2023 - Adaptivity Issues

[will.astill]

I'm trying to understand why I'm having so much difficulty using Adaptivity so I thought I'd make a very simple part and film my workflow to see what I'm doing wrong.

I think there is some behaviour that I would consider "buggy" and other behaviour that seems weird to me.

I can do this stuff in Solidworks really easily and it seems crazy how unreliable it is in Inventor.

I created a simple back plate, with a top and a bottom using two very slightly different workflows.

• Workflow 1 When creating the "Top" part I didn't click the "Constrain sketch plane to selected face or plane" option (01:30).
• Workflow 2 When creating the "Bottom" part I did click the "Constrain sketch plane to selected face or plane" option (02:38).

The "buggy" behaviour seems to be that Inventor creates either one "Flush" mate (3:42) using workflow 1 and two conflicting "Flush" mates (3:38) using workflow 2.

I had a quick go at manually sorting the issue by deleting one of the mates on the bottom part but just ended up breaking everything. I need to dig into this a bit more.

The "weird" behaviour is in the way that sketch constraints seem to be applied.
I made all the parts using what I'd consider to be good practice. In the "Top" part, I even manually added all the constraints so that I knew exactly what was there (1:55 to 2:25).

However, when I then edit the size of the "back" part (3:50 to 4:00), everything fails and Inventor complains that cross part associations have failed.
Inspecting the "top" part sketch (5:38) reveals that this is because Inventor has applied the most mental constraints possible for the projected geometry (see image below).

Can someone tell me what I'm doing wrong please? Or is this a big in the way Inventor behaves that needs fixing?

[tobias]

After modifying the main sketch, I would do an update

(Quick Acces toolbar, top of your screen)

If it doesn't move I would do a rebuild all (under the manage tab)

Nevertheless, I would avoid adaptivity.

[Frederick_Law]

Don't need adaptivity to make parts that update when parameter changes.

Master Sketch/Skeleton will update every parts and assemblies.

What are you trying to do with adaptivity?

[johnsonshiue]

Hi Will,

If possible, please share the Inventor files with me johnson.shiue@autodesk.com. I suspect there could be corrupted or conflicting sketch constraints.

Many thanks!

[will.astill]

Hi there,

I seek only knowledge...

I occasionally use adaptivity (for example in pipe runs or when making concept designs). It never works properly and so I'm trying to understand its nuances so I can use it more effectively when I need to.

The box back and top exercise is just a simple model that allows me to try different techniques and observe the effects.

[will.astill]

Hi Johnson.

It's ok. The model is an experiment and the behaviour is repeatable rather than corruption.

If you take a look at my video and previous posts you'll see that the workflow used to create the adaptive sketch makes a difference to the constraints applied automatically by the software.

I'd say this isn't a good implementation in the routine that the software uses to creates a part origin but at least it's repeatable and therefore predictable for users.

[Frederick_Law]

---

@will.astill wrote:

Hi there,

I seek only knowledge...

I occasionally use adaptivity (for example in pipe runs or when making concept designs). It never works properly and so I'm trying to understand its nuances so I can use it more effectively when I need to.

The box back and top exercise is just a simple model that allows me to try different techniques and observe the effects.

---

Most users use adaptive as reference.  That's not what it was design for.

The tutorial for adaptive was to solve part dimension by moving the part (a linkage) around.

Once the solution is found, adaptive is turned off.

With a catchy name, Adaptive attract new users and every part is adaptive.  Believing it'll solve all the problems.

Think about it, with 100 unknown dimensions to solve, how many possible solutions will it be?

Would be nice to find new way to use adaptive to solve new problems.

Maybe add some ML.

[will.astill]

Hi - I just noticed that somehow I managed to cross post this.

The other thread is here: https://forums.autodesk.com/t5/inventor-forum/adaptive/m-p/12338715#M912475

[will.astill]

This is a pretty typical workflow for me using adaptivity.
Is there a better way to achieve this that I'm not aware of?

1. Design a complex assembly with holes for pipework and cable routing (usually using assembly-level features).
2. Create an adaptive part or assembly to create a pipework or conduit run that uses the hole centres from the assembly features for reference.
3. Continue working and altering the top level assembly file as the design evolves, allowing the adaptive parts to update and re-route the cables and pipes automatically.

I know that there is a pipework module but I always thought that this just effectively automated the process I described above.

[Frederick_Law]

I don't think you need adaptive to do that.

Frame Gen can update complete frame without adaptive.

I'm building cabinet with different size: 3 width, 5 depth, 7 height in ModelState without adaptive.

I had built conveyor with over 20 configs with different size without adaptive.

In this model, width, length, height can change.

Front and rear section length is adjustable.

Front and rear section height is adjustable.

Center section length is adjustable.

All driven by a Master Sketch.

Depend on how big your project is, you might need multiple Master in different area.

[Frederick_Law]

Do you have a simplified assembly you can share?

[will.astill]

Unfortunately I can't share a typical model from my stock because I don't own the IP.

I design railway wagons and modifications so my reference parts are either supplied by a customer or modelled from existing drawings under an NDA.

It'll take me a while, but I'll design my own underframe and bogie assembly to share when I have some free time as it would be good to understand how you do the master sketch and whether it could work with what we do. I'll share that here and we can carry on the discussion then.

In the mean-time can you recommend any tutorials using the master sketch approach you mentioned?

To be continued...

[Frederick_Law]

Imported model could be problem.  If they got update often.

https://www.cad-notes.com/master-skeleton/

https://www.youtube.com/watch?v=1dkTw89bj00

https://www.youtube.com/watch?v=JaiVSNIDMSM

[Frederick_Law]

@CGBenner might have better advice on T&P

[CGBenner]

@will.astill @Frederick_Law 

Hi.  I have never used Tube & Pipe in a skeletal modeling environment, but THIS CLASS from Autodesk University 2016 covers the topic.  Maybe there are some good pointers in here.

[SBix26]

@will.astill

I think that you are using adaptivity differently than it was intended.  I rarely use it, but when I do, it is to connect features from one part to another part in the context of the assembly, not to connect to assembly features.

Can you add the clearance holes as part features rather than assembly features?  This actually seems like a good application for skeletal modeling, where a "skeleton" part consisting only of sketches is used to control hole placement in individual parts and also to control tubing or conduit runs.  This technique is far more robust than adaptivity.

---

Sam B

Inventor Pro 2024.1.1 | Windows 10 Home 22H2

[johnsonshiue]

Hi! In general, to avoid confusing behaviors, adaptive sketch should be as free as possible. It should be driven by the adaptive driver. If the adaptive sketch have local constraints, it will be like a car being driven by two drivers. The solving behavior will be unexpected. Also, it will require the solve to be sequential. But Inventor the sketch solve and assembly constraint solve are bi-directional respectively. The solvers dictate what to solve first within its own context.

Many thanks!

[will.astill]

Thanks all for the advice and pointers.

I'll have a good read and practice and post back with more questions probably.