2008 there were about 25,000 - 35,000 SWI installations in the UK of which 60% were EWI. Costs of SWI are generally high and this is a significant barrier to wider uptake [6].
Placing insulation on inner side of the walls (internal wall insulation-IWI) blocks heat from reaching the existing wall and so reduces the thermal mass of the construction. In addition, this construction could lead to problems with interstitial condensation, water retention and cold bridging. On the other hand, placing insulation on outer side of the walls (external wall insulation-EWI) maintains the thermal mass of the construction therefore the construction loses heat more slowly. The EWI system can also improve structural integrity and weather-tightness of the building. In addition the internal room dimensions are maintained-important when treating small properties-and there is less disruption to residents during installation [6].
Visual impact issues were found to be a defining characteristic of EWI at a recent major conference on the insulation of Scottish solid masonry walls [6]. EWI could have a positive impact if the original appearance of a property is poor, or would struggle to get planning permission for a listed building in a conservation area.
EWI systems were found to use a range of insulation materials varying in thermal conductivities, from mineral wool with =0.036 W/mK [7] to phenolic foam = 0.02 W/mK [8].
EWI tends to be more expensive than IWI. EWI for an individual mid-terrace house as part of a multiple installation campaign is estimated to cost £5908 [9]. The cost of materials can vary by around 20%, and for EWI materials account for 30% - 40% of total install costs [6]. Additional costs for EWI may include scaffolding, re-siting pipes and moving of wall appendages like satellite dishes.
District Heating (DH)—where heat is produced centrally and hot water is piped to buildings-can improve the efficiency of energy use and provide the flexibility to accommodate heat from a variety of sources such as low carbon biomass and Energy from Waste.
In the UK it achieves only low market penetration: 2% versus Finland (49%) and Denmark (60%). Poyry identified the main economic barrier facing new DH projects in the UK as being high upfront capital costs [1]. For District Heating Network (DHN) development, the civil engineering work required for laying pipes attributes roughly half of these costs. Civils costs include traffic management for closed roads, excavating, backfilling and resurfacing trenches and rerouting of services. If these costs were addressed, and under the correct market conditions, DH in the UK could supply 14% of heat demand [1].
The conventional DH Local Network design for a terraced street would have the mains pipe buried under the road, with individual branches connecting each house [1]. An obvious way of reducing the Civils cost of DHN is to route the mains pipe so that the road is not dug up. This could be achieved by routing pipes through gardens or roof-spaces [1]. However pre-1919 small terraced houses, where the door opens on to the street, don’t have the option of garden-routing and many will require installation of SWI. Instead, the mains pipe could be routed through EWI. This is the concept of the DHWall [10] (Figures 1, 2).
The proposed benefits from the DHWall are:
• Avoided construction work = reductions in capital costs
Figure 1. The DHWall concept. Tf = 80˚C and Tf = 55˚C. Mains pipe diameters, i.d. = 32.6 mm. Insulation thickness was 120 mm or 65 mm depending on material.
Figure 2. Peak space heating days for the EXISTING and EWI models.
• Shorter pipework lengths = reductions in capital costs and heat losses.
• Reduced demand = reduced pipe diameters = reduced costs.
• Enables EWI installation.
This project aims to demonstrate the proposed benefits and confirm initial design parameters as well as highlight potential design issues. The methodologies used to develop each model will be accounted in next section. The findings will be discussed then conclusions and recommendations for further work given.
2. Methodology
The methodology includes selection of a study case and development of models to calculate quantitatively the benefits of the DHwall system comparing against the conventional DH approach for that case.
The study case was terraced street in the northern England, a cluster of AECOM’s DH development (Figure 3). Its selection was based following reasons:
• Near to Staffordshire University (an anchor load).
• Reasonable street length.
• Uniform number of terraced houses either side of the street.
• Relatively plain existing façades.
• Obvious dwelling boundaries on OS Map.
The first three reasons can be thought of as DH suitability criterion. It was thought the visual impact of applying EWI to terraces would be less for those terraces that do not have intricate brick facades. Some terrace designs were noted to have passageways from street to back garden, these created unobvious dwelling boundaries on the OS Map. Such a design represents a deviation from the standard terraced house. The last two can therefore be considered to be EWI suitability criterion.
The RockWool BrickShield system was chosen for the wall insulation from a survey due to the following reasons:
• An insulation thickness of 100 mm, the expected minimum thickness is required to treat solid brick walls [7].
• The manufacturer had installed systems in Stoke-onTrent and technical data was available.
Figure 3. Small terraced houses in Stoke-on-Trent, UK.
• The product is faced with brickslips, which could minimise visual impact.
Two models were developed for predicting values for the quantitative comparison:
1) A building thermal model-to quantify heat demand of a terrace with and without EWI.
2) A network model-to find the network requirements to transport the heat to the terrace predicted by the building model.
2.1. The Building Thermal Model to Calculate Heat Loads of the Dwelling
The heat demands in this study included space heating and domestic hot water supply, both are time dependent variables throughout a year and calculation had to include their peak values as well as annual ones. In addition, the model should be able to calculate the effects of embedding hot water pipes into the fabric of the building model, to enable future design of the DHWall should the DHWall concept prove viable. Hence Dynamic thermal modeling-one of the three levels of CIBSE recognised modeling-was used in this exercise [11], and the program was IES-VE-Pro [12].
As a commercially available dynamic modelling software IES-VE-Pro is widely used in both research and commercial design. It treats a building as a multi-zonal system, models all major thermal processes and solves the coupled heat and mass transfer equations with an accuracy that primarily depends on the quality of input data [13]. Hence it allows a “virtual building” to be built which simulates its thermal performance. In addition to the prediction of thermal properties, heating and cooling loads and energy consumptions, it also offers a diverse range of analyses such as the plants efficiency and carbon emissions.
2.1.1. The Geometry of Case Study House
The OS Map gave a dimensioned footprint of the house. The width of 3.7 m was used to classify the house as being a “small” terraced house; having no ground floor hallway. Assuming the house to be built pre-1919, plans for houses that were built in 1900 (Figure 4) were used to determine the internal layout of rooms [14]. Only one house was dynamically simulated but the adjacent terraces were all included in the model so that their influence on solar gains were included in the heat calculation.
2.1.2. Constructions of the External Walls
Two thermal models were created differing only by the constructions assigned to the front external wall. One is the existing conditions (without insulation) whilst the other is after renovation (with insulation on the outer side
Figure 4. Floor plans and cross-section for four houses built in Norwich in 1900 [14].
of the wall). Table 1 shows an overview of U-values used for the major building elements.
A typical U-value of 2.33 W/m2K for existing external walls was used [15]. The author reports the in-situ Uvalues for existing buildings and compares these to Uvalues calculated by BuildDesk v3.4, where the constructions are based on in-situ thicknesses and thermal conductivity values from BS/EN 12524. The author reveals that in 73% of cases the U-values are over-estimated in the calculating software. Therefore a decision was made to use the in-situ U-value for existing external wall constructions. This was to replicate real life as much as possible (Table 2).
Building Regulations Part L1B states that the improved U-value must achieve 0.30 W/m2K when the area to be renovated is greater than 50% of the surface of that element [16]. The IES model assumes that 100% of the surface area of the front external wall is covered in EWI and so a target U-value of 0.3 W/m2K was used.
The selected insulation slab has a thermal conductivity of 0.036 W/mK. However the BrickShield system uses Rockwool Façade Ultra insulation [17] and it is assumed this has the same thermal conductivity. The insulation thickness was tweaked to obtain the target U-value shown in Table 2.
2.1.3. Other Details
We assumed the building has undergone draft stripping work. Infiltration rates for the EXISTING and EWI models were therefore set to 0.25 ACH, half of the suggested infiltration rate in 2000 Building Regulations.
The standard National Calculation Methodology (NCM) domestic profiles for occupancy, lighting gains and misc
Table 1. U-values assigned to building model constructions.
gains were applied to the relevant rooms.
The heating system, that is the radiator circuit, was set at 90% efficiency. The heating systems were applied to rooms as Table 3. To enable an adequate comparison with a DH supplied Heat Exchanger (HEX) the DHW system had no water storage.
The Fuel and Electricity Order 1980 prohibits the use of fuels or electricity to heat premises above 19˚C [11]. Therefore the internal design temperature was set to 19˚C and Birmingham weather data was used.
2.1.4. Verification of IES Modelling Results
Normally a computer model needs validation to confirm its robustness. Verification was carried out to compare the total annual gas demand, the key output of the sample house with the benchmark data from DECC in (Figure 5), as a full scale validation of this type was difficult.
The total gas demand was calculated using:
Table 2. U-values used for external walls. The thickness and thermal conductivities were tweaked to obtain overall target Uvalues. For EXISTING model U-value of 2.33 W/m2K from [15] and for EWI model 0.3 W/m2K from [16].
Table 3. Heating systems applied to different spaces (DHW system based on floor area so cupboard and stairs included).
Figure 5. Gas consumption of the EXISTING model compared to LLSOA gas consumption data for the area [22] and UK-wide consumption statistics for houses having similar characteristics [23].
(1)
The annual efficiency was set to 85%, the average found in a field trial measuring the actual performance of condensing boilers [18].
2.2. The Network Model
2.2.1. Methods to Calculate Heat Loads in DH Design
In district heating, local networks are linked to district networks which may also serve larger buildings. To finding the heat load a DHN is designed to serve, the following steps are to be taken. Geographic Information Systems (GIS) heat maps are used to identify the zones that could be served by local networks. For example DECC’s recently released heat map can be used to identify areas of high heat demand density [19]. The viable heat density threshold is expressed as an average annual power density of 3000 kW/km2 [1,20]. Within those areas anchor loads-those buildings having a high heat demand, and that are key to development of a DHN-are identified. The peak hourly loads (kW) and annual energy use (kWh) for the identified zones are then found. To do this, energy use factors-annual energy consumption factor (kWh/m2) and hourly energy load factor (kW/m2 and floor areas for each land use in the zone are calculated [21]. Energy use factors are locality specific if they are estimated from local consumption data. The average domestic gas consumption is 13,576 kWh for the LLSOA covering the case study area [22].
Historically, annual space and water heating factors for different building types found in tables could then be used to find the energy factors. However the average building type varies in age, building materials and operating characteristics considerably among localities so locality-specific data is the first choice. In addition, energy consumption data of high resolution is now made publicly available by DECC as part of the drive to achieve national energy policy objectives [23].
2.2.2. Network Layouts
Two network layouts were developed for the project: conventional DH (Figure 6) and DHWall (Figure 7). Based on the physical street parameters shown in Table 4 two pipe sizes were designed for:
• the street-pipe-supplying all heat demand of street;
• the house-pipe-serving each individual house.
Figure 6. Conventional DH layout with streetpipe buried under road and housepipes branching off to individual houses. The flow pipe is indicated by red and the return pipe by blue.
Figure 7. DHWall with streetpipe integrated into EWI and housepipe dropping down to serve individual houses.
Table 4. Guildford Street characteristics.
2.2.3. Design Peak Load
The following describes the procedure for converting the domestic heat demand to heat required from the network.
Domestic hot water demand is required at different times for different houses. To account for this the standard practice is to apply a diversity factor to the DHW load [21,24]. A diversity formula that works in practice is the Danish diversity factor [25]. For 46 houses the individual house diversified DHW demand is 4.34 kW. However this would only be required for a short amount of time, for example a ten minute shower. As the Heat Exchanger (HEX) prioritises DHW it would be able to cover such loads with little impact on its ability to supply space heat. Therefore the following calculations are done on space heating demand only.
Figure 8 shows full carrying capacity is only required on a few days per year. To design for this load would result in an oversized system. In line with manufacturer’s guidance [26], a smaller peak design load was selected, using a “well-sizing factor” of 80%. Inspection of Figure 8 found this to result in only 20 hours per year of undersupply. Table 5 shows the evolution of the peak space heating demand to the heat required from the network. The peak load for the house was then multiplied by the number of houses to find peak load for a street.