🏡 Smart Energy Optimizer - IBM TechXchange Hackathon 2025

Real-time Smart Home Energy Management with IBM watsonx.ai Integration

Dashboard	Devices

Analytics	AI Optimization

🎯 Project Overview

Smart Energy Optimizer is an intelligent home energy management system that leverages IBM watsonx.ai to provide real-time energy monitoring, AI-powered predictions, and smart optimization recommendations. Built for the IBM TechXchange Hackathon 2025.

🌟 Key Features

• 🤖 Real IBM watsonx.ai Integration - Granite-3-8b-instruct model for energy predictions
• 📱 Native iOS App - Beautiful SwiftUI interface with glassmorphism design
• ⚡ Real-time Monitoring - Live energy consumption tracking with WebSocket updates
• 🔮 AI Predictions - 24-hour energy forecasts with peak/off-peak analysis
• 🎯 Smart Recommendations - AI-powered optimization suggestions
• 🏠 Device Control - Remote control of smart home devices
• 📊 Live Analytics - Real-time charts and energy flow visualization

🏗️ System Architecture

┌─────────────────┐    ┌──────────────────┐    ┌─────────────────┐
│   iOS App       │    │   Backend API    │    │ IBM watsonx.ai  │
│   (SwiftUI)     │◄──►│   (Node.js)      │◄──►│ (Granite Model) │
└─────────────────┘    └──────────────────┘    └─────────────────┘
         │                       │                       │
         │              ┌────────▼────────┐              │
         │              │ Device Simulator │              │
         │              │  (Real-time)     │              │
         │              └─────────────────┘              │
         │                                               │
         └───────────── WebSocket Real-time ─────────────┘

🚀 Getting Started

Prerequisites

• Node.js 18+
• Xcode 15+ (for iOS development)
• IBM Cloud account with watsonx.ai access

Backend Setup

1. Clone the repository

   git clone https://github.com/Amith71965/smart-energy-optimizer.git
   cd smart-energy-optimizer

2. Install backend dependencies

   cd backend
   npm install

3. Configure IBM watsonx.ai

   cp .env.example .env
   # Edit .env with your IBM Cloud credentials:
   # WATSONX_API_KEY=your_api_key
   # WATSONX_PROJECT_ID=your_project_id
   # WATSONX_URL=https://us-south.ml.cloud.ibm.com

4. Start the backend server

   npm start
   # Server runs on http://localhost:3000

5. Start device simulators (in new terminal)

   cd simulators
   npm install
   npm start

iOS App Setup

1. Open Xcode project

   cd ios-app
   open ios-app.xcodeproj

2. Configure network settings

   • Ensure Info.plist allows localhost connections
   • Backend URL: http://localhost:3000
   • WebSocket URL: ws://localhost:3000

3. Build and run

   • Select iPhone simulator
   • Press ⌘+R to build and run

📱 iOS App Features

Dashboard

• Real-time energy usage ring (kW display)
• Current cost tracking
• Device status grid with live updates
• Live energy flow chart
• Today's summary with efficiency scoring

Device Management

• HVAC temperature control (60-85°F)
• Lighting brightness control (0-100%)
• Appliance on/off toggle
• Real-time power consumption display

AI Analytics

• 24-hour energy predictions
• Peak/off-peak period identification
• Cost optimization recommendations
• Efficiency trend analysis

Real-time Features

• WebSocket live updates (30-second intervals)
• Smooth animations and transitions
• Haptic feedback for interactions
• Auto-reconnection handling

🤖 IBM watsonx.ai Integration

AI Agents

1. Monitor Agent - Real-time energy pattern analysis
2. Prediction Agent - 24-hour consumption forecasting
3. Optimization Agent - Smart recommendation generation

Model Details

• Model: ibm/granite-3-8b-instruct
• Authentication: IBM Cloud IAM with auto-refresh
• Response Time: < 2 seconds average
• Fallback: Graceful degradation with statistical models

AI Capabilities

• Energy consumption prediction with 70% confidence
• Peak hour identification (7-9 AM, 6-9 PM)
• Cost optimization recommendations
• Device usage pattern analysis

🔧 API Endpoints

Core APIs

• GET /health - System health and AI status
• GET /api/devices - Device list and status
• GET /api/energy/current - Real-time energy readings
• GET /api/energy/summary - Dashboard summary data

AI APIs

• GET /api/predictions - 24-hour AI forecasts
• GET /api/optimization/recommendations - AI suggestions

Device Control

• POST /api/devices/:id/control - Device control actions

WebSocket

• ws://localhost:3000 - Real-time data streaming

🎨 Design System

UI Features

• Glassmorphism design with ultra-thin materials
• Dark text on light backgrounds for accessibility
• Smooth animations with spring physics
• Haptic feedback for user interactions

🏆 Hackathon Highlights

Innovation

• ✅ Real IBM watsonx.ai integration
• ✅ 3 specialized AI agents working in concert
• ✅ Real-time WebSocket data streaming
• ✅ Professional iOS app with native performance

Technical Excellence

• ✅ Granite-3-8b-instruct model integration
• ✅ SwiftUI with modern iOS development patterns
• ✅ RESTful API design with proper error handling
• ✅ Real-time data visualization

User Experience

• ✅ Intuitive energy management interface
• ✅ Actionable AI recommendations
• ✅ Real-time feedback and control
• ✅ Professional design and animations

📈 Performance Metrics

• Backend Response Time: < 100ms for most endpoints
• AI Processing Time: < 2 seconds for predictions
• Data Update Frequency: Every 30 seconds
• WebSocket Latency: < 50ms
• iOS App: 60fps smooth animations

🛠️ Technology Stack

Backend

• Node.js with Express.js
• WebSocket for real-time communication
• IBM watsonx.ai SDK
• Device simulation with realistic patterns

iOS App

• SwiftUI 5.0 with @Observable pattern
• Combine framework for reactive programming
• Network framework for connection monitoring
• Native iOS performance optimization

AI/ML

• IBM watsonx.ai platform
• Granite-3-8b-instruct model
• Statistical fallback algorithms
• Real-time data processing

📄 License

MIT License - Built for IBM TechXchange Hackathon 2025

👨‍💻 Author

Amit Kumar Yadav K Aryan Yadav Srikar

• IBM TechXchange Hackathon 2025 Participant
• Smart Energy Optimization with watsonx.ai

---