Value engineering of quay wall at Teesport No. 2 roll-on/roll-off

c′

drained cohesion of soil

c_u

undrained shear strength of soil

c_w

interface cohesion of cohesive soil

E′

drained Young’s modulus of soil

E_c

Young’s modulus of concrete

E_s

Young’s modulus of steel

E_u

undrained Young’s modulus of soil

I

second moment of inertia of combi-wall and sheet pile counterfort (SPC) wall

k₀

coefficient of earth pressure at rest

N_SPT

standard penetration test blow count

γ

unit weight of soil

Δγ_s

incremental shear strain

δ

interface friction angle of the granular soil

ν′

drained Poisson’s ratio

σ_h′

horizontal effective stress

σ_v′

vertical effective stress

φ′

effective friction angle of the soil

Teesport, owned and operated by PD Ports, is a deep-sea port located on the southern bank of the River Tees in Middlesbrough, northeast England. It handles over 28 Mt of cargo annually and is a crucial logistical hub for various industries, including steel, petrochemicals and manufacturing. In 2020, PD Ports sought to attract and secure new shipping contracts between northeast England and continental Europe, but this was restricted by the condition and arrangement of the existing berth. To address this, PD Ports committed to deepen the berths, construct new quay walls and construct an outer harbour basin, which included the No. 2 roll-on/roll-off (Ro-Ro) linkspan replacement.

The original No. 2 Ro-Ro consisted of a floating linkspan deck supported on an existing bankseat and guide pile. Adjacent to the linkspan was a sheet-piled quay wall with a retained height of 4.4–11.4 m, constructed in phases between 1961 and 1989 (Figure 1). The berth provided Ro-Ro services for vessels up to 195 m long. Condition surveys performed on the No. 2 Ro-Ro structures showed them to be in poor condition and that they were approaching the end of their design life. It was decided to relocate the bankseat 25 m further inland, replace the quay walls to accommodate a deeper berth and install a new floating linkspan to accommodate larger vessels of up to 220 m in length.

The project delivery team included PD Ports as the ultimate client, Royal HaskoningDHV as the client's technical advisor, Land and Water Services as the principal contractor and Tony Gee and Partners as the temporary works designer and the designer of the permanent quay wall. The supply chain included Ravestein as the pontoon fabricator and Bachy Soletanche as the bored concrete piling contractor.

The success of the project was dependent on effective collaboration between the delivery team and the client team to overcome key challenges and constraints, as set out below.

• There was a requirement to keep the existing berth operational for as long as possible. This required meticulous pre-planning and temporary works design to support the existing piled wall and ensure operational continuity without compromising safety or efficiency.

• There was a very constrained construction schedule to meet the firm dateline, as the linkspan was being fabricated and shipped over from the Netherlands and could not be stored at Teesport once delivered.

• The breakout of the Ukraine war in spring 2022 led to a significant overnight increase in steel price and affected its availability.

• There were significant space constraints as a warehouse was located near the proposed quay wall, with live services situated in between them.

This paper presents the design development of the quay wall design, the analysis that supported the selected solution and the collaboration of the project delivery team. This led to a value engineered solution that delivered a net saving in the region of £0.75 million on the £5 million project and was instrumental in unlocking the entire scheme.

The ground conditions at the Teesport No. 2 Ro-Ro site consist of Made Ground overlying estuarine and marine alluvium underlain by Mercia Mudstone.

The site investigation for the project was undertaken in March 2017 and comprised three cable percussion boreholes with rotary follow-on. In situ standard penetration testing (SPT) was carried out and samples were recovered for contamination and geotechnical sampling, including undrained triaxial testing, unconfined compressive strength testing, point load index testing and various index tests.

The Mercia Mudstone was generally logged as extremely weak to weak, completely to highly weathered and moderately to highly fractured calcareous mudstone. Weathering grades, in line with Ciria 570 (Chandler and Forster, 2001), ranged from Grade I to Grade IVb. The mudstone was interbedded with weak to medium strong siltstone between 20.7 and 30.6 m below ground level. There was thus a risk of pile refusal upon encountering harder strata within the mudstone below −12.90 m chart datum (CD). The geotechnical properties of these materials are summarised in Table 1. The tidal water level was recorded to range between 0.00 m CD (lowest astronomical tide) and 6.10 m CD (highest astronomical tide).

The reference design consisted of a suspended concrete quay supported by cased bored concrete piles, with a grout mattress placed on the slope underneath to avoid severe scour caused by propellor wash from vessel movement. Evaluation of the reference design identified savings that could be made in the construction and maintenance cost. These are detailed below.

• Removal of the existing retaining wall and construction of the new structure using land-based plant rather than marine-based plant as required by the reference design could generate significant upfront cost and programme savings.

• Eliminating the need to work directly in or over the water would offer the contractor improvements from a health and safety perspective.

• The grout mattress on the slope would need to be installed by divers, making it costly to install and introducing health and safety hazards within a live port environment.

• The existing anchored sheet pile wall would need to be removed prior to excavation of the slope, which could compromise the warehouse foundations and adjacent services and suspend shipping operations to avoid washout of the exposed slope material.

• From PD Ports’ maintenance experience, waves and prop wash had travelled up the embankments below suspended decks, continually splashing the underside of the concrete suspended deck. Over time, the seawater degraded the bottom face of the concrete, which is difficult to inspect and repair. This was the key reason that, in the tender documents, PD Ports required that the quay had a design life of 50 years.

During the tender period, the reference design and specification were comprehensively reviewed. It was concluded that if the reference design could be value engineered, including the partial redesign of the adjacent quay wall, it would yield a potential saving to the client of approximately £245 000 and reduce the amount of downtime required for operational shipping.

To achieve this, the project delivery team focused on a combi-wall solution as it would provide better durability and could easily surpass the 50-year design life with only occasional maintenance. The following alternative combi-wall options were considered during the tender design (Figure 2). Each had its own construction challenges.

• Combi-wall tied back with an anchorage wall. The challenge with this solution was that due to the proximity of the existing warehouse and live services, there was limited space for the anchorage wall. This meant that it would need to be placed close to the combi-wall, leading to interaction of the active and passive wedges resulting in high forces on the combi-wall and tie rod, and increased wall displacement.

• Combi-wall anchored back using ground anchors. The challenge with this solution was that the ground anchors would have needed to be installed with a steep inclination to avoid clashing with the piled foundations of the existing warehouse. Preliminary analysis suggested that the anchor design was very inefficient, which meant a large number of closely spaced anchors that would be difficult to install and financially prohibitive.

Following contract award, the design team sought to develop the tender design to overcome the challenges faced by the combi-wall solution. During a workshop, where the limitations of the tender alternatives and the site constraints were reviewed, a solution of using a series of sheet pile counterfort (SPC) walls, each tied into the combi-wall through reinforced concrete capping beams, was selected. This solution allowed a smaller wall footprint that could fit in the area in front of the warehouse (see Figure 3). The counterfort walls could be installed between existing tie rods (this enabled the sequential construction of the new wall), with the decommissioning of existing ties safely planned and executed once the new wall superseded the functionality of the existing wall. It was recognised, however, that this solution would present design challenges of its own. Although the use of cross/buttress walls to reduce wall and ground displacements had been studied previously (e.g. Ou et al., 2008; Soccodato et al., 2015) and was a well-established technique, the design team was not aware of any use of a counterfort sheet pile wall to provide anchorage/buttress restraint. While first-principle hand calculations had established the viability of the solution, the design team was aware that substantial verification would be needed to reassure the client’s technical advisor that the design basis was sound.

Building on the first-principle calculations, two-dimensional (2D) proprietary software was used to investigate the system's behaviour.

The combi-wall and the SPC wall acting as the return wall were assessed using Geosolve Wallap.

The following idealised failure modes were assessed for the geotechnical stability (see Figure 4).

• The forces leading to the direct pull-out of SPC walls from the soil block. Only the interface friction of the SPC walls located outside the active wedge of the combi-wall was considered for the pull-out resistance as there was a concern that the soil within the active wedge would displace together with the combi-wall, reducing the interface friction.

• The forces leading to block sliding failure at the toe of the SPC walls.

• The forces leading to block sliding failure at dredge level.

• The forces leading to overturning pull-out of sheets from the soil block. The overturning resistance of the SPC walls was considered to be provided by the vertical tension/compression in the counterfort walls. It was anticipated that compression resistance (towards the front of counterfort walls) would largely be provided by the end bearing of the sheets whereas the tension (towards the rear of counterfort walls) would be provided by interface friction of the counterfort walls located outside the active wedge of the combi-wall.

• Surcharge and active earth pressure leading to sliding (block) failure.

• Surcharge and active earth pressure leading to overturning failure. Only the interface friction of the SPC walls located outside the active wedge of the combi-wall was considered for the pull-out resistance.

The restraint provided by the counterfort walls was applied to the 2D wall model as a lateral restraint at the top of the wall to replicate the point of connection from the capping beam. Due to uncertainty about the effectiveness of the counterfort walls in providing sufficient lateral restraint, as a proportion of the sheet piles was located within the active wedge of the combi-wall, parametric studies were performed to evaluate how the structural forces and displacements in the wall varied with both ‘soft’ and ‘stiff’ buttress restraint behaviour. In addition, various groundwater levels (GWL) were considered for the combi-wall design, namely

• GWL at 5.55 m CD at both active and passive sides

• GWL at 4.90 m CD and 0.38 m CD at passive and active sides, respectively

• GWL at 3.15 m CD and 5.55 m CD at passive and active sides, respectively

• GWL at 3.15 m CD at both active and passive sides.

Finally, parametric studies were also performed with pore water pressure balancing at the toe of the combi-wall, considering that there is a high degree of fracturing within the mudstone.

These assessments demonstrated that the requirements of the Eurocode 7 (BSI, 2004) design approach 1 combination 1 (DA1-C1) and design approach 1 combination 2 (DA1-C2) were satisfied. The combi-wall’s shear force and bending moment of various key construction stages are shown in Figure 5.

The factor of safety (FoS) against global instability was investigated using Rocscience Slide2. Two scenarios were considered when computing the FoS against global instability, namely slope failures in between SPC walls and adjacent to the SPC walls. The Eurocode 7 DA1-C2 FoS was estimated to be 1.08–1.33. Higher values of the FoS were computed at the location of SPC walls as the critical slip path was pushed beyond the SPC walls.

The wall solution established comprised 1.42 m dia., 25.8 mm thick tubular piles, spaced at 2.885 m centres, with AZ46-700 infill sheet piles. The toe levels of the combi-wall varied between −14 m CD and −16 m CD, depending on the design dredge level. The quay wall capping beam was 2.0 m wide by 2.0 m deep, with the top of the capping beam at the ground level of 7.4 m CD. The SPC walls comprised ten pairs of AZ46-700. These counterfort walls were spaced at approximately 5.5 m. The toe level of the SPC walls was −10.6 m CD. The SPC walls capping beams were 2.0 m wide × 1.5 m deep.

The construction sequencing of the new quay wall was meticulously planned to reduce the downtime of the existing berth. Construction of the combi-wall was thus to commence prior to the removal of the existing sheet pile structure. The combi-wall and SPC walls were to be installed from behind the existing retaining wall using land-based plant. The sequence comprised pre-auguring of sheet piles only, followed by impact driving to depth. So as not to impact the characteristic properties of the mudstone, as softening of the top of the Mercia Mudstone under exposed marine conditions was historically observed at the site, pre-auguring and internal auguring (or reaming) of the tubular wall piles was not proposed. The sequence was as follows.

• Construct a working platform in front of the existing warehouse and live services.

• Install additional temporary ties between the existing sheet pile wall and the existing anchorage wall, where required.

• Remove existing clashing ties as piling works progressed.

• Install the tubular piles forming part of the combi-wall. The existing tied sheet pile quay wall was kept in place.

• Construct additional temporary ties between the existing sheet pile wall and tubular piles.

• Remove existing tie rods between the existing sheet pile quay wall and the existing anchorage wall.

• Install infill sheets between the tubular piles that form part of the combi-wall.

• Install SPC walls.

• Cast in situ combi-wall capping beam and SPC walls capping beams.

• Remove remaining ties.

• Remove existing sheet pile quay wall and dredge to final dredge level.

While the assessments using separate 2D propriety software showed favourable results, the design team was concerned that, with the complex soil–structure interaction, there was a risk that a 3D failure mechanism had been overlooked. Therefore, a Plaxis3D model was created to confirm that the first-principle calculations and the Wallap and the Slide2 modelling were conservative and the soil–structure interaction had been adequately interpreted.

Figure 6 illustrates the idealised geometry for the combi-wall system. A strip of 5.5 m width was modelled in the Plaxis3D analysis, with a refined mesh generated adjacent to the combi-wall system to avoid ‘element locking’. This mesh comprised approximately 110 000 elements and 175 000 nodes.

The combi-wall and SPC wall were modelled as plate elements with a flexural stiffness of E_sI, where E_s is the Young’s modulus of steel (taken as 200 GPa) and I is the second moment of area of the steel tube per metre run, ignoring the contribution of infill sheets. Although there was a gap between the combi-wall and SPC wall, these walls were connected through capping beams modelled using linear elastic solid elements with the short-term and long-term Young’s modulus of concrete (E_c) taken as 35 GPa and 17.5 GPa, respectively. The connection of these capping beams was modelled as a fully fixed connection (i.e. moment transfer allowed).

As a ‘first-order’ approximation, a linear elastic–perfectly plastic Mohr–Coulomb model was used to represent the behaviour of all soil and rock layers (Brinkgreve et al., 2012). The characteristic geotechnical design parameters used are summarised in Table 1. To take account of pre-boring during wall installation, the wall and soil interface properties were lowered to c_w/c_u = 0.25 for the cohesive materials and δ/φ′ = 0.333 for the granular materials.

The surcharge was applied as a uniform distributed load at ground level. The mooring load was modelled as a point load at the top of the combi-wall capping beam.

The assumed sequence in the Plaxis3D model was as follows.

• Initial stress condition, with total stress parameters. Although there was an existing sheet pile quay wall in front of the proposed wall, it was not considered in the model for conservatism. It was assumed that the ground level in front of and behind the proposed quay wall was the same.

• Wish-in-place combi-wall, SPC walls and the capping beams.

• Excavate in front of the proposed quay wall to −10 m CD.

• Softening of the extremely weak mudstone between −10 m CD and −10.75 m CD and the installation of the permanent berm in front of the quay wall.

• Consolidation.

• Long-term condition.

  • Soil properties switch to effective stress parameters.

  • Young’s modulus of capping beams switched to long-term Young’s modulus of 17.5 GN/m².

  • Combi-wall and counterfort wall properties switched to the corroded properties (i.e. tube dimensions of ∅1420 × 22.15 mm, considering a thickness reduction of 1.825 mm on the inside face and 1.825 mm on the outside face).

• Applied Eurocode 7 DA1-2 partial factors for actions and material strength.

• Restore the long-term condition stage (i.e. stage f), apply a 15 kN/m² surcharge behind the quay wall. Consider tidal lag with GWLs of 0.38 m CD and 4.90 m CD in front and behind the combi-wall.

• Apply Eurocode 7 DA1-2 partial factors for actions and material strength.

• Restore the consolidation stage (i.e. stage e), apply tidal lag (similar to stage g) and bollard load at the top of the quay wall capping beam.

• Applied Eurocode 7 DA1-C2 partial factors for actions and material strength.

Figure 5 shows the combi-wall shear force and bending moment profiles at construction stages h to k, respectively, showing that Plaxis estimated lower combi-wall structural forces. As the Plaxis3D model had the capability of capturing the more complicated soil–structure interaction, it demonstrated that the SPC walls could provide the lateral and rotational restraints through interface skin friction even though partially located within the active edge of the combi-wall.

The moment restraint at the capping beam, conservatively not accounted for in the Wallap analyses, was seen to help reduce the maximum combi-wall bending moment. The Plaxis model confirmed that the combi-wall capping beam would be subjected to high torsional moments at the connections between the front capping and the counterforts. It was determined that these connections would need to form the load transfer point between the combi-wall and counterforts. While a clutched connection between the tubular piles and counterfort sheet piles could be installed (and was included where possible as a guide for alignment of the sheet piles), there were not easily identifiable methods for verifying the integrity of such a connection, and it was therefore not considered appropriate to rely on such a detail as the load path into the counterfort walls. The forces in the capping beam, including torsional moments, were accounted for in the concrete reinforcement design, with the combi-wall capping beam and connected counterfort capping beams assessed separately in a structural Lusas model that considered the interaction of adjacent counterforts.

The Plaxis-estimated vertical effective stress (σ_v′) and horizontal effective stress (σ_h′) behind the combi-wall at selected construction stages are shown in Figure 7. It was observed that σ_v′ adjacent to the counterfort sheet pile walls was lower than σ_v′ in between the counterfort sheet pile walls between −7 m CD to 6 m CD (Figure 7(a)). Below −7 m CD, a higher increment of σ_v′ was observed adjacent to the counterfort sheet pile walls compared to that in between the counterfort sheet pile walls. This suggested that a small percentage of the vertical surcharge was transferred to a lower level through the counterfort walls. A similar trend was observed for σ_h′ (Figure 7(b)). However, the lower σ_h′ adjacent to the SPC walls could be due to the reduced vertical effective stress and the potential beneficial effect that the SPC walls would provide lateral restraint to the soil within the active wedge of the combi-wall.

The Plaxis model also demonstrated that, in stage h, the proposed combi-wall FoS against global instability (of ≈1.04, with Eurocode 7 DA1-2 partial factors applied) complied with the Eurocode 7 requirement, although this was lower than the Slide2 analysis. As shown in Figure 8, the Plaxis model automatically considered pull-out of the SPC wall, generating a smaller critical slip path.

These results reinforced the overall findings of the first-principle calculations and 2D modelling outcomes, and confirmed that the proposed design was geotechnically stable and code compliant.

During construction of the quay wall, the counterfort walls’ sheet piles were initially pre-augered, followed by vibratory installation methods using an ABI TM20 leader rig and MRZV 18VV and 30VV hammers. These sheet piles were then driven to the target level using a BSP CX110 hydraulic hammer.

As detailed earlier, pre-auger was not carried out for the installation of the tubular piles due to concerns about its impact on rock properties. Initially, the piles were driven using a Juntann HHK12S impact hammer, which met practical refusal at −8.3 m CD to −10.7 m CD. Subsequently, a BSP CG300 impact hammer was used, which also met practical refusal, although slightly lower at approximately −9.6 m CD to −11.2 m CD, which was approximately 1.7 m above the weak to medium strong siltstone layers.

Due to programme and cost constraints, there was a need for the principal contractor, piling contractor and designer to quickly develop a remedial solution to provide the necessary pile embedment. Based on previous project experience, the solution developed comprised extending the piles with a 1200 mm dia. auger to −16 m CD and −17 m CD, and then inserting a 1084 mm dia. 34.1 mm thick tube into the bore, to overlap between 2.5 and 4 m with the refused 1420 mm dia. tubes. The bore and original tube were then infilled with C32/40 concrete, extending from the base of the bore to the underside of the pile cap sockets, to create a structural connection.

Instrumentation was proposed to verify the quay wall design and protect the construction workers, public, existing infrastructures, structures and buried services

• 14 3D reflective prism survey targets on the combi-wall structure

• four 3D reflective prism survey targets on the existing suspended deck structure

• 22 retro-reflective survey targets at the existing warehouse

• four tiltmeters at the seaward side of combi-wall structure

• two tiltmeters at the seaward side of the existing return wall.

It was anticipated that monitoring of the structures would be accompanied by appropriate reporting of the tidal level, tidal direction (i.e. rising or falling) and water level within the dock at the time the readings were taken. However, instead, manual survey monitoring was performed with the 3D reflective prism survey targets due to the long procurement time and challenging environment for installing automated instrumentation (i.e. busy construction plant/ship movements meant that the targets would always be obstructed and there was no safe place to install the survey station). The manual survey monitoring confirmed that no red trigger levels were breached during construction. The red trigger levels were set as predicted SLS movement. Where no meaningful SLS movement prediction could be made, the equivalent of SLS movements was taken to be the value of movement above which there would be concern in terms of structural stability.

PD Ports committed to deepen Teesport's existing berths and construct new quay walls, including the No. 2 Ro-Ro replacement to attract and secure new shipping contracts between northeast England and continental Europe. A key challenge for the project was to maintain port operations for as long as possible, which necessitated a very tight construction programme. The project delivery team embarked on a novel design solution to minimise the construction, reduce maintenance and overcome restrictive site constraints.

This solution consisted of a combi-wall system with lateral restraint provided by SPC walls. The combi-wall system was initially designed using first-principle assessments and 2D proprietary software. Although the assessments showed favourable results due to the complex 3D nature of soil–structure interaction, it was necessary to investigate the design further using Plaxis to confirm that the 2D assessments were conservative and a failure mechanism had not been overlooked. This model confirmed the first-principle assumptions and showed that the SPCs provided adequate lateral and rotational restraint, even though part of the SPC walls is located within the combi-wall active wedge. Furthermore, the model revealed that a small percentage of the vertical load was transferred to a lower level, helping to reduce the average active pressure acting on the combi-wall. The model also showed that, while the capping beams lining the front wall to the buttress walls provided moment restraint at the capping beam level and so reduced the maximum bending moments in the combi-wall, large torsional forces were developed that needed to be accounted for in the connection detail.

During construction, the tubular piles of the combi-wall refused at approximately −11 m CD, necessitating the development of a remedial solution to advance the piles. This comprised drilling below the installed piles, installing smaller diameter tubes over the extended zone and infilling the base of the bore to the underside of the pile cap sockets with concrete.

Manual survey monitoring performed on site indicated that no red trigger level was breached during construction. This verified the combi-wall system design and the findings of Plaxis model.

The project team embraced the challenge of developing a novel design solution to overcome construction and logistical constraints. The solution unlocked the scheme by providing a saving of 5% of the overall construction cost, minimising port operation downtime and reducing future maintenance activities.

The authors express their sincere gratitude to PD Ports for granting permission to publish this paper. The willingness of PD Ports to share information was crucial for the development of the findings of this work. The authors also acknowledge colleagues who contributed to this project.