‘Modern’ Wave Energy Converters (WECs) have been designed since the 1970s, using a mix of theoretical, numerical and experimental modelling approaches. Each approach, as the word ‘modelling’ suggests, aims to emulate the dominant characteristics of a ‘real’, full-scale WEC.
However, and owing to the very essence of modelling, each approach also carries its own hypothesis and limitations – leading to modelling uncertainty.
While the above may be widely accepted at a conceptual level, few studies have to date dedicated themselves to dissecting the implications of such uncertainties. Additionally, the propagation of such uncertainties and their potential effects in the estimation of key design loads for WECs remains a relatively untouched research topic.
To this, a WEC designer needs to add the effects of e.g. variability of a WEC response due to the inherent variability to the inputs, linked to either environmental conditions and / or machine states.
In short: it is not easy! A WEC design load envelope should, in principle, also address all of the above detailed effects.
The immaturity of the sector and the lack of consensus in design practices – especially in topics such as load variability and uncertainty – is perhaps clear when we realise that the first Design Load Case (DLC) tables for WECs, which provide clear combinations of design situations, environmental conditions and machine states to be addressed in the WEC design process, have only been proposed in the last five years (see e.g. Wave Energy Scotland’s Structural Forces and Stresses Landscaping Study, conducted by Arup and Cruz Atcheson Consulting Engineers).
But why is knowing your loads very well so important?
Noting the mere magnitude and load variability, WEC design is a challenging topic even if there was no uncertainty in the dominant loads. To design a WEC without detailed knowledge of the load uncertainty across a range of critical DLCs is somewhat the equivalent of driving in a very winding road blindfolded – it will make a potentially hard task considerably harder!
An early-stage yet detailed characterisation of the loading envelope affecting a WEC will allow the identification of ‘hot-spots’, flag potential design showstoppers to stimulate early re-design, at a stage where this is relatively inexpensive (in the overall context of the entire design life-cycle costs).
Ultimately, such approach may lead to the early failure of unsuitable components, sub-systems and overall designs at a ‘drawing board’ stage – before ‘design locks’ are applied and a (unfeasible) full-scale WEC is created and / or build.
In this context, failing fast is a bonus!
What can be done about it?
At present there is a data gap for the characterisation of key WEC sub-systems (e.g. power take-off, control system, …) down to component level. Equally, there is a lack of knowledge with regard to robust methodologies to obtain datasets capable of such characterisation – addressing key items such as scale effects, model coupling, hybrid testing procedures (i.e. the use of both numerical and experimental models in a controlled environment), accelerated testing approaches, etc.
But in principle, if such sub-systems can be characterised, at relevant scale, in a controlled yet relevant environment, detailed knowledge of sub-system behaviour can be gathered and later incorporated in more accurate models of a WEC – contributing to a reduction of uncertainty in the estimated WEC response.
The earlier that is achieved in the design process, the more likely the design will be successful.
What are we doing about it? Next generation wave energy testing
Projects like IMPACT aim to tackle some (or all!) of these challenges.
New test rigs
By conceptualising, designing and building test rigs suitable for hybrid testing of key WEC sub-systems, a EU-based facility to derive key datasets will be made available to the WEC community, allowing a thorough characterisation of the loading envelope across a range of DLCs.
New design methods
Additionally, and to fulfil its overall objective of delivering such facility, some of the key activities in IMPACT directly address the detailed characterisation of the loading envelope affecting a WEC – including the main sources of uncertainty in the estimation of relevant load metrics. For example, and in addition to modelling a range of WEC types across multiple DLCs, the analysis conducted by Yavin Four Consultants follows a full design life-cycle philosophy, covering load estimation aspects spanning from the pre-processing to the post-processing stages.
Ongoing research at Yavin Four Consultants is targeting a detailed characterisation of the uncertainty in the estimation of WEC load metrics such as e.g. the 50-year return load associated with key DLCs. Building on previous work conducted in e.g. the RiaSoR 2 project, Yavin Four Consultants have identified variables that may have a dominant contribution to the overall uncertainty in a load estimate.
Variables such as the post-processing method(s) used to derive the load metrics, including e.g. the type of extreme value distribution used, may have particular importance; and at present, there is no clear guidance on how to evaluate the quality of the distribution fit – which may in turn allow WEC designers to apply any type of distribution regardless of its suitability to characterise a particular phenomena.
This may in turn lead to (grossly) misleading estimates, which may set the entire design on a path towards failure, either technically or economically (see example above).
Ultimately, the process to characterise such key load metrics can form part of a new WEC design methodology – that in turn will hopefully contribute to a faster, more accurate WEC design process, less likely to lead to catastrophic failures at large (or full) scale.
Interested? Subscribe to the IMPACT newsletter below and follow us on LinkedIn.