🔋 Battery Temperature Forecasting Web App

Forecast future battery temperature trends over upcoming charge/discharge cycles using Machine Learning + Flask + Chart.js.

📌 Project Overview

This project implements an end-to-end ML pipeline and interactive web app to:

✅ Analyze NASA B0005 battery dataset.

✅ Train a Linear Regression model for time series forecasting.

✅ Deploy the model via a Flask REST API.

✅ Build a simple UI Dashboard with Chart.js for visualization.

✅ Allow user to input temperature value and forecast next 50 future cycles.

🚗 Why Battery Temperature Matters

• Battery temperature affects capacity fade, thermal degradation, and safety.

• Accurate forecasting can improve:

  • Battery life cycle management
  • Thermal control system optimization
  • EV reliability & efficiency

🗂 Directory Structure

Battery_forecast/
│
├── model/
│   ├── battery_temp_model.pkl      # Trained Linear Regression model
│   ├── scaler.pkl                  # Fitted MinMaxScaler
│
├── static/
│   └── index.html                  # Interactive UI (Chart.js + HTML + JS)
│
├── app.py                          # Flask REST API
│
├── B0005_discharge.csv              # NASA Battery Dataset (discharge cycles)
│
├── requirements.txt                # Project dependencies
│
└── README.md                       # Project documentation

🚀 How to Run the Project

1️⃣ Clone the Repository

git clone https://github.com/Kshitijasharma/Thermal-Battery-Predictions
cd Thermal-Battery-Predictions

2️⃣ Install Dependencies

pip install -r requirements.txt

3️⃣ Train and Save Model (Optional)

If not already trained, run the model training script (train_model.py or inside a notebook):

python train_model.py

Or use the provided battery_temp_model.pkl and scaler.pkl.

4️⃣ Run the Flask App

python app.py

5️⃣ Open Web App

Go to:

http://127.0.0.1:5000

⚙️ How It Works

• User inputs a starting temperature value in UI.
• The app sends this value to /predict endpoint via POST request.
• Flask API uses Linear Regression model to forecast next 50 cycle temperatures.
• Forecasted temperatures are returned as JSON.
• UI displays the result as an interactive line chart using Chart.js.

🛠 Technologies Used

✅ Python ✅ Flask ✅ Scikit-learn ✅ Pandas / Numpy ✅ MinMaxScaler ✅ Chart.js ✅ HTML + JavaScript

🚧 Future Work

• I am working on implementing LSTM / GRU for more advanced time series forecasting
• Adding Voltage and Current multivariate forecasting
• Will deploy on Heroku / AWS / Render
• Improving UI (Bootstrap / React)
• Adding real-time data streaming support

📚 Dataset Reference

• NASA Ames Prognostics Data Repository https://www.nasa.gov/content/prognostics-center-of-excellence-data-set-repository

Dataset: B0005 Battery aging dataset

---

✨ Example Result

---

If you like this project, give it a ⭐️ on GitHub! 🚀