This repository contains the implementation of the Heating Tunnel Project, a comprehensive application of unsupervised machine learning techniques for industrial process optimization. The project identifies operational inefficiencies and potential sources of quality defects in a manufacturing heating tunnel system using clustering analysis.
The goal of this project was to analyze operational energy data from a heating tunnel system, detect patterns, and uncover anomalies leading to quality defects and increased energy consumption. Using the CRISP-DM methodology, the project encompasses all phases from business understanding to deployment.
- Analyze archived operational data to identify inefficiencies in energy consumption.
- Implement unsupervised clustering algorithms to group operational patterns.
- Draw actionable insights for improving manufacturing consistency and efficiency.
- Data Preprocessing: Data cleaning, standardization, and preparation for analysis.
- Machine Learning: Implementation of the Time Series KMeans algorithm for clustering.
- Evaluation Metrics: Use of Silhouette and Calinski-Harabasz scores for cluster validation.
- Visualization: Graphical representation of energy consumption patterns and clustering results.
- Programming Language: Python
- Libraries:
tslearn
for time series clusteringnumpy
andpandas
for data manipulationmatplotlib
for data visualizationscikit-learn
for evaluation metrics
- Tools: Jupyter Notebook / Google Colab
Heating_Tunnel_Project/
│
├── data/ # Raw and processed data files
├── notebooks/ # Jupyter notebooks for each phase of the project
│ ├── Phase_1_Business_Understanding.ipynb
│ ├── Phase_2_Data_Understanding.ipynb
│ ├── Phase_3_Data_Preprocessing.ipynb
│ ├── Phase_4_Model_Implementation.ipynb
│ └── Phase_5_Evaluation_and_Deployment.ipynb
├── outputs/ # Visualizations and clustering results
├── requirements.txt # Required Python libraries
└── README.md # Project documentation
- Clone this repository:
git clone https://github.com/aksh-ay06/Operational-Efficiency-Analysis-Machine-Learning-for-Heating-Tunnel-Optimization.git
- Navigate to the project directory:
cd Operational-Efficiency-Analysis-Machine-Learning-for-Heating-Tunnel-Optimization
- Install the required dependencies:
pip install -r requirements.txt
- Open the Jupyter Notebooks in the directory to follow each project phase:
jupyter Project_3(Smart_manufacturing).ipynb
The project successfully identified two distinct operational patterns and highlighted inefficiencies related to:
- Operator-dependent parameter variations.
- Maintenance inconsistencies in pneumatic systems.
- Process control limitations due to the absence of automation.
- Silhouette Score: 0.65–0.66
- Calinski-Harabasz Score: 63.03–82.04
- [Akshay Patel]
Graduate Student in Industrial Engineering
Email: [akshaypatelnitb6@gmail.com]