I am curious and I know people here might know the answer. Over the weekend the guys building the Pennsy T1 put several detail CAD drawings on their Facebook Page http://www.facebook.com/t1locomotive?fref=ts

My question is this, has anyone in the US in the past 30 years or so ever designed a complete locomotive from the ground up and run simulation testing on the CAD Model? I believe Kloke used drawings from the National park Service to build both of his locomotives, I am not sure what the guys building the Lyon did?

Can others give examples of how they may have used computer modeling on their steam locomotives and did they find any design flaws in doing so? How about 3D printing to take the discussion one step further?

My son, 10stewi, does this all the time. He has a large library of live steam models in Autocad. He is also modeling all the workshop buildings at Rungsted Kyst, Denmark, for our planned facility expansion.

Could be useful, could be eyewash, no way to tell from what is presented. If it's just pretty pictures, it's eyewash. Need material call-outs (steel call-outs w/o heat treatment specs doesn't cut it for me) and tolerances to be truly useful.

What is shown are not drawings, they are simply renderings for presentation purposes... these parts may be rough "pretty pictures", or they may be properly engineered with complete material and treatment specs, tolerances, etc. Without seeing completed engineering drawings (if used, not necessarily required in today's manufacturing environment) or prowling through the actual 3D CAD model's annotation, one cannot tell.

Presumably, the T-1 guys are developing fully engineered CAD parts and testing the design with FEA software to reduce the risk of error. Are the parts completed to that level yet? Only they will know.

I suspect there has been a lot of CAD work used on steam locomotive PARTS... entire locomotives are so rare it's doubtful.

As for 3D printing being used, it's not very useful at all for full size locomotive parts. Perhaps patterns for small castings, otherwise cost and printer size will work against you.

However, a CNC router or mill could make short work of machining casting patterns, working directly from the CAD data, and the finished casting could be machined directly from the CAD data.

Same with CNC flame cut or water jet parts and any parts hogged out of chunks of metal, no matter how large.

Steve Hunter

I have not modeled a complete steam locomotive but we have modeled a 20 hp Stanley engine, a Robert's Marine engine and a number of other unique pieces.

As posted previously without proper fully annotated 2D drawings a 3D model's value will be limited - though not useless. This is a point I stress continually with my students. While the 3D models always captures the attention its the 2D drawings that allow the piece to be fabricated correctly. The 3D model is only half the equation.

The value of the 3D model is to test accuracy, tolerances, function and strength. Once those are established than traditional 2D drawings are created that are linked to the 3Dmodel - in other words if you revise the model than it revises the item depicted in the 2D drawings.

Yes, you can send the 3D model to a shop for CNC fabrication and/or 3D printing but you still need the 2D drawings to check against, determine material quantities and a whole host of other things, tolerance, finish, etc.- unless the person on the receiving end is experienced with 3D CAD it will be difficult for them to determine even basic dimensions. We just did this recently with a set of connection rods we reverse engineered. The shop preforming the work received both the 3D files and a set of 2D drawings.

Best regards,

Terry

You mention "simulating", quite a few years ago I did some 3D renderings of various steam locomotives on the Amiga computer. They were nothing more than pretty pictures, I had no aspirations to have them be actual design documents, just eye candy and a way to learn CAD. I did, however, use some drawings from Live Steam magazine as a basis for my 3D models, and it worked quite well. The drawings ("Lucky 7") were intended for use in building a live steam locomotive. For my purposes, they were close enough.

At one point, the 3D program I was using came out with a motion simulation function. They did the usual "design a car" engine stuff with it, showing moving pistons, so of course I said "Well, hey, why not a steam engine?"

Starting small, I "built" a simple stationary engine with a flywheel, piston and main rod. I told the program to have the piston move back and forth, started the simulation and watched. It worked, just as it should, for exactly 1/4 revolution. Then it hit top dead center and just sat there and quivered slightly. My lovely flywheel did no good, since the program was nowhere near complex enough to consider things like mass and circular momentum. At top dead center, there was no sideways force, and things stopped. The program designers had anticipate this sort of thing, thus the "quivering", but the motion wasn't large enough to get it off of center.

I solved the problem by making the flywheel rotate and having the piston go along for the ride. Looked great, and you couldn't see the difference in the video. I ran out of spare time, so nothing more than the "proof of concept" renderings were ever done.

What you see so far are just 3D renderings for presentation purposes. The initial design work will be done by translating the drawing data (with the materials data, tolerancing, etc. derived from research) into 2D and then 3D files in a good CAD program like SolidWorks, and then import details into a multiphysics environment like COMSOL to develop the physical simulations. Once the 3D model has been refined, it will become possible to do other testing in areas like CFD analysis to optimize circulation or front-end design.

Here is a useful psge from COMSOL that shows how a model can be imported from a 3D CAD package, or a kinetics modeling program like Inventor, into the multiphysics simulation environment.

https://www.comsol.com/cad-import-module


Now, in the meantime, there are two other areas where the T1 people will be doing "3D modeling". One is for 3D renderings like the ones on the Web site so far; these are just that: renderings in a program like Maya. The other is the generation of simulation models for one or more of the 'rail simulator' programs, some of which have very intricate and reasonably 'complete' physics engines.

Let me make it clear to some of the people on here who might get this confused: No one doing engineering for the T1 Trust is confusing the physics-engine model created for a rail simulator with the 3D models that will be used to develop and refine the actual design of the locomotive.

Hi,

3D printing is still known by some as Rapid Prototyping. Rapid Prototyping can be used (as only one of many uses) for assembling new model prototypes (not prototype as in model railroading but prototype as in the first of its kind - I use the term "engineering prototype" myself).

Rapid Prototyping/3D printing is an additive process using CNC. It adds and melds new material to the new creation. This would be good for patterns made on CAD.

CNC machining is a subtractive process. It takes a block of metal and removes material. Some my recall the joke kit from the 1960s and 1970s for you to make your own brass locomotive. You got a block of brass and were told to remove anything that was not part of the model of the prototype.

In my attempt to support myself after loosing an engineering job, I use 2D cad for all three views. I found eMachineshop on line and used their proprietary software to make 3D models and have brass parts machined. It is very useful in places and is expensive . If you order 300 parts, then the price per part becomes very reasonable.

I also use Shapeways for 3D printing. It has its niche also. It is usually less expensive than eMachine shop. They can also 3D print some metals like Brass Silver, and Gold. The eMachineshp CAD has an export function to an STL file that Shapeways can use. It does not work all the time but considering the economics, I can not purchase a 3D CAD system yet. FREECAD does import the STL file but I've not been able to merge STL objects into a final part. The only help is a video by FreeCAD and with a dial-up link, I can not use it. Documentation for FreeCAD is not good enough FOR ME.

In HO scale, I've made PRR electric models in brass (using eMachineshop) and part of an HOn30 steam locomotive. In HOn3, I've been using Shapeways for some early D&RG 4-wheel rolling stock and a PRR N scale passenger car.

So the process works with either approach. For full size work as with the T1, other factors come into play.

I offer my experience only as a suggestion to not exclude one method but for you to consider both.

As far as simulations, I've not tried that. For myself and my models, the simulation would be a waste for the simple items I am making. Others would need to weigh in on this.

Food for thought.

Doug vV

It is my understanding from a recent presentation the T1 CMO gave at the PRRT&HS annual meeting that they have taken the orginial blue prints from the Pennsylvania (2D) drawings and converted them into exact 3d models all in Solidworks. The CMO said after the 3d model is complete they take up to 8 hours to measure and check each model against the orginial drawing to make sure it is 100% accurate.

From my limited expirerence with Solidworks, I know you can easily take a 3d model and convert it into working 2D drawings. I know Solidworks also has many simulation plug-ins for their software.

The CMO also said the metallurgical properties are assigned in Solidworks modela so down the road simulation can be conducted.

There is some video clips online of the talk, it looks like cell phone video but very interesting. Some audio is hard to understand.

So it would appear to me that the renderings you see above and on their Facebook page are much more than pretty pictures and are actual working 3d models that can be used for simulation.

I wanted to get back to my original question, has anyone else ever done this extensive CAD modeling for a working steam loco here in USA? Or Canada? It is looking like the answer is no, but I think someone had to do it before this.

I am interested in building historical rolling stock and locomotives Solidworks, just for the objective of the model. Right now, I am working on a 1880s era caboose of composite construction. The point is to just create the model for the same purpose that one might fabricate a model out of plastic or metal. I might explore making rendered print images as one form of expression of the cad model. So all I have to do is acquire the dimensions of the parts, and model to those dimensions.

But as others have said, to actually use the cad model for building an operating locomotive, you would need to develop all of the fits and tolerances. You would also have to develop assembly procedures, notes, and specifications, and choose materials, and finishes.

I am sure that most live steam locomotive builders in various scales would tend to develop all their designs in 3D cad these days. But I don’t know of anyone who has modeled an entire modern steam locomotive as large and complex as the T1. There would be no reason to do it if you were not planning to build one. For me to model the T1 just for the sake of a model, as I am doing with the caboose, would take me a mighty long time to finish. It would probably take years of full time work. And that would require obtaining all of the dimensions by pre-existing drawings. If that is not all available, I suppose some new design work would be needed to develop the proper parts and part features. That would also require a lot of time.

Modeling the T1 for actual 1:1 scale construction, and then doing all of the engineering work with fits, tolerances, and materials would be a staggeringly large project. It would be interesting to know if there is an estimate of man-hours required for this work.

of course these original engines never had these technologies yet they were built.

old guys with heavy glasses leaning over large drawing tables hump backed with slide rules...or less....


well thats an exageration...8-)

I think there is a reason that the trust is starting their CAD modeling with the wheels. On their website they mention that the cast frame may have to be broken up or fabricated from plate. This means that they aren't simply making parts from old drawings, and will have to do some real analysis before real money is spent. Since they are using Solidworks (there was an announcement on their Facebook page this morning that they had six licenses donated by Dassault) it would make a lot of sense to start with what is already fixed in location and design around that. As someone who uses Solidworks every single day (I have it open right now on my other screen), I can tell you this is by far the easiest way start designing a new plate or multi-section casting in a modern CAD program.

With today's parametric 3D CAD software 3D models are made first. From those, 2D drawings that are linked to the master 3D models are made. 2D drawings are crucial for tolerances, heat treats, finishes, etc, but A LOT of machine shops and metal fabricators want 3D CAD models so they don't have to remodel something in CAM when they put a part on their CNC machines.

I see a huge benefit with modeling the entire locomotive for fit of parts. I'd be very surprised if every single part on a locomotive was actually made to print. Would it shock you if someone assembling the locomotive modified a finished part to fit in an assembly and didn't tell engineering they were doing so? I'd be shocked if this wasn't the case. Additionally, does the Trust know they are working off of the latest set of prints? Do they have all of the engineering change notices for the T1? If they do, do they know they have a complete set? Fits of parts can be checked before metal is cut and money is spent reducing rework and scrap costs.

I'm not sure how much FEA and simulation you should expect to see just because they are modeling the locomotive. FEA is an add on in Solidworks called Simulation. Simulation does static stress, static thermal, static pressure vessel, static buckling, and fatigue. Solidworks Motion is a stripped down version of ADAMS that can be used for some dynamic simulation, but it is BUGGY. Anything of value would need a full blown version of ADAMS. Solidworks Flow is pretty bad, and I don't think they have anything for combustion. Each one of these add-ons are about $10,000.

However, with a number of 3D CAD systems, all the tolerances, material grade and treatment, finish notes, etc. are carried in the 3D model itself and a great many workplaces are putting tremendous pressure on their engineering staff to eliminate most paper drawings.

Canada's National Research Council is one such employer... I can tell you I was very relieved to retire early, as I am extremely uncomfortable with this practice and produced finished drawings to the end, in addition to the 3D models.

True, our Pro Engineer Wildfire (now Creo) software was fully able to go paperless, but I don't like trusting that a toolmaker will notice every single parameter on his screen, where it is easy to highlight important things on a proper drawing.

It all boils down to cost... how much does it cost to produce detailed drawings, and how much could one error cost.

Steve Hunter

Hi,

I recall the indignation by politicians over a hammer costing $500.00. They made the military the "bad guy" for padding their pockets. The truth was the Military standard was to be able track the materials back to the raw materials and the mine or forest it came from.

The $5.00 hammer was a $5.00 hammer with $495 in overhead paperwork for tracking.

Simulations are not bad, but the idea that you have to test everything might need to be balanced by "proven" designs. There were many T1's in operation.

Just an item to think about.

Doug vV

I completely agree with this. A lot of time can be wasted evaluating stuff that is already proven and/or won't be changed. The operational history of all the miles logged of all the T1's and all the locomotives that shared parts and designs will tell you way more than most simulations ever will.