Batch Optimization: Optimizing Erosion Resistant Coating Parameters

Overview

This project focuses on optimizing the erosion resistance of protective coatings using Atmospheric Plasma Spraying (APS) through batch optimization techniques.

Problem Context

Background

Atmospheric Plasma Spraying (APS) is used to deposit protective coatings on material substrates. The goal is to apply a high-hardness metallic coating to improve erosion resistance, measured by mass loss under sand blasting.

Optimization Challenge

• Testing requires shipping samples to a specialized laboratory
• Each test takes approximately 4 days
• Budget: 30 experiments total
• Solution: Batch optimization approach (3 samples per batch)
• Total optimization time reduced from 120 to 40 days

Design Space Parameters

Parameter	Range	Units
Primary Gas Flow Rate	30 - 80	SLPM
Secondary Gas Flow Rate	10 - 50	SLPM
Gun Current	300 - 800	A
Carrier Gas Flow Rate	2 - 10	SLPM
Power Feed Rate	10 - 100	g/min
Spray Distance	50 - 150	mm

Device Constraint

To protect the APS device, the following constraint must be satisfied:

device_stress_index = primary_gas_flow_rate + secondary_gas_flow_rate + gun_current <= 750

Tasks

Task A: Optimization Setup

Use Honegumi to set up and run the optimization problem:

• Configure parameters and constraints
• Implement batch processing
• Handle device stress constraints

Task B: Results Analysis

Report:

• Optimal parameters
• Associated erosion rate
• Device stress index

Task C: Stress Analysis

Calculate the number of solutions in the bottom 15% of trials that have a stress index > 700.

Task D: Batch Performance Analysis

For non-Sobol batches:

• Calculate how many experiments per batch were lower than the previous best
• Compute the average of these improvements

Task E: Batch Diversity Analysis

For non-Sobol batches:

• Calculate diversity using Euclidean distance between parameters
• Identify most and least diverse batches

Implementation Details

• Uses synthetic objective function measure_erosion() from utils.py
• Implements proper error handling and validation
• Provides comprehensive logging
• Ensures reproducibility through seed setting

Project Structure

.
├── qSOBO_assignment.py    # Main optimization implementation
├── utils.py              # Utility functions and measurements
├── requirements.txt      # Project dependencies
└── README.md            # Project documentation

Getting Started

1. Install dependencies:

pip install -r requirements.txt

2. Run the optimization:

python qSOBO_assignment.py

Notes

• First 3 batches use Sobol sampling
• Subsequent batches use Bayesian optimization
• All parameters are validated before optimization
• Device stress constraint is enforced throughout optimization