Solid Fluid System Solutions  
Home Software About Hardware Firmware
Folder Icon Sheet1_GA


BSI Cork

Outline of the design process

In a normal development cycle for most "clean sheet" engineered products, the typical approach to design would be top down. One might start with some sort of specification, and from it a general outline of the design target would be formed. In turn, the top level design forms the framework for lower level specifications. These then become final components, or outlines for the next lower level of design. By this method a large project can be made managable, by division of labour.

In practice one might already have some existing product, component parts of which might be included in the "clean sheet" design. Also, prior knowledge, might lead to the conclusion that a particular area of the design effort has inherent uncertainty, and requires specific and special attention. In these cases the bottom up design approach would typically be used.

The idea of pre-existing components and areas of uncertainty are very similar, because of their necassary use of the bottom up approach. It's effect on the project managment process can be detrimental. Bottom up design necessitates integration effort. Whilst uncertainty can lead to large cost and delay, it may not. Perhaps more importantly, existing components do not necassarily reduce cost and delay. Irrespective of the complexities, the ability of the bottom up design approach to deliver the overall project goal, must not be underestimated. It is the key mechanism by which progress is made.

Visualising the design process

Typically when Richard Trevithick invented the first steam railway locomotive, he would likely have been working with a bottom up design approach. This was likely not because he already had parts of a steam locomotive, but because the whole of what he was trying to do, had never been done before.

For him key components like boilers, cylinders and valve gear, may have existed, but they didn't exist with the form, name, or context which we can today place upon them. Trevithick, then, would have brought these smokey, foggy concepts together and made them work in sympathy with each other. The result culminated in a steam railway locomotive. Yay!

Stroudley, by contrast, knew he needed a steam locomotive, with given functional components, but elements of it would have been individually developed to take the final form.

The key difference between these approaches, is that Stroudley, with his top down approach, had the opportunity to merely specify the function and relationship of the components. He could leave the details to others if he chose. Trevithick was not in this position. He had to devise the whole locomotive becase he didn't know how the components would relate beforehand. Once he'd finished he was at the top of his design, and able to declare as such.

Relating the design process to Gladstone

Drawing up Gladstone, is somewhat different from either of these approaches. It already exists, but it's not a simple case of measuring the real thing and dividing all the dimensions down to suit the 5" gauge secondary requirement. Of course, the primary requirement is a visual likeness with the real locomotive. The other requirements, then, are what make the process slightly different.

In this case, the "other" requirements are; That the locomotive should be functional, having the ability to haul a useful load. It should still be fired by coal. Still boil water. And still use steam to haul the load. All of these things are really an extension of the primary, visual likeness requirement. Essentially they extend that requirement into a functional likeness.

The challenging requirement, is that it must be possible to manufacture the design. Clearly if the locomotive cannot be made, there is no point in drawing it. In practice, this requirement is the one which challenges the simple copy and scale approach, even if it were possible in and of itself. Many components such as valves, bearings and mechanisims simply cannot be directly scaled. Even if it were possible to manufacture them, they would not necassarily function in the way that they would need to. Some aspects of the design, like for example the fire and water, exhibit natural phenomena, and simply do not scale. In miniature they are simply a part of what happens on a larger scale.

These constraints make the design of a working model locomotive unusual in the context of design as a whole. The entire locomotive is essentially a reused component since an attempt is being made to produce a scaled version of the existing locomotive. On the other hand, it's a clean sheet because it doesn't already exist in scale. To complete the confusion, no individual part can be depended upon to function at it's new scale.

If pushed for a descriptive name, I tend to think of this process as "parallel design". By nature one is trying to work through the whole of a locomotive design, but redesigning each individual component from top to bottom. Each component must be smaller, but functional, and where possible look the same as the original.

Design approach

The approach I will adopt, is one that takes the most possible account of top down design, but continuously considers the practicality of each and every component. Unlike a typical "clean sheet" design research must be rigorous, in the extreme.

Unfortunately I don't have an original drawing to work to, but if it were possible to measure each and every component I would. Normally "clean sheet" design would only need research that identifies the functional principles, and the rest can be made up anew. With a model it's not possible to work that way, or the visual similarity requirement fails.

To satisfy the desire for a top down approach, the process starts at the general arrangement drawing. Once the general arrangement is complete, it will be possible to break out into subcomponents like the boiler, cylinders and valve gear. In each of these lower level groups it will be possible to have a smaller degree of concern for visual similarity, since for example, it is not easy to see between the frames when the model is complete. In equal measure it's more important that what's between the frames works, than looks right.

Drawings will be placed in folders, with one folder for each drawing sheet, on the left. Each session that I put in on the drawings will produce a new page with a date and a version number, detailing what's been done, with a downloadable PDF. The drawing displayed directly on the web page is limited mainly by the PNG imaging standard used to present the images. This is why the PDF version is available, by clicking the drawing. Assuming you have acrobat the PDF should load directly in your browser. It gives more opportunity for interaction, and to zoom and look at parts of the drawing. Sadly the PDF generator that I use is not so good at producing curves, and whilst curves always line up on the original CAD drawings, in PDF fromat they may not. In fact, this task is obviously enough of a problem, that some small curves actally appear as rectangles at extreme zoom levels. The PDF files do offer a reasonable likeness to the original drawing, and they're certainly better than a standard pixel based image.

If you need to get acrobat, then you can do so here.

The drawings are scaled in the CAD system, to make dimensioning easier. Drawings published on the internet will not have dimensions until I decide I am ready to release them, and then if I decide to release them for free.

Copyright © Solid Fluid 2007-2015
Last modified: SolFlu  Mon, 27 Jul 2009 05:11:50 GMT