The concept of digital twins is growing in a number of industries. In offshore wind, it’s claimed, a digital twin could keep a windfarm operator informed about the performance of an asset and potential technical issues without an engineer having to access it
Much of the focus on cost reduction in the offshore wind industry has been on developing larger, more powerful turbines, but as more offshore windfarms enter service attention has been brought to bear on operations and maintenance (O&M) and how significant O&M cost reduction really is.
One of the main cost drivers in offshore wind O&M is that most of the assets are far offshore. A windfarm consists of large numbers of discrete assets, each of which needs to be inspected.
But it takes time to transport personnel to a windfarm and provide them with a means of access to turbines. Weather conditions can prevent them from accessing a turbine or turbines at all. How much less costly might it be if a virtual model or ‘digital twin’ of assets offshore could be used to provide data about turbines’ performance and provide advance warning about potential issues?
Using digital twins, it is suggested, operators could be provided with up-to-the-minute data on how assets are performing rather than having to schedule regular shutdowns for inspection, the outcomes of which may result in unwelcome detection of faults that have developed unnoticed.
Just this kind of cost reduction potential is being addressed in a number of projects, among them the application of digital twin technology by Octue in the UK, a start-up founded in 2013 that is applying ‘intelligent digital twins’ to the challenges of windfarm O&M and working with the Offshore Renewable Energy Catapult to do so.
Speaking at Global Offshore Wind 2018 in Manchester in June, Octue technical director Tom Clark explained that the main aim of the company’s intelligent twin for offshore wind turbines is to monitor and predict their behaviour, integrating many data sources and connecting many analyses together.
Key to the digital twin concept it is working on for offshore wind is what Dr Clark described as ‘frictionless learning’ and automated learning techniques that would allow learning to be translated seamlessly across an asset.
“The problem is,” he told delegates at the event, “collaborating means sharing sensitive data. We need to find a way to keep data controlled and ‘siloed.’ The good news is that, with the right system architecture, we can collaborate securely on data analysis and processing.
“Essentially, we’re looking at three key problems in digitalising a fleet of assets: the difficulty of collaborating between teams; how to share knowledge and training between assets; and how to keep sensitive data secure while enabling collaboration.”
If a solution to these challenges can be found, he suggested, data from different assets could be shared without running into commercial issues.
Dr Clark described using ‘component twins’ in which artificial intelligence (AI), simulation or mathematical modelling are used to predict and improve asset performance. Over time, this would enable better and better predictions of performance.
“Digital twins can be linked together and interchanged,” he explained, “and models can be improved, refined and changed. This would enable the transfer of generalised learning across assets and enable collaboration to build an overarching model.”
With Octue’s approach to digital twins, a network of components allowing reuse of component modules and transfer of learning between assets would enable detailed data to be built up that could predict performance, not just monitor and record it. And all of this would take place within a secure collaborative environment.
Something similar is being attempted in the WindTwin project, which is aiming to develop a digital platform that will ‘virtualise’ offshore wind assets as digital twins.
WindTwin is being developed specifically for the wind turbine industry. Like the work Octue is doing, it aims to reduce costs associated with operating and maintaining offshore wind turbines. As with Octue’s work, the project partners also anticipate that digital twins can feed into the design of new products and support the development of more reliable wind turbines.
The project will harness digital twin technology in the form of a high-fidelity, digital software platform combining operational data with virtual system model data. To facilitate this, a sensor network utilising optimised signal processing and condition monitoring algorithms is being applied to live wind turbines to collect operational data which will interface with the digital twin. The output will be collated and processed data providing a description of a wind turbine’s dynamic behaviour and physical state during real-time, real-world operation.
A good example of how WindTwin might work is condition monitoring on a gearbox, which would be applied using sensors installed on the real-world turbine. The data collected would be processed and transferred to the digital twin continuously, resulting in a close to real-world digital twin of the wind turbine showing real-time performance.
The partners behind the WindTwin project, which is funded by Innovate UK, anticipate that virtual models will allow windfarm operators to predict failure and plan maintenance, thus reducing maintenance costs and downtime.
The application of the WindTwin platform will include using data and knowledge-based tools and simulated testing of wind turbines before manufacturing. It will also use continuous predictive and preventive maintenance, condition monitoring of turbines, and power setting scenario analysis and analysis of wear and tear at different power outputs.
The partners in the WindTwin project believe the digital platform could reduce maintenance costs by up to 30% for end users and operators. They anticipate that early detection of defects will increase reliability by 99.5% and will reduce losses due to downtime by 70%.
Like Octue, the partners in WindTwin – ESI UK Limited, Dashboard Limited, Agility3 Modelling and Simulation, The Welding Institute (TWI) and Brunel University – are using a range of techniques to develop digital twins for offshore wind, including advanced sensors, high-performance cloud computing, system fault and degradation modelling, data analytics, and 3D visualisation to present data to windfarm owners and turbine manufacturers.
The data they collect is being integrated into a high-fidelity digital platform using advanced models that ensure a digital twin accurately represents the condition of the real-world asset in real-time. A turbine’s dynamic behaviour, performance and degradation effects are assessed by passing the data through mathematical models that compare it against given parameters, then display these changes or outputs through a virtual interface.
TWI senior project leader condition and structural health monitoring Ángela Angulo said she believed the data provided by the WindTwin digital software platform has the potential to provide the wind turbine industry with many benefits. “It will enable operators to diagnose performance variation in a windfarm right down to the component level, anticipate degradation and failures and enable the industry to implement condition-based maintenance instead of schedule-based strategies,” she said.
Other companies that specialise in this kind of technology are also targeting the offshore wind industry and other industries where large numbers of individual units and components need to be continuously monitored.
Among them is US-based ANSYS, which released the latest version of its software, ANSYS 19.1, in May 2018. The software is designed to enable product developers to accelerate development by rapidly building, validating and deploying simulation-based digital twins.
ANSYS 19.1 includes ANSYS Twin Builder, a product enabling customers to build, validate and deploy simulation-based digital twins. The company anticipates that it will have potential applications in the offshore oil and gas, energy, aerospace and defence industries.
“Traditional preventive maintenance for industrial assets leads to expensive and potentially unnecessary maintenance costs,” said ANSYS, echoing remarks by Dr Clark and the companies in the WindTwin project. “Those costs can be greatly reduced with a digital twin.
“The resulting intelligence and predictive maintenance insights enable engineers to analyse machines in real-world operating conditions and make informed decisions that substantially improve performance – reducing risk, avoiding unplanned downtime and enhancing product development with extremely accurate, individualised feedback about product behaviour.”