Logo Lanfrica

asiimwe-dev/SPE-Enerython-CS1

Domain:

digital infrastructureenvironment and energy

Record type:

project
Creator:
asi
Host:
This repository is part of the SPE-Energython 2026 Project for designing a 10 - 20 MW Data Center in the tropical Africa and it covers IT Architecture and Load Profiling. # Data Center Architecture & Load Profiling **SPE AMTS Energython 2026** A computational engine for sizing, profiling, and benchmarking a 10–20 MW natural gas-powered data center in tropical Africa. This repository implements the **Computer Scientist #1** role — Data Center Architecture & Load Profiling — producing the technical deliverables that feed the team's techno-economic model. --- ## The Challenge The SPE AMTS Energython 2026 requires a cross-disciplinary team to design a techno-economic model for a multi-megawatt data center co-located with natural gas generation in tropical Africa. The facility must serve heterogeneous compute workloads — enterprise cloud, high-performance computing, and AI training — while operating in ambient temperatures of 35–42°C with high humidity. This repository solves the **demand side** of the equation: how much IT capacity the facility can support, what the power consumption looks like over a year, and how efficient the cooling architecture will be under tropical conditions. ## Our Solution: 3-Phase Deployment Strategy We model a phased deployment that scales from 10 MW to 20 MW gross generation, with PUE improving as chillers reach optimal utilization. This approach reduces upfront CAPEX risk while designing for maximum long-term revenue. | Phase | Gross MW | PUE | IT MW | Racks | Purpose | | ----------- | -------- | ---- | ----- | ----- | ----------------------------------- | | **Phase 1** | 10.0 | 1.30 | 7.69 | 250 | Initial deployment, prove the model | | **Phase 2** | 15.0 | 1.25 | 12.00 | 360 | Scale to meet demand | | **Phase 3** | 20.0 | 1.20 | 16.67 | 490 | Full build-out, maximum revenue | **Why PUE improves with scale**: At partial load (Phase 1, 50% chiller utilization), cooling systems run below optimal efficiency. At full load (Phase 3, 95%+ utilization), hybrid liquid cooling achieves peak COP. ### Deliverables | Deliverable …