Autodesk Forge (Platform Services), Point Cloud to RCS Converter, Scheduler Service

ABOUT HELIX:

HELIXRE, the first company to emerge from Google X, is leading the charge to bring 'digital twinning' to everyone. They're changing the world one structure at a time. They produce accurate digital equivalents of entire buildings in a fraction of the time and cost.

THE CHALLENGE:

1. It was necessary to effectively manage and maintain various services and jobs on GCP (Google Cloud Platform).

2. Downloading and uploading of huge files on Autodesk Forge (Platform Services) resulted in JVM OOM exception, that had to be tackled.

3. Another challenge was to get the IFC and Revit file's model data (tessellation data and metadata/properties) in two separate custom files to store in the database so that it could be queried faster.

4. The Helix team desired to generate a 3D Revit model from entity data. The problem was that the entire Revit model could not be created from the input data.

5. Another challenge was to reschedule the jobs which were terminated due to out of memory. 

6. At times config maps remained even on the deletion of jobs. The problem had to be resolved.

THE PROJECT:

Helix was looking for:

• A generic solution to convert CAD model to another format:

          - IFC to OBJ and metadata
          - Revit to OBJ and metadata

• They also wanted a Revit Add-in to be created.

• Extraction of point cloud data (points, color, intensity and normals) and converting in RCS (Autodesk Reality capture scan) format.

The ProtoTech team wrote a short running service (job) to run on cloud. This job effectively converts one CAD (Revit) model to another(OBJ).
Both the teams, ProtoTech and Helix, collaborated closely to find the best solution for their business needs.

THE SOLUTION:

1. Autodesk Forge APIs were used to convert Revit models to OBJ.

2. ProtoTech team serialized tessellation data in obj format using 3rd party SDK. 

3. Fetched the metadata/properties and serialized it in custom JSON format.

4. Implemented an OBJ and JSON serializer using Teigha BIM SDK, to export tessellation data in obj format and metadata/properties in custom JSON format. 

5. Also exported view data to pdf and SVG format.

6. Deserialized structured data from entity data and created each Revit entity to generate a Revit model using Forge Design Automation for Revit API.

7. Gathered the LAS, LAZ, or EPT data (points, color, intensity, and normals) and converted it to RCS using Autodesk Recap SDK.

8. Implemented logic to reschedule the jobs with optimal higher memory resources on failure to ensure successful completion of jobs

9. Implemented logic to delete orphaned config maps ensuring the smooth running of the GKE node

Transforming large building infrastructure, which is currently not available in digital form, provides a large boost to quick maintenance, better planning for the addition of new features, etc. The optimization done on processing of large point clouds has sped up work with quality maintained in a cost-effective and scalable manner.

"We have got a winner here! I personally appreciate the dedication and teamwork of the ProtoTech developers for making the challenges achievable. Good Job Guys!"

– Prem Kumar, Helix.re