Over the coming pages I shall be describing the scheme carried out by MWH as part of the Virtual Design House (VDH) for Dwr Cymru Welsh Water Alliance (South East Team). The scheme is being implemented as part of the AMP3/4 Barry Bathing Waters Project. The scheme demonstrates the benefits that can be gained by challenge, focus on environmental benefit, enhanced hydraulic modelling, and integrated solution development.

The Barry West catchment area comprises the western half of Barry, Barry Island, Rhoose, Fontygary and Tredogan. The catchment has a total area of approximately 531ha and population of 33,000. The southern edge of the catchment is bounded by the Bristol Channel with the Barry East catchment to the east. Barry West comprises five areas/subcatchments: Porthkerry and Rhoose; Inland Barry; Knap; Barry Island and Docks; Barry Town and Broad Street.

The sewerage network within Barry is fully combined, although newer developments around the town are utilising separate surface water systems. The systems in Rhoose and Fontygary are separate but there is a considerable storm response within the catchments, due to misconnections, poor workmanship and other issues.

There are 23 CSOs and 27 pumping stations in the Barry West catchment. They range from small unscreened ‘hole in the wall’ type structures to large complex pumping stations with associated off-line storage tanks and mechanical screens.

The catchment drains to the Terminal Pumping Station at Barry Town. This in turn pumps to Cog Moors WwTW for treatment. The pass-forward flow to Cog Moors from Barry Town Pumping Station is approximately 770 l/s. There are four main receiving waters in the Barry West catchment: Nant Talwg, Bristol Channel, Old Barry Harbour and Barry Docks. All the assets spill to these receiving waters, either directly or via surface water systems.

Environmental standards

There are three EU-designated bathing waters in the Barry area: Jackson Bay and Whitmore Bay are immediately off Barry Island; Cold Knap is a short distance to the west. Other ‘non-identified’ bathing waters are located around the adjoining coastline.

The spills from these assets, plus the assets in Aberthaw, Barry East, Penarth and Cardiff, were initially perceived to cause impact on the three bathing beaches above. Thus the default solution criterion was to restrict each of the spills to three spills per bathing season agglomerated over the wider catchment area.

Service drivers

The project also needed to provide solutions to 38 DG5 flooding properties plus a further 10 SEF locations. All solutions needed to meet

1-in-30 return period flooding.

Maintenance drivers

Many of the pumping stations in the catchment had maintenance issues, and the operation of some assets (in particular Barry Town Pumping Station) was overly complex, thus increasing operational risk.

Development drivers

The catchment also has significant growth predicted in the Porthkerry catchment and in the brownfield areas around the docks. The dock area has been earmarked for residential and commercial development equivalent to 10,000 population.

Catchment strategy

A delivery team was set up for this project, loosely referred to as the Virtual Design House. This consisted of designers from MWH, Arup, Hyder Consulting (HCL) and Black & Veatch (BV). This team would agree the overall strategy and manage the delivery of the project.

To deliver this project for the best value against a tight budget and timescale, the VDH catchment strategy had three main thrusts:

Enhanced hydraulic modelling

Hydraulic models for all the VDH catchments had been developed in Hydroworks. These models had been verified using asset-and-flow survey data obtained in 1999. The VDH agreed that a full review of each of the models should be implemented. The main recommendations of the review, undertaken by MWH, were: each model needed to be re-verified utilising the new UK runoff model; the ground infiltration model should be used to represent any shortfall in infiltration; the resolution of each model should be improved.

Regulatory challenge and coastal modelling

The default environmental driver for not only this catchment but all the VDH catchments (Barry West, Barry East, Penarth and Cardiff) was the standard bathing waters driver. The three-spills-per-bathing-season driver would have forced onerous solutions, and the agglomeration of this driver across multiple catchments would constrict the project delivery and possibly drive up the solution size even further.

In order to challenge the driver, a dispersion model of the Severn Estuary was constructed so as to undertake an impact analysis using :the results of the re-verified hydraulic models. The result of the impact analysis was to stratify the assets in terms of impact on the bathing waters (Table 1).

By classifying each asset in terms of its environmental significance we could pinpoint the assets that caused real impact to the beaches and thus target the investment to maximise the environmental benefit. This strategy was agreed with the Environment Agency and a revised list of assets and drivers developed.

Integrated solution development

Designers working for the VDH were empowered to develop integrated solutions that met the environmental drivers and also tackled the flooding, rehabilitation and future development needs of each catchment.

In terms of Barry West the need for integrated solutions was clear: there was significant existing in-catchment flooding, development drivers and limited construction space at a number of key assets.

As part of the VDH, MWH was assigned Barry West. The other catchment, Barry East and Penarth, fell to HCL and BV to complete, with Arup tasked to resolve Aberthaw (and key assets in Cardiff).

To enhance the Barry West model, an asset data collection exercise was done. The data collected was from a number of sources: Dwr Cymru Welsh Water’s AIS (main GIS) system; existing and new asset surveys; as-built drawings from key locations; local operator knowledge; original Linen sewer record plans.

The data was managed through MWH’s Data Manager software and used to amend the original model database. The significant number of pumping stations in the catchment required particular attention to be paid to pump performance over the head ranges in each sump. The model build exercise added significant detail to the model and increased the confidence for solution development.

In addition to the existing 70 monitor flow surveys, a new flow and impermeable area survey was procured, to be targeted at areas of high infiltration, slow response and ‘opportunity’.

For example, a large section of Barry Island had been redeveloped for housing since the first survey. These developments had new separate systems, whilst the original land use had been combined. Further flow monitoring showed the net effect to be a reduction of 500 l/s storm runoff contribution during the critical one-in-five event.

Once the enhancement and verification exercise was completed the verification reports were submitted to the EA for approval.

Impact analysis

The reverified hydraulic model was re-run in Infoworks over the 11 bathing seasons contained in a rainfall dataset recorded at Rhoose (Cardiff Airport). The results were then input in the coastal model with all the other VDH catchments to allow the impact analysis to be undertaken. A summary of the assets and revised initial drivers is given in Table 2 on the next page.

Out of the 30 assets needing improvement, only three retained their original bathing waters driver, 14 were reduced to local drivers and the remaining 14 would require further work to determine an appropriate driver.

This split allowed an early release of ‘unagglomerated’ local driver work that could be progressed immediately. These schemes could be designed and implemented whilst schemes were being developed for the more tricky undetermined and bathing water assets.

With the reclassification of the assets, schemes could be developed for the assets in the Porthkerry and Inland Barry sub-catchments, which consisted mainly of pumping stations. Each would need to meet Formula A compliance, DWF failure protection and have adequate aesthetic control. In all cases we were able to meet these drivers by optimising the operational levels, fitting power generators (where needed) and retrofitting 6mm bar screens on the overflows. The larger integrated schemes are documented below.

Knap subcatchment

Knap Pumping Station serves a subcatchment of 3,500 population and had a Formula A/PFF of 65/100 l/s. It drains a steep catchment which is partially separate. Within the catchment there were four DG5 properties and another CSO located within an environmentally sensitive wood. The discharge from the Knap was directly onto one of the bathing beaches and thus a bathing waters driver was appropriate. Model runs indicated that the asset would require approximately 1,200m3 to meet the agglomerated three spills requirement.

Within this subcatchment two key improvements were made to the model. The first was the inclusion of a 380m-long twin 300dia inverted siphon with a fall of approximately 1 in 14, serving the Porthkerry subcatchment to the west; the pumped flows from this area pass (with the help of gravity) through the siphon directly to the Barry Town Terminal Pumping Station.

The second improvement was a bifurcation that also sent flows directly to Barry Town; it had become abandoned due to a downstream sewer collapse, causing all flows to pass to the Knap Pumping Station. The inclusion of both these assets in the model allowed us to utilise them in the solution. The resulting scheme consisted of relatively minor works.

Firstly, we assessed the available capacity of the inverted siphon and disconnected the incoming sewer from an in-catchment CSO at Marine Drive. This allowed us to decommission completely the

CSO without the need to undertake any significant construction work in the protected woodland. We took advantage of this new connection and diverted the maximum allowable flow into the new sewer from other parts of the upstream systems, thus maximising the benefit at both the Knap Pumping Station and the DG5 properties in the catchment.

Secondly, we rehabilitated the abandoned bifurcation. By doing this we allowed flows up to one-in-one year storms to pass directly to Barry Town rather than the Knap Pumping Station. Again, we took advantage of the new connection and upsizing to divert flows up to one-in-one year storm events to the bifurcation. The above works halved the population acting at Knap Pumping Station and reduced the storage needed there from 1,200m3 to 450m3.

Thirdly, to design out the storage completely we optimised the station’s operation. This was achieved by changing the impellers in the foul pumps to pump approximately 140l/s from 84l/s. The solution also utilised the 200m3 storage available in the storm sump by raising the it as high as possible and repairing a broken scavenger pump. Thus flow would enter the storm sump but only be pumped to the bathing waters when the sump was full; if, during a storm, the sump did not fully fill, the scavenger pump would return it to the foul sump.

The above packages of works reduced the spills at this asset from 46 annually to <1 (thus meeting the bathing drivers), reduced DWF pumping volume by 40% and solved a number of DG5 flooding problems without the construction of any additional storage. Holton Road

At Holton Road, Barry’s main shopping street, we had to deal with 11 CSOs and 36 DG5 properties. The area has experienced serious flooding for over 20 years, due to the condition of both the foul and surface water sewers and the steepness of the catchment.

The CSOs in the area all spilt to a surface water system which drained into the dock.

A major difficulty in this scheme was finding available sites to locate any new CSO chambers. We considered multiple new assets and pump-away schemes, but both were unviable. Thus we chose to build a single new overflow with major resewering to allow the abandonment of all the others. Significant disruption would be caused to the businesses and extensive negotiation with the council, third parties and the public were required. However, as the scheme removed significant flooding, it was supported by council and residents alike.

The PFF flow from the new structure was designed to offer protection to the downstream system so that, during extreme events, its throttle pipe would become submerged, reducing PFF. This offered significantly more flood protection for the downstream system. The performance of the new overflow was more efficient than the old assets in terms of passing more DWF forward before spilling, so, despite passing less peak flow forward, the volumes and frequencies of spill also dropped (Table 3).

We achieved a low-spill (Barry Town Pumping Station

Barry Town is the terminal pumping station for the catchment and the centre of the integrated catchment approach; it is due for improvement in AMP4. The site has a complex of six individual pumping stations either lifting flows into the site, pumping forward to treatment, or pumping to sea via the long sea outfall or short sea outfall.

The site was earmarked as a location for any required storage, due to the availability of land and existing facilities at the site. The nature of the in-catchment works had increased the volume and peak flows arriving at the site and thus its spills, initially via the LSO then the SSO. This was accepted by the EA as a temporary case due to the large reductions at other more sensitive in-catchment sites like the Knap.

To develop solutions for this complex asset, the strategy needed several approaches: increase flow to treatment to meet future Formula A, 770 to 820 l/s; integrate scheme with Broad St and development needs around docks; optimise the site’s assets and controls; develop a strategy to allow best use of existing long sea outfall.

In order to achieve an integrated solution with Broad Street, we are proposing to build a new tank sewer draining either to a new overflow to the dock or to Barry Town. This allows us to offer capacity for the new development, resolve the flooding, and ‘peak lop’ the flows arriving at Barry Town, reducing the need for storage at that site by about 2,000m3. The new sewer is required to convey the flow anyway and had to have a large diameter to meet the required capacity. By oversizing this section we achieve the required volume to offer a new three spills/bathing season CSO spilling either to the dock or to sea. The actual route will be agreed by the VDH team with Dwr Cymru Welsh Water, Associated British Ports (ABP) and the developers.

There was significant scope to improve the existing operation at Barry Town. For example, the existing tank on the site was designed to drain continuously, which caused the flow to recirculate around the site until it could find a way out either to treatment or to sea. The simple addition of a motorised penstock on the tank to control its emptying will smooth the flows significantly. Additional works include construction of a weir wall at the SSO highpoint utilising the 800m3 storage available in the SSO rising main that needs to be filled before spillage to sea.

The 1.5km-long sea outfall at Barry Town had been designated an undetermined impact on the bathing waters, whilst the SSO was significant due to its discharge proximity to the beach. We could construct additional volume at the site to limit the spills down the LSO to 3/BS; however the asset would become redundant since its use would be in parallel with the SSO. So rather than provide the conventional three-spills solution, we chose to re-run the impact analysis to determine whether a reduced pump rate could achieve a non-significant impact.

This rate was determined to be approximately150 l/s. At 150 l/s the LSO could operate in its current configuration without any limit to its spills. The proposed arrangement provides a non-significant solution for the LSO and a three-spills solution for the SSO.


The Barry West UID Project represents an excellent demonstration of the scheme benefits which can be achieved through the use of detailed hydraulic models, impact analysis (via quality modelling) and integrated solution development.

The use of the coastal model to determine each asset’s impact allowed an early release of low-risk work which could be achieved whilst solution development priority was given to the more complex assets. Had these schemes been developed in isolation with no regard for the bigger catchment picture, the solutions would have cost Welsh Water significantly more (both in capital and operational terms) and increased the construction risk, whilst delivering less environmental benefit.


This article has been reproduced with permission from MWH, AMEC Construction and Welsh Water. Views expressed are the author’s and not necessarily those of the organisations.

Action inspires action. Stay ahead of the curve with sustainability and energy newsletters from edie