Integrated Factory Modelling - Facts & Next Generation

ROBERT OSTERMANN: Dear ladies and gentlemen, I'd like to welcome you here in the session about integrated factory modeling. And I want to share with you some facts and considerations about future methods, what I call integrated factory modeling next generation. My name is Robert Ostermann. And I am a developer for planning methods and processes concerning the digital factory at Magna Steyr Fahrzeugtechnik Graz.

In this presentation, you will hear about my mission at Magna. I'll preview talk about the challenges we face. And I'll also give you some insights and facts about using Autodesk Toolchain to solve these challenges. Finally, I'll give you an overview about tools and methods we want to implement to improve our workflows in future.

My mission as a developer for methods and processes at Magna is to deliver better workflows with modern technologies and systems, systems that are more collaborative in using CAD models from different planning disciplines and more manageable in terms of collaboration.

This breaks the mold regarding usual CAD thinking, especially in the use of systems that tend to implement everything into one CAD environment, no matter the cost for data conversion and breaks in communication they cause. It also needs a change in communication regarding how CAD models are integrated into holistic factor models.

Because the communication of issues, change requests, requests for information, and so on need to take place directly on the integrating factory model and not separated from all discipline. Henry Ford formulated very appropriately when he said, if you always do what you've always done, you'll always get what you've always got.

Before we start and have a closer look at our workflows, I want to point out that Autodesk provides software with unique capabilities, which are driven by users in the market. But most importantly, software can collaborate and interact. Nevertheless, Autodesk software isn't always as collaborative as we would wish. Because the market doesn't seem to demand it, especially when it comes to collaboration between different planning disciplines.

But compare it to our idea of a perfect factory planning world. It's much better than combining all demands into one monolithic software product. To better understand our focus on developing methods, on this slide on the left side, there is a status quo, where you can see that the different views on-- to our factor models have too many manual processes.

For example, a fixture was designed in 3D and delivered as a neutral file format each step file, and manually simplified to the element with properties copied manually from the 3D file has been used in the 2D layout. So if the position of this element is changed in the layout, for example, according to a point cloud, which means 3D reality data, this change won't be automatically reflected in the 3D factory model.

Also, interfaces to other systems with different purposes were difficult because some housekeeping in the CAD environments had to be done. In particular, managing these gaps would need expensive and sluggish PLM, product lifecycle management systems, which were developed for product CAD models that have huge gaps in handling a whole factory structure. This leads to our view on the right, where the workflow isn't broken by working in one of these views and where one connector to different systems exists.

To establish this previously seen idea of a building and building infrastructure CAD environment, it's easy to find a software tool on the market. But most of them have huge gaps concerning data management, the use and management of point clouds, 3D reality data, and the collaboration with machinery design in particular.

At Magna, we've developed our methods using Autodesk Revit. Because what you can see here in this video is that working in 2D is like working in 3D at the same time. Point clouds for quality checks and integrations, where no CAD model exists, can be referenced as one format in all Autodesk systems. And builds of material are always live and easy to configure right.

The use of data management systems like Autodesk Vault to manage libraries and file dependencies is nearly seamlessly integrated, but still has some gaps due to market demands. There are also very efficient workflows to review and analyze models compared to the point clouds.

In this video, you can see possibilities to visualize models in point clouds, how to measure and markup differences between models and reality. Also, you can see how seamlessly the review on cloud management and CAD environment can work together due to the use of Autodesk vector design utilities. And finally, you can see how easy it is to update the model according to the point cloud.

The use of Revit makes it possible to revise models intelligently in model and plan views, and to integrate reality data point clouds. But there are still more when it comes to integration, and collaboration, and whole factory model, something I'll show you in a few minutes.

To establish the previous idea of production facilities in the CAD environment, it took us some more time to develop all the necessary methods. Many available software tools on the market have huge gaps when it comes to setting up and managing automatic design reductions, and level of detail, and creating parametric assets. Most of them also have huge problems concerning data management, the use of management of point clouds, 3D reality data, and the collaboration with building design in particular.

To be able to reference different file formats from other vendors, but not as a file conversion, and to automatically generate different levels of detail, which fully makes sense for a factory model, we switched from Siemens FactoryCAD to Autodesk Factory Design utilities eight years ago.

We also established-- and this is the first time we have ever shown this-- a workflow to map and inherit properties from the so-called AnyCAD file to the intelligent 2D, 3D factory asset, and make the entire process visible in the Autodesk Vault data management system.

What we are also very proud of is that we can easily create parametric assets in the corresponding user interfaces without recreating geometry, even if it comes from other vendors. We use the capabilities from Autodesk Inventor-- and mainly, it's iLogic and Invent Trigger functionalities-- to create our own and required standard libraries.

Here in this slide is an example of a linear unit from KUKA, provided as a STEP file, and showing some variations of specifications. Here we can get an overview about what style formats can be referenced. So there's no need for conversion. And it can be used in the previous shown workflows for unique and parametric design data.

This slide shows in more detail what it means to automatically create a reduced level of detail representation from reference file formats. If you look closer, you can see that small parts, small camphors, fillets, and holes, as well as internal voids, have been removed. Therefore, the generated 2D and 3D representations are very lean and efficient concerning system requirements.

Ladies and gentlemen, participants of this Autodesk University, I'm extremely pleased to show you for the first time ever, all these dependencies from the original AnyCAD file, referenced at the full level of detail, automatically reduce and convert it to an intelligent 2D, 3D asset with inherited data properties, are fully traceable in Autodesk Vault data management system.

This is one of the ways of breaking the mold I mentioned earlier. This is a complete break with previous MicroStation and AutoCAD thinking, and raises data flow to an amazing new level of consistency. But that's not everything concerning consistency. In this video, you can see the inherited properties of an object in a 3D factory layout.

One of them is a direct link to its full detailed model in the vault data management system, which can be reviewed with all its part properties. Also, you can jump from one of these assets directly into the Vault data management system and analyze from which AnyCAD data file it has been generated from, in this case, a Kittila assembly.

There are also very efficient workflows to drive the position of objects in a complex model with high system demands, which I'd like to share with you. In this video, you can see a point cloud in the so-called Master Layout, where a simple rectangle hosts and drives the position of a 3D element in the model, a fixture in this case.

After reviewing and analyzing the Master Layout compared to the point cloud, this positional frame of the fixture is moved forward to the left, which also moves the 3D element in the detailed 3D model. The implementation of the Autodesk factories and utilities makes it possible to realize our vision.

The workflows aren't broken, whether working from 3D to 2D or vice versa. Everything stays synchronized between these views. Properties are inherited from its origin. And the use, reuse, the change of the cap data is fully traceable, especially in combination of implementing Autodesk's data management system.

Reuse of the factory design utilities makes it possible to revise models intelligently in model and plan views, and to integrate reality data, point clouds. But there are still more when it comes to the integration and collaboration in the whole factory model, something I'll show you in the next couple of minutes.

Here is the core question of this presentation. In an Autodesk software environment, is it possible to collaborate between building and building infrastructure and production facilities while everything is being managed in one data management system? And by the way, a data management system should have at least configurable data categories, data life cycles, data property mappings, data usage, and, where used, functionalities, and of course, data permissions and the role management.

Let's start with the question of whether it is possible to implement a common data structure in the authoring system, the CAD system, extract and re-use this information in a factory model, which consists of all planning disciplines, and map this information to the data management system.

Here in this slide is an example of what this could look like for a building model. It's used library parts and detailed documentation in our data management environment. On top left is a CAD model of a building. On the lower left and marked with number one, you can see the second level out of four, which defines our hierarchical data structure.

Highlighted in red and marked with number three is the fourth data structure level, which shows all flat foundations. On the right, in our data management system, you can see the library parts and the detailed documentation structured by the same data properties.

This slide shows an example of a robot cell. On the top left is a CAD model. On the lower left, highlighted in red and marked with number three, you can see the fourth data structure level, which shows all the fixed handling equipments and fixtures. On the right, in our data management system, you can see the parts of this CAD assembly and the full level of detail of the documentation CAD model, structured with the same data properties.

On this video, you can see this common data structure in action. First, I will select one of the top levels, labeled protection resources. And this will highlight all of them on the canvas. Then I quickly cycle through level two, namely machinery, production facilities, and fixtures, equipment, and production infrastructure. And then I switch over to the other top level called building resources, from where I will select level two, called building structure, followed by level three, vertical building structure.

And finally, I will select level four, the instructional columns, which will highlight all the elements of this category on the canvas. To make methods of model coordination more obvious, I've structured functionalities into four levels. Level one, I call this low-level coordination. It's a manual or PDM/PLM triggered file conversion [INAUDIBLE] JT, STEP, IFC, or DWFx.

Level two, basic coordination is achieved when an automatic cache file with update functionality exists, but there are no constraints between objects in file references and there is no native data management, which means direct file usage and, where used, functionality.

Level three, advanced coordination is achieved when a direct file reference can be used, and the manual or automatic update functionality exists, constraints between objects in the file references can be set up, and native data management is possible. Level four, real-time digital coordination is achieved when a model can be linked live and tracked on mobile devices, and also if a live communication in the model environment is possible.

So if we look at the model coordination with Autodesk Navisworks, it's a mix of several levels, highlighted in red. You could use files generated from level one, also directly implement several file formats, which automatically generate cache files where you can trigger updates. The combination of these converted or directly implemented files can be managed in the Autodesk Vault data management system.

This video shows the use of a converted JT and an IFC file, combined with a native Autodesk Inventor factory file and an Autodesk Revit file in the Navisworks coordination model. You could review, check for clashes, or document model tasks. Also, you can see the complete file structure can be managed in all this data management system. And in the case of a native file reference, where factory assets are used, you can trace back to the original full level of detail where this asset originated.

In the case of using a production facility model in the Autodesk Revit building environment, we have a good level two basic coordination which has some benefits in performance. The current challenge is the constraints between the objects of the production model and the objects of the building model aren't possible yet.

It is more or less common in the AC industry, due to the lack of requests for data management workflows, that the coordination file must be controlled manually in the file checkout and access process. If you have any further questions, feel free to ask me for more details during the Q&A session.

Using a building model in the factory environment for protection models is straightforward and efficient in terms of using the file reference. Setting up constraints between objects is possible. And managing the model is fully integrated in the Autodesk Vault data management system.

To help you better understand, here's a video where you can watch how I create a new duct and an additional air supply in the context of a robot cell in Revit. What's really amazing is to see the performance, the visual quality, and the level of detail of how the production facility model can be used. I'd like to point out that this is the Inventor factory model. And it is referenced as a coordination model. So no need for manual data conversion.

In the second video, you can see how I reference a defined model state of the building infrastructure directly into Inventor factory, and how it can be updated after the building infrastructure model is enhanced by a new duct and an additional air supply. Again, it's amazing to see the performance, the visual quality, and the level of detail the models can be used for different planning purposes.

To summarize, what you've seen is that it's possible to seamlessly work together within the different specialized tool sets from Autodesk without any need for additional file conversion. The models can be coordinated from different points of view. For example, from a building point of view, a production point of view, as well as a factory coordinative view.

It's always very important in factory planning processes to be able to combine these models and model coordination files with point clouds, checking the quality or coordinating with the real environment when things aren't available as a digital CAD model. Nevertheless, there is still potential for innovation to do things better.

Currently, we are investigating how cloud platforms can leverage the processes of model coordination in terms of communication and data mobility, something which is important not just in times of COVID-19. The more this should accelerate planning processes and decisions, and shake off the burden of sluggish, expensive, and time-consuming PLM processes where they are needed.

Now I'd like to show you videos of three interesting platforms. Autodesk BIM360, Unity Reflect, and Cintoo. This video shows how data from the vault data management system can be synchronized to the BIM360 cloud platform. In BIM360, the model of a supplier, who is in the middle of finding a solution for an additional air supply, is automatically checked for collisions. The workflow also shows how an issue is created and tracked until a solution for the problem has been found.

This video shows at the bottom left how a duct and an air supply is changed and fine tuned in the Revit CAD environment. In the background, you can see a model coordination of the building structure, the building infrastructure, and the production facility within the Unity Reflect viewer and the live connection to these models. A few seconds after changing the model, the coordination view is also updated to wherever you are globally. Now that's impressive.

The last video here shows a comparison between a CAD model and the point cloud on the Cintoo cloud platform. The nice thing here is that the point cloud is being matched and that this match can be viewed as a comparison map to check the quality of the CAD model regarding content, object positions, and deviations. Both platforms you need to reflect in Cintoo can use CAD data directly from the BIM360 platform in order to keep data flow consistent.

OLIVER NEUMANN: Thank you, Robert. That's a really impressive example, how our idea of an integrated factory model can become true. And as you've just explained in detail, it's all about convergence, convergence of industries, the emergence of disciplines and data, which will lead to a more-- to more integration, better collaboration, and drive decisions early in the process.

In a typical planning process, you'll have many experts with different expertise in various areas. And all of them deliver their value to the big central thing, the integrated model, yeah. In combination with a central platform, well-known as a common data environment or CDE, which collects and manage all existing data, we all-- and I really mean, we all together can become more successful for sure.

And with this, I'd like to thank Robert once again for his exciting presentation of Magna's realization of an integrated factory model. And I finally want to invite you to our question and answer session, which starts exactly now. Thank you all. Please stay safe and see you soon. Bye bye.