What does digitalization hold for early-phase engineering risk-reduction in offshore wind?
In this blog post we give you insights in current approaches on how to deal with uncertainties in new offshore wind markets.
Despite decades of challenges, trials, and setbacks, large offshore wind farms have been successfully installed throughout the North, Baltic, and Irish seas. In the recent years, many new geographical markets in the US, APAC and Europe have opened, welcoming the industry with new targets, regulatory frameworks and challenges.
While northern European soils provide good support for foundations with their dense clays and bedrock, soils, the Taiwan strait for example is composed mostly of loose sands. While the North Sea can have regular, strong storms, the wind speeds are significantly less than the comparatively rare hurricanes in the United States or typhoons in Asia. In addition, Earthquakes are a new challenge for offshore wind turbines entirely.
In this blog post we give you insights in current approaches on how to deal with uncertainties in new offshore wind markets. And as we are continuously developing new tools that can better suit our customers’ specific needs, we also invite you to take part in a short survey. The survey is particularly concerned with uncertainties involved in engineering and design challenges faced in the early phase of projects. Tell us what you think.
Dealing with uncertainties in new markets
Despite the challenges and differences in these new offshore wind markets much of the engineering involved is fundamentally the same. So how are we leveraging DNV’s extensive experience to reduce risks in new markets? Specifically, how do we bring our engineering knowledge to the customer and reduce technical design risks coming from early project uncertainties, logistics and design choices? And how are we bringing innovative technologies and digitalization to address this?
DNV currently has three main approaches dealing with these uncertainties.
1. Expert assessment
DNV have experts with a vast range and depth of experience on projects around the world. This has been demonstrated throughout the evolution of the wind industry as DNV takes thought leadership positions based on our strong individual expertise. When faced with new situations DNV experts can assess the methodology applied, evaluate against the latest literature and industry knowledge, and provide a clear judgement. DNV experts are constantly working to address new challenges in the industry with projects such as the ACE Joint Industry Project which has helped to standardize the approach to design for cyclones and earthquakes. These projects help us to continually update our Recommended Practices and Standards which are the basis for our renewables certification services, one of the most accepted ways to reduce risks of wind energy worldwide.
Our “expert assessment” approach has been the ‘gold standard’ solution for many years and remains the core of DNV’s value proposition for our customers. This thorough kind of judgement takes time, and often our customers need a quicker answer to their challenges with only slightly lower fidelity. In these cases, the other two methods on this list may be more fitting.
2. Engineering model tools such as Renewables.Architect
Renewables.Architect is a powerful, yet flexible software framework which aids in Levelized Cost of Energy reduction. The software uses a library of engineering models and numerical algorithms to quickly provide automated component design and comparative costs based on high level site and technology selection parameters. The models run internal optimisation algorithms to automate component sizing based on the most important design constraints and limit states and systems level optimisation, sensitivity analysis and uncertainty propagation algorithms can be used to understand the most important design parameters and the level of risk in the project. The models in Renewables.Architect cover the turbine mechanical, electrical, and structural components, foundations, balance of plant, installation and operations and energy yield and project finances.
Find out more
By building this tool and others like it, DNV knowledge is codified in a software format which can quickly and easily be applied to specific projects. DNV can provide quick, independent views early in a project Lifecyle.
3. Data driven methods
This third solution is the newest approach and has significant potential for the future, which is to apply a data-driven approach. Data can be used, given consent from data owners as well as careful and thoughtful handling, to provide real-world predictions to clients. So how could we use data, in a conscientious manner, to provide early-phase value to clients in new markets and reduce project risk?
An exciting new approach we are exploring is a data-driven method to predict project outcomes based on a set of engineering inputs. The key for any effective prediction is fitting a model between some set of input parameters that are known early in the process with an output parameter that they have a correlation to. Furthermore, if the level of uncertainty in the inputs is known, then the output could also be a distribution, which would quantify the level of certainty.
How do we apply European data to new markets? This is especially challenging, and it is less clear how effective this might be. However, there are novel approaches in machine learning, such as transfer learning which can re-use learning from other models and apply it to new situations. As mentioned in the beginning, although new market conditions are different to European conditions, much of engineering involved is the same. A data driven tool based on a combination of European and new-market conditions could be built which takes advantage of existing historical European data in a way that requires much less new-market data to provide predictions with greater accuracy.
All this together could provide new opportunities to de-risk projects in a more automated way.
DNV is always mindful of our clients’ expectations when data is provided to us and would never use data in a way that the owners do not expressly authorize. When data-driven tools are presented to clients it is after significant processing to maintain anonymity and there is no chance of reverse engineering. But this data presents the possibility to apply new methodologies.
Example use case of data-driven tool
Taking a hypothetical case, if a tool were to be constructed which could make predictions on the US East coast of how input parameters affect relative cost of the project. These inputs could include soil parameters that might be encountered, metocean conditions, cable lengths, distance from the nearest port, costs to charter vessels – a wide range of variables could be included, the model could be as simple or as complex as required. Each of these parameters would have an expected value and some uncertainty. Machine learning could be used in combination with a Monte-Carlo based approach to determine not only an output prediction for relative cost, but to also give a distribution for the output, relating the input uncertainty to output uncertainty. Further, a sensitivity study could be conducted, specific to the site location, for how each input affects the expected output or the output uncertainty and so identify input values that are most critical and quantify the financial value in reducing their uncertainty. Initially this approach could be used to provide confidence in the two other approaches mentioned. This would already deliver value as a ‘sanity check’ and give more confidence. A key advantage of machine learning is that it is often able to identify patterns and relations between variables which might otherwise be missed. Going forward, as this method becomes more developed the possibilities are broad.
Future directions for risk mitigation of offshore wind markets
As explained above, DNV is continuously working to help our customers to reduce risks and is actively working on new market areas. By utilizing our range of experience in new and innovative ways we are developing new tools that can better suit our customers’ specific needs.
Have you been involved in any early-phase engineering decisions and have an interest in using new digital methods to reduce project risk and uncertainty? Do you see any of these approaches working for you and want to know more? We are always excited to discuss this, please contact us.
Tell us what you think!
We also invite you to participate in our survey to further provide the best solutions to reduce uncertainties.
Offshore Wind: What are your most important early-stage design challenges?