Socket Bridge

Socket Bridge Connection

Description

Connect to Abaqus CAE via socket bridge for curing simulation. Invoke when user needs to use socket bridge, connect to Abaqus, or execute Python in Abaqus kernel.

Content

Socket Bridge Overview

The socket bridge provides a TCP connection to Abaqus CAE at 127.0.0.1:48152, allowing remote execution of Python code in the Abaqus kernel.

Prerequisites

• Abaqus CAE must be open and running
• The socket bridge plugin must be active in Abaqus

Connection Check

Use the check_abaqus_connection MCP tool to verify the bridge status before executing any commands.

Available MCP Tools

run_python Usage

When using run_python, set the result variable to return data back to the caller:

# Example: get list of steps
result = []
for step_name in odb.steps.keys():
    result.append(step_name)

Abaqus Python Environment Quirks

The Abaqus Python environment has several important quirks to be aware of:

Python Version

• Abaqus 2020+ uses Python 2.7 based environment
• Do not use Python 3 specific syntax

Unicode Issues

• print() statements with unicode may cause errors
• Use result = ... to return data instead of printing

Imports

• Must import constants: from abaqusConstants import *
• This provides access to all Abaqus constants (OFF, ON, RUNNING, etc.)

String Formatting

• Use the % operator for string formatting
• Do NOT use .format() or f-strings (not supported in Python 2.7)

# Correct
text = '%d nodes found' % node_count

# Incorrect (will fail)
text = '{} nodes found'.format(node_count)
text = f'{node_count} nodes found'

Dictionary Keys

• Plain strings work reliably as dictionary keys
• Unicode keys may fail in some operations

Standard Workflow

1. Check connection: Use check_abaqus_connection to verify bridge is active
2. Set working directory: Use set_workdir to configure the work directory
3. Create jobs: Use run_python to create jobs from input files
4. Submit jobs: Use submit_job or run_python with submit()
5. Monitor: Use monitor_job_status to check for errors or convergence issues
6. Copy ODB: Copy the ODB file from scratch to work directory
7. Extract data: Use inspect_odb or run_python to read results