OFFSHORE FOUNDATION

FIELD OF THE INVENTION

The present invention relates to a foundation and a method of installation for an offshore structure.

BACKGROUND TO THE INVENTION

There are many circumstances when a structure may need to be installed in an offshore environment. These structures often are securely supported on offshore foundations resting on the seafloor.

Offshore wind power generators are installed in bodies of water such as lakes and seas for producing electricity from wind. Offshore wind power, despite its growth, remains relatively more expensive than onshore wind power, partly because of the massive foundation required to support a wind power structure in an offshore environment.

An offshore wind power foundation supports the wind power plant which typically includes a tower and a nacelle which includes a turbine and a rotor. The foundation must resist a challenging combination of large horizontal forces and moments generated by waves, wind and the motion of the nacelle. Several different foundations are currently employed in offshore wind power plants depending on site conditions and water depth.

In shallow waters, gravity foundations are mainly used. Gravity foundations typically consist of a cement foundation ("footing") that rests on top of the seafloor simply under gravity. The footing provides support for a pile or tower extending vertically above the sea surface on top of which the wind power tower is fixed. Another foundation structure, the so-called "guyed" system, combines a central tower secured with the help of a plurality of pre-stressed cables anchored at a distance from the tower. Through such an arrangement, the moment forces are mainly undertaken by the cables thus reducing the overall tower requirements. The guyed systems are more suitable for onshore wind power but may also be used in shallow waters.

In medium water depths a single elongated pile, commonly referred to as a monopile, is hammered into the seafloor. The wind power tower is fixed on top of the monopile. Monopiles are well accepted in the industry mainly due to their proven track record in oil and gas offshore installations. For instance, in the UK the vast majority of offshore wind power generators are supported on monopiles. Typical monopile dimensions may range from about 30 to about 40 meters in length, and from about 4 to about 6 meters in diameter. The process of hammering such large monopiles into the seafloor entails great engineering difficulties. These difficulties increase considerably as the dimension requirements for the monopile increase for larger wind power generators or applications in even deeper bodies of water.

Another foundation, often used in medium depths is the so-called "suction caisson". A suction caisson is typically a skirted circular footing which is installed into the seafloor by pumping water out from the footing to create a vacuum force that drives the vertical wall of the footing into the seafloor. Suction caissons reduce the need for hammering, but require heavy, airtight structures and rather involved installation procedures.

In deeper waters, multi-footing structures are typically employed consisting of a plurality of piles and/or suction caissons. Such foundations can resist the increased forces and moments due to wind and wave loading experienced in deeper waters but they are generally more expensive to build and install.

Various other foundations have been proposed in the literature.

Examples of proposed monopile foundations are described in patent documents, DE202005004739, CN2002265837, and KR20120001384.

In an effort to reduce the size requirements for the monopile structures, hybrid, monopile-type foundations have been proposed. For example, a hybrid, monopile-type foundation which combines a single pile with a very rigid and heavily loaded foundation, made of reinforced concrete is proposed in Stone KJL, Newson Ta and Sandon J. (2007), "An investigation of the performance of a hybrid monopile-footing foundation for offshore structures", Proceedings of 6 ^th International Conference on Offshore Site Investigation and Geotechnics, London: SUT, 391-396. See also, Stone KJL, Newson Ta and El Marassi M. (2010), "An investigation of a monopiled-footing foundation, International Conference on Physical Modelling in Geotechnics, ICPMG2010, Rotterdam: Balkema, 829-833.

Also, European patent application EP19881219A1 of Jaroszewicz describes a hybrid monopile type foundation which includes a single pile, a resistance plate surrounding the single pile, and gravel or other loose and hard material fill the space between the external surface of the single pile and the internal surface of the resistance plate. Jaroszewicz recommends that the resistance plate be made of heavy concrete.

Both hybrid structures proposed by Stone et al. and Jaroszewcz remain large and heavy and thus are expensive to transport and install. SUMMARY OF THE INVENTION

According to a first aspect of the present invention there is provided an offshore foundation for supporting a structure, the offshore foundation comprising a footing and an elongated pile. When installed at the offshore location, the footing may rest on the seafloor with the pile extending vertically through the footing. The footing may be free to move in the vertical direction around the pile. This freedom of movement or decoupling in the vertical direction of the two main structures of the foundation is advantageous because it may allow repositioning of the footing on top of the seafloor to accommodate long-term consolidation settlements of the seafloor.

However, the footing and the pile may maintain an effective engagement in the lateral direction to permit adequate transmission of horizontal or lateral loads and moments from the footing to the pile thus reducing the structural requirements for the footing. Through such decoupling of vertical and horizontal loads, the accumulation of rotation during long-term cyclic loading can be effectively reduced, allowing for a substantial reduction of the length of the pile. Since the vertical loads are undertaken directly by the surface foundation, the soil around the pile is subjected to limited initial plastic deformation, and hence there is less accumulation of plastic deformation. Moreover, the resistance of the pile may be increased due to the increase of vertical stresses, and hence of the horizontal as well, directly underneath the footing. Through such an arrangement, the footing may undertake roughly 100% of the vertical loads, and 30% to 50% of the lateral loads.

The footing may include a mounting section for positioning a structure to be supported thereon. The footing may be constructed to support substantially all vertical loads of the structure of a complete offshore wind power installation, including a tower and a nacelle. Thus, unlike monopile foundations, the loads of the structure, such as dead loads or any vertical forces generated by the nacelle, may be supported substantially entirely by the footing and not by the pile. Therefore, one important function of the footing is primarily to support the vertical loads of the structure. As a result, the dimensions and/or other structural requirements for the pile may be reduced substantially.

The footing may comprise a base structure defining a bore to accommodate the pile. A bore wall structure may surround the bore and may extend above the base structure in a vertical or substantially vertical orientation. The bore wall structure may comprise a bore plate. The bore wall structure may provide additional contact surface for lateral engagement with the pile. The height of the bore wall structure may be selected in order to provide optimum contact surface with the pile in the lateral direction. The bore wall structure may comprise a plurality of sections and may comprise a top section having a mounting section for positioning the structure to be supported thereon. The bore wall structure may extend above the water surface which may permit a supported structure to be readily mounted thereon.

The bore wall structure may extend below the base structure to provide a section that may be inserted into the seafloor.

The mounting section of the footing may comprise an annular section surrounding the bore wall structure and may abut the bore wall structure or be an integral part of the bore wall structure.

The mounting section may be placed away from the bore wall structure to allow for the diameter of the base of the structure that is supported on the footing to be larger than the diameter of the bore. This may be advantageous in allowing a reduction in the diameter of the pile without limiting the design requirements of the overall structure.

The footing may comprise an outer wall structure defining the outer perimeter of the footing. The bore wall structure, the base structure, and the outer wall structure may form a compartment or pocket within which ballast material may be added to increase the weight of the footing and securely anchor it on the seafloor.

The footing may comprise one or more stiffener elements that are used to increase its rigidity. The shape and/or size of the stiffener elements and/or their number may vary depending upon the requirements of each installation. The footing may comprise a plurality of stiffener elements that may be spaced apart at regular intervals around the bore wall structure and may extend radially from the bore wall structure to the outer wall structure. The stiffener elements may be in the form of vertical or substantially vertical plates that together with the bore wall structure, the base structure, and the outer wall structure may define a plurality of compartments or pockets within the top surface of the footing. Ballast material may be added inside one or more of these compartments.

The bottom surface of the base structure of the footing may comprise a flat or substantially flat portion in order for the footing to establish good contact with the seafloor. The footing may be installed directly on the soil surface, or on a levelled gravel bed of appropriate thickness. As the footing is able to move freely in the vertical direction any settlement or change on the surface of the seafloor below the footing will cause the footing to move to re-establish contact with the seafloor.

The bottom surface of the base of the footing may comprise one or more protruded sections that may be shaped to provide additional anchoring or binding to the seafloor. For example, the bottom surface of the footing may comprise a plurality of short protrusions, spikes or similar devices that may be inserted into the seafloor. Such spikes may be advantageous in securely positioning the footing, and reducing the overall lateral load exerted on the pile via its engagement with the footing.

When the footing is positioned on the seafloor the base structure may rest on the seafloor with the pile disposed in a vertical or substantially vertical orientation through the bore. The shape, dimensions and structural requirements of the pile may vary based on the design considerations of each installation. Partly because the pile is not required to support all or a significant proportion of the vertical loads of a supported structure the overall size of the pile may be significantly reduced. Also, the structural requirements of the pile may be reduced allowing for a lighter and more flexible foundation. Reductions in the required length of the pile, all other factors being the same, of at least 10, preferably of at least 20, and more preferably of at least 40 percent may be achieved. Reducing the length of the pile may be advantageous as it may reduce the overall material, construction, transportation costs relating to the pile. Also, installation by hammering (pile driving) may be more easily facilitated due to the smaller length, thus substantially reducing the installation cost.

The pile may be an elongated pipe (i.e. a thin-walled section) or a solid rod. The pile may have a substantially cylindrical cross section but other cross sectional shapes may be employed such as oval, tear drop, or polygonal.

The diameter of the cross section of the pile may be substantially constant along the entire length of the pile. Alternatively, the pile may include a taper becoming gradually narrower toward an end, for example a bottom end thereof. The pile may have a more pronounced tapered lower end to facilitate insertion into the seafloor. The pile may have an upper section that is sized to fit within the bore of the footing.

The length of the pile may be sufficient to provide for an adequate length of the pile to be inserted within the seafloor to provide effective anchoring and support for the footing so that the overall foundation may withstand any lateral forces and moments exerted on the foundation. The upper portion of the pile may extend above the seafloor inside the bore of the base structure of the footing and the bore wall structure.

The upper portion of the pile may extend above the top of the bore wall structure to provide lateral engagement with the tower structure.

The upper portion of the pile may comprise a section that may be configured to facilitate connecting the pile with machinery to hammer the pile into the seafloor.

A clearance may be maintained between the pile and the bore wall structure to permit substantially free movement of the footing around the pile in the vertical direction while maintaining an effective lateral engagement between the pile and the footing. Thus, no significant interference may be present in the vertical direction between the footing and the pile. The clearance may be sufficiently small to prevent relative displacement and rotation between the pile and the footing, as well as foreign materials such as sand, or gravel from entering the clearance. Such foreign matter may, over time, damage the foundation.

According to one embodiment the clearance has to be very tight and may range from about 0.5 to about 3 millimetres, for example from about 1 to 3mm. According to one embodiment of the present invention, the pile may be sized to allow the footing to slide along the pile in the vertical direction while obtaining maximum engagement in the lateral direction.

The clearance may have a special low-friction coating to allow relative displacement in the vertical sense, minimizing relative horizontal displacement and rotation. Lubrication may be provided in the clearance area to ensure the proper functioning of the foundation and also prevent foreign matter from entering into the clearance. A gasket, made of rubber or any other suitable material, may be installed at the top of the clearance to protect the interface.

The footing and the pile may be made of any suitable material that may be used in offshore foundations including steel, steel alloys, stainless steel alloys, aluminium, aluminium alloys, steel or fibre reinforced concrete, or any synthetic based material such as a carbon based synthetic material. Preferred materials may include high strength steel typically used with for offshore structures. The foundation may be used to support many different types of offshore structures including wind power installations, drilling rigs, natural gas platforms and oil platforms.

Another aspect of the present invention relates to an offshore structure comprising:

a foundation according to any of the aspects of the present invention; and an offshore structure mounted on the foundation.

The offshore structure may include a tower mounted on top of the footing of the foundation. The tower may be placed on a mounting section of the footing especially designed to receive the wind tower. The tower may be securely fixed on the mounting section of the footing using any type of conventional connectors. The connection may be made below or above the sea level.

The tower may be an integral part of the bore wall structure. A nacelle including a turbine and a rotor may be securely mounted on top of the tower at a desired height.

Yet another aspect of the present invention is directed to a method for installing the foundation according to any other aspect of the present invention in an offshore location such as a seafloor or the floor of a lake. The method may comprise positioning the foundation on the seafloor with the pile positioned through the footing, and driving the pile into the seafloor in a substantially vertical orientation.

The footing and the pile may be lowered to the desired seafloor location with the help of a derrick from a vessel. Divers or ROVs may assist in the exact positioning of the footing and the pile on the seafloor. The pile may be driven into the seafloor before or after the footing is positioned on the seafloor.

The footing may be assembled onshore or on the vessel before being lowered to the seafloor. The footing may be assembled around the pile with the help of divers and/or ROVs.

Ballast material may be added inside the footing before and/or after the footing is securely positioned on the seafloor.

After the foundation is installed in place a structure to be supported, such as a tower of a wind power installation, may be positioned on top of the footing. The remaining components of an offshore structure such as a nacelle of a wind power structure may then be fixed on the tower.

According to an embodiment of the present invention there is provided an installation method, the method comprising:

levelling of the seabed; installation of a footing on the seabed; backfilling of the footing with ballast material; and

driving a pile into the seabed using the footing as a guide.

The footing and the pile may have any of the features as discussed in relation to any other aspect of the present invention.

The pile may be driven using a transition section which is mounted as an extension to the pile. The transition section may have sufficient length so that a top end of the transition section may extend above the sea level at all times during driving of the pile into the sea bed to thus avoid underwater pile driving. The transition section may be or comprise one or more sections. For example, the transition section may comprise a first section attached to the pile at a lower end and having sufficient length to extend above the sea level to allow initiation of the driving of the pile above sea level. As the pile is driven into the sea bed one or more extensions may be added to the transition section to always keep an upper end of the transition section which engages a driving device above sea level.

Couplers may be used to securely connect the pile with a lower end of the transition section. The lower end of the transition section may be designed so that that when the pile reaches a predetermined depth inside the sea bed the lower end of the transition section will reach the footing and rest on the footing. For example the lower end of the transition section may rest on the bore wall structure of the footing signifying the end of the pile driving.

Upon completion of the driving, the couplers may be removed to decouple the pile from the transition section. Thus, the transition section will rest on top of the bore wall of the footing. The installation may be completed by mounting a tower on top of the transition section and any other equipment such as a nacelle. In this manner, substantially all vertical loads from the transition section and the tower are fully transferred to the footing.

The footing may be securely connected to the transition section via many well- known methods. For example various fasteners may be used which could be secured with the help of divers or ROV's. Alternatively, an auto-clamping or auto-locking system may be implemented which allows secure attachment of the lower end of the transition section to a mating section of the footing, preferably with limited or without any underwater work.

The transition section may comprise an elongated member having an upper end section and a lower end section. The lower end section of the transition section may be adapted to be readily secured to an upper end to the pile. The upper end section of the transition section may be adapted to operatively connect with a suitable driving device. The transition section may also comprise a footing connection member. The footing connection member may be adapted to connect with a corresponding member of the footing such as the bore wall of the footing once the pile driving inside the sea bed is completed. The footing connection member may be an integral part of the lower section of the transition section. The footing connection member may be an insert or add on member that may attach to the lower end section of the transition section.

The upper end section of the transition section may be an integral part of the transition section.

The upper end section of the transition section may be an insert or add-on that could be readily removed to add an extension to the transition section to always keep the upper end of the transition section above sea level as the pile is driven inside the sea bed.

The transition section may be adapted to receive an extension. For example, the upper end section of the transition section may be an insert or add on to allow temporary removal of the upper end section, addition of an extension section and remounting of the upper end section to resume driving of the pile. In another contemplated embodiment the upper end transition section is adapted to receive an extension wherein said extension has a suitable upper end section to allow connection to the driving device. As many extensions as may be needed may be added to ensure underwater driving is avoided.

The part of the transition section between said lower and upper end sections may be of any suitable shape or geometry.

The transition section may be removed after installation of the pile.

The transition section may form part of the tower section of the installation.

Driving the pile through the bore wall structure of the footing may be facilitated by allowing a clearance between the two. The clearance between the two parts may be grouted afterwards with suitable materials to enhance transfer of lateral loads and moments without preventing freedom of movement between the footing in the vertical direction.

According to another embodiment of the installation method, a transition section may be connected to the footing prior to installation. In this manner, the clearance between the pile and an upper opening of the transition section may be large enough at the top, to facilitate pile driving initiation, being progressively reduced to achieve a tight fit with the footing at the bottom, allowing transfer of lateral loads and moments without the need for grouting. The tighter fit may also prevent accumulation of foreign matter between the footing and the pile which in turn may allow better accommodation of long- term consolidation settlements because the footing is free to move vertically relative to the pile. Pile driving may be accomplished under water or above water. Initiating pile driving above sea level may be preferred as it may be easier controllable. An adapter extension may be used to extend the pile above sea level as may be needed to avoid underwater pile driving. The adapter extension may be removed after completion. Alternatively, the adapter extension may remain as part of the completed installation thus enhancing lateral interaction between the pile and the footing through the interaction between the adapter extension and the transition section. In all cases, it is important to maintain effective decoupling in the vertical direction between the footing and the pile.

The present invention method is advantageous compared to existing methods and may be employed in shallow, medium depth and deeper bodies of water. The present invention method is simpler and more economical than heretofore methods partly because the foundation may be significantly bulky, and lighter than existing foundations capable to support the same level of vertical and lateral loads. Therefore, the foundation is simpler and more economical to transport and install.

BRIEF DESCRIPTION OF THE DRAWINGS These and other aspects of the present invention will now be described, by way of non-limiting examples only, with reference to the accompanying drawings, in which:

Figure 1A is a diagrammatic illustration of a foundation for an offshore wind power generator, according to one embodiment of the present invention;

Figure 1 B is a diagrammatic top view of the foundation of Figure 1 A;

Figure 2 is a diagrammatic illustration of a portion of an offshore installation showing a portion of the foundation of Figure 1A with a portion of a tower positioned thereon;

Figures 3A and 3B show the finite element models used to computationally compare the performance of a conventional monopile foundation 100 (Figure 3A) with an embodiment of the present invention foundation 10 (Figure 3B);

Figure 4 is an illustration of a finite element model of the embodiment of the present invention foundation 10 shown in Figure 3B;

Figures 5A and 5B is a comparison of the conventional monopile foundation 100 (Figure 5A) with the embodiment of the present invention foundation 10 (Figure 5B) in terms of deformed finite element mesh with plastic strain contours after 60 cycles of cyclic wave-induced loading; and

Figure 6 is a comparison of the moment-rotation response of the conventional monopile foundation (line 102) with an embodiment of the present invention foundation (line 110).

Figures 7a to 7d provide a simplified diagrammatic illustration of an installation method, according to an embodiment of the invention.

Figures 8a to 8d provide a simplified diagrammatic illustration of an installation method, according to an embodiment of the present invention.

Figure 9 shows a diagrammatic illustration of the hybrid system according to an embodiment of the present invention..

Figure 10 shows a finite element illustration of the pile section of the hybrid system according to an embodiment of the present invention.

Figure 11 shows a graphical illustration of an embodiment of the invention against physical model tests

Figure 12 shows a graphical illustration of an embodiment of the invention against vertical loadings

Figure 13 shows a diagrammatic illustration of the loaded pile section of the hybrid system according to an embodiment of the present invention.

Figure 14 shows a graphical illustration of the stiffness degradation according to an embodiment of the invention Figure 15 shows a graphical illustration of the stiffness degradation of the system components according to an embodiment of the invention

Figure 16 shows a graphical illustration of the response of the hybrid system under various loadings according to one embodiment of the invention

Figure 17 shows a graphical illustration of the moment response of the hybrid system under various loadings according to one embodiment of the invention

Figure 18 shows a graphical illustration of the moment response of several hybrid systems under loading according to various embodiments of the invention

Figure 19 shows a graphical illustration of a comparison in stiffness degradation of several systems according to various embodiments of the invention

Figure 20 shows a graphical illustration of the possible service life of the system according to one embodiment of the invention

Figure 21 shows a diagrammatical illustration of a finite element model of the footing of the hybrid system according to one embodiment of the invention

Figure 22 shows a diagrammatic illustration of a finite element model of the stresses and deformation of the footing and pile of the system after loading, according to one embodiment of the invention

DETAILED DESCRIPTION OF THE DRAWINGS

The present invention provides a foundation which may be employed to support many different offshore structures and may not be limited to the support of wind power installations. An exemplary application of the present invention foundation is provided herein in respect to an offshore wind power installation.

Figure 1A illustrates a foundation 10 for an offshore wind power generator (not shown) according to one embodiment of the present invention. The foundation 10 comprises a footing 12 and a pile 14. The footing 12 comprises a base structure 16 having a centrally located bore 18, a bore wall structure 20, an outer wall structure 22 and stiffener elements 24 which extend radially between the bore wall structure 20 and the outer wall structure 22.

The base structure is resting on top of the seafloor 26. The pile 14 is inserted into the seafloor 26 for anchoring the foundation 10 at the seafloor 26. A top portion 14a of the pile 14 extends above the seafloor 26 through the bore 18 and the bore wall structure 20. The footing 12 is free to move around the pile 14 in the vertical direction thus allowing the footing 12 to reposition itself on top of the seafloor 26 to accommodate any long-term settlements of the seafloor 26. The footing 12 and the pile 14 maintain an effective engagement in the lateral direction to ensure adequate transmission of any horizontal or lateral forces and moments from the footing 12 to the pile 14. This is advantageous as it may reduce the overall size and structural requirements for the footing 12, and the length of the pile 14.

Figure 1 B shows a top view of the foundation 10 of Figure 1A. The stiffener elements 24 are spaced apart at regular intervals around the bore wall structure 20 and extend radially from the bore wall structure 20 to the outer wall structure 22. The stiffener elements 24 comprise vertical plates and define together with the bore wall structure 20 and the outer wall structure 22 a plurality of compartments 30 within the top surface of the footing 10.

A clearance 28 is allowed between the pile 14 and the bore wall structure 20. The clearance 28 to be effective should permit substantially free movement of the footing 10 around the pile 14 in the vertical direction while maintaining sufficient engagement between the pile and the footing in the lateral direction.

Figure 2 is side view of a portion of the foundation 10 shown in Figures 1A and 1 B. Figure 2 shows a portion of a wind power generator tower 32 positioned on top of the bore wall structure 20. Thus the wind power generator tower 32 rests on the footing and not on the pile. This is advantageous as it may significantly reduce the size and structural requirements for the pile. Besides the material savings, and transportation costs associated with a reduced size pile, significant advantages arise with respect to the installation of the pile. Smaller sized piles are substantially easier to hammer into the seafloor. Thus, the foundation may be employed in applications that were heretofore limited by the installation limitations of very large piles.

The single pile 14 extends within the bore wall structure 20 all the way up to the top of the bore wall structure 20 to provide sufficient contact area for an effective engagement in the lateral direction with the footing 12.

Ballast material 34 such as rocks and/or gravel is placed within compartments

30 formed between the stiffener elements 24 the bore wall structure 20 and the outer wall structure 22. The ballast material 34 provides adequate weight to the footing 12 forcing it to establish good contact with the seafloor 26. The ballast material 34 renders the footing structure heavier, more solid, and resistant to horizontal forces and moments. Besides rocks and gravel, other materials can be used as ballast material including cement bricks, metal bars, and sand.

Referring now to Figures 7a to 7d an installation method will be described according to one embodiment of the invention. Figures 7a to 7d have many features in common with Figures 1A to 2 thus for ease of reference we refer to similar features with the same reference numerals augmented by 100.

The method comprises first levelling of the seabed 126 as shown in Figure 7A. Then the footing 112 is installed on the levelled sea bed 126 and backfilled with ballast material 134 as shown in Figure 7b. Then a pile 114 is driven into the seabed 126 using the footing 112 as a guide through the pile's centrally located bore 118. The pile 114 is driven using a transition section 150 which is coupled to the pile 114 as an extension. The transition section 150 comprises a cylindrical pipe 152 having a lower end section 153 coupled to a top end of the pile 114 via couplers 154. Any suitable couplers may be used.

The pipe 152 also has an upper end section 156 that is adapted to mate with suitable driving means for driving the pile 114 into the seabed 126 by generating a substantially vertical driving force as indicated by arrow D. The upper end section 156 may be of any suitable configuration dependent upon the driving device used. The cylindrical pipe 152 may be of sufficient length to ensure underwater driving is avoided as shown in Figures 7c and 7d. Driving is continued until lower end section 153 of the cylindrical pipe 152 reaches the bore wall structure 120 of the footing 112. When this happens the driving is completed and then the couplers 154 are removed to decouple the transition section 150 from the pile 114. As discussed earlier, such decoupling ensures that the vertical loads are primarily transmitted to the footing.

The bore wall 120 of the footing 112 is then securely attached to the lower end section 153 of the transition section 150 via one or more well-known techniques including use of fasteners 158 that are secured either by divers or remotely controlled vehicles. Alternatively, an auto-locking system may be used to reduce or completely eliminate underwater work.

Driving of the pile 114 through the bore wall structure 120 requires a clearance between the pile and the bore wall structure. Depending upon the width of the clearance some grouting may be used to improve transfer of the lateral loads between the pile and the footing, however care should be exercised such that the grouting does not interfere with the decoupling between the footing and the pile in the vertical direction.

Referring now to Figures 8a to 8d, another embodiment of the installation method is illustrated. Many features of Figures 8a to 8d are similar to the features of Figures 7a to 7d thus for ease of reference similar features are represented by the same numerals augmented by 100. According to this embodiment a transition section 250 is connected to the footing 214 prior to installation of the pile 214 the transition section 250 may be preinstalled on the footing 212 before the footing is installed on the seabed 226. Alternatively, the transition section 250 may be installed and secured to the footing 212 after the footing is installed on the seabed 226. The footing 212 with the transition section 250 is installed to the levelled seabed 226 as shown in Figure 8a. Then the footing is backfilled with ballast 234 as shown in Figure 8b. The transition section 250 comprises a cylindrical pipe 252 having a lower end section securely mounted to the bore wall structure 220 of the footing 212. An upper end section of the transition section 256 is configured to mate with suitable driving means (not shown). In this embodiment, the transition section 25 is a hollow pipe having a trough bore to allow the introduction of the pile through the transition section 250. The upper end section 256 may have a larger diameter than the lower end section creating a slight tapering from the first end to the second end. Thus, the clearance between the pile 214 and the bore of the transition section 250 may be larger at the upper end section 256 in order to facilitate the initial driving of the pile, and being progressively reduced to achieve a tight fit with the footing at the bottom. In this manner, a tight fit between the bore wall structure of the footing and the pile may be accomplished thus improving the transmission of lateral loads and moments between the two.

An adapter extension 260 as shown in Figures 8c and 8d may also be employed to further increase the length of the transition section as may be needed in order to eliminate the need for underwater driving.

Comparative Example

A foundation 10 according to one embodiment of the present invention as shown in Figure 1 is compared with a conventional monopile foundation employing finite element analysis. In accordance of this comparative example, the foundation 10 of Figure 1 is designed for supporting a 3.5 MW wind turbine. The wind turbine and related structure, i.e. the wind tower, and the nacelle have a total mass of 420 tonnes. The rotor of the wind turbine is positioned at a height of 80 m from the foundation level. A typical clayey soil of undrained shear strength = 60 kPa is considered for simulating the seafloor 26.

A conventional solution consisting of a monopile foundation 100 would require a pile having a 5 metre diameter and 30 metres length based on the analysis. The hybrid foundation 10 embodiment as shown in Figures 1A and 1 B requires a pile 14 having the same 5m diameter as the conventional solution but with a reduced length of only 15 metres. Therefore a reduction of 50% in the overall length of the pile is achieved. The footing 12 used in the analysis has a diameter of 14 meters and it has 8 stiffener elements extending radially from the bore wall structure 20 to the outer wall structure 22. The stiffener elements 24, the bore wall structure 20 and the outer wall structure 22 all are vertical plates having a constant width cross section of 0.02 meters. The clearance between the pile 14 and the bore wall structure 20 is 1 mm. Ballast material 34 comprising gravel is used having a total weight of 240 tonnes (taking account of the buoyancy force). All the components of the footing 10, i.e. the base structure 16, the bore wall structure 20, the outer wall structure 22, and the stiffeners are made of high strength steel, typically used for offshore structures.

The analysis comprised three steps: (a) application of vertical loads; (b) monotonic wind loading P = 1000 kN at 80 m height; and (c) cyclic wave loading application of forced controlled cyclic loading to simulate the cyclic wave loading P = + 2000 kN at T = 10 sec, 60 cycles. The imposed wind loading of 1000 kN has been computed according to widely accepted design standards, such as the American Petroleum Institute Recommended Practice for Fixed Offshore Structures (API RP2A, 1993). The +2000 kN wave loading has been computed according to the simplified procedure described in AASHTO 1992, assuming a water depth of 15 m.

The finite element models of a conventional monopile foundation 100 and the present invention hybrid foundation 10 of Figure 1 are shown in Figures 3A and 3B respectively. The soil 26 is modelled with hexahedral brick type elements. Details of the model of the hybrid foundation model are shown in Figure 4. The stiffener elements 24, the bore wall structure 20 and the outer wall structure 22 are modelled with shell elements, while the ballast material 34 is modelled with hexahedral brick type elements. The pile 14 is a thin-walled steel of 5 cm thickness. The pile is modelled with beam elements, circumscribed by eight-noded hexahedral continuum elements of nearly zero stiffness. The nodes of the beam elements representing the pile are rigidly connected with the circumferential solid element nodes at the same height. Thus, each pile cross-section behaves as a rigid disc. Such a modelling technique allows direct computation of pile internal forces (through the beam elements), and realistic simulation of the 3D geometry of the soil-pile interface.

The results of the analysis are shown in Figures 5A, 5B and 6. More specifically, Figures 5A and 5B compare the results of the conventional monopile foundation 100 (Figure 5A) and the hybrid foundation 10 (Figure 5B) in terms of deformed finite element mesh with plastic strain contours, after 60 cycles of wave- induced loading.

Figure 6 compares the moment-rotation response measured in MNm of the two foundations. Line 102 shows the moment-rotation response of the conventional monopile 100 having a diameter of 5 metres and length of 30 metres, while line 110 shows that of the hybrid foundation having a pile with a diameter of 5 metres, length of 15 metres and a footing of 14 metres in diameter. Wind loading (step 2 of the analysis) develops a moment of 80 MNm. The subsequent wave loading (step 3 of the analysis) leads to an alternating moment of + 16 MNm. As shown in Fig. 6, when subjected to wind loading the hybrid foundation 10 is characterized by a stiffer and more elastic response, compared to the conventional monopile foundation 100, and its rotation is roughly 20% lower. Thanks to this more elastic response, when subjected to subsequent cyclic wave loading, the hybrid foundation 10 accumulates smaller permanent rotation. Although in reality the wind turbine is subjected to millions of wave loading cycles, the finite element analysis results are indicative of the superior performance of the invention 10 compared to the conventional monopile 100. _.

Importantly, this performance is achieved with a 50% shorter pile. Depending on the hourly rental rates of the vessels used to carry and operate the hammering machinery for the installation of the pile, the reduction of the required pile driving by 50% may lead to significant cost savings. Generally, the installation usually constitutes more than 60% of the foundation cost. Hence, reducing the installation cost by 50% may lead to cost savings of the order of 30%.

It should be understood that the embodiments described herein are merely exemplary and that various modifications may be made thereto without departing from the scope of the invention as defined in the following claims.

Further Examples

A numerical study is conducted to investigate the efficiency of the hybrid foundation. Three typical OWT assemblies are considered. Their key geometric parameters are summarized in the following Table.

Table 1. Key geometric parameters of the examined wind turbines.

Type d _R (m) ho (m) m _NR (Mg) m _T (Mg) d _T (m) tT (mm)

2 MW 65 60 200 128 4.0 20

3.5 MW 90 80 220 195 4.0 23

5 MW 110 90 350 271 4.9 23

Referring to Figure 9 a hybrid foundation 400 is composed of a circular footing 412 of diameter D and a pile 414 of diameter d, length L, and wall thickness tp. A 40 m deep clayey stratum 426 of undrained shear strength S _u is considered. Two characteristic cases are examined: (a) a homogeneous profile of S _u = 60 kPa and v- 0.49, and (b) a non- homogeneous profile with S _u - 30 kPa at seabed, linearly increasing with a gradient k =3 kPa/m. To reduce problem variables, pile diameter and thickness are kept constant: d = 5 m and t _p = 80 mm (considered representative for the cases examined). The problem is analysed by employing 3D finite element (FE) modelling. As illustrated in Fig. 10a and 10b, taking advantage of problem symmetry, half of the foundation is modelled. The hybrid foundation 400 (Fig. 10a) is comparatively assessed against an L = 30 m monopile, which used as a benchmark (Fig. 10b). The footing 412 and the soil 426 are modelled with hexahedral continuum elements, assuming elastic and inelastic response, respectively. The footing 412 is idealized as a rigid cylindrical disk of 2 m height, and unit weight corresponding to the surcharge load q. Unless otherwise stated, q = 30 kPa is considered, corresponding to the submerged weight of the lightweight steel footing and roughly 2.5 m of rubble.

The piles are modelled with elastic beam elements, circumscribed by "dummy" (i.e., of nearly zero stiffness) continuum elements. The nodes of the beam elements representing the pile are rigidly connected to the circumferential solid element nodes at the same height. This way, each pile cross-section behaves as a rigid disk. Such a modelling technique allows direct computation of pile internal forces (through the beam elements), and realistic simulation of the 3D geometry of the soil-pile interface. The tower is also modelled with elastic beam elements, and a concentrated mass is used for the rotor- nacelle assembly.

Special contact elements are employed to model the pile-soil and footing-soil interfaces, allowing detachment and sliding. The interfaces are tensionless, allowing detachment. With respect to shear resistance, two alternatives are considered: (a) Coulomb-type friction, assuming a coefficient of friction μ = 0.5, and (b) interface shear resistance equal to a fraction of the undrained shear strength aSu, assuming a = 0.4. Second order effects are also accounted for in the analyses including the superstructure. The connection footing-pile connection is simulated with very stiff horizontal and rotational springs to model the transfer of lateral and moment loads, but allowing relative movement in the vertical sense.

Soil modelling

Nonlinear soil behaviour is modelled with a kinematic hardening model, incorporating a Von Mises failure criterion and associative flow rule. The model is considered appropriate for clay under undrained conditions, and has been extended for sand with a simple Abaqus [2010] user subroutine that accounts for normal pressure dependence. See Anastasopoulos et al. "Simplified Constitutive model for simulation of cyclic response of shallow foundations: validation against laboratory tests" Journal of Geotechnical and Geoenvironmental Engineering, ASCE, 137(12):1154-1168, 2011]. The model is briefly described, focusing on clay. The evolution of stresses is defined as:

σ = σ ₀ + a (1) Where σ ₀ corresponds to the stress at zero plastic strain, and a is the "backstress", determining the kinematic evolution of the yield surface in the stress space. The evolution of the kinematic component of the yield stress is defined as: a = C— (σ— ά)^ ¹ — γαεΡ ¹ (2)

where C is the initial kinematic hardening modulus [C = a _y / _y = E = 2(1 + v)G ₀ ] and y is a parameter determining the rate of decrease of the kinematic hardening with increasing plastic deformation. In the case of clay, the maximum yield stress can be defined as: a _y = ¾ _u (3)

And since a _y = C/ γ+σ ₀ , parameter / can be expressed as:

Parameter a ₀ controls the initiation of nonlinear behaviour and is defined as a fraction y of the yield stress o _y . σ ₀ = γσ _ν (5)

Finally, C corresponds to the Young's modulus for very small strains.

The model has been validated against physical model tests (centrifuge, 1 g, and large-scale) for: (a) surface and embedded foundations subjected to cyclic loading and seismic shaking, (b) piles subjected to cyclic loading; (c) bar-mat retaining walls subjected to seismic shaking; and (d) circular tunnels subjected to seismic shaking. Despite its simplicity, the model was shown to perform very well, in many cases outperforming more sophisticated models. Moment Capacity Due to their large height, OWTs are typically subjected to large lateral load (H) and overturning moment (M) due to wind and wave loading. Thanks to their light construction, the vertical loads (V) of the superstructure are relatively low. From the foundation point-of-view, this leads to a challenging combination of very large overturning moments with relatively low vertical loading: the M/V ratio is of the order of 25 for a typical 3.5 MW OWT. Moreover, especially for the case of wind loading, the moment to shear ratio is quite large: M/H = 80m for a typical 3.5 MW OWT (Table 1). Therefore, a logical first step is to explore the factors affecting the moment capacity of the hybrid foundation.

For this purpose the FE model is subjected to monotonic displacement controlled pushover loading (Fig. 11a to 11d). To focus on the foundation, the superstructure is ignored and pure rotation is applied at the top of the footing. The presented results refer to the homogeneous soil profile. First, a hybrid foundation with an L = 15m (d = 5m) monopile is considered, varying the footing diameter D from 11 m to 20m (Fig. 11a). The results are compared to a monopile of the same length. Then, the footing diameter is kept constant to D = 15m, varying L from 10m to 25m (Fig. 11 b). The results are compared to a footing of the same D.

As illustrated in Fig. 11a, the addition of the footing leads to a pronounced increase of the moment capacity M _ult . For example, adding a D = 15m footing to the L = 15m monopile leads to 60% increase of _u/f : 210 MNm as opposed to 130 MNm. It is worth noting that the increase of M _ulf is larger than the superposition of the two components. As shown in Fig. 12b, the moment capacity of the D = 15m footing alone is roughly 50 MNm, which means that the superposition yields M _ult = 180 MNm. This reveals the dual beneficial role of the footing, which not only contributes with its moment capacity (as an individual foundation element), but also enhances the capacity of the monopile acting as a rotational restraint.

Although the addition of the footing is beneficial, the increase of L has a more pronounced effect (Fig. 11 b). The contribution of the two foundation components is better depicted in Figs. 11c and 11d, where M _ult is plotted as a function of D and L, respectively. The mean rate of capacity increase per metre of pile length is approximately 3 times larger than the rate per metre of footing diameter. In terms of materials, the increase of D is more costly compared to L, but the latter is also associated with increased pile driving cost. The effect of vertical loading V

The vertical load V acting on a footing is well known to be an important factor. This is also the case for the hybrid foundation examined herein. Through the proposed footing- pile connection, the surcharge load q and the vertical loads of the superstructure are transferred entirely to the footing (at least with the first two installation options). This way, the vertical loads are fully exploited to enhance the performance of the footing. It is therefore of interest to explore the effect of V on M _uU . Therefore, applying displacement probes of constant rotation/settlement (θ/w), M-V failure envelopes are computed for: (a) hybrid foundations of L = 15 m, varying the footing diameter D (Fig 12a); and (b) hybrid foundations of D = 15m, varying the pile length L (Fig. 12b). As expected, the increase of foundations of D = 15m, varying the pile length L (Fig. 12b). As expected, the increase of V is beneficial up to a certain limit, beyond which its further increase leads to reduction of M _M . This resembles the behaviour of shallow footings, where the vertical loads are beneficial up to 50% of the bearing capacity under purely vertical loading V _ult . This is due to the footing-pile connection which allows relative movement in the vertical sense.

If the two foundation components were rigidly connected, the vertical loads would be undertaken almost purely by the pile due to its much larger vertical stiffness. An in the absence of vertical loading, the footing cannot offer any moment (or shear) resistance. This is clearly seen in Fig. 12a, where all M-V envelopes cross the vertical axis {V = 0) at exactly the same point, which corresponds to the moment capacity (M _ult = 130 MNm) of the L = 15m pile alone. In the same graph, the vertical dotted lines indicate the vertical loading for a 3.5 MW OWT, including q = 30kPa. Although q is not constant, the corresponding vertical load V is different due to the increase in D. While for the smaller D = 11 m footing there is no point in increasing V, for larger diameters such an increase is beneficial. However, the vertical load required to fully exploit the moment capacity of the hybrid footing is excessively large, and would not be economical. For example, for D = 20m a vertical load V∞ 48 MN would be required to maximise M _u i _t , and given the superstructure dead load of 4.2 MN, q∞ 140 kPa would be required. This value is lower than that of the tests of Stone et al. [2010], as the foundation is on clay instead of sand. Figure 12b plots the M-V envelopes for D = 15m hybrid foundations, varying L. For V - 0, the footing is inactive and M _ult is a function of L only. While M _ult is significantly affected by L, this is not the case of the vertical bearing capacity V _ult . This is due to the special footing-pile connection, which transfers the vertical loads only to the footing. Interestingly, while V _u u is exactly the same for L = 20 and 25m and very close to the theoretical value of (π + 3) Su A, it is slightly lower for the shorter L = 15m pile. The increase of L has an obvious positive effect of M _ult , in accord with the previous results.

Further insights are derived and exploring the failure mechanisms in function of the normalized vertical load V/V _u it. Few such examples are presented in Fig. 13 for a D = 15m and L - 20m hybrid foundation (plotting the deformed mesh with superimposed plastic strain contours). As shown in Fig. 13a, in the absence of vertical loading (VA/ _Ulf - 0), the footing is inactive and the imposed rotation leads to detachment from the soil. Consequently, the failure mechanism is that of the monopile (point A in Fig. 12b). Being short (Ud = 4), the pile rotates more-or-less as a rigid body generating passive stresses at its head and toe. As expected, the increase of VA _Ult leads to activation of the footing, which plays an increasingly important role. For VA/ _Ult = 0.25 (Fig. 13b), the footing is still lightly loaded and its response is uplifting dominated. The combined failure mechanism mobilizes a larger soil mass, leading to the previously discussed increased _u/i (point B in Fig. 12b). Substantial soil yielding can be observed at the right edge of the footing. The zone of plastic deformation generated by the footing meets the corresponding plastic zone generated by the pile at mid-length.

Further increase of the vertical load to VA _Uit = 0.50 (Fig. 13c) leads to further expansion of the failure mechanism, and to a corresponding increase of _u , (point C in Fig. 12b). The plastic zone that originates from the right edge of the footing extends all the way to the pile tip. The response of the footing entails very limited uplifting and full mobilization of its moment capacity. Although the response of the hybrid footing is optimized in terms of M _ulu the required vertical load is unrealistically large and such a solution would not be economical. Finally, for V/V _utt = 0.75 (Fig. 13d), the response of the footing becomes sinking - dominated. Being heavily loaded, the footing generates substantial plastic deformation at its right edge, leading to a reduction of M _ult (point D in Fig. 13b).

Rocking Stiffness

As previously discussed, the accumulation of deformation under cyclic loading governs the design of OWT foundations. Due to their large height, foundation rotation 0 is critical. The latter is a function of the rocking stiffness KR and the rate of deformation accumulation during cyclic loading. While cyclic loading is studied later on, this section focuses on KR and its comparative assessment to the reference case of an n L = 30m monopile. The FE model is subjected to displacement-controlled loading, applying rotation at the top of the footing. The horizontal displacement is restrained in order to compute the pure rocking stiffness.

The secant rocking stiffness KR is plotted as a function of the imposed 0, leading to the stiffness degradation charts of Fig. 14. The latter are computed for: (a) hybrid foundations of L = 15m {q = 30 kPa), varying D (Fig. 14a); and (b) hybrid foundations of D = 15m (q = 30 kPa), varying L (Fig. 14b). In both cases, the results are compared to the L = 30m reference monopile. All curves are characterised by an initial linear segment (0 < 10 ^"4 ), which corresponds to the small strain rocking stiffness KR, 0, followed by the stiffness degradation for larger values of 0. In terms of KR,0, all hybrid foundations outperform the reference monopile.

As shown in Fig. 14a, the addition of a D = 11 m footing to an L = 15m pile leads to a stiffer initial response compared to the reference monopile. KR, 0 is controlled by the diameter D of the footing. For example, the increase of D from 11 m to 15m leads to an increase of KR,0 by about 20%: 196 GNm as opposed to 163 GNm (Fig. 14a). In contrast, KR,0 by about 20%: 196 GNm as opposed to 163 GNm (Fig. 14a). In contrast, KR,0 is practically insensitive to L (Fig. 14b). The latter only affects the rocking stiffness for larger values of rotation. The increase of L starts playing a role for 0 > 2 x 10 ^"3 , which is however outside the typical rotation range of OWTs. Compared to the footing, the pile requires larger deformations to mobilise its lateral capacity. Therefore, although the increase of L is beneficial in terms of M _M (and KR for large 0), it does not offer any appreciable advantage with respect to KR,0. It may be concluded that the two foundation components are complementary to each other: while the footing is efficient in terms of KR,0. It may be concluded that the two foundation components are complementary to each other: while the footing is efficient in terms of KR,0, the pile is dominant with respect to M _ult

To derive further insights on the contribution of the two component, the D = 15m, L = 15m, q = 30kPa hybrid foundation is analysed in more detail. Figure 15 compares the rocking stiffness degradation curve of the hybrid system to its components and their superposition. As for M _ul the hybrid foundation attains larger KR,0 compared to the superposition. This is attributable to the previously discussed beneficial restraint offered by the footing to the pile. Nevertheless, the difference is not substantial and can be neglected for design purposes: KR,0 can be conservatively estimated through superposition of the stiffnesses of the two components.

Monotonic pushover loading

This section explores the performance of the hybrid foundation under more realistic loading conditions. Focusing on a 3.5 MW OWT (which is currently the most commonly used system) founded at 15m water depth (which is very common for current offshore sites), the FE model is subjected to displacement-controlled monotonic pushover loading at two different heights: (a) hO = 80m, corresponding to wind loading; and (b) hW = 8m, corresponding to wave loading. In contrast to M-V loading, the pushover analysis (M-H-V loading) accounts for the effect of the M/H ratio and the associated coupling between the translational and rotational degrees of freedom. Failure envelopes in the M-H plane are also computed to fully describe foundation performance. A constant vertical load V (equal to the dead load of the superstructure) is applied first, followed by lateral pushover varying the height of application of the lateral displacement, and hence the M/H ratio.

The results of the analyses are summarised in Fig. 16, focusing on hybrid foundations of D = 15m and q = 30 kPa. The M-H failure envelopes are plotted in Fig. 16a as a function of pile embedment length L. The increase of L leads to rotation and expansion of the failure envelopes. Their shape is typical of embedded foundations, resembling more of an inclined "bay leaf" rather than an ellipse, and revealing once more the dominant role of the pile. This shape is representative of the beneficial role of the coupling between M and H acting to the opposite direction.

Figure 16b focuses on the L = 15m hybrid foundation, highlighting the effect of M/H. The increase of M/H leads to a steeper load path, and hence to an increase of the moment capacity and to a decrease of the lateral capacity (H _ult ) of the foundation. As a result, in the case of wind loading {M/H = 80m) M _M reaches 165 MNm, combined with Hun = 2.1 MN. The latter is much larger for wave loading (H _un = 11.4MN), but is combined with a substantially lower _u/f = 63 MNm. Figure 16c compares the moment- rotation (M-O) response for the two loading conditions. Besides M _uH , the M/H ratio plays a major role in terms of rocking stiffness. Focusing on the initial part of the M-0 diagram (Fig. 17d), it becomes clear that the increase of M/H leads to an increase of KR,0.

The monotonic M-0 response of the hybrid foundation is compared to the reference L = 30m monopile in Fig. 17. The latter is considered as a reasonable solution for a 3.5 MW OWT at 15m depth (for the specific soil conditions). The comparison is performed for a D = 15m, L = 15m, q = 30kPa hybrid foundation, which can be cost efficient as the embedded pile length is reduced by 50% compared to the reference monopile. The results are presented in Fig. 17a for wind loading. The Z.=30m monopile grossly outperforms the hybrid foundation in terms of M _uf (Fig. 17a): 420 MNm as opposed to 165 MNm. The same conclusion is drawn for wave loading (Fig. 17b). However, focusing on the operational loading range, the situation is completely reversed (see zoomed diagrams on the right). For a wind design load of 1 MN [according to API, 2000], the design moment due to wind loading does not exceed 80 MNm (HO = 80m). At this regime, the hybrid foundation exhibits much stiffer response and its inferior moment capacity is irrelevant. Exactly the same conclusion is drawn for wave loading, in which case the design moment is 16 MNm (2MN at HW = 8m). While a large embedment is required for a monopile to offer the required lateral stiffness, the hybrid foundation achieves the same and better results with 50% reduced pile length. Cyclic Loading

OWTs are typically subjected to:

(a) Relatively slow cyclic loading due to environmental loads, included wind and waves; and

(b) Dynamic loading arising from operational loads, including rotor loading (at a frequency commonly referred to as 1P) and blade passing (at a frequency of 3P for 3-bladed turbines).

To avoid resonance that may lead to dynamic amplification and non-tolerable tilting, the design must ensure that the natural frequency of the foundation-structure system is not close to the dominant frequencies of the external excitations.

A modal analysis was conducted to explore the dominant modes of oscillation of the examined wind turbines along with the examined foundations. For the 3.5 MW system, the first Eigen-frequency was found to range from 0.24 Hz (hybrid foundation of D = 15m and L = 15m) to 0.26 Hz (L = 30m monopile). A value of this order may fall within the limits of the "soft-stiff design range (i.e. between 1P and 3P). While the acceptance of these values is a function of the specific site conditions and characteristics of the OWT, within the context of this study the key conclusion is the insensitivity of the dominant frequency to the foundation type. Hence, it is reasonable to assume that all foundations examined are equally efficient in terms of dynamic operational loads: resonance is avoided in all cases.

Based on this assumption, the long-term performance can be assessed considering slow-cyclic loading due to wind and waves. The latter have a period of the order of 10 seconds, and can therefore be simulated in a quasi-static manner. The same applies to wind loading, which is acknowledged to be of much larger period. Three cyclic loading scenarios were considered: (i) cyclic loading due to wind only; (ii) cyclic loading due to waves only; and (iii) monotonic loading due to wind followed by cyclic loading due to waves. The third scenario is used for the comparison, being considered the most representative. It is also more conservative, as it generates oneway cyclic loading which has been shown to lead to larger accumulation of plastic deformation and stiffness degradation.

As for monotonic loading, the comparison is made for a 3.5 MW wind turbine founded at 15m water depth. Based on the literature and current codes of practice [API, 2000], a design wind load of 1 MN is assumed, combined with a ±2 MN design wave load. The actual flexural stiffness of the tower is considered to account for P-6 effects. Considering 20 years of service life, an OWT is expected to sustain roughly 10 ⁸ cycles of loading. However, with the current computing power simulating such a large number of cycles is not feasible. Therefore, two cyclic loading scenarios are applied, one with 10 cycles of wave loading and the other with 40 cycles. While the first is used for parametric analyses the latter is used for specific cases to investigate the accuracy of extrapolation. Figure 18a compares the cyclic M-0 response of 3 representative q = 30 kPa hybrid foundations to the reference L = 30m monopile. A key conclusion is that all hybrid foundations considered outperform the monopile. After 10 cycles of loading, the monopile has accumulated 0.0022 rad of rotation, followed by the smallest (D = 15m, L = 15m) hybrid foundation (0.0018 rad). The initial, almost linear, segment of the M-0 curves is due to the wind loading. The comparisons are made considering the loads and moments acting at seabed level (e.g. the moment due to wind loading is equal to 82 MNm). The second segment corresponds to cyclic wave loading, and is mostly responsible for the accumulation of rotation. It is worth observing the decrease in stiffness as soon as wave loading is applied, which is due to the previously discussed effect of the M/H ratio (Fig. 16). The D = 20m and L = 15m hybrid foundation exhibits the stiffest response followed by the D = 15m, L = 20m system. Hence, in terms of cyclic performance D is more important than L.

Figure 18b illustrates the effect of q in the cyclic Μ-Θ response of a D = 20m, L = 15m hybrid foundation. Since the performance is totally acceptable for q = 30kPa, there is no point in exploring larger values. As expected, the reduction of q leads to an increase of the accumulated Θ. Still though, even with q = 10kPa the D = 20m, L = 15m hybrid foundation outperforms the smaller D = 15m, L = 15m system and the L = 30m monopile. It is also worth observing that the initial stiffness is not affected by q. As the rotation increases, a larger q is required to compensate for the reduced stresses at one side of the footing. But this does not affect the initial stiffness. At this initial stage, even a relatively small q is enough for the footing to maintain full contact with the soil, and therefore the initial stiffness is not affected.

Figure 19 compares the performance of the hybrid foundations to the reference monopile in terms of stiffness degradation. The normalised secant rocking stiffness KR/KR,0 is plotted as a function of load cycles N. The hybrid foundations experience less intense stiffness degradation compared to the reference monopile. The D - 20m, L = 15m hybrid foundation exhibits the best performance, followed by the D - 15m, L - 20m system. The smaller D = 15m, L = 15m hybrid foundation slightly outperforms the monopile. In contrast to the larger ones, KR/KR,0 is not yet stabilised at the end of cycled loading, following the same trend with the monopile. Hence, it is concluded that the smaller hybrid foundation offers the same (and slightly better) level of performance with the reference monopile.

As already mentioned, OWT foundations are expected to sustain a very large number of load cycles: N ^¾ 10 ⁸ for 20 years of service life. To explore in more detail the rate of rotation accumulation was a function N, the analysis is repeated for selected cases applying 40 cycles of wave loading. Figure 20a presents the results of one such analysis, referring to the D = 15m, L = 15m hybrid foundation. The maximum accumulated rotations 9 _max is plotted as a function of N. The accumulation rate tends to diminish after the 30 ^tfl cycle, and 9 _max more or less stabilises to 0.0019 rad.

An attempt is made to predict 9 _max for the entire service life. A simple logarithmic extrapolation is performed for this purpose, the results of which are shown in Fig. 20b. If the extrapolation is made on the basis of 10 cycles only, the final 9 _max (for N=10 ⁸ ) is underestimated by roughly 8%. The key conclusion is that the hybrid foundation outperforms the reference monopile: 9 _max reaches 0.0027 rad as opposed to 0.0042 rad for the monopile. Still though, both systems are within acceptable rotation limits (9 _max < 0.01 rad). Although such a simple extrapolation is not claimed to be accurate, its comparative value is essential. The role of the fill material in buckling prevention

Based on the results of the previous section, it is concluded that a D = 15m, L = 15m, and q = 30kPa hybrid foundation offers the same (and even slightly better) level of performance with the L = 30m reference monopile. The decrease of L by 50% has obvious economic implications, both in terms of materials and pile driving cost. On the other hand, the steel needed to construct the footing is not negligible and should be taken into consideration. Besides the installation, the cost savings largely depend on the design of the footing. As with most thin-walled steel structures, buckling is expected to be a critical issue. Although such detailed design does not fall within the scope of the paper, a brief study is conducted to derive preliminary insights.

A more detailed FE model is developed for this purpose, modelling the footing with shell elements (Fig. 21). The steel is modelled with an elasto-plastic constitutive model with Von-Mises failure criterion. A steel grade S355 is considered, having Young's modulus E = 210 GPa and yield strength fy - 335 MPa. As previously discussed, the compartments formed between the stiffeners and the peripheral and bottom plates are filled with rubble to increase the surcharge load q acting of the footing. Besides acting as a ballast, the fill material provides confinement and lateral restraint to the stiffeners, and can therefore be beneficial for buckling prevention. To explore such a potential, the fill material is modelled:

(a) by applying a surcharge load q at the bottom of the footing; and

(b) more rigorously with continuum elements.

While the dead load of the footing is exactly the same, the confinement offered by the fill material is accounted for only in the second case. A parametric analysis was conducted, varying the thickness t of plates and stiffeners. To highlight the role of confinement offered by the fill, an extreme case is presented, referring to t = 15mm (Fig. 22). The detailed FE model is subjected to monotonic pushover loading, as previously described. As illustrated in Fig. 23a, when the confinement offered by the fill is ignored, the t = 15mm radial stiffeners sustain buckling failure. Moreover, as revealed by the contours of Mises stresses, there are several areas where the steel exceeds its yield strength fy = 335 MPa. Obviously, such a performance cannot be acceptable and an increase of t would be required. However, when the lateral restraint offered by the fill is accounted for, the behaviour of the footing is totally acceptable (Fig. 15b). Mises stresses are kept at much lower levels, and buckling is prevented due to the confinement offered by the rubble fill.

The key conclusions are as follows:

1) The addition of the footing to the monopile leads to a pronounced increase of the moment capacity M _M . The increase of M _M is larger than the superposition of the two foundation components, as the footing is not only contributing with its moment capacity, but is also enhancing the moment capacity of the monopile acting as a rotational restraint.

2) The vertical load acting on the footing plays a crucial role. Through the proposed footing-pile connection, the vertical loads of the superstructure are transferred solely to the footing (at least with the first two installation options). This way, the vertical loads are fully exploited to improve its performance.

3) A detailed comparison was performed for a 3.5 MW OWT at 15m depth. A hybrid foundation of D = 15m, L = 15m, and q = 30 kPa was compared to the reference monopile. In terms of M _M , the L =30m monopile outperforms the hybrid foundation. However, focusing on the operational loading range, the hybrid foundation exhibits a much stiffer response, outperforming the monopile while having 50% reduced length.

4) Subjected to cyclic loading, the hybrid foundation experiences less intense stiffness degradation compared to the reference monopile. Based on a simple logarithmic extrapolation to the service life of the OWT (10 ⁸ cycles of leading), the D = 15m, L = 15m hybrid foundation outperforms the reference monopile.

5) The decrease of the length of the pile by 50% has obvious economic implications, both in terms of materials and pile driving cost. Besides the installation procedure, the cost savings largely depend on the design of the footing and buckling is a critical issue. The rubble fill has been shown to provide confinement and lateral restraint to the stiffeners, and is therefore beneficial for buckling prevention. It should be understood that the embodiments described herein are merely exemplary and that various modifications may be made thereto without departing from the scope of the invention as defined in the following claims.