IEEE Future Directions
Related Articles
Over the last two days I got a deep-dive into the world of Digital Twins, courtesy of Mevea yearly seminar. You can browse the material presented from their website starting next week. You can get a mile high view of Mevea use of Digital Twin technology in the video clip below. The seminar was attended by over 200 companies from around the world, showing the high level of interest for this technology.
The big message that came across is that Digital Twins are for real, well past the hyper cycle curve I discussed in yesterday post. They are a mature technology, which does not mean the technology is not evolving, actually, there are some application areas, like the ones I mentioned yesterday, that arre still in research phase.
One application of Digital Twins that was mentioned over and over, through very specific examples, is their use as a training tool.
Siemens presented the use of a simulator for crane operators. In the first image you see an operator training with the simulator.
A lot of emphases was put on the simulation of the environment, including the sky and various position of the sun with different possible reflections as sun-rays hit the container and the ship loading area.
Something I did not know was that Siemens has created a simulation platform that uses directly the software operating the crane. In this way the operator trains on the real software he will be using in the field. If a change is made to this software the change is mirrored in the simulator since it uses the same software.
Siemens is using digital twins throughout the whole lifecycle, it actually starts creating the digital twin “mock-up” in the specification phase. The same CAD that is used to create the model of the machine in the specification phase is also creating that machine digital twin mock up. This is further refined as specs are refined and it keeps growing through the manufacturing phase. The manufactured product is associated with the generic digital twin of that type of machine that is instantiated to the specific product instance. Instances may differ even though they represent the same product since each instance has a specific digital thread that has recorded the various phases of manufacturing (including info on the supply chain, the specific components, what robots/human worker has participated in the manufacturing…) and of course any specific customisation required by the customer. This digital thread will continue to expand (and diverge from other instances of that product) as shadowing will bring in data from the actual use.
Siemens it their talk emphasised particularly the advantage of using digital twins for training. These are complex machines operating in complex environment and training operators through simulators is both cheaper and more effective. As an example, through a simulator it is possible to inject errors that might occur and train the operator both to avoid them and to recover from them. Doing that in real life will lead to safety hazards and increase cost.
Another aspects emphasised by Siemens and other companies like Sandvik, is the high level of stress these machines are subject to, a stress that may reach peaks in a random way. Hence the usefulness of analysing “shadowing” data and activate analyses that may lead to proactive maintenance. This is saving cost and most importantly decreases down time.
