RAMP Load Demand Simulation – Streamlit Application

This is a Streamlit-based interface that makes the bottom-up, appliance-level load demand modelling capabilities of the RAMP framework accessible through a structured graphical workflow.

RAMP generates realistic, high-resolution electricity demand time series by simulating user behaviour, appliance operation, and stochastic variability at minute-scale resolution. Traditionally, this workflow requires detailed Excel templates and Python scripting. This application exposes the same pipeline in a transparent, guided, and reproducible environment.

The app is organised into three complementary pages:

1. Inputs Editor – build full RAMP Excel inputs directly in the interface
2. RAMP Simulation – configure temporal structure and run the stochastic engine
3. SSA Archetypes – generate demand profiles from empirically derived SSA archetypes

Together, they provide a complete workflow from appliance definition → stochastic simulation → export-ready profiles for energy system modelling.

---

Application Structure

1. Inputs Editor

This page allows users to build full RAMP-compatible Excel input files directly inside the app, avoiding manual editing of the full template.

Features

• Define multiple user categories (e.g., Households, Clinics, Schools)
• Add:
  • Standard appliances (functioning windows + total daily usage)
  • Duty-cycle appliances (cycle segments and admissible cycle windows)
• Validate appliance timing via hourly heatmap visualisation
• Automatically compile a full RAMP template workbook
• Download a ready-to-run Excel file

The editor uses user-friendly (“pretty”) column names, internally mapped to canonical RAMP parameter names through a structured schema layer. Missing optional parameters are automatically filled with consistent defaults.

This page is recommended as the starting point for new projects.

---

2. RAMP Simulation

This page runs the full stochastic RAMP workflow.

Step 1 – Temporal Structure

Users define the structure of the synthetic year:

• Number of seasons (1–4)
• Custom month allocation per season
• Optional within-week differentiation:
  • Weekday vs Weekend
  • Custom week-classes (up to 7)

Each (Season × Week-class) combination becomes an archetype, representing a stochastic pool of daily profiles.

Step 2 – Upload Full RAMP Inputs

Users upload full RAMP Excel files:

• Single-archetype mode
  One input governs the entire year.

• Multi-archetype mode
  One full Excel file per archetype (Season × Week-class).

Unlike earlier simplified approaches, this version expects the complete RAMP template structure, validated internally through a schema layer.

Step 3 – Stochastic Simulation

RAMP generates:

• Independent synthetic daily profiles per archetype
• A synthetic 365-day year assembled by sampling from archetype pools according to the calendar structure

This captures:

• Intra-archetype stochastic variability
• Seasonal and weekly behavioural patterns

Step 4 – Visualisation & Export

Outputs include:

• Aggregated and per-category minute-resolution profiles (365 × 1440)
• Average daily curves with min–max variability envelopes
• Day-level inspection
• Filtering by season / week-class (if enabled)

Export options:

• profile_aggregated.csv (minute resolution)
• profile_aggregated_hourly.csv (8760 hourly series)
• Per-category minute profiles

The hourly profile is directly usable in optimisation, dispatch, techno-economic, and planning models.

---

3. SSA Archetypes

This page provides a rapid scenario-building pathway based on empirically derived Sub-Saharan Africa (SSA) demand archetypes.

Instead of appliance inventories, users define the settlement composition:

• Households by tier (1–5)
• Schools
• Health facilities (5 tiers)

Climatic context is selected via latitude / cooling regime logic.

The methodology is based on:

N. Stevanato, I. Sangiorgio, R. Mereu, E. Colombo, Archetypes of Rural Users in Sub-Saharan Africa for Load Demand Estimation, 2023 IEEE PES/IAS PowerAfrica.
https://doi.org/10.1109/POWERAFRICA57932.2023.10363287

The profiles in config/archetypes_library/ correspond to the open dataset v1.0.0 on Zenodo (https://doi.org/10.5281/zenodo.22832973), which also provides the RAMP inputs, 1-minute household profiles and the definitions of household and health-facility tiers. Compared with earlier versions of this app, v1.0.0 updates health-facility tiers 2–5 (2025 correction of the X-ray and cooling modelling) and household archetypes NC_F1 tiers 1–3 (re-simulated in 2026); see the dataset CHANGELOG.

This mode outputs:

• Aggregated hourly (8760) profile
• Per-user category profiles

It prioritises accessibility, speed, and structured scenario design for data-scarce environments.

---

Inputs & Outputs

Inputs

• Full RAMP Excel templates (built via the Inputs Editor or externally prepared)
• SSA archetype composition (tiers, schools, hospitals)

Outputs

All outputs are written in CSV format:

File	Description
profile_aggregated.csv	365×1440 aggregated RAMP profile
profile_aggregated_hourly.csv	8760 hourly profile (RAMP)
profile_<user>.csv	Per-category minute-resolution profile (RAMP)
ssa_aggregated_hourly.csv	Hourly aggregated SSA profile
ssa_<user>.csv	Per-category SSA profile

These files are directly compatible with:

• Mini-grid optimisation models
• Dispatch simulations
• Techno-economic assessments
• Power flow studies
• SDG7 electrification planning analyses

---

Installation

Conda (recommended)

conda env create -f environment.yml
conda activate ramp_streamlit

Running

streamlit run app.py

Then open:

http://localhost:8501

---

Contact

Alessandro Onori
📧 alessandro.onori@polimi.it

Technical Advisors

• Riccardo Mereu — Politecnico di Milano
• Emanuela Colombo — Politecnico di Milano

---

License

European Union Public Licence (EUPL v1.1).