
General Electric, as already mentioned, has been working for several years applying DTs to the energy production, specifically using them to monitor and control wind turbines. Wind farms are costly and complex systems where efficiency can be increased by fine tweaking of the blades angle and this in turns alters the flow of the air (wind blowing across the wind farm). Hence the fine tuning has to take into account the impact on other wind mills to achieve not a local best but a global optimisation. Also, monitoring is important to enable proactive maintenance, rather than having to resort to recovery maintenance. GE equipped wind farms all over the world have digital twins for each single wind mill, digital twins mimicking the processes and a digital twin for the whole farm. These digital twins are “hosted” on Amazon AWS Cloud providing both a local presence and a centralised hub (in the cyberspace there are no distances).
Digital Twins are talking with one another both among the ones mirroring equipment in a specific wind farm and across wind farms. Machine learning is used to create knowledge and to fine tune processes and operation/maintenance decision as part of the GE Assets Performance Management Software -APM-.
Interestingly, the creation of digital twins to mirror local conditions, processes and equipment (in addition to the ones provided by GE that is already delivered with its associated DT), can be done using a Digital Twin library provided by GE that has reduced the time to create a customised DT by 75%.
Operation data from the wind farms are reporting a 40% decrease in reactive maintenance, thanks to the use of Digital Twins.
A further interesting feature of GE DTs is that they can be used as knowledge repository. When a staff turnover occur the DTs can be used for training the new staff and they can also be used to let the new staff get in touch with experienced ones located in other parts of the world, a very smart use of DTs showing the convergence of product-service-knowledge.
This extended use of the Digital Twin has some aspects that would place it at stage 5.