tl;dr Now we’ve got the data from the laser, the obvious thing to do is to print the steam engine or parts of it at a scale of up to 1:1
Taking the laser data as a starting point the next part of the journey was for Digital Surveys to produce an engineering model of Winifred.
Above is the model they produced. The quality of the model is exceptional and will be of great use in the reassembly of Winifred during restoration.
The ability to remove swathes of the model to see clearly certain details will also prove useful for those restoring the locomotive. We are now working at how we pull some of the more historical detailed data into the model, such as the damage to the dome.
However clearly nowadays one of the things that you want to do with geometrical data the moment you have it is to print it.
An order was made to Sculpteo for a copy of Winifred at 1:25 scale and there was then what felt like a very long week waiting for it to arrive.
What emerged from the bubble wrap was something that felt very different to a model. It feels far more like data made physical, which is of course what it is.
Even printed in the cheapest white plastic and with all of the inherent faults of the medium and process it is an exquisite model and so much more. It is an artefact of digital reproduction and the industrial revolution; both the old one powered by humans and the new one which has more computers and robots.
Even though it lost a few of the more fragile parts in the process and in transit and it is full of the striation of its manufacture if unequivocally is Winifred with the bent brake handle and hand bent pipework.
Obviously we took the model to Wales on the latest trip to compare it to the real thing.
The other thing we took to Wales was a part of the model at 1:1 scale which the lovely people at MakieLab had very kindly printed out for us on their MakerBot. We did this in part to see how the latch compared to the real thing to test the dimensionality of the model and we did it in part just because we could.
What it showed us clearly was that we were, at the macroscopic level, dimensionally accurate. What it also showed was that if we did want to print out individual items from the locomotive at 1:1 scale we should probably do it with far closer reference to the mesh data from the laser scan, since that provides the more complex geometry which has occurred through wear and use to an item such at the latch, the model of which is held by Julian Birley (Winifred’s keeper) in the photo above.
The printed latch would function perfectly, but if we were creating an exhibit or a real part for the loco based on the data, clearly we’d want to make it match the scanned data far closer.
Many thanks for this part of the project go to Peter Setterfield from Digital Surveys who has worked wonders with the 3D modelling from the data, to Alice Taylor, Ed Sludden and Sulka Haro from MakieLab and to Bre Pettis from MakerBot for being so overenthusiastic about what we were up to when I told him at the 3D Printshow in London (always useful when you are embarking on seemingly insane side projects).
There are more photos of the 3D scans, the 3D model and the prints here.