Pump & Valve Predictive Maintenance

This repository demonstrates a machine learning approach to Predictive Maintenance (PdM) for industrial pumps and valves. By leveraging sensor data, the project aims to predict equipment failures before they occur, helping to reduce unplanned downtime and maintenance costs in industrial operations.

📌 Project Overview

In industries like Oil & Gas, pump failure can lead to significant production losses. This project uses historical sensor readings to train a classification model that identifies the state of the equipment (Normal vs. Failure/Warning).

Key Objectives:

• Failure Prediction: Classify equipment status based on real-time sensor inputs.
• Feature Importance: Identify which sensors (pressure, temperature, vibration) are the strongest indicators of imminent failure.
• Optimization: Provide actionable insights for maintenance scheduling.

📂 Repository Structure

• pump_data.csv: Dataset containing sensor readings and corresponding equipment status labels.
• Predictive_Maintenance_Pumps_Valves.ipynb: The main Jupyter Notebook containing data preprocessing, exploratory analysis, model training, and evaluation.
• readme.md: Project documentation.

🛠️ Tech Stack

• Language: Python 3.x
• Data Analysis: Pandas, NumPy
• Machine Learning: Scikit-learn (Random Forest, XGBoost, etc.)
• Visualization: Matplotlib, Seaborn

🚀 Workflow

1. Data Preprocessing: * Handling missing values and sensor noise.

• Feature scaling and encoding of categorical status labels.

2. Exploratory Data Analysis (EDA):

• Correlation heatmaps to see relationships between sensors.
• Time-series plots of sensor data leading up to failure events.

3. Modeling:

• Training a Random Forest Classifier to handle non-linear sensor relationships.
• Evaluating performance using Precision, Recall, and F1-Score (crucial for imbalanced failure data).

4. Insights:

• Generating a Feature Importance plot to show which parameters (e.g., "Discharge Pressure" or "Motor Vibration") contribute most to failure prediction.

📊 Business Impact

• Reduced Downtime: Shift from reactive "fix-it-when-broken" to proactive maintenance.
• Cost Savings: Minimize emergency repair costs and extend the lifespan of expensive machinery.
• Safety: Prevent catastrophic failures that could lead to environmental or safety hazards.

📈 Getting Started

1. Clone this repository:

git clone https://github.com/anmolshukla13/Pump-Predictive-Maintenance-.git

2. Install dependencies:

pip install pandas scikit-learn matplotlib seaborn

3. Open the notebook:

jupyter notebook Predictive_Maintenance_Pumps_Valves.ipynb

---

Author: Anmol Shukla