🚀 SCRAP 2 SPARK (S2S)

AI-Powered Battery Intelligence Platform for Second-Life EV Battery Decisioning

SCRAP 2 SPARK is an AI-powered platform that analyzes lithium-ion battery telemetry to predict Remaining Useful Life (RUL), assess battery health, recommend second-life deployment strategies, and estimate sustainability impact.

🔗 Quick Links

• 🌐 Live Demo: https://sdg-hackthon.vercel.app/
• ⚡ Backend API: https://sdg-hackathon.onrender.com/
• 📂 Repository: https://github.com/pulkittaneja09/sdg-hackthon

Overview

Electric vehicle batteries often retain significant usable capacity even after reaching the end of their primary automotive lifecycle. Identifying whether these batteries should be reused or recycled requires analyzing complex telemetry data, making manual evaluation difficult and time-consuming.

SCRAP 2 SPARK provides an AI-powered decision support system that predicts battery health, estimates Remaining Useful Life (RUL), evaluates operational risk, and recommends appropriate second-life applications using machine learning.

Features

• Battery telemetry upload using CSV files
• Remaining Useful Life (RUL) prediction
• Battery health assessment
• Risk classification
• Second-life deployment recommendation
• Sustainability impact analysis
• Interactive analytics dashboard
• Production-ready REST API built with FastAPI

Workflow

Battery Telemetry CSV
        │
        ▼
Data Validation
        │
        ▼
Feature Engineering
        │
        ▼
Random Forest Model
        │
        ▼
RUL Prediction
        │
        ▼
Risk Assessment
        │
        ▼
Deployment Recommendation
        │
        ▼
Analytics Dashboard

Machine Learning Pipeline

Model

Random Forest Regressor

Input Features

• Cycle Count
• Capacity
• Voltage
• Current
• Temperature
• Time

Feature Engineering

• Capacity degradation trend
• Voltage stability
• Temperature variance
• Current consistency

Outputs

• Remaining Useful Life (RUL)
• Confidence Score
• Battery Grade
• Risk Category
• Deployment Recommendation

Why Random Forest?

Random Forest performs well on structured battery telemetry data, handles nonlinear relationships, minimizes overfitting, and produces reliable predictions with minimal preprocessing.

Tech Stack

Frontend

• React
• TypeScript
• Vite
• Tailwind CSS
• Axios

Backend

• FastAPI
• Python
• Pandas
• NumPy
• Scikit-learn
• Uvicorn

Machine Learning

• Random Forest Regressor

Deployment

• Vercel
• Render

System Architecture

                 React Frontend
                        │
                  REST API Calls
                        │
                        ▼
                FastAPI Backend
                        │
      ┌─────────────────┼────────────────┐
      │                 │                │
      ▼                 ▼                ▼
 Data Processing  Feature Engineering  ML Model
                        │
                        ▼
          Recommendation Engine
                        │
                        ▼
         Sustainability Calculator
                        │
                        ▼
            Interactive Dashboard

Project Structure

SCRAP-2-SPARK
│
├── frontend/
│   ├── src/
│   ├── components/
│   ├── pages/
│   └── assets/
│
├── backend/
│   ├── api/
│   ├── services/
│   ├── models/
│   ├── utils/
│   └── main.py
│
└── README.md

Local Setup

Prerequisites

• Python 3.10+
• Node.js 18+

Backend

cd backend
pip install -r requirements.txt
uvicorn main:app --reload

Runs on:

http://localhost:8000

Frontend

cd frontend
npm install
npm run dev

Runs on:

http://localhost:5173

Deployment

Layer	Platform
Frontend	Vercel
Backend	Render

Environment Variable

VITE_API_URL=https://sdg-hackathon.onrender.com

Future Improvements

• Multi-battery batch analysis
• Explainable AI (SHAP)
• IoT telemetry integration
• Battery Passport support
• Authentication and user management
• Enterprise analytics dashboard

License

This project is licensed under the MIT License.