Cost-driven design and optimization of steel support structures for offshore energy platforms: a review of CAPEX cost models

1 Introduction

Steel structures are widely used as the predominant support structure for offshore oil and gas platforms and offshore wind turbines. These structures—ranging from fixed jackets and monopiles to floating semi-submersibles and spar buoys—require massive amounts of steel, often hundreds of tonnes per unit. In addition, these steel structures demand complex manufacturing processes and challenging installation operations, ultimately resulting in substantial project costs. These engineering and economic challenges drive engineers to explore more optimized and cost-effective solutions. To this end, various mathematical cost models have been proposed to estimate the expenditure of steel support structures for offshore energy platforms during their life cycle.

The overall cost of steel offshore structures can be classified into three main categories: capital expenditure (CAPEX), operational expenditure (OPEX), and decommissioning expenditure (DECEX). CAPEX represents all the initial investment required to construct and install the structure (Maienza et al., 2020). OPEX incorporates expenses to maintain the structure throughout its operational lifespan, through regular maintenance and inspections. DECEX accounts for costs associated with end-of-life removal and site restoration. The typical cost breakdown shows CAPEX makes up around 70% of total costs in offshore structures, while OPEX and DECEX roughly account for 20%–30% and 5%, respectively (Maienza et al., 2020). This clearly indicates CAPEX is the most critical cost component and a major focus for design and optimization efforts. As shown in Figure 1, CAPEX in steel offshore structures primarily consists of four main cost components: material procurement, manufacturing, transportation, and installation (Hajinezhadian and Behnam, 2023). In other words, CAPEX in offshore structures is the cost involved in sourcing raw materials (steel), fabricating and assembling the structure, and finally transporting and installing it offshore.

Due to the importance of CAPEX in the project economics, this paper reviews numerous studies that have focused on the optimization of CAPEX in steel support structures for offshore energy platforms. These optimization efforts rely heavily on the underlying cost models used to estimate expenses. However, a wide range of cost modelling approaches exists in the literature, varying in terms of methodology, assumptions, accuracy, and level of detail. This review paper systematically examines existing CAPEX cost models for steel support structures for offshore energy platforms to identify their strengths, limitations, and sources of uncertainty. Unlike individual studies that focus on a specific structure type or cost component, this review is the first to bring together all CAPEX cost modelling approaches used for steel offshore structures, both detailed and simplified. The paper classifies cost modelling approaches systematically and identifies the gaps that only become visible when the literature is examined as a whole. Moreover, the paper provides a quantitative picture of how completely CAPEX is modelled across the field. It offers a new framework that combines cost model completeness, modelling detail, and structural complexity. This framework allows direct, quantitative comparison of CAPEX cost models across all reviewed studies. Together with these findings, the review highlights key gaps to guide future research toward more accurate cost estimation methods.

This paper is structured as follows: Section 2 describes the literature review methodology. Section 3 provides a detailed examination of CAPEX cost models in steel offshore structures, covering material, manufacturing, transportation, and installation costs. Section 4 discusses the findings and suggests directions for future work, and finally, Section 5 concludes the paper.

2 Literature review methodology

This review paper adopts a structured literature survey methodology to systematically examine cost-driven design and optimization studies of steel offshore support structures, with particular emphasis on capital expenditure (CAPEX) cost models. The methodology is designed to capture both the evolution of cost modelling approaches and their application in the optimization of fixed and floating offshore structures, while ensuring transparency and reproducibility of the review process.

2.1 Scope and focus of the review

The scope of this review is restricted to steel offshore support structures, including both fixed foundations (monopiles and jackets) and floating foundations (spar buoys, semi-submersibles, tension-leg platforms, and barge-type floaters). The review explicitly focuses on the support structure level only. This review focuses on CAPEX cost models, covering the four main cost components:

• Material costs
• Manufacturing costs
• Transportation costs
• Installation costs

It is worth noting that most manufacturing cost models used for offshore steel structures have been originally developed for conventional welded steel structures. For this reason, this review also includes some of those foundational models from the general steel structures literature, as they help explain where the offshore-specific approaches come from and how they have developed over time. However, these studies are not included in the statistical analysis and discussion in Section 4.

2.2 Literature search strategy

The literature search is carried out using three main academic databases: Google Scholar, Scopus, and Web of Science. Additional studies are found by following citations from highly cited papers, both backward and forward. Search terms are grouped into three categories, as summarised in Table 1. Search terms within each category were combined using OR. Terms from the Structure type and General terms categories were then paired with those from the Cost components category using AND, forming queries such as: (“jacket” OR “semi-submersible” OR “spar buoy”) AND (“cost model” OR “CAPEX” OR “manufacturing cost”). In additional searches, broader cost terms were replaced with more specific phrases such as “fabrication cost” or “welding cost” to capture studies that did not use standard CAPEX terminology. The first literature search was conducted in January 2025, using all databases described in this section. The search was updated periodically, with a final update in February 2026, to include newly published relevant studies.

TABLE 1. Search terms used in the literature search.

Databases: Google Scholar, Scopus, Web of Science.

The screening process starts with a title and abstract review to identify candidate studies, followed by a full-text manual assessment by the author to confirm they meet the inclusion criteria defined in Section 2.3. Screening was performed by the first author, with periodic verification by the co-authors. Studies that met the exclusion criteria were removed at each stage. Ultimately, 99 studies were retained for review. Figure 2, PRISMA flow diagram, shows the identification, screening, and inclusion of studies for this review.

2.3 Selection criteria

Most of the reviewed studies were published between 2000 and 2026. The starting point aligns with the commissioning of Horns Rev 1 in Denmark in 2002, the world’s first large-scale commercial offshore wind farm at 160 MW. This marked a key transition from early demonstration projects to commercial-scale deployment. As offshore wind grew alongside the maturing oil and gas sector, the scale and cost of these projects increased, and the need for reliable cost estimation methods became more important. Some earlier works were also included, where they introduced foundational cost estimation methods that later studies built upon. A study is included if it:

• Directly addressed steel offshore support structures
• Proposed or applied a cost model in the context of structural design or optimization
• Provided enough detail to understand how the cost model worked
• Was published as a peer-reviewed journal article, conference paper, or other credible academic publication in English

A study is excluded if it:

• Only covered topsides, towers, turbines, electrical systems, or other non-structural components
• Mainly focused on concrete or composite structures, unless steel was the primary load-bearing material
• Reported only high-level economic indicators like levelized cost of energy (LCOE) without linking them to structural parameters or cost breakdowns. Where LCOE was calculated using a detailed CAPEX breakdown, only the separable CAPEX components were extracted.
• Addressed only OPEX or DECEX without any CAPEX consideration
• Did not provide sufficient detail on the cost model formulation

Duplicate records identified across databases were removed prior to screening. Where a conference paper and a journal article reported the same study, only the more detailed version was retained.

Following the screening process, a total of 99 studies were identified as meeting the inclusion criteria and are reviewed in this paper. As shown in Figure 3, the number of relevant publications grew noticeably after 2016, with the years between 2022 and 2025 having the highest output.

Table 2 shows the journals in which the reviewed studies were published. Ocean Engineering and Marine Structures account for the largest share, which is expected given their direct focus on offshore structural topics. High-impact energy journals such as Applied Energy and Renewable Energy are also present, reflecting the growing interest in cost modelling within the offshore wind sector. Conference papers were also drawn from leading venues in the offshore engineering field, including OMAE and ISOPE, as well as other relevant proceedings. The complete classification of all 99 reviewed studies is available in a supplementary dataset for reference (Asgari Motlagh et al., 2026).

TABLE 2. Journals and performance indicators (2024) for the reviewed literature.

3 CAPEX cost models in steel offshore structures

3.1 Overview

Given the substantial volume of materials used in steel support structures for offshore energy platforms, efficient design and optimization of these structures has always been of interest to researchers and industry. As an example, steel jacket support structures alone can account for 19% of total offshore wind farm costs, which shows the importance of cost-driven design for these structures (Tian et al., 2022). Accurate cost models help designers make better decisions in early design stages, when small changes can still have a big impact on final costs. This section reviews the CAPEX cost models used in the literature for both fixed and floating steel offshore structures. The review begins with material cost models, as they are the most common due to their simplicity. Manufacturing costs are then discussed, covering both fundamental approaches developed for general welded steel structures and their adaptation to offshore applications. Finally, transportation and installation cost models are reviewed.

3.2 Material costs

Among the key factors in the design and optimization of steel offshore structures, the effective use of material, e.g., steel, itself stands as one of the most significant considerations in the cost models. Indeed, the most basic cost model for steel offshore structures simply correlates total cost with structural weight, as material quantity is directly proportional to overall costs (Qian et al., 2024).

3.2.1 Material cost models in fixed offshore structures

Many studies simplify the cost function by focusing solely on minimizing structural mass, using it as a proxy for minimizing material cost in fixed offshore jacket structures (Chew et al., 2015; Savsani et al., 2021; Bárcena Pasamontes et al., 2014). Although these studies often do not explicitly focus on cost, they implicitly target CAPEX reduction through structural weight minimization. In the same vein, Benítez-Suárez et al. (2025), Syarifudin et al. (2025), and Cruz et al. (2024) adopted simplified cost models expressed as the sum of all masses of structural elements for fixed jacket structures. The mass of each element is calculated as the product of material density, cross-sectional area, and element length, and hence, the general form of the cost model is as follows Equation 1:where ρ i stands for the density of material (steel), while A i x and l i indicate the cross-sectional area and length of each element, respectively. As offshore jacket structures generally consist of tubular elements, the cross-sectional area (A) in the cost model is calculated using the outer diameter (D i) and thickness (t i) of each structural element (Figure 4), as expressed in Equation 2 (Asgari Motlagh et al., 2021; Couceiro et al., 2020; Prastianto RW. et al., 2024):

f x = ∑ i = 1 n e ρ i A i x l i (1)

A i = π 4 D i 2 − D i − 2 t i 2 (2)

Moreover, steel grade is an important material property that defines the yield strength and mechanical behaviour of structural members. Although S355 is commonly used, jacket optimization studies based on material cost models have also adopted other steel grades, such as D36 or steels with yield strengths as low as 260 MPa. This variation in steel grade, alongside design variables such as member diameter and wall thickness, can directly affect the required member dimensions, structural weight, and consequently material cost. Table 3 summarises the steel grades and design variables reported in selected jacket optimization studies.

TABLE 3. Steel grades and design variables used in material-based cost models for jacket optimization studies.

Sandal et al. (2018a) and Sandal et al. (2018b) also employed simple cost models that used mass minimization as the sole objective function for jacket structures, explicitly acknowledging this as a simplification and noting that their methodology can certainly be extended with a more realistic cost model considering manufacturing, transportation, and installation expenses. Stolpe and Sandal (2018) also used a similar mass minimization methodology for jacket optimization with discrete design variables (diameter and thickness from catalogues). The same approach was also used by Asgari Motlagh et al. (2021), Oest et al. (2018), Prastianto R. et al. (2024), Kaveh and Sabeti (2018a), and Han et al. (2024), who all formulated their cost models as the sum of element weights in the design optimization of steel offshore jacket platforms.

Weight-based cost models have also been adopted for other types of fixed offshore structures, such as steel monopile foundations. McWilliam et al. (2021) employed mass minimization as their primary objective function in monopile optimization, while Ishtiyak and Sarkar (2025) used cross-sectional area minimization as their cost proxy, explicitly acknowledging that reducing material usage directly reduces costs. Similarly, Yu et al. (2022) and Lee et al. (2014) adopted volume and mass minimization, respectively, as their sole cost objectives in monopile design optimization. The same approach has also been applied for tripod sub-structures for structural mass minimization as a cost proxy (Yang et al., 2015).

3.2.2 Material cost models in floating offshore structures

Similar to fixed offshore jacket structures, simplified weight-based cost models–that solely focus on the mass of steel as the only cost factor–have been widely implemented across floating offshore structures (Bai et al., 2023; Pan et al., 2025; Mao et al., 2022; Tian et al., 2021; Qiu et al., 2019; Park et al., 2015; Abdelmoteleb et al., 2022; Drabo et al., 2024; Gallardo et al., 2024). Chen et al. (2024) developed an optimization framework for comparing double-pontoon and ring-pontoon types of semi-submersible platforms. They simply estimated structural weight through volume calculations multiplied by component-specific weight coefficients as shown in Equation 3:where V D, V C, and V P stand for volumes of deck, columns, and pontoons, respectively. Besides P w D, P w C, and P w P are the lightweight coefficients of the deck, column, and pontoon (260, 210, and 280 kg/m³, respectively). Similarly, Wang B. et al. (2025) used structural weight as a proxy for cost in the optimization of a semi-submersible drilling platform column, optimizing plate thicknesses of outer plates, longitudinal beams, double-shell, and bulkheads. Jang et al. (2019) applied the same weight-based approach to optimize a semi-submersible central column, considering both plate thickness and stiffener sections as design variables. Using a similar approach, Hsu et al. (2022) utilized steel weight as an indicator of cost in the parametric optimization of a 15 MW tri-column semi-submersible platform, applying the steel-to-displacement ratio as an efficiency indicator.

L W = V D . P w D + V C . P w C + V P . P w P (3)

On the other hand, unlike the material-based cost models used for fixed jacket structures, cost models for steel floating offshore structures also account for additional material used in other components such as mooring systems, anchors, and ballast (Lai et al., 2025). In the studies of Dou et al. (2020) and Polli et al. (2021), the objective function combined mass-proportional steel costs for the spar-buoy floater (2.5 €/kg), mooring lines cost (40.0 €/m), and ballast (set at zero), while using fixed steel thickness (6 cm) to simplify the optimization process. In the same vein, Faraggiana et al. (2022) modelled the costs of an OC3 spar buoy concept offshore wind turbine using an objective function that minimized the sum of structural and ballast material costs, as follows Equation 4:

f c o s t = m s t r u c t . c s t r u c t + m b a l l a s t . c b a l l a s t (4)

In their cost model, m s t r u c t and m b a l l a s t are the masses of steel structure and concrete ballast, while c s t r u c t (3,000 €/ton) and c b a l l a s t (500 €/ton) are their specific costs. The steel mass is calculated using a constant thickness of 7 cm for the platform walls, while ballast mass is computed based on hydrostatic balance requirements. Ojo et al. (2025) applied NURBS-based shape optimization to OC3 spar variants, minimizing structural steel mass (ρ_steel × Volume) with fixed 4 cm wall thickness as the cost objective and assuming negligible ballast costs.

In a similar approach, Wang J. et al. (2025) developed a cost-optimization model for a semi-submersible floating offshore wind turbine, using material-specific coefficients (2.7 €/kg for steel, 3.5 €/kg for mooring) with the constraint (Equation 5):where M p t is the steel mass of the platform, m m l is the mooring line mass per unit length, L m l is line length, and C o is the original design’s reference cost. It is noted that mooring costs are reported with different units and values across studies, ranging from 40.0 €/m (Dou et al., 2020; Polli et al., 2021) to 3.5 €/kg (Wang J. et al., 2025). Such differences can be attributed to variations in assumed chain diameter, market conditions, or whether installation costs are included in the reported rate. Bai et al. (2025) adopted the same steel cost coefficient (2.7 €/kg) as a cost constraint for a 15 MW semi-submersible platform, where platform steel mass excludes the tower, nacelle, and ballast. To optimize a TetraSpar-type floater and tower, Pollini et al. (2023) proposed a cost model using a mass-proportional approach where total cost is obtained by summing the costs of floater steel mass (c f m f, tower mass (c t m t, mooring line mass (c m m m, and keel connecting cable mass (c k m k, using the following Equation 6:

M p t . 2.7 + 3 . m m l . L m l . 3.5 ≤ C o (5)

c v = c f m f v + c t m t v + c m m m v + c k m k v (6)

Fylling and Berthelsen (2011) proposed a cost model for spar buoy-type floating wind turbine structures for the minimization of the total material costs, including the spar buoy, mooring lines, and power cable. The cost model of the spar buoy scaled proportionally with material mass, incorporating geometric parameters (Equation 7):where F i , 0 is the initial cost for section i, D i is diameter, Z i is vertical position, L i is length, and the exponents p d i and p z i control diameter and depth dependency, respectively. In addition, Fylling and Berthelsen (2011) calculated mooring system costs as follows Equation 8:where C i is the cost per unit mass of line segment i, m i is mass per unit length, and l i is the segment length. Li et al. (2025) optimized the mooring system considering only the total mooring chain mass as the objective function, excluding other structural components (Equation 9):where L is the mooring chain length, D is the chain diameter, and ρ is the chain linear density (mass per unit length). In the studies of Sugita (2014) and Sugita and Suzuki (2016), a cost model was developed for the weight optimization of a tension leg platform (TLP), including the hull and tendon weights. The structural weight estimation was conducted considering plate thickness and stiffener size of the TLP hull, while frames were made of high-tensile steel (50 ksi yield strength). Their cost model accounted for all key components, including pontoons, columns, bulkheads, and access shafts.

F s p a r = ∑ F i , 0 D i D i , 0 p d i Z i Z i , 0 p z i L i L i , 0 (7)

F m o o r = ∑ l i n e a l l ∑ s e g m . a l l C i m i l i (8)

m x = L π D 2 ρ 4 (9)

3.2.3 Application of material cost models

Although the simplified weight-based cost models applied for different steel offshore structures ignore manufacturing, transportation, and installation costs, they provide a practical first approximation for preliminary design optimization, particularly in comparative studies of different structural configurations and optimization algorithms. Accordingly, simplified weight-based cost models were utilized to compare different optimization methods, such as Enhanced Colliding Bodies Optimization (ECBO) and Genetic Algorithm GA (Shabakhty et al., 2024), Colliding Bodies Optimization (CBO) and ECBO (Kaveh and Sabeti, 2018b), GA, Particle Swarm Optimization (PSO), and marine predator-based algorithms (Han et al., 2024), Parametric Optimization and Numerical Genetic Algorithm optimization (Wang et al., 2024), GA, Teaching Learning-Based Optimization (TLBO), and several other state-of-the-art algorithms from CEC (Savsani et al., 2021), GA and Differential Evolution with various constraint-handling techniques (Cruz et al., 2024), PSO, Simulated Annealing (SA), and Ant Colony Optimization (ACO) (Drabo et al., 2024), and gradient-based Sequential Quadratic Programming algorithms (SNOPT and SLSQP) (Rohrer et al., 2026). Additionally, being straightforward, weight-based cost models were applied to investigate different structural configurations for steel offshore structures (Benítez-Suárez et al., 2025; Bai et al., 2025; Tran et al., 2022; Dhaneswara et al., 2023; Chew et al., 2013; Fernández Rodríguez et al., 2025; Zhou et al., 2025).

Furthermore, due to the simplicity of weight-based cost models, they are combined with other performance-based objective functions, such as platform stability (static heeling angle) and heave natural period (Bai et al., 2023; Drabo et al., 2024), maximum Response Amplitude Operator (RAO) in each degree of freedom (Pan et al., 2025) maximum heave or roll responses (Mao et al., 2022; Tian et al., 2021; Qiu et al., 2019; Park et al., 2015), maximum dynamic tendon tension (Zhang et al., 2018), and static pitch under maximum thrust (Abdelmoteleb et al., 2022), maximum tower top displacement (Chang and Peng, 2026), and minimum collision depth (Wang B. et al., 2025), in multi-objective optimization frameworks. Although these non-economic objectives can indirectly influence the lifecycle costs, they mainly focus on the technical performance and safety aspects of the steel offshore structure rather than direct economic considerations.

Similarly, weight-based approaches are also well-suited for topology optimization of steel offshore structures. Topology optimization focuses on distributing material (steel) efficiently within a design space rather than adjusting member dimensions. While no explicit cost model is generally applied, the mass of steel serves as an implicit cost proxy in topology optimization. The mass-based topology optimization has been applied to various offshore steel structures, such as jacket structures (Tian et al., 2022; Zhou et al., 2025; Marjan and Huang, 2023; Guo et al., 2025), and semi-submersibles (Chen et al., 2026). While these simplified weight-based approaches provide a useful starting point, more detailed cost models that account for geometric complexity, welding, and manufacturing processes are discussed in the following sections.

3.3 Manufacturing costs

While material costs form the foundation of cost models in steel offshore structures, many other factors must be incorporated for a more comprehensive economic assessment. Among these, manufacturing expenses are the most significant. Manufacturing processes such as blasting, cutting, welding, and painting can make up more than 30% of the total production cost of welded steel beams (Mela et al., 2014). For steel offshore wind turbines, in particular, welding and bending account for 46% of manufacturing costs. This shows that manufacturing expenses, especially welding and forming, should be included in design optimization rather than relying only on steel mass as a cost proxy (Nekeman, 2021).

Many approaches considering manufacturing costs in steel offshore structures evolved from methodologies originally developed for conventional steel structures. These fundamental methodologies established basic cost estimation approaches and calculation frameworks that were later applied for offshore steel structures. Examining these foundation models provides insight into how researchers systematically adapted manufacturing cost considerations to steel offshore structures.

3.3.1 Fundamental manufacturing cost models in steel structures

In the realm of cost calculation of welded steel structures, Jármai and Farkas (1999) presented an estimation approach, focusing on the manufacturing costs. The manufacturing costs included costs of welding, cutting, preparation, assembly, tacking, and painting. The general form of manufacturing costs of welded steel structures can be expressed as:

K = K m + K f = k m ρ V + k f T 1 + T 2 + T 3 + T 4 + T 5 + T 6 + T 7 ⏞ ∑ i T i (10)

In Equation 10, K m and K f are the material and manufacturing cost factors, ρ and V are material density and volume, and T i are the individual production times. These times cover seven activities: T 1 is preparation, assembly, and tacking, dependent on a difficulty factor, the number of elements, and structural mass; T 2 + T 3 cover welding time and additional actions such as electrode changes and deslagging, dependent on weld size, weld length, and welding technology; T 4 is plate flattening, dependent on plate thickness and area; T 5 + T 6 cover surface preparation and painting, both dependent on surface area and difficulty factors; and T 7 is cutting and edge grinding, dependent on plate thickness and cutting length. Their case studies showed manufacturing costs reach 29%–35% of total costs for welded box beams and 46%–71% for stiffened plates. Switching from SMAW to CO₂ welding achieved savings of 13% for box beams and 32% for stiffened plates, demonstrating that welding technology selection has a larger impact on structures with more extensive welding requirements. Building on the Jármai-Farkas approach, Gu et al. (2023) applied it to the topology optimization of welded frame structures, combining material cost (C m) with manufacturing costs for preparation (C 1), cutting (C 2), welding (C 3), and painting (C 4) (Equation 11):

M Z = C m Z + ∑ i C i Z (11)

Their welding cost (C 3) was calculated as a linear function of welding length, with a weighting factor depending on the welding technology. Pavlovčič et al. (2004) also developed a comprehensive cost model for steel frames covering material, welding, cutting, painting, surface preparation, flange aligning, joint costs, transportation, and erection (Equation 12):

C = C s t e e l − e l e m . + ∑ n e l C w e l d i n g + ∑ n e l C c u t t i n g + ∑ n e l C p a i n t i n g + ∑ n e l C s u r f a c e p r e p r a t i o n + ∑ n e l C f l a n g e a l i g n i n g + C j o i n t s + C t r a n s p o r t a t i o n + C e r e c t i o n (12)

They further broke welding costs down into three sub-components: manufacturing cost (covering labour, equipment, and electrode changes), welding material cost (consumables such as wire, shielding gas, flux, and energy), and assembly cost (handling, tack welding, and preparation), each expressed as a function of weld length, weld size, element volume, and relevant cost factors. Jin et al. (2017) developed a multi-objective optimization model for steel structures considering welding costs and constructability of beam-column connections. Their cost model combined material cost, production/assembly cost, and welding cost (Equation 13):

C = C S t e e l + C f a c t o r y + C w e l d i n g (13)

Welding cost was calculated by multiplying the unit weld cost by the equivalent length of a standard 6 mm fillet weld, with factory and site welding treated separately, and each depending on unit weld cost and weld length. Jármai (2023) later extended their earlier work (Jármai and Farkas, 1999) with a dual methodology combining local (direct manufacturing costs) and global (environmental and life cycle) approaches. He found that laser welding reduces overall costs by 24%, and laser cutting reduces them by 12% compared to acetylene cutting. Mela et al. (2014) developed a parallel cost optimization methodology for welded high-strength steel beams, with a cost function covering material (C M), blasting (C B), cutting (C C), sawing (C S), welding (C W), painting (C P), transportation (C T), and erection (C E) (Equation 14):

C = C M x + C B x + C C x + C S x + C W x + C P x + C T x + C E x (14)

Each cost term was calculated using a generic equation that accounts for both productive and non-productive time, along with unit costs for labour, equipment, maintenance, real estate, consumables, and energy (Equation 15):where T N k and T P k are non-productive and productive times, and the unit costs cover labour (c L k), equipment instalment (c e q k), equipment maintenance (c M k), real estate (c R E k), and real estate maintenance (c S e k), divided by utilization ratio (u k), while c C k and c E n k are unit costs of time-dependent consumables and energy, and C C k is the total cost of time-independent consumables. Individual cost terms were then formulated as functions of geometric parameters: blasting depends on beam length; cutting on plate thickness and cut length; and welding on weld length and weld size. Based on this methodology, Haapio (2012) developed a feature-based costing method (FBCM) for skeletal steel structures using the same cost-centre framework. Miler et al. (2022) directly compared the Jármai-Farkas and Mela-Heinisuo methods on two welded structures. Both produced similar total cost values and were equally effective for ranking design alternatives. However, they differed in cost distribution: the Jármai-Farkas method predicted higher welding costs, while the Mela-Heinisuo method estimated higher cutting and material costs. Miler et al. concluded that both methods are reliable for total cost calculation and design ranking but show inconsistencies in estimating individual manufacturing process costs.

C k = T N k + T P k c L k + c e q k + c M k + c R E k + c S e k u k + T P k c C k + c E n k + C C k (15)

3.3.2 Manufacturing costs in steel offshore structures

The cost estimation approach developed by Jármai and Farkas was later applied to obtain the minimum cost design of a welded orthogonally stiffened cylindrical shell member of an offshore fixed platform truss (Jármai et al., 2006). The cylindrical shell member was subjected to axial compression and external pressure. In this study, the cylindrical-shell member was orthogonally stiffened using ring stiffeners of box cross-section and stringers of halved rolled I-section.

As expressed in Equation 16, the cost function included the cost of material (K M), cost of manufacturing such as forming of plate elements into cylindrical shapes (K F i), welding, and painting (K P) (Jármai et al., 2006)

K = K M + ∑ i K F i + K P (16)

In Equation 16, the cost of manufacturing (K F i) encompassed six parts: K F 0 is the cost of forming plate elements into cylindrical shapes, K F 1 is the cost associated with the welding the shell segments from two curved plate elements, K F 2 stands for the expense of welding small shell segments to form the whole unstiffened shell, K F 3 indicates the cost of welding plate elements to form ring-stiffeners, K F 4 is the cost related to the welding of ring-stiffeners into the whole shell, and finally, K F 5 is the expense of stringers welding into the shell. The cost of forming a shell element into a cylindrical shape can be calculated using Equation 17 (Jármai et al., 2006; Farkas et al., 2004; Simões et al., 2006; Farkas, 2011). In Equation 17, R is the shell radius, and t is the shell thickness.

K F 0 = k F Θ e μ μ = 6.582513 − 4.527217 t − 0.5 + 0.009541996 2 R 0.5 (17)

Zhou et al. (2021) conducted a sensitivity analysis to study the impacts of substructure configuration on manufacturing cost and hydrodynamics of a four-column semi-submersible floating wind turbine. In their study, the manufacturing cost estimation was based on the study of Jármai and Farkas (1999) and included the time of preparation, assembly, and tacking (T w 1), the real welding time (T w 2), and the additional manufacturing action time (T w 3). Zhou et al. assumed that the raw steel plate with a dimension of 10 m (length) × 4 m (width) was applied for the manufacturing of the semi-submersible structure. Zhou et al. estimated the welding cost by multiplying the welding time by the labour price. A similar approach was utilized to estimate the painting and mooring line costs. Finally, the total manufacturing cost was estimated as given in Equation 18 (Zhou et al., 2021).where m p t f m, m b l s t, and m m o o r are the mass of platform, ballast, and mooring, respectively, while A stands for the platform surface area. Besides, P s t e e l, P b l s t, P l a b o u r, P p a i n t, and P m o o r indicates the price of steel, ballast, labour, painting, and mooring price, respectively. In addition, they considered 27% of the total manufacturing cost as the overhead cost.

T C = m p t f m P s t e e l + m b l s t P b l s t + T w 1 + T w 2 + T w 3 P l a b o u r + A P p a i n t + m m o o r P m o o r / 0.73 (18)

Utilizing Jármai and Farkas’s cost estimation approach, Hegseth et al. (2020) developed an integrated design optimization framework for spar floating wind turbines that considers both material and manufacturing costs. Their cost model did not include installation, maintenance, and decommissioning expenses. They calculated the cost of the platform using Equation 19:where K m and K f are the steel cost factor and the manufacturing cost per unit time, respectively, while M s p a r indicates the steel mass of the hull and T i is the time spent at the manufacturing stage i, expressed as a function of the geometry. In their model, the costs of manufacturing comprised expenses related to forming plate elements into shells, assembly, welding, and painting. In addition, they did not consider the cost of ballast in their study. Using different design variables, the study took into account the scantling design of the hull of the spar platform, i.e., circumferential T-ring stiffeners. In another study, Hegseth et al. (2021) applied the Jármai and Farkas cost estimation approach, as expressed in Equation 19, for a 10 MW spar floating wind turbine, with a steel cost factor K m = 2.7 €/kg and K m/K f ratio of 1.0. They neglected concrete ballast cost due to its lower price compared to structural steel. Ferri and Marino (2023) developed a three-stage optimization framework for a 10 MW floating offshore wind turbine designed specifically for Mediterranean Sea conditions, considering objective functions, including fatigue damage over significant Serviceability Limit State (SLS) load cases, the maximum stresses on the tower under a 50-year return period ULS LC, and the overall cost of the floating foundation (floater + moorings + anchors). Although total costs comprised manufacturing, transport, and installation, due to the complexity of quantifying installation costs, only manufacturing and assembly costs were considered in their cost function, as stated in Equation 20.

C s p a r = K m M s p a r + K f ∑ i T i (19)

M C = M P l a t f o r m P s t e e l + M b a l a s t P c o n c r e t e + T 1 + 1.3 T 2 P l a b o u r + M M o o r i n g P c h a i n + M A n c h o r P A n c h o r + A P p a i n t (20)

In the cost estimation proposed by Ferri and Marino (2023), T 1 and T 2 are construction times associated with the processes of assembling and welding, respectively, and are evaluated according to the study of Jármai and Farkas (1999).

In addition to the Jármai and Farkas cost estimation approach that has been widely used for steel offshore structures, other comprehensive cost models have been developed that incorporate multiple manufacturing cost components. To optimize the steel jacket structure for offshore wind turbines, Häfele et al. (2018), Häfele et al. (2019), and Häfele and Rolfes (2016) developed a framework to calculate the total costs based on expenses of material, manufacturing, coating, transition piece, and transport (Equation 21):where a j represents unit costs and c j are corresponding cost factors. The manufacturing costs consisted of cumulative weld volume in three main components based on the joint configurations in the jacket topology (Equation 22):

C t o t a l x = ∑ j a j c j x (21)

C manufacturing = C brace − to − brace and brace − to − leg weld volume + C mud brace − to − leg weld volume + C Leg − to − leg weld volume (22)

The cumulative volume of all structural welds (C m a n u f a c t u r i n g) is calculated based on the weld root thickness, 45° weld angles, and perimeter lengths of elliptical projections onto connected chord surfaces. Additionally, the coating costs are determined by the total outer surface area of all tubular members, accounting for corrosion protection requirements applied during the manufacturing process. Similarly, Burak and Mengshoel (2021) estimated welding cost based on the contact area between connected tubular elements in their multi-objective jacket optimization. This weld weight was combined with material weight to form a simplified manufacturing cost objective. While these studies focused on jacket structures with geometry-based welding cost calculations, Yeter et al. (2017) applied a more straightforward manufacturing cost approach for monopile support structures. Their model simplified the manufacturing cost estimation into three main components (Equation 23):where steel costs (C s t e e l) depended on volume, price (2.5 €/kg), and manufacturing amplification factors; welding costs (C w e l d i n g) considered welding material costs (15 €/kg) and welding profiles; and labour costs (C l a b o u r) were based on welding profile and speed requirements. Taking a different approach, Maienza et al. (2020) developed a parametric cost model for steel floating platforms, including semi-submersible, spar buoy, and tension leg platforms. Instead of calculating welding costs directly, they estimated the direct labour cost C p , 2 based on working hours as follows Equation 24:where h L , i represents labour hours for preparation, prefabrication, painting, and corrosion protection, and c L is the hourly labour rate. An additional 5% overhead was included for engineering and administrative activities. Similarly, Castro-Santos et al. (2016) and Cas et al. (2016) applied an Activity-Based Cost model for floating steel platform manufacturing in shipyards. Their model calculated the platform manufacturing cost as Equation 25:where C M a t is the materials cost, C D L is the direct labour cost, C A C T covers indirect activity costs, and Bi is the industrial profit. Both material and labour costs were functions of platform mass and surface areas. Besides, for mooring and anchoring manufacturing, costs were calculated as Equations 26, 27: where p M o o r is mass per meter, L M o o r is the mooring length, C M o o r is unit cost (€/kg), L P is the number of mooring lines, m A n c h is anchor mass and C A n c h is anchor unit cost (€/kg).

f c o s t = C s t e e l + C w e l d i n g + C l a b o u r (23)

C p , 2 = ∑ i h L , i . c L (24)

C P l a t f o r m = C M a t + C D L + C A C T . 1 + B i (25)

C m o o r i n g = p M o o r . L M o o r . C M o o r . L P (26)

C a n c h o r i n g = L P . m A n c h . C A n c h (27)

3.3.2.1 Simplified manufacturing cost models

While detailed manufacturing-based approaches provide precision for specific structural configurations, simplified manufacturing cost estimation methods have also been developed for broader applications. Myhr et al. (2014), Ioannou et al. (2020), and Baudino Bessone et al. (2022) estimated manufacturing costs for steel offshore structures using Manufacturing Complexity Factors (MCF). These studies calculated manufacturing costs by multiplying material costs by complexity factors (MCF) to account for welding, assembly processes, and manufacturing suitability. Using the monopile as a benchmark, they derived different MCF values for steel floating concepts, such as spar buoys and semi-submersibles. Similarly, Ghigo et al. (2020) applied a similar simplified approach in their steel platforms optimization study, using form factors of 1.1 for spar-buoy and 1.3 for Hexafloat designs multiplied by material costs to account for the complexity of the structure and the cost of manufacturing. Oberti (2022) later applied the MCF approach in a genetic algorithm-based optimization study of multiple floating platform designs for the NREL 5 MW turbine.

Some researchers preferred even simpler methods based on cost per unit mass. Lefebvre and Collu (2012) used this approach for their tri-floater offshore platform design. They applied different costs per tonne for plates versus trusses and beams. Karimi et al. (2017) simplified this further by applying a single cost of $2.5/kg for all structural mass. This combined the material and manufacturing costs into one factor without separating different manufacturing activities. Shafiee et al. (2016) also combined material and manufacturing costs using established models from Dicorato et al. (2011). Similarly, Martinez and Iglesias (2021) and Martinez and Iglesias (2022) used a fixed cost of 8 M€ per steel semi-submersible platform (WindFloat-type). This value combines material and manufacturing costs without separating individual manufacturing activities. Castro-Santos et al. (2018) defined manufacturing cost in their framework but obtained values from concept designers. Díaz and Guedes Soares (2023) also combined material and manufacturing costs into capacity-based parametric equations for steel floating platforms, including TLP, spar, and semi-submersible. Their model expressed substructure production costs as linear functions of wind farm capacity, adapted from Myhr et al. (2014). Lerch et al. (2018) also simplified their approach by receiving substructure manufacturing costs directly from designers as inputs for sensitivity analysis. Table 4 summarises the MCF values reported across the reviewed studies for different offshore structure types, providing an overview of the complexity factors applied in each case.

TABLE 4. Manufacturing complexity factors (MCF) for steel offshore structures.

3.4 Transportation and installation costs

Transportation and installation costs represent a considerable portion of capital expenditure for steel offshore structures, typically ranging from 10% to 16% of total CAPEX depending on project characteristics and foundation type (Lacal-Arántegui et al., 2018). The complexity of offshore installation and transportation makes them nonstandard operations, such that the associated costs vary significantly based on foundation type, support structure design, and site conditions (Jiang, 2021). For transportation and installation, cost models are mainly driven by the time required to complete operations and the daily rates of the vessels involved (Sarker and Faiz, 2017).

3.4.1 Transportation cost models

Transportation plays a significant role in steel offshore support structure deployment, with costs highly dependent on distance, vessel type, and structural characteristics. Compared to fixed offshore structures, floating offshore structures can typically be transported and towed to the site by tugboats. This significantly reduces the need for specialized vessels and lowers overall costs (Sykes et al., 2023). However, for fixed offshore wind foundations such as jackets and monopiles, transportation typically requires specialized barges and heavy-lift vessels, with costs varying based on foundation weight, distance to port, and vessel availability, as the installation vessel may either carry foundations directly or rely on separate barge transport systems (Kaiser and Snyder, 2012). Figure 5 illustrates common vessel types used in the transportation and installation of steel offshore structures.

The majority of research studying transportation cost models in steel offshore structures has focused on vessel-based approaches. Using vessel-day rates, Shafiee et al. (2016) developed a transportation cost model for fixed-bottom jacket-type steel foundations. However, their cost model combined transportation and installation into a single cost term.

Castro-Santos et al. (2016) and Cas et al. (2016) developed a more detailed cost model for calculating the transportation costs of steel floating platforms (C p − t r a n s), based on the number of vessels (N v e s s e l), the number of platforms (N A), the cost of the vessel in €/day (C v e s s e l), and the cost of moving the vessel in € (C M o v) (Equation 28):

C p − t r a n s = N v e s s e l × N A × T A 1 × C v e s s e l + C M o v (28)

Additionally, the coefficient T A 1 can be calculated using the following Equation 29:where, N S B v e s s e l is the number of steel offshore platforms per vessel, t l o a d stands for the time spent loading the offshore platform at port (h), d s h i p y a r d − p o r t is the distance from the shipyard to port (m), v v e s s e l is the speed of the vessel used to transport the offshore platforms (m/s), and K t i m e is a bad weather loss coefficient. Maienza et al. (2020) used a similar approach to model transportation costs for floating platforms. They calculated costs based on vessel daily rates, travel time, and mobilization costs. Their comparison across platform types showed that semi-submersibles have the lowest transport costs, as they can be assembled at port and towed directly. Spar buoys and tension leg platforms require barges and multiple trips, leading to higher costs. As shown in Table 5 (Equations 30–32), for the mobilization cost of vessels (C M o v), Kaiser and Snyder (2012) provided detailed estimations that vary depending on the transport type, including tug transport, self-propelled transport, or heavy-lift vessel transport. Tug transport is typically used for derrick barges, heavy-lift vessels, and jack-up barges. Self-propelled transport is utilized for turbine installation vessels capable of mobilizing without vessel support. Heavy-lift vessel transport is employed for transporting installation vessels and derrick barges using semi-submersible vessels.

T A 1 = N S B v e s s e l · t l o a d + 2 3600 s 1 h · d s h i p y a r d − p o r t v v e s s e l · 1 K t i m e · N A N S B v e s s e l 24 s 1 d a y (29)

TABLE 5. Mobilization cost of vessels (Kaiser and Snyder, 2012).

X: tow distance in nautical miles

V t, V i, V s: speed (knots) of the tug, vessel, and semi-submersible

D: installation vessel displacement (tons)

T: tug dayrate per horsepower ($/hp.day)

G: cost of fuel per gallon ($/gal)

I: installation vessel dayrate ($/day)

R: replacement cost of the installation vessel ($)

N t, N s: insurance premium (%)

P i, P s: installed power vessel and semi-submersible heavy-lift (hp)

S: semi-submersible dayrate ($/day)

Castro-Santos et al.’s methodology was later applied to calculate the transport and installation costs of floating offshore wind foundations across different scenarios (IT1–IT4), where transport and installation are treated as integrated operations (Castro-Santos et al., 2018). Baudino Bessone et al. (2022) applied the IT1 scenario for semi-submersibles. They calculated transportation costs by multiplying vessel day rates by towing duration. For towing from the shipyard to port, they used one large tugboat (€28,321/day) and two small tugboats (€4,135/day). For towing from port to site, two AHTS vessels (€39,891/day) and two small tugboats were used. They also included shipyard storage costs based on the semi-submersible footprint area. Building on the vessel-based approach, Myhr et al. (2014) studied a comprehensive transportation cost methodology for both fixed (monopile and jacket foundations) and floating (Spar-Buoy, Semi-Submersible, Tension-Leg platforms) steel offshore structures. To estimate the transportation costs, they considered vessel mobilization, towing distance, towing speed, vessel day-rates, and weather-dependent delays.

In the study of Lerch et al. (2018), although no explicit cost model was presented for the transportation of steel TLPs, they correlated the transportation costs with vessel-specific parameters such as the day rate and fuel consumption, the rental time, and usage of the vessel, as well as mobilization and demobilization costs. They noted that the TLP concept requires a special purpose-built barge for offshore transportation due to its instability. Díaz and Guedes Soares (2023) also combined transportation and installation costs for floating platforms. Their model uses distance to coastal facilities (DCF) to account for transportation. They provided different coefficients for semi-submersibles, spar buoys, and TLPs. Similarly, Martinez and Iglesias (2021) and Martinez and Iglesias (2022) combined transportation and installation costs for floating semi-submersibles. They estimated the transportation cost based on travel time. The travel time was calculated as twice the distance to the site divided by vessel speed (2 d / V), to account for the round trip. This time was then multiplied by the vessel’s daily rate. This approach is simpler but does not allow separate estimation of transportation costs.

In contrast to detailed vessel-based methodologies, some studies have adopted simplified transportation cost models for computational efficiency in their structural optimization studies. Häfele et al. (2018) and Häfele et al. (2019) modelled transportation costs of steel jacket structures, C 5 x, using a mass-dependent approach (Equation 33):where C 1 x is the total jacket mass. Their approach assumes linear mass dependency, reflecting that barges have limited transport capacity, partially related to mass or deck space coverage. This mass-dependent approach for transport costs was also adopted in the study of Häfele and Rolfes (2016), with a suggested unit cost of 2.0 €/kg based on practical experience. However, this represents a large simplification as transport cost determination is acknowledged to be very difficult due to dependencies on unsteady market conditions, economies of scale, and environmental conditions. Similarly, Ghigo et al. (2020) bundled transportation within installation costs using power-based unit rates: 2.5 m€/MW for spars and 1.5 m€/MW for Hexafloat platforms, reflecting that towable floating structures need fewer specialized vessels than designs requiring offshore assembly.

C 5 x = C 1 x (33)

3.4.2 Installation cost models

Installation costs represent a significant portion of capital expenditure for steel offshore structures. The installation method depends largely on the foundation type (Kaiser and Snyder, 2012). For steel fixed foundations, monopiles are driven into the seabed using a hydraulic hammer, while jackets are lifted into place and secured with piles driven at each corner (Kaiser and Snyder, 2012). For steel floating platforms, installation complexity varies with platform geometry, mooring, and anchoring system (Sykes et al., 2023). This diversity in installation methods makes developing generalized cost models challenging. In addition to foundation type, installation cost models typically consider vessel type, hire rate, installation duration, and weather windows (Sykes et al., 2023).

Despite the challenges in developing generalized models for installation costs, several studies have proposed cost models for both fixed and floating steel offshore structures based on vessel hire rates, installation duration, and site-specific conditions. For the steel fixed offshore structures (jacket type), Shafiee et al. (2016) combined transportation and installation costs (C s s − t r a n s) into a single term, based on vessel-days required (N v − d) and daily vessel hire rate (V r) (Equation 34):

C s s − t r a n s = N v − d × V r (34)

Kaiser and Snyder (2012) split the installation cost of steel fixed foundations for offshore wind turbines into two models: barge, where foundations are brought to a stationary installation vessel, and self-transport, where the same vessel carries and installs them. These two installation configurations are illustrated schematically in Figure 6.

Barge model (Equation 35):

C i n s t a l l a t i o n = I + M × 1 W × N u n i t s 24 × V D R + S D R (35)

Self-transport model (Equation 36):where I is installation time per foundation (h), which includes pile driving and transition piece grouting, M is intra-field movement time (h), W is the weather uptime factor, N u n i t s is the number of foundations, V D R is vessel dayrate ($/day), S D R is spread dayrate ($/day), D is distance to port (nm), S is vessel speed (kn), V C is vessel capacity (units/trip), and L is loading time per unit (h).

C i n s t a l l a t i o n = 2 D S + V C . L + V C . I + V C . M × 1 W × N u n i t s V C 24 × V D R + S D R (36)

For floating steel platforms, Castro-Santos et al. (2016) and Cas et al. (2016) proposed a methodology to estimate the installation cost of semi-submersibles as follows Equation 37:where N v e s s e l T P L A T is the number of vessels, t i n s t a l l a t i o n P L A T is installation time (hours), N A is the number of platforms, and C c r a n e s h i p y a r d is crane cost (€/day). This applies to platforms that float and are towed directly to the site (e.g., semi-submersibles), where installation requires only a shipyard crane. Castro-Santos et al. (2018) also modelled installation cost for platforms carried on barges (e.g., TLPs and spars), which require a floating crane for offshore installation (Equation 38):where T A is the time coefficient (days), C I F P is floating crane cost (€/day), and C m o v I F P is the mobilization cost (€). For mooring and anchoring installation, Castro-Santos et al. (2016) and Cas et al. (2016) used an Anchor Handling Vessel (AHV) (Equation 39):where C A H V is vessel cost (€/day), C D L is direct labour (€/day), C p u m p s & d i v e r s is the pumps and divers cost (€/day), N a n c h o r s is the total number of anchors, and T i n s t A H V is the installation rate (anchors/day). Similarly, Maienza et al. (2020) estimated mooring installation cost using the AHV daily rate, direct labour cost, number of mooring lines, and installation rate. Ivanov et al. (2025) applied a similar vessel-based approach, assuming an AHV rate of $250,000/day based on Japanese floating wind project data. In another study, Castro-Santos et al. (2018) also considered a barge-tug combination for mooring and anchoring installation (Equation 40):where C b a r g e is barge cost (€/day), C t u g is tug cost (€/day), and T i n s t B a r g e is the installation rate (anchors/day). Myhr et al. (2014) proposed an installation cost model for both fixed and floating steel offshore wind turbines. They calculated costs based on vessel day rates, operation durations, personnel costs, and weather windows. For fixed foundations, vessel mobilisation time and capacity are considered. For floating platforms, upending time, ballast loading, and mooring line attachment time are included, while for mooring systems, installation time per anchor type with depth-dependent adjustments are applied. Baudino Bessone et al. (2022) estimated installation costs for steel semi-submersibles using equipment rates and operation durations. They applied a quayside crane rate of €833/h for turbine assembly and 14 h for on-site installation, with 50% workability for marine operations. Their model also included a mooring hook-up and cable-laying vessel for dynamic cable connection. To calculate the installation costs, Martinez and Iglesias (2021) and Martinez and Iglesias (2022) also multiplied the vessel day rate by installation and transit time for semi-submersible tow-to-site operations. They added 240 k€/turbine for mooring installation. Using a similar approach, Lerch et al. (2018) calculated installation costs for steel TLPs based on vessel day rates, divers, and auxiliary equipment. They noted that platform instability requires a purpose-built barge. Extending this framework, Díaz and Guedes Soares (2023) developed installation cost models for floating steel platforms for offshore wind farms incorporating water depth (W D), distance to coastal facilities (D C F), and wind farm capacity (W F C), as follows Equations 41–43:

C i n s t a l l a t i o n = N v e s s e l T P L A T × t i n s t a l l a t i o n P L A T 24 h d a y × N A × C c r a n e s h i p y a r d (37)

C i n s t a l l a t i o n = T A + C I F P + C m o v I F P (38)

C i n s t a l l a t i o n = C A H V + C D L + C p u m p s & d i v e r s × N a n c h o r s T i n s t A H V (39)

C i n s t a l l a t i o n = C b a r g e + C t u g + C D L + C p u m p s & d i v e r s × N a n c h o r s T i n s t B a r g e (40)

TLPs : C i n s t = 99746000 + 21625 × W D + 75650 × D C F / 600000 × W F C k W (41)

Semi − Submersibles : C i n s t = 23658000 + 11625 × W D + 35450 × D C F / 600000 × W F C k W (42)

Spar Buoys : C i n s t = 94577688 + 9850 × W D + 175081 × D C F / 600000 × W F C k W (43)

In Equations 38–40, the coefficients differ based on each platform’s installation method. Mooring costs are included in the structure installation via the water depth term.

Similar to transportation cost models, a number of studies used simpler approaches for installation costs, instead of detailed vessel-based models. For floating steel platforms, some studies applied fixed per-unit rates. Ghigo et al. (2020) estimated installation costs using fixed rates based on turbine capacity. They noted that spar is more expensive than semi-submersibles due to offshore assembly needs. Karimi et al. (2017) applied rates for anchor installation: $5,000/anchor for drag embedment, $8,000/anchor for vertical-load, and $11,000/anchor for suction pile. They selected the anchor type based on the mooring line angle. For fixed steel jackets, on the other hand, Häfele and Rolfes (2016), Häfele et al. (2018), and Häfele et al. (2019) assumed installation cost depends on the number of piles (Equation 44):where a is the unit cost (€/leg) and N L is the number of legs. They assumed €200,000/leg, covering pile foundation and offshore installation.

C i n s t a l l a t i o n = a × N L (44)

Furthermore, several studies considered installation costs as part of their economic analysis, but without developing detailed cost models. In this regard, Zhao et al. (2026) noted that clump weights raise mooring installation expenses due to extra handling. Lai et al. (2025) also included installation cost as a CAPEX component for semi-submersibles; however, they did not provide any explicit formula.

3.5 Uncertainty and market variability

Even detailed CAPEX models for steel offshore structures depend on steel prices, labour rates, and vessel day rates. Most of the reviewed studies nonetheless treat these specific cost inputs as fixed and unchanging.

However, a small number of studies did test this uncertainty directly within their own cost models. Häfele et al. (2018) modelled several CAPEX unit costs as normally distributed random variables within a probabilistic cost model. This produced a distribution of possible total costs for each candidate jacket design in the study. A ±10% bounded range of investment cost was tested by Díaz and Guedes Soares (2023) through sensitivity analysis. Myhr et al. (2014) varied material, vessel, and mooring costs using generic high and low percentage scenarios. The One-at-a-time (OAT) method was applied by Lerch et al. (2018) to identify parameters most influencing floating wind LCOE. The same OAT method was used by Martinez and Iglesias (2021), who then mapped best and worst cases regionally. Percentage-based cost changes were tested by Karimi et al. (2017) where anchor cost changed the optimal choice, not material cost. Ioannou et al. (2020) tested wall thickness at ±10% and ±20%, and manufacturing complexity at ±20%, against manufacturing cost. Baudino Bessone et al. (2022) similarly tested wall thickness and manufacturing complexity, both at ±20%, against manufacturing cost. Across these studies, material or substructure cost most often emerged as a strong driver, though the specific factor varied.

Overall, only a limited number of the reviewed studies have carried out this analysis so far. Most of these studies relied on simple sensitivity analysis, while probabilistic modelling and scenario-based assessment appeared far less often. Robust optimization under cost uncertainty, however, remains largely unexplored for offshore support structures so far. Wider use of these approaches is recommended for future CAPEX modelling of offshore structures. This should extend to factors such as manufacturing-yard capability, weather downtime, and exchange rates as well.

4 Discussion and future work

4.1 Discussion and synthesis

As noted in Section 2, the number of publications on CAPEX cost modelling for steel offshore structures has grown considerably since 2016, reflecting the increasing economic importance of these structures. As illustrated in Figure 7, offshore wind energy covers 77% of the reviewed CAPEX cost modelling studies. Oil and gas platforms make up 18%, while other applications account for just 5%. This trend can be attributed to the growing industry interest in renewable energy alongside the relatively early stage of offshore wind technology.

Additionally, Figure 8 shows the distribution of studies by structure type for steel offshore platforms in this review. Steel jackets and semi-submersibles have received the most attention, probably because of the more sophisticated manufacturing process. These structures require more material and significant welding and assembly effort, making detailed cost models more important during design. Besides, jackets are common in both oil and gas and offshore wind, while semi-submersibles are quickly becoming the top choice for floating offshore wind. This widespread use makes their cost modelling even more critical. On the other hand, spar buoys, monopiles, and tension leg platforms appear less frequently in CAPEX cost modelling research. This gap shows that cost estimation methods for these steel structures need more work. Furthermore, although semi-submersibles have been explored more, floating technology is still developing, and new floating platform concepts appear often. Understanding the CAPEX costs of these newer designs remains an open area for research.

In addition, the reviewed studies show that material costs are the most widely included component, appearing in 98% of all studies (Figure 9). However, as shown in Figure 9, 66.7% of studies rely on material cost alone, without incorporating any other CAPEX component. This is expected since weight-based models are the simplest approach, requiring only structural mass and steel price (Section 3.2). Due to their simplicity, these models have also been used to compare different optimization algorithms, investigate various structural configurations, combine with performance-based objectives in multi-objective frameworks, and conduct topology optimization. However, they do not capture the full picture of CAPEX. Manufacturing costs appear in 29% of studies, despite making up more than 30% of total costs in welded steel structures (Section 3.3). This is likely because these models need detailed information about welding lengths, assembly times, and labour rates. Only 13.1% of studies combine material and manufacturing costs without extending to transportation or installation. Transportation and installation costs have received the least attention, appearing in 16% and 20% of studies, respectively. These costs depend on vessel day rates, weather conditions, and distance to port, making them harder to model (Section 3.4). Notably, only 14.1% of the reviewed studies include all four CAPEX components together, which highlights a clear gap in the current literature.

Furthermore, the distribution of CAPEX cost categories varies when looking at different steel offshore structures (Figure 10). The values behind Figure 9 were calculated within each structure type’s own subset of studies. Compared to floating steel platforms, fixed steel structures like jackets and monopiles show less research on transportation and installation costs. This is notable because, as discussed in Section 3.4, installing fixed steel foundations involves complex operations like pile driving and heavy lifts that require specialized vessels. Floating steel platforms, on the other hand, can often be towed to the site, which is generally simpler. This suggests that transportation and installation cost models for fixed steel structures could use more attention.

Beyond cost categories, the structural components considered in cost models also differ between fixed and floating steel structures. Fixed structures like jackets and monopiles mainly consist of tubular steel members and piles, while floating steel platforms include several components such as the steel hull (columns and pontoons), mooring lines, anchors, and internal structures like stiffeners and bulkheads. Figure 11 shows the inclusion frequency of these components across the reviewed studies on floating steel platforms, as well as how often they are considered in combination. The steel hull has been included in 96% of CAPEX cost calculations, which is expected since the hull typically involves the largest volume of steel material. Mooring lines and anchors have been considered in 50% and 30% of CAPEX calculations, respectively. The most common combination is hull, moorings, and anchors together, appearing in 28% of studies, while 42% of studies considered only the hull with no other components. Internal steel structures like stiffeners and bulkheads have been included in only 12% of cases, always in combination with the hull. This is a notable gap because these components play an important role in the structural performance and integrity of floating steel platforms. Ignoring them in cost models may reduce the accuracy of CAPEX estimations and lead to less reliable optimization results.

Looking at manufacturing costs, around half (45%) of the studies that considered manufacturing expenses in steel offshore structures used simplified methods. These include Manufacturing Complexity Factors (MCF) that multiply material costs by a complexity coefficient (Myhr et al., 2014; Ioannou et al., 2020; Baudino Bessone et al., 2022; Ghigo et al., 2020), fixed cost rates per kilogram of structural mass (Lefebvre and Collu, 2012; Karimi et al., 2017), and fixed platform costs or values obtained directly from designers (Castro-Santos et al., 2016; Cas et al., 2016; Martinez and Iglesias, 2021; Martinez and Iglesias, 2022; Lerch et al., 2018). While these simplified methods are easier to apply, they do not capture how design changes affect manufacturing costs. On the other hand, the remaining studies used detailed manufacturing cost models, as summarised in Table 6. Most of these are based on the time-based manufacturing approach originally developed by Jármai and Farkas (Jármai and Farkas, 1999) for welded steel structures (Section 3.3.1). This method calculates costs based on the time required for each activity, including forming, assembly, tacking, welding, cutting, and painting. Several studies adapted this approach for floating steel platforms such as spars (Hegseth et al., 2020; Hegseth et al., 2021) and semi-submersibles (Zhou et al., 2021; Ferri and Marino, 2023). Other detailed approaches include Activity-Based Cost models that consider direct labour and activity costs (Maienza et al., 2020; Castro-Santos et al., 2016; Cas et al., 2016). For fixed steel structures like jackets and monopiles, welding-based methods have been used, estimating manufacturing costs from parameters like weld volume, weld weight, and welding speed (Häfele et al., 2018; Häfele et al., 2019; Häfele and Rolfes, 2016; Burak and Mengshoel, 2021; Yeter et al., 2017).

TABLE 6. Summary of detailed manufacturing cost models for steel offshore structures.

In addition, Figure 12 shows how often each manufacturing operation has been considered in detailed CAPEX cost models. Welding appears in 75% of studies, the highest among all operations. This is expected, given its major role in the manufacturing costs of welded steel structures. Painting and coating follow at 44% and 25%, respectively, which is reasonable given the corrosion protection demands of offshore environments. Forming has been included in 31% of studies, the highest after welding and painting, reflecting the fact that forming is among the main drivers of manufacturing costs. Assembly and cutting appear in 25% and 6% of studies, respectively, and both have received comparatively less attention in detailed manufacturing cost models. Since welding and forming together account for a large share of total manufacturing expenses, keeping them as key considerations in future CAPEX cost models remains important.

On the other hand, transportation and installation costs are often treated together in the literature, with many studies combining them into a single cost term rather than separating them explicitly. As shown in Figure 13, vessel and equipment-based models are the most common approach, used in about 60% of studies. These models rely on various operational parameters to estimate costs, as illustrated in Figure 14. Vessel day rate and time/duration appear in almost all vessel-based models, as these are the primary drivers of transportation and installation of steel offshore structures. Distance and vessel speed are also frequently included since they directly affect travel time and fuel consumption. Other parameters, such as mooring and anchoring operations, weather conditions, workability windows, mobilization costs, crane and equipment rates, and labour and personnel costs, are also considered in many of these models. While vessel and equipment-based models require more input data and are more complex to implement, they provide more accurate estimates since they capture the actual operations involved in transporting and installing steel offshore structures.

Besides, mass and component-based approaches account for around 20% of studies, offering a simpler alternative to vessel-based models. These methods estimate costs based on structural mass or number of components, such as applying a fixed cost rate per kilogram of steel or a fixed cost per foundation leg. Capacity-based models, used in about 10% of studies, take an even more simplified approach by relating costs directly to turbine or wind farm capacity in MW. The remaining 10% of studies mention transportation and installation costs as part of their CAPEX analysis but do not provide detailed formulations or explicit cost models. Table 7 provides a summary of transportation and installation cost models for steel offshore structures, along with the parameters considered in each study.

TABLE 7. Summary of transportation and installation cost models for steel offshore structures.

Overall, the approaches used to model the CAPEX of steel offshore structures differ in the detail they capture and the data they require. Table 8 compares these approaches in terms of their required inputs, underlying assumptions, expected accuracy, and computational complexity, grouped by the design stage each one suits best. Mass-based, Manufacturing Complexity Factor, Activity-Based Costing, and capacity-based models use inputs such as structural mass, material cost, surface area, or project capacity, all of which are available early in design. Time-based manufacturing, weld-volume methods, and vessel-day-rate models use inputs such as weld length, weld volume, welding technology, and vessel day rates, which are typically defined later in design. Across these approaches, accuracy and complexity move together: the early-stage methods offer lower accuracy at lower computational cost, while the detailed methods offer higher accuracy at higher computational cost.

TABLE 8. Comparative synthesis of CAPEX cost-modelling approaches.

In addition, the reviewed literature shows that dominant cost drivers and modelling detail also differ between fixed and floating support structures. Table 9 compares these two structure groups directly, covering their cost drivers, common simplifications, and recommended modelling detail. For fixed structures, the main cost drivers are the leg and brace members, along with piles. Installing these fixed structures then adds further cost, mainly through heavy lifts and specialized vessels. For floating structures, the hull, mooring lines, anchors, and internal structures are the main cost drivers. These floating structures are then towed to the site, which is generally a simpler operation.

TABLE 9. Comparative cost characteristics of fixed and floating structures.

Additionally, to evaluate the quality of CAPEX cost models across the reviewed studies on steel offshore structures, a composite Cost Model Completeness and Detail Index (CMCDI) was developed as the product of two indices (Equation 45):

C M C D I i = M D I i × S D I i (45)

The Modelling Detail Index (MDI) measures how accurately the included cost components were modelled (Equation 46):where N D e t a i l e d and N S i m p l i f i e d are the number of cost components modelled in detail and in a simplified manner, respectively. A score of 1.0 indicates that all included components were modelled in detail, while 0 means all were simplified. The Structural Detail Index (SDI) accounts for the differing structural complexity of fixed and floating platforms. For floating platforms, this complexity includes hull-related elements, mooring and anchoring systems, and internal structures. Three component groups were therefore considered. Each group was scored as 1 if reported in the study and 0 if not. The SDI is the average of these three scores (Equation 47):

M D I i = N D e t a i l e d N D e t a i l e d + N S i m p l i f i e d (46)

S D I i f l o a t i n g = H u l l i + M o o r i n g / A n c h o r i n g i + I n t e r n a l S t r u c t u r e s i 3 (47)

A score of 0 means none of the groups were included, and 1 means all three were covered. Fixed structures, by contrast, are considerably simpler overall in form, mainly consisting of tubular steel members. For jackets, the pile is a separate element, and only a few studies included it. Some of these studies covered material cost only (Syarifudin et al., 2025; Zheng et al., 2023; Sandal et al., 2018b; Tran et al., 2022), others also covered manufacturing cost (Häfele et al., 2018; Häfele et al., 2019; Häfele and Rolfes, 2016), and a few covered installation costs as well (Myhr et al., 2014; Kaiser and Snyder, 2012). A jacket study therefore scores 1.0 if pile detail is included, or 0.5 if it is not. In monopiles, the pile is not a distinct component, as the shaft itself serves this function (McWilliam et al., 2021; Ishtiyak and Sarkar, 2025; Yu et al., 2022; Lee et al., 2014; Yeter et al., 2017). Monopiles are therefore considered structurally simpler overall, and their SDI score remains fixed at 1.0., as expressed in Equation 48:where P i l e i = 1 if pile cost or weight was included in study i and 0 otherwise. The CMCDI ranges from 0 to 1, with higher values reflecting broader and more detailed cost coverage. Figure 15 represents all reviewed studies along the number of CAPEX components considered (x-axis) and the CMCDI score (y-axis). Each bubble represents a group of studies sharing the same CAPEX scope and CMCDI score, coloured by structure type. Bubble size reflects the number of studies in that group, and the percentage inside shows its share of all reviewed papers.

S D I i f i x e d = 1.0 , M o n o p i l e S u b s t r u c t u r e i + P i l e i 2 , J a c k e t (48)

As shown in Figure 15, the majority of studies are concentrated in the left region of the chart, where only material cost is considered with a CMCDI of between 0 and 1.0. This reflects a heavy reliance on weight-based cost proxies across the literature, particularly for jacket and semi-submersible structures. For jackets that model material cost only, most reach a CMCDI of 0.5, since they omit pile detail. Only a small share of these jacket studies reaches a full CMCDI of 1.0 by including pile detail. Moving toward the right, the number of studies drops considerably, and CMCDI values become more varied, indicating that studies covering more cost components do not always model them in detail. The top-right corner of the chart, representing the highest CAPEX coverage combined with fully detailed modelling, contains no studies at all, highlighting the most significant gap in the literature. Studies accounting for all four CAPEX components with high CMCDI scores remain rare, representing less than 5% of the reviewed literature, and none of them achieved a fully detailed treatment across all components.

4.2 Limitations and recommendations

This review offers several new insights into how CAPEX costs are currently modelled for steel offshore structures. Across the reviewed literature, cost models consistently trade breadth of coverage for depth of detail, or the reverse. This trade-off appears consistently across structure types, cost components, and modelling methods throughout the review. The points below set out specific recommendations that follow directly from this general pattern observed:

• For floating structures, spar buoys and tension leg platforms remain less studied than semi-submersibles, even as semi-submersibles grow more popular. Extending detailed cost models to these less common concepts, and to newer floating designs, would broaden the field.
• Floating platforms also rarely include internal components, such as stiffeners and bulkheads, in their cost models. Future studies should treat these components as a standard part of the model, not an afterthought.
• For fixed structures, monopiles remain comparatively understudied, while transportation and installation costs receive less attention overall. Applying detailed, vessel-based models instead of simplified capacity-based or mass-based proxies would help close this gap.
• For jacket structures specifically, pile detail is included in only a small share of the reviewed cost calculations. Future studies should model pile and other connection details explicitly, not just the main tubular members.
• Welding is one of the largest cost drivers in manufacturing, yet most detailed models still follow the Jármai and Farkas (1999) method. Other established methods, such as those of Mela et al. (2014), and Pavlovčič et al. (2004), could also be adapted. This would let future studies compare results across several different modelling approaches for offshore structures. Cutting, assembly, and coating should also receive more consistent treatment, to give a clearer picture of manufacturing cost.
• Overall, for both structure types, most reviewed studies still model only material cost, largely leaving out the remaining CAPEX components. Broadening future cost models to cover manufacturing, transportation, and installation together would produce a more accurate estimate of total cost.
• Only a small number of the reviewed studies examined how uncertain cost inputs affect their final estimates. Most of these studies relied on simple sensitivity analysis, while probabilistic modelling and scenario-based assessment appeared far less often. Future studies could also apply robust optimization, building uncertainty directly into the design process itself.

5 Conclusion

This paper reviewed CAPEX cost models for steel support structures used in offshore energy platforms. The complete classification of all 99 reviewed studies is available in the supplementary dataset (Asgari Motlagh et al., 2026). The review covered fixed structures such as jackets and monopiles, as well as floating platforms, including semi-submersibles, spar buoys, and tension leg platforms. The number of publications in this area has grown notably since 2016, which reflects the increasing economic importance of offshore structures, particularly for offshore wind energy. Table 10 summarises the key findings of this review, including the share of studies addressing each CAPEX component, the modelling approaches adopted, and the main gaps identified.

TABLE 10. Summary of CAPEX cost modelling approaches, coverage, and key gaps identified in the reviewed literature.

The most common approach in the literature is to represent cost through structural mass alone, with 66.7% of studies relying on material cost as the sole objective. This is straightforward and useful for early-stage comparisons, but it misses a large portion of actual project costs. Manufacturing expenses, which can reach more than 30% of total production costs in welded steel structures, have been included in 29% of the reviewed studies. Among those, about half relied on simplified methods such as manufacturing complexity factors or fixed cost rates per kilogram of steel. The most widely used detailed approach is the time-based manufacturing method developed by Jármai and Farkas, which has been adapted for both floating and fixed offshore structures. Welding is the most frequently modelled manufacturing operation, followed by forming, which together represent the main drivers of manufacturing costs. Cutting and assembly, however, have received comparatively less attention in detailed cost models.

Transportation and installation costs have been the least studied categories, appearing in only 16% and 20% of studies, respectively, despite together accounting for 10%–16% of total CAPEX. Many studies either combine these two costs into a single term or use simplified capacity-based or mass-based rates, rather than detailed vessel and equipment models. This gap is especially notable for fixed steel structures, which typically require pile driving, heavy lifts, and specialized vessels.

Furthermore, unlike fixed structures, floating steel platforms consist of several structural components beyond the outer hull, including internal elements such as stiffeners and bulkheads that are essential for structural integrity. However, the review shows that the steel hull is included in 96% of CAPEX models for floating platforms, while mooring lines and anchors are considered in 50% and 30% of cases, respectively. Internal structural components such as stiffeners and bulkheads appear in only 12% of cases, always in combination with the hull.

Overall, the field would benefit from cost models that go beyond material weight and capture the full range of CAPEX components. More realistic and complete cost models are needed to support better design decisions, especially as floating offshore wind moves into deeper waters and new platform concepts continue to emerge.

Generative AI statement

The author(s) declared that generative AI was used in the creation of this manuscript. During the preparation of this work the authors used Google Gemini to improve the readability and language of the manuscript. After using this tool, the authors reviewed and edited the content as needed and took full responsibility for the content of the publication.

Any alternative text (alt text) provided alongside figures in this article has been generated by Frontiers with the support of artificial intelligence and reasonable efforts have been made to ensure accuracy, including review by the authors wherever possible. If you identify any issues, please contact us.

Publisher’s note

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.

Footnotes

Edited by: Wen Zhong Shen, Yangzhou University, China

Reviewed by: Depeng Liu, China University of Petroleum, Beijing, China

Kenneth Bisgaard Christensen, SEA Business Academy, Denmark