ahmedthebest31/ecopulse-ai
Domain:
environment and energyRecord type:
project
Creator:
ahm
Host:
AI-powered real-time energy monitoring, anomaly detection, and demand-response optimization — built for Egypt's SGG Smart Grid Initiative (Cycle 4, 2026)
# EcoPulse AI — نبض البيئة الذكي
**Energy Analytics · Peak-Hours Optimization · Carbon Footprint Tracking**
A submission to the **National Smart Green Projects Initiative (SGG)** —
المبادرة الوطنية للمشروعات الخضراء الذكية
Fourth Cycle 2026 · Category 3: Small Local Projects (Individuals / Innovators) · Cairo, Egypt
> **Notice.** This repository is made visible **solely for evaluation** of the SGG
> submission above. Downloading, copying, modifying, redistributing, or any other
> use is **not permitted**. See LICENSE.
---
## 1. The problem
A mid-sized commercial or industrial facility pays an Egyptian seven-tier progressive
tariff while carrying an invisible burden: standby loads running outside peak hours,
equipment degrading toward failure with no early warning, and carbon emissions that are
never measured at all.
## 2. What EcoPulse AI does
Four operational modules run on top of one measured data pipeline:
- **Tiered billing.** Exact electronic billing against the official seven-tier Egyptian
tariff (rates effective September 2024: `0.68 → 2.23 EGP/kWh`), allocated sequentially,
displayed in EGP and USD.
- **Peak detection & load shifting.** Magnitude-based critical-spike detection and a
configurable peak window (`18:00–22:00` default), quantifying the shiftable load.
- **Predictive maintenance.** Gradual micro-surge detection grouped into per-equipment
maintenance alerts before failure happens.
- **AI executive reporting.** Deterministic Gemini-backed summaries (measured numbers
only) with a fully offline bilingual fallback that was live-tested against the real API.
## 3. Architecture
- `data-generator/` — pure-stdlib Python 3.12 simulator (managed by `uv`): one minute of
telemetry per facility over 24 hours, smooth day profiles, forced spikes and micro-surges,
atomic writes, deep config validation, deterministic under a fixed seed.
- `backend-go/` — Go analytics engine built on the standard library only (**zero
third-party dependencies**): REST …