ARTICLE AD BOX
I am experimenting with a Python-based MCP server to connect AI assistant with Aspen Plus.
The idea is that the AI assistant can interpret user instructions and trigger simulation-related tasks through Claude code.
For example, the AI could help:
Analyze simulation workflows
Suggest parameter settings
Trigger simulation runs
Interpret results
My current approach is to use Python as the middle between Claude Code and Aspen Plus.
The Python layer would translate AI's instructions into commands that interact with Aspen Plus.
I am wondering what architecture is commonly used for this type of system.
Is it better to build a modular tool interface that the AI can call (for example, functions like run_simulation() or modify_parameters()), or should the AI interact directly with the Aspen Plus API?
Any suggestions or best practices for integrating AI assistants with engineering simulation software would be appreciated.
