COMSOL Multiphysics Automation with Grok AI

A Python-based automation framework for running parallel COMSOL Multiphysics simulations with configurable parameter sweeps. Built with the help of Grok AI to streamline large-scale photonic device optimization workflows.

This tool automates the process of loading COMSOL models, sweeping refractive index parameters across user-defined wavelength ranges, solving studies in parallel across multiple COMSOL server instances, and saving results — all without manual GUI interaction.

---

🚀 Key Features

• Parallel Multi-Client Execution: Spawns up to 5 independent COMSOL client instances simultaneously using Python's multiprocessing, dramatically reducing total sweep time.
• Automated Parameter Sweeps: Configurable sweep over refractive index (n_ana) with per-point wavelength ranges (wl_start, wl_end).
• Smart Resume (v2): Version 2 detects previously completed simulations and skips them, enabling crash-safe, resumable batch runs.
• Batch Orchestration: Includes a Windows batch script to chain multiple optimization campaigns (e.g., varying metal thickness) sequentially with auto-shutdown.
• Clean Logging: Real-time console output with progress tracking, timing, and emoji status indicators (✅/❌).

---

📁 Repository Structure

Comsol-Automation-using-Grok/
├── README.md                          # This file
├── LICENSE                            # MIT License
├── .gitignore                         # Python & COMSOL ignore rules
│
├── scripts/
│   ├── version_1.py                   # Basic parallel sweep automation
│   ├── version_2.py                   # Enhanced version with resume capability
│   └── run_all_simulations.bat        # Batch runner for sequential campaigns
│
└── docs/
    └── CHANGELOG.md                   # Version history & differences

---

📋 Prerequisites

Requirement	Details
Python	3.8 or higher
COMSOL Multiphysics	Licensed installation with command-line server support
MPh	Python package for COMSOL automation (pip install mph)
OS	Windows (batch scripts use Windows-specific paths)

---

⚡ Quick Start

1. Install dependencies

pip install mph

2. Configure your sweep

Edit the sweep_config list in scripts/version_1.py or scripts/version_2.py:

sweep_config = [
    (1.28, 0.55, 0.65),   # (n_ana, wl_start, wl_end)
    (1.29, 0.55, 0.65),
    # ... add your parameter points
]

3. Update file paths

Set BASE_MODEL and RESULTS_DIR to point to your local COMSOL model and output directory:

BASE_MODEL  = r'path\to\your\model.mph'
RESULTS_DIR = r'path\to\your\Results'

4. Run

# Run a single sweep campaign
python scripts/version_1.py

# Or use v2 with resume support
python scripts/version_2.py

# Or chain multiple campaigns
scripts\run_all_simulations.bat

---

🔄 Version Comparison

Feature	Version 1	Version 2
Parallel execution	✅ 5 clients	✅ 5 clients
Parameter sweep	✅ Full config	✅ Full config
Resume after crash	❌ Restarts all	✅ Skips completed
Progress detection	❌ None	✅ Scans results folder
Dynamic batch sizing	❌ Fixed split	✅ Adapts to remaining

---

🛠️ How It Works

┌─────────────────┐     ┌──────────────┐     ┌──────────────────┐
│  Sweep Config    │────▶│  Split into  │────▶│  5 Parallel      │
│  (15 RI points)  │     │  5 batches   │     │  COMSOL Clients  │
└─────────────────┘     └──────────────┘     └──────┬───────────┘
                                                     │
                              ┌───────────────────────┘
                              ▼
                    ┌──────────────────┐
                    │  For each point: │
                    │  1. Load model   │
                    │  2. Set params   │
                    │  3. Solve study  │
                    │  4. Save .mph    │
                    │  5. Clear memory │
                    └──────────────────┘

Each client runs on a separate COMSOL server port (2036, 2046, 2056, ...) with 2 CPU cores, designed for machines like Ryzen 5 5500U with 16 GB RAM.

---

📌 Notes

• The scripts use mph.start(cores=2, port=...) — adjust cores based on your CPU.
• The batch file (run_all_simulations.bat) includes a shutdown /s /t 30 command at the end. Remove or comment this line if you don't want your PC to shut down after completion.
• Paths in the scripts are configured for the original author's setup. You must update them before running.

---

📄 License

This project is licensed under the MIT License.

---

🤝 Acknowledgments

• Grok AI — AI assistant used during development of these automation scripts.
• MPh — The Python package that makes COMSOL automation possible.
• COMSOL Multiphysics — The simulation platform powering the physics.