Applications of Machine Learning in CAD/PLM


Machine learning becomes much more widespread than before. So finally it’s time to implement it in computer aided design too.

Possible applications

The short list of possible applications of the artificial intelligence within the Mechanical Engineering and Architecture, Engineering, and Construction with the high-level descriptions is provided below. On some of them I am working at the moment. This is not entire list, but allows to understand how useful and cutting-edge this technology is.

Common applications

Systems engineering approach

Implement and support the linkage between: system model, interactions between subsystem and implementation of this system in geometry model.

Workflows

Perform the work breakdown structure and workflow generation, according to the standardized workflows, early performed by users.

Part library categorization

Categorize parts across libraries by their shapes and parameters.

Parametrizing of non-parametric solids

Parametrize STL and IGES, using geometry parsers and library of standard parameters for the parts and assemblies.

Natural language user interface

Interaction with CAD systems using text or voice conversation in natural language: review the model (pan, rotate, zoom, comments, etc.), call common commands and perform input.

Design wizards and generators

Improving the behavior of wizards and generators according to standards and technical specification of overall project.

  • Fasteners generator
  • Frame generator
  • Mold-base generator for casting and molding
  • Hole generator
  • Thread generator

Handling of Agile/Scrum development process

PDM/PLM management according to Agile/Scrum methodologies. Schedule meetings of engineers and other involved persons. Handle the data and update project and parts properties appropriately according to the decisions. Perform the work breakdown structure according to technical specifications and recommendations for further implementation and supervision.

Verification of conformity to standards

Design check of performed design operations according to standards and technical specifications.

Blueprints management

Manage the heritage of scanned drawings and blueprints, performing further using of this data in Civil engineering, AEC and mechanical engineering. This includes: parsing, recognition, data extraction, automated build of 3D models, etc.

Applications for Mechanical engineering

Part selection

Improved part selection according to system model of design, technical specification and standards.

Constraints within assembly

Intelligent setting of constraints with reduced user interaction.

Applications for AEC

Tracing of HWAC, electricity and engineering systems

Perform tracing of HWAC, electricity, engineering systems, structured cabling, etc. systems according to the design.

Reinforcement and structural system generators

Calculating optimal reinforcement and structural systems and appropriate documentation.

Construction process verification

Compare ‘as designed’ vs ‘as built’ using data from 3D scanners, etc.

Exploitation and management

Cost-efficient management of building according to the data from different type of sensors.

Possible Algorithms

The matrix of ML algorithms applied to possible applications is provided below:

FieldClusteringAnomaly detectionClassificationRegression
Systems engineering approach
Workflows
Part library categorization
Parametrizing of non-parametric solids
Natural language user interface
Design wizards and generators
Handling of Agile/Scrum development process
Verification of conformity to standards
Blueprints management
Part selection
Constraints within assembly
Tracing of HWAC, electricity and engineering systems
Reinforcement and structural system generators
Construction process verification
Exploitation and management

Leave a comment

Your email address will not be published. Required fields are marked *