MBD Implementation 10 DOs and DON’Ts – Don’t Stop at Graphical PMI (Part 1)

In my last blog I underlined the importance of 3D PMI organization and presentation. A great story shared by Tim with Waters Corp. in a SOLIDWORKS MBD forum discussion illustrated this point well:

“One point listed that should NOT be discounted is the capturing of organized, thoughtfully presented 3D views. Just recently, our core research team began using MBD, and one of the first parts completed was not presented very well. The views and dimensions were scattered about, in no logical order, and it just looked bad. So bad, in fact, that the machinist ended up creating his own manual sketch on paper to create something he could work with. Not exactly a ringing endorsement of MBD. Fortunately, I was able to clean up the presentation and show the machinist how the data could/should be presented. He was much happier after that.”

Now, let’s go deeper than what meets the eyes. Let’s examine the true meanings of 3D PMI by looking into two key concepts: Graphical PMI and Semantic PMI, or some may call them human-readable PMI for presentation and Machine-readable PMI for representation.

First of all, let’s explain these two concepts:

  • Graphical PMI: this type of 3D annotations is readable to human eyes, but NOT consumable by machines or software. Why? Because it lacks the representational meaning and relies on human interpretation. For example in Figure 1, we can read the handwritten lines and curves as a diameter dimension of 20 plus and minute 0.05 using our interpretation, but is this readable to an NC machine? Not likely. Similarly, the 3D PMI in Figure 2 is just a group of sketch lines and curves instead of real numbers or symbols, which require some serious guesswork by our eyes and brains. In fact, this interpretation is so tricky that it differentiates human intelligence and machine intelligence, at least today, which is why capcha works: software recognition capability is still far behind human recognition, so capcha can prevent malicious automatic online registrations and approve real human attempts. On the flip side, massive manufacturing activities call for streamlined automations to avoid duplicate effort, reduce human errors, save time, and ensure quality.Graphical PMI relies on human interpretations, which blocks automations.
    Graphical PMI example: Handwritten dimensions and tolerances
    Graphical PMI example: Handwritten dimensions and tolerances

     

    Graphical PMI example: 3D PMI in sketch lines and curves.
    Graphical PMI example: 3D PMI in sketch lines and curves.

     

    Capcha images, easy for human recognition, but hard for software.
    Capcha images, easy for human recognition, but hard for software.

     

  • Semantic PMI: simply put, this kind of PMI is not only human-readable, but also machine-readable. It represents the true meaning of design intention beyond presentation, so can be reused by manufacturing automation, such as Computer-Aided Manufacturing (CAM) and Coordinate Measuring Machines (CMMs). This intelligent information reuse from design all the way down to production and sustainment is the number-one benefit of MBD and a cornerstone of 21st century smart manufacturing.The images below from Rolls-Royce shows another difference between Graphical PMI (Left) and Semantic PMI (Right). The annotation looks identical to our eyes, which means they are at the same presentation level, but the semantic PMI on the right not only says there are 16 holes, but also associates and highlights all the correct instances, which drive CAM to machine all of them as a pattern. On the other hand, the graphical PMI on the left leads to only one hole. We can probably guess it’s really specifying all 16 holes, but it’s extremely difficult for software to tell and associate all the correct features, which consequently curbs automation.
    Comparison between human-readable PMI and machine-readable PMI on a Turbine Disk Model (Source: Technical Data Package for the Digital Enterprise, Kong Ma, Rolls-Royce Corp, 2014) 
    Comparison between human-readable PMI and machine-readable PMI on a Turbine Disk Model (Source: Technical Data Package for the Digital Enterprise, Kong Ma, Rolls-Royce Corp, 2014)

     

    The above is a brief touch on these two important concepts. There are more aspects of machine-readable PMI, such as syntax, structure, and interpretation algorithm to be discussed in the future. Next up is Part 2 of this topic on practical applications: how machine-readable 3D PMI is automating downstream production, which is the real power of MBD in my opinion. This is why my recommendation here is Don’t stop at graphical PMI. To learn more about SOLIDWORKS MBD, please visit its product page. Also welcome to discuss with me at Twitter (@OboeWu) or LinkedIn (OboeWu).

Oboe Wu

Oboe Wu

Product portfolio manager of SOLIDWORKS MBD, passionate about smart manufacturing opportunities, Keen listener to customer challenges, Sharp problem solver with 20 years of experiences in engineering, Sleepless father trying best to take care of a baby daughter.