Origin Axis Methodology An Argument for consistency in modeling practices

posted Apr 3, 2014, 6:48 PM by Rick Smith   [ updated Apr 8, 2014, 11:07 AM ]
In a recent discussion thread on Linked In:
Jay Williams of Genesys Systems Integration asks:

"Does anyone use standardized practices for how and when to use certain features of their Inventor modeling package?
i.e. constraining, shrink-wrapping, adaptivity etc.
Our goal is to create stable, easy-to-edit models used for fabrication and layout drawings. The parts lists on these drawings are expected to have descriptions and quantities that are dynamically linked and not manually overwritten. 
We have discovered tools available in the software might do a fast job for it’s designed use but does not always lend itself to editable fabrication drawings without “work-arounds”. Here’s a scenario involving (Inventor) Frame Generator that yielded an unstable, difficult to edit model.
An advanced user created a very complex 3D sketch driven frame. Years later, this frame needed to be edited by a moderate level user. (assume the originator of the model is no longer around) A lot of time was wasted trying to decipher how this frame was constructed. If there had been some guidelines in place on how to use this particular feature or the extent of complexity to which this tool could be used, it would have saved a lot of time. 
We have continuous training and standardization sessions and have definitely improved as a group. Real world is you’re always going to have varying skill levels amongst users through attrition. We are taking the approach of keeping our models as simple as possible through regulating the use of some features in the software. Surely this is not unique to just us. Wanting to hear how others have addressed issues like this.
A little history on how we do business: We have a 30+ user base. We are design-build company without product lines. During the design process models get handed off to other engineers for drawing creation and changes."

While the discussion largely centered around his example of the frame generator and prompted his follow up question:

"Thanks everyone for the input! The frame generator scenario from above was just a case in point where some guidelines for the use of that Inventor feature would have been helpful. I'll ask the question from another angle. Does every one have full reign to create models any way they see fit or are they subject to departmental guidelines or restrictions of any sort? If so, would like to hear what some of the guidelines/policies are."

I believe that what follows is at least in part more focused on an answer to his question and the techniques that may be employed to achieve a solution. See the original discussion thread here.

In 2006 while working for Northrup Grumman and explaining to the team my modeling approach I coined the term for an Origin Axis Methodology and wrote this paper to help explain it in detail. The paper is largely finished but I need to create some better examples and revise it based on the results so It continues to be a work in progress but the opening argument is posted below.  

A Problem:

Those in the field of mechanical design and engineering who are specialized in the use of CAD or Computer Aided Design and Engineering tools such as Autodesk Inventor, SolidWorks, Creo Parametric and others are all familiar with what is probably the first most common phrase through out the industry. "There are a million ways to model a part." 
The problem is that some of them can make the second most common phrase through out the industry "Think of the next guy because there will always be a next guy." And the job of editing the part in question so difficult that it becomes far easier and even necessary for the "the next guy" who has the job of making the inevitable changes to the design as a result of the revision process that occurs in course of product life cycle to start over, wasting valuable time and requiring the associated drawings to be revisited as a result of numerous factors involving part stability or even file association.
This begs the question: Can a company, community, or even the industry as a whole establish a common methodology in which consistency in modeling practices be achieved? 
If so how, what could be used as a common point of reference that all of an organizations designers and engineers could agree on? 
How would consistency in modeling practices be encouraged, enforced or even applied? And this would be just the start of it.

As defined, this is a problem of methodology where in the interpretation of how to go about modeling a part to meet design specifications outlined as a result of factors such as research of commercial parts to be used in the designs, custom form, fit and function requirements, weight, cost of production and more will all play in to the approach an individual might see as best in the process of modeling the part(s), subsequent sub-assemblies and assemblies even the materials and bill of materials may have an impact on how an individual goes about creating the models to be used and later documented in the detailing and drafting processes.
Indeed, modeling methodologies and decisions on how to model something as simple as a box with a hinged lid by six different people will result in six different solutions. 
In fact, even having the same person model something two or three separate times whether given sufficient time to forget how he/she might have done it the first time or not will inevitably result in several different solutions. Sometime as a result of having studied the problem and deciding it would be better this way than that or just as a result of finding other options.
Does this mean that we want to impose on everyone a means where by everyone models a given part the same way time and again? 
No, of course not it is not only impractical but impossible to expect an individual or group of individuals to model or approach a given problem identically. Nor can we expect or want a cad product to impose on the designer stringent rules reducing flexibility and effectiveness of the product. 
However, might it not be possible to employ a set of common ideas, rules, and methods such that an individual or design team might design a part the same way twice or even utilize common ideas to work together to create a consistency in there modeling practices such that if they understand the rules and what to look for, one person in the group can pick up the work of another in the same group or even that of someone outside that group who is familiar with the rules and have a reasonable expectation of knowing how this person likes to model things based on that common set of rules? 
For starters, there are common "features" with in an empty part and assembly file template. Further, there are common associations with in the file and more. 
This article hopes to demonstrate how these commonly overlooked features found in the initial template files of a CAD system can be leveraged to create consistency in modeling practices, employ common techniques and more so that the "next guy" who picks up a model that you worked on several years ago can, if he is paying attention and is good at his job, begin to notice that your designs have commonality in the file(s) where in the next one and the next after that have a consistent modeling methodology and have as a result made it easier and quicker to make the required edits to your work and fix the drawing package efficiently without making a mess of it and/or causing more pain than is already the case after revision G, H, J... etc. thus taking more of an engineering approach to the modeling process, after all isn't that the objective to "engineer" a solution? Create designs that are easily controlled, edited and detailed such that the model is stable, has few if any failures in the course of the product life cycle and the design changes made to it are fast, efficient, stable and update the drawings and documentation quickly and efficiently?  

As discussed previously, methodologies are inherently subjective to the common practices of the individual and as a result are quite likely never the same twice let alone between the work of two or more designers. 
A possible solution exists in the common practices employed by most if not all designers, even those who may employ part if not all of the methods described herein are still "feature based" that is a "feature" of a part is defined by the sketch and then extruded, revolved or used in surfacing in some fashion or other followed by another feature and another each built upon and dependent upon the previous feature.
However, common practices are often employed by the designer just because he/she found that this technique worked for them and is often used such as slots and hole patterns that use a line defined by the previous feature for direction. 
The immediate problem for this particular feature as defined is what if the previous feature containing the line on which the direction of the pattern changes or fails due to other unrelated geometry or any number of design changes especially in that of sheet metal applications and more.
Feature based methodology has been well established by the industry and is in fact the only way to date of creating cad models as we understand them today even with the advent of "direct modeling" in Autodesk Inventor or "synchronous modeling" in Solid Edge however, this kind of methodology has numerous weaknesses starting with the fact that it is completely open to how this person goes about modeling practices as opposed to that person. 
A completely subjective process which holds no consist processes even when the designer might rebuild a part or assembly which they have done before. 
So what do we do?
Let's start by establishing some ground rules and definitions.

Continued...
Contact Rick Smith @ rick.smith@deltades.com for the full White Paper on OAM.

Copyright October 2006


Comments