Open-source implementation of DESSEM (Daily Short-Term Hydrothermal Scheduling Model) in Julia using JuMP.
OpenDESSEM is a Julia-based implementation of Brazil's official day-ahead hydrothermal dispatch optimization model (DESSEM - Modelo de Programação Diária da Operação de Sistemas Hidrotérmicos). This project aims to provide a transparent, reproducible, and extensible version for research, validation, and innovation in short-term electricity market optimization.
- 🏗️ Entity-Driven Architecture - Model dynamically builds itself from system entities
- 💾 Database-Ready - Native PostgreSQL support with SQLite for development
- 🔧 Modular Constraints - Pluggable constraint system for extensibility
- 🌊 Full Hydro Modeling - Cascade reservoirs with time delays and spillage
- 🔥 Thermal Unit Commitment - Complete UC with ramp rates and min up/down times
- ⚡ Network Model - DC-OPF and AC-OPF using PowerModels.jl (proven formulations)
- 📊 Renewable Integration - Wind, solar, and hybrid plants with forecasting
- 🎯 Brazilian Market Compatible - PLD calculation and CCEE output formats
# Clone the repository
git clone https://github.com/your-org/openDESSEM.git
cd openDESSEM
# Install Julia dependencies
julia --project=. -e 'using Pkg; Pkg.instantiate()'using OpenDESSEM
# Load system from database
loader = DatabaseLoader(
connection = LibPQ.Connection("dbname=opendessem"),
scenario_id = "deterministic",
base_date = Date("2024-01-15")
)
system = load_system(loader)
# Create optimization model
model = DessemModel(system, time_periods=168, discretization=:hourly)
# Add constraints
add_constraint!(model, EnergyBalanceConstraint(...))
add_constraint!(model, ThermalUnitCommitmentConstraint(...))
add_constraint!(model, HydroWaterBalanceConstraint(...))
add_constraint!(model, PowerModelsDCOPFConstraint(...)) # Using PowerModels.jl
# Build and solve
create_variables!(model)
build_objective!(model)
solution = optimize!(model, HiGHS.Optimizer)
# Extract results
dispatch = extract_dispatch(solution)
prices = calculate_marginal_costs(solution) # PLD- Documentation Index - Complete documentation map
- Quick Reference - Essential commands and workflows
- Development Guidelines - TDD, commit conventions, code style
- Technical Plan - Architecture and implementation
- Planning Document - Background and problem definition
- Constraint System Guide - Constraint builder documentation
- Architectural Decisions - Design rationale
- Interactive Wizard - Step-by-step system builder
- Examples - Usage examples and workflows
openDESSEM/
├── src/ # Source code
│ ├── core/ # Core model structures
│ ├── entities/ # Entity type definitions
│ ├── constraints/ # Constraint implementations
│ ├── data/ # Data loading (database, files)
│ ├── solvers/ # Solver interfaces
│ └── analysis/ # Results and visualization
├── test/ # Test suites (TDD)
├── database/ # Database schema and migrations
├── examples/ # Example scripts
├── docs/ # Documentation
└── scripts/ # Development utilities
- FCF curve loader from infofcf.dat
- Load shedding and deficit variables
- Production cost objective with scaling
- Terminal water value from FCF curves
- Cascade topology utility with cycle detection
- Inflow data loading from dadvaz.dat
- Water balance with cascade delays and inflows
- Unified solve_model!() API with two-stage pricing
- Lazy loading for optional solvers (Gurobi, CPLEX, GLPK)
- Infeasibility diagnostics with compute_iis!()
- PLD DataFrame output and cost breakdown
- End-to-end integration tests
- All variable extraction (thermal, hydro, renewable, deficit)
- PLD dual extraction from energy balance constraints
- CSV/JSON export with formatting
- Constraint violation reporting with type classification
- Nodal LMP extraction via PowerModels DC-OPF
- Unified pricing pipeline (nodal-first, zonal-fallback)
- Validation types and reference data loaders
- Comparators for cost, PLD, and dispatch
- Multi-format reporters (console, markdown, JSON)
- End-to-end integration test with ONS sample data
- Language: Julia 1.10+
- Optimization: JuMP.jl
- Network Constraints: PowerModels.jl (DC-OPF, AC-OPF formulations)
- Data Parsing: PWF.jl (Brazilian .pwf file support)
- Solvers: HiGHS.jl (open-source), Gurobi.jl (optional)
- Database: PostgreSQL (production), SQLite (development)
- Testing: Test.jl
- Documentation: Documenter.jl
DESSEM is Brazil's official day-ahead hydrothermal dispatch model, operated by ONS (Operador Nacional do Sistema Elétrico) and CCEE (Câmara de Comercialização de Energia Elétrica) since January 2020. It solves the daily energy dispatch problem for:
- Temporal scope: 1-14 days with hourly/half-hourly discretization
- Spatial scope: Brazilian Interconnected System (SIN)
- ~158 hydro plants
- ~109 thermal plants
- 4 submarkets (N, NE, SE/CO, S)
- ~6,450 buses
- ~8,850 transmission lines
For more details, see the planning document.
We welcome contributions! Please see:
- Development Guidelines - TDD, commit conventions, code style
- Quick Reference - Essential commands
- Technical Plan - Architecture overview
# 1. Write tests first (TDD)
# 2. Implement feature
# 3. Run tests
julia --project=test test/runtests.jl
# 4. Clean and check
./scripts/clean_before_commit.sh
julia scripts/pre_commit_check.jl
# 5. Commit
git commit -m "feat(scope): description"This project is licensed under the MIT License - see the LICENSE file for details.
- ONS - Operador Nacional do Sistema Elétrico (Brazilian System Operator)
- CCEE - Câmara de Comercialização de Energia Elétrica (Electricity Trading Chamber)
- CEPEL - Centro de Pesquisas de Energia Elétrica (original DESSEM developers)
- JuMP.dev - Optimization modeling in Julia
- Diniz, A.L., et al. (2020). "Hourly pricing and day-ahead dispatch setting in Brazil: The DESSEM model". Electric Power Systems Research.
- Official DESSEM Documentation
- CCEE Market Rules
- Issues: GitHub Issues
- Discussions: GitHub Discussions
Status: ✅ Phases 1-4 Complete - Ready for Validation
Test Coverage: 2075+ tests with >90% coverage
Sample Data: ONS DS_ONS_102025_RV2D11, CCEE DS_CCEE_102025_SEMREDE_RV1D04
Note: This is an independent open-source project and is not affiliated with or endorsed by ONS, CCEE, or CEPEL.