PowNet v2.1 (Refactoring and Feature Release)

This release of PowNet introduces major refactoring for improved code organization and maintainability, along with several new features to enhance functionality and user experience.

**Key changes: **

• **Simplified Modeling Workflow: ** All modules have been refactored to streamline the modeling process and make the codebase easier to understand and extend. This includes:
  • Reduce the number of function calls required by a user
  • Streamlined solver selection
  • Simplified user inputs by switching from config files to direct arguments in functions and classes
• **Enhanced Modeling/Solver Capabilities: **
  • Enabled model updating at every simulation step instead of building an instance from scratch
  • Added rounding heuristics to improve the optimization problem-solving process
• **New Analysis Features: **
  • Fixed modeling in a rolling horizon fashion
  • Allowed aggregation of generators at a node
  • Added the stochastic module for generating synthetic time series
  • Added locational marginal price (LMP) calculation and plotting functionality
  • Separated shortfall and curtailment penalty factors for more granular control over penalties
• **Improved User Experience: **
  • Increased flexibility in selecting input and output folders. This allows users to install PowNet as a package and work within their own project directories
  • Added input checking to help users identify and correct potential errors in their input data

**Other improvements: **

• **Increased Code Reliability: ** Added unit tests to ensure code quality
• **Bug Fixes: ** Fixed the handling of line loss calculations.
• **Improved Codebase: ** Revised folder structure for better organization.
• **Enhanced Documentation: **
  • Revised CSV field descriptions for clarity
  • Added API documentation
• **Code Cleanup: ** Deleted unused files to reduce clutter.

PowNet v2.0

This is an entirely new version of PowNet that includes the following key features:

• The underlying algebraic modeling language is gurobipy, which replaces Pyomo. gurobipy is faster than Pyomo and has lower memory requirements.
• A model created with PowNet can be solved with Gurobi as well as the free, open-source, solver HiGHS.
• The mathematical formulation of PowNet is based on the 'strong' MILP formulations proposed by Knueven et al. (2020).
• DC power flow is represented through different methods proposed by Hörsch et al. (2018)). These include both voltage angle and cycle flow formulations.
• The implementation of PowNet is modular, so it allows users to include new formulations for a given constraint.
• Models for a given country are built fully from .csv files.
• Visualization of the model output has been standardised w.r.t. previous versions.

PowNet v1.3

Changes since v1.2

• Fix: Ramp Rate Constraints at Segment B.9 are modified to fix ramping constraints #7
• Fix: Up/Down Time Constraints at Segment B.9 are modified to fix minUp/minDown constraint violation within the same horizon. Note that the constraint violation across horizon is not fixed, but we checked that the violation is minimal in a run #8
• Enhancement: Derate factors of dispatchable units are now hourly inputs (instead of yearly) and are provided in separate csv files saved as ‘deratef_xxxx’, where ‘xxxx’ denotes the simulation year #9
• Enhancement: Options for applying time limit to optimization is added
• Enhancement: Input and output files are organized into different folders
• Enhancement: Additional hydropower data are provided in input/hydro folder

Release notes

• Third release, date: 15-Apr-2021.
• Archived in Zenodo with doi: 10.5281/zenodo.4688309
• Contributors: Ng Jia Yi, AFM Kamal Chowdhury, Stefano Galelli

PowNet v1.2

Changes since v1.1

• Fix: Logical Constraints at Segment B.9 are modified to fix a bug that used to initialize the binary decision variables (on and switch) for each horizon (e.g., 24 hours).
• Enhancement: Derate factors of dispatchable units – the climate-dependent variable which can vary for each year in simulation over multiple years – are now provided in separate csv files with column heading as ‘deratef_xxxx’, where ‘xxxx’ denotes the simulation year (e.g., 2016, 2017 etc.).
• Enhancement: Storing process of outputs in the solver script slightly reduced the computational requirements.

Release notes

• Second release, date: 08-Sep-2020.
• Archived in Zenodo with doi: 10.5281/zenodo.4020167.
• Contributors: Rachel Koh, AFM Kamal Chowdhury, Stefano Galelli

v1.1

Release notes

• First release, date: 27-Sep-2019.
• Archived in Zenodo with doi: 10.5281/zenodo.3756750.
• Contributors: AFM Kamal Chowdhury, Jordan Kern, Thanh Duc Dang, Stefano Galelli, Arijit Bagchi