PGAz - Power Grid Analyzer

Software

We are actively developing two tools for power system analysis designed for research and educational purposes: (i) PGAz on the MATLAB platform, and (ii) PyGAz on the Python platform, both developed primarily for research and educational purposes. The PGAz project is led by Sorasak Jaingeawkum under the supervision of Tossaporn Surinkaew and Professor Issarachai Ngamroo. The PyGAz project is led by Nanthachai Inonn under the supervision of Tossaporn Surinkaew and Professor Issarachai Ngamroo.

About PGAz

Based on our previous experience, fully open-source development has shown limitations in responsiveness, as new models often remain at a research or conceptual stage and are not available for timely practical use. Such approaches may also be unsuitable for certain application areas and present coordination challenges. Meanwhile, existing open-source and commercial power system analysis tools remain prevalent and can be extended to meet evolving needs. Therefore, PGAz has not yet been released as a fully open-source tool. At this stage, the updated models and algorithms in PGAz are shared exclusively with the core team and collaborators, while the platform remains under active research and development. PGAz provides six core features:

Power Flow (PF)
Continuation PF (CPF) and Optimal PF (OPF)
Quasi-Dynamic Simulation (QDS)
Small-Signal Stability Analysis (SSSA)
Time-Domain Simulation (TS)
Bifurcation Analysis (BA)

Download PGAz

PGAz for educational purpose is available for download below. Please ensure that MATLAB is installed and that all required basic MATLAB packages are available. Installation instructions and an overview of the basic functions and features can be found here.

Download PGAz version 1.2	Release Date: 10th March 2026
Download PGAz version 1.1	Release Date: 3rd November 2025
Download PGAz version 1.0	Release Date: 29th September 2025

PGAz Updates

25 March 2026 — Sparse matrix techniques were applied to the Jacobian matrix in numerical solvers to accelerate PF, CPF, OPF, and TS. An adaptive sparsity pattern based on the actual system topology was employed to reduce computational overhead and improve robustness (e.g., mitigating convergence failures) in large-scale systems, thanks to Dr Nawakun Triruttanapiruk from KMUTNB. Verified using "case13659pegase" and other large systems transcoded from MATPOWER. [Sorasak]
10 March 2026 — Release PGAz version 1.2 for educational and learning purposes, bugs fixed, major improvement in GUI, and several models and test systems included. [Sorasak]
14 February 2026 — Major revisions were made to PGAz GUI, algorithms, model representation in PGAz Library, result logging systems, etc. We also removed demo versions from the website. [Sorasak]
22th December 2025 — Since 20 December 2025, updated versions of PGAz are NO longer open-source or publicly available. Development still continues, but updates and newer versions are shared only within the team and collaborators. For learning and educational uses, we believe that PGAz version 1.1 is sufficient.
7th December 2025 — Added two tutorial videos: (i) the user-defined model and (ii) the PGAz Simulink Library (see Download PGAz). [Sorasak]
1st December 2025 — Improved the PGAz Simulink Library by updating icon colors (blue for DC, black for AC), adding an optional nonlinear TS with the linear TS as default, remodeling the DC fault, and fixing various bugs. [Sorasak]
28th November 2025 — Added tutorial videos for CPF, OPF, SSSA, and TS (see Download PGAz) [Sorasak]
26th November 2025 — Added a tutorial video for PF analysis (see Download PGAz), fixed minor and major bugs, and wrapped up the remaining unfinished/unverified models [Sorasak]
3rd November 2025 — PGAz v1.1 has been released! This update includes several bug fixes, improved solvers and algorithms, and the addition of new and more stable models. Note: TS for nonlinear DAEs is not included yet in this version — it still needs more testing to improve robustness and trade-off between computational time and simulation accuracy [Sorasak]
28th October 2025 — The following major update are made:
- Update and correct the models and solvers for TS so they can handle large disturbances accurately (such as faults, line trips, and breaker operations) by performing time integration on a system of nonlinear DAEs. This replaces the old model, which was based on linearized DAEs and therefore only accurate for small disturbances. With this update, TS can now return more accurate responses under large disturbances. However, TS is available for both linear and nonlinear DAEs with a trade-off between computational time and simulation accuracy
- We are wrapping up these updates with minor and major bug fixes, model/solver verification, and adding several new models. These will be included in PGAz v1.1, a more stable version, coming soon
[Sorasak]
26th October 2025 — Start writing and developing detailed documentation for PGAz [Sorasak]
24th October 2025 — Add bifurcation analysis. Update/remodel the ANN-PSS and ANN-POD by including a washout filter (like that in a conventional PSS), modeling adaptive gain with ANN, and adding filter blocks between hidden layers to better show their dynamics and controller order. Can also adjust the time constants in the model. These models are referred to as "ANN-PSS Type I" and "ANN-POD Type I". [Sorasak]
14th October 2025 — Add the equal-area criterion analysis window after running TS [Sorasak]
8th October 2025 — Develop an algorithm that can perform AC and DC PF simultaneously and effectively to improve accuracy and performance, and revisit and revise all models and solvers based on these conditions. Add the exponential linear unit (ELU) activation function for ANN-PSS and ANN-POD as another activation option. Also apply the concept of multi-slack and multi-area conditions for PF and OPF with new solvers to run more effectively when the system is reconfigured under these conditions. Also verify this as well [Sorasak]
4th October 2025 — Include a tutorial video link (see Download PGAz). Future tutorials on basic functions and features in English will also be accessible through this link [Sorasak]
3rd October 2025 — The following major updates are made:
- Revise the ANN-POD and ANN-PSS by adding a first-order washout filter and a first-order output filter. The states match the number of hidden units. We also look at using long short-term memory networks with reinforcement learning concept for alternatives of these models (thanks to Wendy, N. Jarajarutouth and Tawan, P. Bunyanupong)
- Revisit and reverify PV, EV, BESS, and DC-AC converter models, and incorporate/fix control platforms (thanks to Klao, P. Sripaibulsup and First, P. Dambunma)
- Revisit and reverify FACTS devices, and upgrade control platforms by adding grid-synchronization and resynchronization to support the connection and reconnection of IBRs and renewable sources (thanks to Mai, P. Vongkichpanit and Pu, P. Jarusuttiruk)
[Sorasak]
30th September 2025 — Add options for activation functions in ANN-POD and ANN-PSS, which include ReLU, Tanh, Sigmoid, None (no activation function), Leaky ReLU, hyperbolic tangent, and linear [Sorasak]
29th September 2025 — PGAz v1.0 is officially released! It is compatible with MATLAB 2017 through the latest versions, including the very recent version MATLAB 2025a and 2025b. In < — PGAz v1.0, you will find five core features — PF, OPF, QDS, SSSA, and TS — along with a range of basic and advanced power system components, PGAz Simulink library, and a set of examples. Thanks to Pu, P. Jarusuttiruk for assisting with the rechecking and verification of this version [Sorasak]
28th September 2025 — Start developing Thailand's power system (transmission level) in PGAz as a prototype for future testing and validation [Sorasak]
19th September 2025 — Final check prior to the release of PGAz 1.0.0 [Sorasak]
15th September 2025 — Major fix IBR and converter models: bugs and errors in the IBR model and its controls are found, and are now resolved. An option is also added to include a damping controller in the P loop, the Q loop, or in both PQ control loops of the converter control. [Sorasak]
8th September 2025 — Remodel ANN-PSS and ANN-POD, and introduce a separate ANN training model that simplifies synchronized training of installed ANN-PODs and/or ANN-PSSs. Also, add this model into PGAz-Simulink library [Sorasak]
5th September 2025 — The following major updates are made:
- Add measurement models
- Add EV-specific battery models
- Add simple/basic control platforms for GFM and GFL
- Add seperated MPPT algorithms for wind and PV
- Add Basic MPC and fuzzy logic controllers
- Add separated small disturbances (impulse, step, sine, or user-defined) that can be applied internally to the models
- Review all models to update connection port types and make sure everything works correctly
These developments are still under review and are not included in PGAz 1.0.0 yet, but are expected in the next version. [Sorasak]
29th August 2025 — Now it is time to finish and tidy up the code, fix small and major bugs, debug, and give all solvers, algorithms, and models a final check to get ready for the release of PGAz 1.0.0 [Sorasak]
27th August 2025 — Disturbances like step and impulse inputs are now built as separate object-oriented classes and added to the Simulink library, instead of being embedded in the model. This makes it easier to use different types of disturbances with different models and state variables, and also improves how this disturbance connections are handled in the Simulink library [Sorasak]
24th August 2025 — Enable the QDS to simulate system reconfiguration events by modeling breakers, lines, generators, renewables, IBRs, and other related devices (if applicable) with on/off statuses that operate according to a defined time schedule [Sorasak]
19th August 2025 — Enhance the QDS GUI, and add the 3D plot to visualize PF, CPF, and OPF in three dimensions. Additionally, fix the model’s connectivity in the PGAz Simulink Library [Sorasak]
16th August 2025 — Revisit the wind and solar PV models and improve their dynamics so they can link smoothly with the converter model, verify in SSSA and plan to further check their transient response in TS [Sorasak]
10th August 2025 — Enhance the usability of the user-defined GUI to make it more accessible and easier to extend and develop custom models. [Sorasak]
9th August 2025 — Wrapping up the main GUI for PGAz 1.0.0, and also checking compatibility with MATLAB 2025a [Sorasak]
7th August 2025 — Add a stochastic solar PV power model with cloud cover uncertainty to PV model for QDS [Sorasak]
6th August 2025 — PGAz now includes support for the QDS feature, with the main GUI capable of presenting corresponding results and is suitable for long-term analysis. QDS-related variables, parameters, and functions are updated and added in generator and load models. Additionally, a new 3D visualization for PF, CPF, and OPF results is added in the main GUI to provide a more comprehensive understanding of the results [Sorasak]
4th August 2025 — Add basic randomization functions and key parameter settings into generator and load models to enable QDS analysis; implement a dedicated QDS function and incorporate QDS into the main GUI [Sorasak]
29th July 2025 — Develop dynamic models of GFL control that incorporate three modes—PQ, droop control, and VSG-based—and implement a mechanism for selecting among these modes within the GFL framework. Additionally, verify the correctness and functionality of these implementations [Sorasak]
23th July 2025 — Enhance functionalities and models in PGAz:
- Enable use of CPF and OPF results as initial conditions for SSSA and TS analyses
- Update static models of FACTS devices (SVC Type I and Type II, TCSC, STATCOM, and UPFC), PV and EV systems, and DC-AC / AC-DC / DC-DC converters for PF and OPF studies
- Revisit the state-space equations and formulate the linearized ANN-based PSS model for further use in SSSA and TS.
- Initiate modeling of HVDC and multi-terminal HVDC systems
[Sorasak]
22th July 2025 — Start developing a quasi-dynamic simulation (QDS) module for analyzing slow dynamic behavior. Also revise the representative icons used for model visualization in the PGAz library.
12th July 2025 — Develop a GUI for creating user-defined models.
1st July 2025 — Kick off three main projects:
- Develop FACTS device models for renewable energy systems and IBRs
- Develop solar-powered EV system
- Develop ANN-based PSS/POD for SG/IBRs
27th June 2025 — Add a "user-defined model" feature that allows users to add or define new state and algebraic variables within the original model.
4th June 2025 — The following updates are made:
- Apply the concept of multiple slack buses for GFM IBRs in systems with more than one slack bus. Now, PGAz can solve both the PF and OPF problems under this configuration.
- GFM IBRs can operate as reference units or slack buses in multi-area systems and coordinate effectively with other GFM and GFL resources. The OPF constraints for GFM and GFL units are incorporated accordingly.
- Initiate developing intelligent, neural network-based PSS and POD for SG and IBR.
- Incorporate a learning-based PSO solver to effectively address complex OPF problems with nonlinear constraints arising from the integration of GFL and GFM resources.
23th May 2025 — The following updates are made:
- Revise the models to enable individual selection of OPF constraints.
- Add adaptive algorithms/solvers for more effectively handling nonlinear OPF problems.
- Implement slight improvements, bug fixes, and GUI enhancements for the latest version of the PGAz demo.
9th May 2025 — The following updates are made:
- Add a feature that allows users to define their own input for PSS using state and system variables, with support for wide-area input signals.
- Start developing a POD and integrate its input signal into the control loops of IBRs operating in both GFM and GFL modes.
- Start developing dynamic models for switches and breakers.
- Minor improvements are made to the @Class files to enable easier handling of dynamic models.
8th May 2025 — Add new OPF constraints for GFL and GFM units. Start developing a feature that allows GFM units to act as slack buses, also try to develop model that can support both single and multiple slack bus configurations.
5th May 2025 — Develop an IBR model with a selectable control mode (grid-forming (GFM) or grid-following (GFL) mode), and verify it in PF (GFM and GFL properties), SSSA, and TS (Effects of GFM and GFL controls).
30th April 2025 — The following updates are made:
- Add feature that can compute PQ injection from IBRs, including compatibility with EVs and future integration of other devices.
- Introduce two additional methods for solving OPF, i.e., Bee Algorithm and Ant Colony Optimization for flexible system-specific optimization.
- Improve PF, OPF, SSSA report formats, GUI interfaces, and using command-line usage.
26th April 2025 — PGAz demo version for small-signal stability analysis and time-domain simulation, is now released! PGAz demo is compatible with MATLAB 2024. We hope you enjoy using this demo and exploring its features. Your feedback is invaluable to us – please share your thoughts!
1st April 2025 — Fix bugs and prepare for the release of a new Demo version with SSSA.
17th March 2025 — Develop GUI for SSSA and TS.
7th March 2025 — Finish verifying SG, AVR, TG, PSS, and IBR models. Integrate SG, AVR, TG, PSS, and IBR into PGAz-Simulink library. Add SSSA and TS functions in PGAz GUI.
28th February 2025 — Model IBR in grid-following mode, add a phase-locked loop (PLL) for IBRs. Also verify PLL in SSSA and TS.
25th February 2025 — The following updates are made:
- Finish verifying SG, AVR, TG, and PSS in SSSA and TS
- Initially, develop the first-order delay for the PSS model.
- Create a new delay @Class used for controller @Class.
- Verify the 4th-order IBR in SSSA and TS
- Start developing a new @Class for a symmetric fault model for TS
- Start verifying SG, AVR, TG, IBR, and PSS models in multi-machine systems
22th February 2025 —Develop automatic tuning algorithms that enable users to determine optimal PSS parameters.
18th February 2025 — Add PSS Tyep II and initially verify in SSSA.
16th February 2025 — Add AVR Type I and Type II and verify them in SSSA.
14th February 2025 — Develop a consecutive perturbation function applied to the state/output/other variables to verify the SSSA model, observe the system's time-domain response, and perform an initial stability assessment.
12th February 2025 — Verify IBR model in standard test systems.
11th February 2025 — Incorporate dynamic model of a 4th-order IBR for SSSA and TS.
10th February 2025 — Add three additional numerical integration solvers (Crank-Nicolson, Adams-Bashforth, and RK45) methods for solving time-domain response.
10th February 2025 — Add three numerical integration solvers (Euler's, 4th-order Runge-Kutta, and ODE45 methods) for solving time-domain response.
8th February 2025 — Start verifying synchronous machine models, AVR, and TG and comparing with PSAT and PST.
4th February 2025 — Verify synchronous generator models in single-machine infinite-bus, and two-are four-machine systems.
29th January 2025 — Update small-signal models for 2nd, 3rd, 4th, 5th, 6th, and 8th-order synchronous generators with AVR and TG models. Enhance compatibility for integration with the PGAz-Simulink library. Generate a new @Class to analyze SSSA and work with command-line usage.
15th January 2025 — The following updates are made:
- Incorporate dynamic modeling of generators and other components, and start developing the SSSA function to evaluate system dynamics and stability.
- For OPF analysis, new constraints considering microgrid components are added.
- Resolve existing bugs and improve the visualization of the models.
14th January 2025 — Initiate modeling new @Class for microgrid components such as controllable load, hydrogen electrolyzer, solar-powered rooftop load, distribution lines, and solar-powered electric vehicles.
12th January 2025 — Developing PGAz-Simulink library has been completed, including basic models for simulations of PFs and OPFs. PGAz now features a user-friendly interface and graphical network editor in Simulink.
11th January 2025 — Tasks from 10th January 2025 have been completed. Currently verifying their correctness and checking for any errors (if any) to ensure that the conversion process is accurate and compatible.
10th January 2025 — Develop a script to convert the [.slx] file (PGAz-Simulink library) back into [.m] file for further analysis within the PGAz GUI interface.
4th January 2025 — More updates are made:
- Initiate modeling on the DC side and converters to enable analyses using PGAz, which can account for effects on both the DC and AC sides.
- Integrate of these models into the PGAz-Simulink library for user interface.
4th January 2025 — The following updates are made:
- Utilize @Class to create the PGAz model.
- Create a dedicated PGAz domain.
- Create PGAz-Simulink library
- Enable seamless connection of PGAz components in Simulink.
1st January 2025 — Happy New Year 2025!
22nd December 2024 — Fixed bugs in PF constraints that affected results in terms of reactive power and resolved issues with the interpretability of PF results, also continue developing the PGAz Simulink library for the user interface.
10th December 2024 — Develop a PGAz data conversion tool to automatically transform MATPOWER data format into PGAz format, done.
9th December 2024 — Develop and integrate models of inverter-based resources (IBRs), including wind generators, solar photovoltaic systems, electric vehicles, and battery energy storage systems, for PF analysis.
4th December 2024 — Initiate the development of the PGAz library to assist users in modeling systems and their components in Simulink.
1st December 2024 — PGAz demo version is now released for PF and OPF analyses! PGAz demo is compatible with MATLAB 2024. We hope you enjoy using this demo and exploring its features. Your feedback is invaluable to us – please share your thoughts!
30th November 2024 — Currently fixing bugs before releasing the PGAz demo version with the PF feature. Stay tuned for the upcoming release!
20th November 2024 — Converting large-scale systems, including 1354_peg, 1888_rte, 3022_goc, 4917_goc, and 6515_rte from MATPOWER, into the PGAz format, followed by performing PF and OPF verification on these systems.
Mid of November 2024 — Released PGAz Part I with PF and OPF features, enabled command-line usage for PF and OPF, fixed related bugs, and initiated the development of a custom solver with constraints for PF and OPF problems.
End of October 2024 — Other PF and OPF bugs have been fixed.
Mid of October 2024/2 — Verified PF and OPF against IEEE benchmark test cases successfully completed.
Mid of October 2024/1 — The IEEE benchmark data has been converted and used to validate our PF and OPF solutions.
October 2024 — Optimal Power Flow (OPF) was initially developed for PGAz, with both minor and major bug fixes implemented. Furthermore, ten fundamental constraints have been formulated for OPF problems, with alternative approaches provided for selecting these constraints.
September 2024 — Various solvers have been developed to address the PF problems, and they have been successfully implemented.
August 2024 — Power Flow (PF) was initially developed for PGAz.
End of July 2024 — Persistent errors and bugs occurred repeatedly, causing ongoing issues.
Mid of July 2024 — Developing a Graphical User Interface (GUI) of PGAz started here
July 2024 — Kick off the PGAz project!

About PyGAz

Owing to its accessibility, versatility, costless, and open-source nature, Python has become a leading programming language for developing advanced tools across engineering, education, and scientific research. We are developing the Python-based Power Grid Analyzer (PyGAz), for research-oriented power system analysis. PyGAz is developed using Python 3.x. The PyGAz project is under ongoing development.

Download PyGAz

PyGAz is currently under development and will be available soon.

Download PyGAz

PyGAz Updates

26th March 2026 — Added new features to PyGAz, such as CPF and PV curve analysis, also improved the GUI and overall system, fixed bugs, and started including basic dynamic models (i.e., SG and start developing IBR model) with current injection coupling, developed disturbance models, including abrupt changes within devices, faults, and tripping events, and implemented time-integration methods such as RK4, recheck the robustness of solvers, etc. Big thanks to P'Nui Sorasak for reviewing, providing ideas, and guiding the development. [Nonthachai]
18th March 2026 — DC PF validated; finalization and benchmark testing next, followed by DC OPF development. [Nonthachai]
11th March 2026 — PyGAz PF module is now completed, including verification with standard IEEE test systems, incorporation of a DC PF formulation and basic constraints and limits, as well as basic solvers. Also update on OPF. [Nonthachai]
6th March 2026 — The PyGAz project has undergone a major revision, including significant improvements to its foundational framework, structural design, GUI, and GNE, with development continued by Nanthachai. [Tossaporn]
26th October 2025 — Develop small-signal models for SGs and their associated control devices, such as AVR, TG, and PSS. Try to develop the models and couple their dynamics to the system through PQ injections based on the system admittance matrix, now done. Again, revisit static PF and OPF models and solvers to ensure their accuracy, flexibility, and correctness [Tossaporn]
27th September 2025 — I decide to go back and fix minor bugs in PF/OPF and ensure that the solvers are robust enough to handle problems fairly without bias during implementation. [Tossaporn]
1st September 2025 — Begin developing the OPF and remodel the related system models to support OPF calculations. Fix minor bugs in how PF calculates the objective or losses, and generalize this for all solvers [Tossaporn]
21th August 2025 — Recode major models and functions into an object-oriented structure, improve PF format export, standardize the PF objective, and generalize the objective function for the PF problem [Tossaporn]
15th August 2025 — Run into cascading errors during the full object-oriented programming conversion, pause this for now, and go back to updating the PF solver, making other solvers more robust by adding an initial-guess algorithm, and adding supervised machine learning to the new PF solver [Tossaporn]
14th August 2025 — A good time to refactor PyGAz into an object-oriented structure and plan for the future model, defining all core classes, enhancing solver robustness, fixing bugs along the way, and testing incrementally. [Tossaporn]
12th August 2025 — Before moving on, to make things more reusable and easier to manage, I think I should organize the functions and classes into a proper object-oriented design—it will really help with future development of PyGAz. So, it is better to fix the current issues and set everything up using object-oriented programming for now [Tossaporn]
10th August 2025 — PF analysis in PyGAz is now complete! Still updating in case any bugs occur. Running PF in PyGAz now has 4 basic (GS, NR, FD, IPM) solvers, 3 stochastic (PSO, GA, SDE) solvers, and 1 intelligent-based solver (Pi-ANN, thanks to the super handy “torch” library from PyTorch). Tested and verified from IEEE 3-bus all the way up to IEEE 300-bus systems. Next steps: will improve the stability of PF execution, exporting PF results, develop a conversion function for importing systems from other sources/tools, and then proceed to OPF [Tossaporn]
9th August 2025 — Run into some PF issues while trying different solvers. Thorough verification is essential, normalizing the PF objective greatly improved performance and is now complete. However, the parameters for each solver still require manual tuning, which can be somewhat subjective [Tossaporn]
8th August 2025 — PF analysis in PyGAz is partly done. PF are tested susing three basic algorithms, Gauss-Seidel, Newton-Raphson, and Fast Decoupled Load Flow, verified in IEEE 3-bus, 5-bus, ... up to 57-bus systems. Still, there is some more works to do to make them run more smoothly and robustly. The main GUI still runs into a few problems—like not being able to export results properly—and overall stability needs more polish. Also thinking ahead, we plan to develop a graphical network editor with a user-friendly interface [Tossaporn]
6th August 2025 — The initial version of PyGAz's main GUI is complete. Updates will be made regularly as core functions are integrated. At this stage, I think it is applicable to focus on and revisit developing basic simulation features such as PF, OPF, QDS, SSSA, and TS [Tossaporn]
3rd August 2025 — Started developing and designing the main GUI for PyGAz, [Tossaporn]
30th July 2025 — Kick off the PyGAz project! [Tossaporn]

Subscriptions

PGAz and PyGAz are supported by KMITL.
PGAz and PyGAz actively welcome partnerships, sponsorships, and funding supports.