Why is my intelligence not enough?

Discuss all Oil & Gas Industry related issues here.
User avatar
Many
I have made <0 posts
I have made <0 posts
Posts: 5
Joined: Fri May 29, 2015 11:54 am
9
Full Name: Manyi
Company Details: Matyas
Company Position Title: Freelancer
Country: Hungary
Linkedin Profile: Yes
Location: Hungary

Re: Why is my intelligence not enough?

Post by Many »

Phill I agree with you on the point of meta-data attached to geometry. If a peace of geometry has an ID it should be up to us to populate the database if we need one. For something as simple as that it is a shame we need to put up with the compatibility issues that come with the use of big CAD brands.
Data rich 3D model is the bridge between laser scanning and clients...
My LinkedIn profile and portfolio: https://www.linkedin.com/profile/view?id=426432432
PAJohnston
I have made <0 posts
I have made <0 posts
Posts: 2
Joined: Mon Feb 20, 2023 5:52 pm
1
Full Name: PAJohnston
Company Details: Kenosha EPC
Company Position Title: Piping Department Manager
Country: United States
Linkedin Profile: Yes
Been thanked: 1 time

Re: Why is my intelligence not enough?

Post by PAJohnston »

Modeling in cyclone really has limited to zero value for plant "BIM" design applications (depending on the facility type and end goal of course). These plant design programs have gotten better at modeling to a cloud while the point clouds have also gotten exponentially cleaner and more dense (illustrative with correct RGB colors). The rub that someone who is model-converting directly from the point cloud is missing is that the real world is NEVER "plumb-straight-or-true", whereas ALL usable "smart" plant modeling software that has all the database relationships and detailed materials embedded really WANTs everything in their models to be Plumb Straight & True (orthographic). Plant CAD objects should be placed in "idealized" locations and orientations, if not you can cripple these large facility models ability to compute with precise real world positions rather than accurate real world positions, for your end purpose. Brownfield plant design (in my personal opinion) should just use the point cloud as the model, there should be very little (just tie points maybe) that is actually needed to be converted/modeled from the cloud into 3d solids.

In the world of digital twin "smart plants", you are then potentially hanging on even more data than the normal BIM model would embed, so try to have a model optimized for computing (minimal mathematical rounding errors due to organic placements, misalignments, slopes/sags, just ruins everyone's day in the Plant Design world)
Last edited by PAJohnston on Fri May 26, 2023 1:04 am, edited 1 time in total.
Graham
V.I.P Member
V.I.P Member
Posts: 183
Joined: Mon Apr 27, 2009 3:38 pm
15
Full Name: GrahamM
Company Details: AVEVA
Company Position Title: Software Engineer
Has thanked: 1 time
Been thanked: 6 times

Re: Why is my intelligence not enough?

Post by Graham »

The previous posting covered a lot of the issues but I will go on as I have develop piping applications at some point in my career.

Intelligent models are as about as varied a point cloud survey registrations, they have the uses. For some a mesh is good enough for other more is required. The details comes as users move towards fabrications as there is where the detail comes in, (hint intelligence)

At one extreme you have primitive based models - using cylinders, cones, torus, extrusions. Then there is grouping of elements as pipes, elbows, valves. Then topological connection between elements so a pipe starts with a flange, then elbow, flange, valve, flange, flange (no need to model the straights as they are implied if the items are in line). Then connectivity model this insures matching components can be connected no screw fist on weld neck flanges.

If you are still awake there are bolting applications to ensure the correct bolting is used.

Transferring data from one system is not a matter of copy and paste of elements, but those nicely modelled valves you did well there are specification items, so need to be replaced with referenced objects.

Not only does everyone like there models to run north, south or east, west there really like there measurements to be to in 50mm increments. As this at the end of the day they are going to produce fabrication drawing by the thousand and real world iso are a horror story.

This is not to say modelling is not required, just everyone is on a journey a lot of the time a sketch model is adequate to discuss, to model new work use the point cloud.

When it comes to BIM you will see AI being used to auto classify the point cloud based on element type, then look to segment the data and identify the segmented items and tag the point cloud. This way users can BIM users can view the data as a bubble click on the item which will be tagged and other data on the items can be drilled down to.

I don't mean to jump down anyone's throat on this just to shine a light in those areas you have not got too.
User avatar
smacl
Global Moderator
Global Moderator
Posts: 1448
Joined: Tue Jan 25, 2011 5:12 pm
13
Full Name: Shane MacLaughlin
Company Details: Atlas Computers Ltd
Company Position Title: Managing Director
Country: Ireland
Linkedin Profile: Yes
Location: Ireland
Has thanked: 643 times
Been thanked: 686 times
Contact:

Re: Why is my intelligence not enough?

Post by smacl »

Oatfedgoat wrote: Sat Jul 26, 2014 10:17 am
Phill wrote:Bill
Matt & Matt, right with you here. You only need to export a circular sweep from Revit to IFC and bring it straight back in again to see the shortcommings of IFC.

Cheers

Phill
My understanding from the people that I speak to is that it is not the IFC schema itself that is flawed, it is the import/export implementation of it by the software vendors.
I know that Autodesk are putting a huge amount of effort into IFC currently. I am sure that many others are too.
I believe that currently ArchiCAD is currently very good at handling IFC compared to other vendors.
The cynic in me thinks that it is not in Autodesk's interest for strong seamless support for open data standards such as IFC. There is a big customer lock-in factor associated with using closed proprietary formats.
Shane MacLaughlin
Atlas Computers Ltd
www.atlascomputers.ie

SCC Point Cloud module
User avatar
gordonired
V.I.P Member
V.I.P Member
Posts: 153
Joined: Thu Mar 09, 2017 4:27 pm
7
Full Name: Gordon
Company Details: Roska DBO
Company Position Title: GIS Coordinator
Country: Canada
Linkedin Profile: Yes
Location: Alberta
Has thanked: 27 times
Been thanked: 38 times

Re: Why is my intelligence not enough?

Post by gordonired »

I can give an opinion on this, regarding the Cadworx system from Hexagon. Best piping design software I have used (the competitor was Autodesk Plant3d). That being said, I don't believe that it has changed very much over the last 15 years. I have gone back to "old" forums and see posts in 2009 that are still very relevant. The backside of the system is the same, it's only the GUI that has been updated, and a few extras thrown in. Each patch note over the last 3 years has had pretty much zero substantial changes.

I would imagine one of the big reasons these don't play nicely, is because the code is so dated, unknown to newer devs that were 5 when it first was written, yada yada yada. They have zero reasons to make it play nicely outside of their ecosystem, as it would likely be a dated dinosaur that would take so much more work to bring up to our standards. Even converting just the plain model, polygons, to IFC you needed to pay for a separate program until recently. Intelligence being passed along is the last of their concerns. My two cents.
User avatar
gord
I have made 30-40 posts
I have made 30-40 posts
Posts: 30
Joined: Tue Dec 22, 2020 9:30 pm
3
Full Name: Gord Anderson
Company Details: tiyuti
Company Position Title: developer
Country: Australia
Linkedin Profile: No
Has thanked: 39 times
Been thanked: 10 times

Re: Why is my intelligence not enough?

Post by gord »

Phill wrote: Wed Jul 23, 2014 2:20 pm
...

show me the difference between a pipe modelled in their package as opposed to one in Cyclone, or one in Autocad with the layer being used for the catalogue name.

...

Most cad/engineering packages can import at the least a DXF… why, because it is simple, readable and anyone with half a programming brain could take a DXF import it to their package in one form or another. Can’t someone do the same for piping, steelwork etc?

...

Phill

You've clearly hit a nerve, and its not just you .. the world really is mad :]

Lets say I'm an independent software dev / firm, who is about to add some code to my [ paid or opensource ] pointcloud modelling tool so it can be used to model pipes/beams/slabs/floorplans directly from the point-cloud, and export in a format that is easy to import and use by hardcore Revit/autocad power users ...

There are some great open formats to consider exporting to : .OBJ, .DXF, .IFC .csv, .json

- IFC has a rich standardised metadata model, but complex to support / implement ..
- DXF is simpler, everywhere, unkillable ..

With DXF, I need to know what conventions to use, so that its as easy as possible to link/paste into CAD tools and model over :

Take piping for example, would the following conventions work ? :

- layer name is catalog : eg "pipe_water_NPS_68_400psi" [ lol : ]
- polylines follow the center of the pipe
- diameter polyline circles at segment changes, for visual reference ?
- text note at flange/valve .. or a big X polyline ?
- polyline curve approximates bend radius .. ?
- name is a uniq id that matches up to some other metadata list I export [ csv / json extra file, with linear meters, etc ] ?
- do I export an actual low-rez quad mesh of the cylinder pipe surface in another layer, for visual purposes ?


I suspect there is a de-facto standard professionals use in the field - a kind of pipe shorthand - and if people like me knew that [ by lurking in these forums :] it would help developers make software that is smoother to use, more interoperable.

A few simple example DXFs - for pipes, floorplans / slabs, roofing, rebar - would go a long way to clarify best practice...
then devs and modellers and surveyors can mimic the same style, and we'd have better interop, less hassle.


Stepping back, I think this community has vast depth of knowledge, industry experience and, dare I say, it power .. so collectively, theres a lot we can do
to improve the quality of our tools [ be it pipe linework conventions, pointcloud archive and streaming formats, quality control best practices, or even working conditions ]

gord
gord anderson
quato.xyz
Post Reply

Return to “Oil & Gas Industry”