Malaysia Energy System Model 2025–2050

A Calliope-based long-term capacity expansion model covering Malaysia's three power grids: Peninsular Malaysia (TNB), Sabah (SESB), and Sarawak (Sarawak Energy) — from 2025 to 2050.

Aligned with Malaysia's National Energy Transition Roadmap (NETR): 31% RE by 2025 → 40% by 2035 → 70% RE by 2050.

---

Features

• Multi-region myopic rolling optimisation across 6 milestone years (2025–2050)
• Three physically separate grids modelled independently, with optional Sabah–Sarawak HVDC interconnector
• Streamlit dashboard for interactive scenario configuration, simulation launch, and results exploration
• Time-series clustering (24 representative days from 8,760 hourly timesteps) for tractable LP solves
• Technology cost learning curves sourced from IRENA 2023

---

Project Structure

malaysia_calliope/
├── model/
│   ├── techs/           # Technology definitions (supply, storage, demand, transmission)
│   ├── locations/       # Regional node definitions (peninsular, sabah, sarawak)
│   ├── timeseries/      # Clustered hourly time-series CSVs
│   ├── scenarios/       # Scenario YAML overrides
│   └── overrides/       # Optional policy & cost overrides
├── scripts/
│   ├── fetch_data.py    # NASA POWER API + synthetic profile generation
│   ├── time_cluster.py  # K-means time clustering (24 representative days)
│   ├── build_model.py   # Assemble Calliope model per region + milestone year
│   ├── myopic_chain.py  # Myopic rolling optimisation chain
│   ├── run_parallel.py  # Parallel regional runs
│   ├── postprocess.py   # Post-processing: NetCDF → summary DataFrames
│   └── time_cluster.py  # Temporal aggregation utility
├── dashboard/
│   ├── app.py           # Streamlit entry point
│   ├── pages/           # Dashboard pages: Inputs, Scenarios, Run, Results, Data
│   └── utils/           # Chart helpers and config writers
├── data/
│   ├── raw/             # Raw input data (capacity, fuel prices, demand)
│   └── processed/       # Cleaned inputs ready for model ingestion
├── results/
│   ├── baseline/        # Baseline scenario outputs (NetCDF + JSON summaries)
│   └── netr_target/     # NETR-aligned scenario outputs
└── requirements.txt

---

Quick Start

1. Install dependencies

pip install -r requirements.txt

2. Install a solver

The model requires a linear programming solver.

GLPK (free, works out-of-the-box):

conda install -c conda-forge glpk

HiGHS (faster for large models — recommended):

pip install highspy

macOS note: After installing GLPK via conda, remove quarantine with:

xattr -d com.apple.quarantine $(which glpsol)

3. Fetch and prepare input data

# Use synthetic solar profiles (no NASA API key needed)
python scripts/fetch_data.py --no-nasa

# Generate 24 representative days via k-means clustering
python scripts/time_cluster.py

4. Run a scenario

# Baseline scenario — all three regions in parallel
python scripts/run_parallel.py --scenario baseline

# NETR target scenario
python scripts/run_parallel.py --scenario netr_target

5. Launch the dashboard

streamlit run dashboard/app.py

Open http://localhost:8501 in your browser.

---

The Three Grids

Region	Operator	2024 Capacity	Generation Mix
Peninsular Malaysia	TNB	26,152 MW	Coal 49%, Gas 42%, Hydro 5%, Solar 5%
Sabah	SESB	~1,500 MW	Gas >80%, Hydro 4%, Solar growing
Sarawak	Sarawak Energy	5,745 MW	Hydro 62%, Gas 21%, Coal 16%

The three grids are physically separate with no existing interconnection. The model includes a Sabah–Sarawak HVDC cable as an investable option from 2035.

---

Scenarios

Scenario	Description
baseline	Business as usual. No new coal post-2030. Moderate RE growth.
netr_target	NETR-aligned. 70% RE by 2050. Coal phased out by 2040. Carbon price trajectory applied.
accelerated_re	Aggressive decarbonisation. Coal exits by 2035. 80%+ RE by 2050.

---

Dashboard Pages

Page	Description
Inputs	Set demand growth, technology CAPEX/OPEX, fuel prices, carbon price, and RE targets
Scenarios	Choose a base scenario and configure optional overrides
Run	Launch simulation and monitor progress
Results	Explore capacity mix, generation dispatch, system costs, emissions, and battery KPIs
Data	Browse and download raw model inputs and outputs

---

Runtime

Setting	Time
Temporal resolution	24 representative days (k-means from 8,760 hourly timesteps)
Regions per run	3 (solved in parallel)
Milestone years	6 (2025, 2030, 2035, 2040, 2045, 2050)
Estimated runtime (GLPK, 8-core laptop)	60–90 min per scenario
Estimated runtime (HiGHS/Gurobi)	20–40 min per scenario

---

Data Sources

Data	Source
Installed capacity	Energy Commission Malaysia Statistics Handbook
Solar irradiance	NASA POWER API (power.larc.nasa.gov)
Technology costs	IRENA Renewable Power Generation Costs 2023
Fuel prices	World Bank Commodity Markets
Policy targets	Malaysia NETR (Ministry of Economy, 2023)

---

Requirements

• Python 3.8+
• calliope >= 0.6.10
• streamlit >= 1.30
• plotly, xarray, pandas, scikit-learn, numpy, scipy, netCDF4, pyyaml, requests
• LP solver: GLPK (via conda) or HiGHS (pip install highspy)

---

License

MIT License. See LICENSE for details.