PSLTDSim 0.1.0

Power System Long-Term Dynamic Simulator

PSLTDSim = Power System Long-Term Dynamic Simulator

Purpose: Use Python and PSLF to simulate long-term power system dynamics.

Examples of long-term dynamics:

1. A 20 minute wind ramp that forces a governor action response that in turn affects the area control error (ACE) which is later corrected by automatic generator control (AGC [a.k.a. load frequency control or LFC]).
2. A Generator is tripped in an area that is importing power. Governor action from other areas will respond before AGC works to restore tie-line balance.
3. A multi-hour simulation of a daily load cycle and forcast used to assure AGC algorithm operates within NERC mandates.

Warning: Work in progress - everything may be broken. Nothing guaranteed.

To install package from git repository: "pip install -e ." from this directory with administrative privleges. If python 3 32 bit is not default python: "py -3-32 -m pip install -e ." Requires IronPython (32 bit) to be on system path.

What this code does:

-Uses an agent-based approach to power system modeling.

-Creates a time sequence of power flows while accounting for combined system frequency, all generator mechanical powers and associated staes along with balancing authority states.

-Uses the PSLF system model format and power flow solver.

-Communicates with PSLF via Ironpython.

-Python 3 communicates with Ironpython via AMQP.

-Reads PSLF dyd files and creates python equivalents of one governor (tgov1).

-Implements a generic governor for models that are not fully created in Python (i.e. anything besides tgov1).

-Allows for noise, step, and ramp type perturbances

What this code doesn't do:

-Use definite time controllers to change the status of agents in the system according to user programmable logic.

-Handle tripping of any generator (power-flow solution mysteriously diverges).

Notes:

• Requires GE PSLF Python API and valid PSLF licencse. (sorry)
• To workaround GE API issues: Erlang, RabbitMQ, and Ironpython are required.
• Additional Python packages include:
  • Python 3: Scipy ( for Numpy and Matplotlib) and Pika (for AMQP)
  • Ironpython 2.7 32 bit: Pika (Note: Ironpython must be 32 bit to work with GE PSLF middleware)

Recent Progress:

• Code refactored to:
  • Utilize Py3<->AMQP<->IPY<-> PSLF workarounds
  • Allow for generic automation using agent current value dictionaries
  • Enable easier code packaging
  • Clarify code structure
• Code creates python mirror of PSLF areas, buses, generators, loads, branch sections, and shunts.
• Code has been written to:
  • Check accuracy of mirror
  • Ignore islanded objects
  • Allow for optional Vsched or Vinit of bus Voltage setting
  • Parse .dyd files (more than one)
  • Parse .ltd files
  • Handle H (inertia)
  • Add noise, step, and ramp perturbances to most any agent value
  • Set area wide governor deadbands
  • Incorporate balancing authority actions into simulation
  • Parse EIA data for multi-hour demand/forcast simulation
  • Distribute changes in Pacc according to inertia
  • Exectute Combined Swing equation
  • Step dynamic modles (tgov1 and genericGov)
  • Log values of interest
  • Generate data dictionary
  • Import/Export Mirror or data data dictionary via shelve
  • Export Model or data data dictionary from Python 3.x to MATLAB .mat file
  • Generate terminal output
  • Generate plots via matplotlib
  • Run multiple simulations in a batch style and handle errors in a non-show stopping way
• MATLAB scripts created to verify and validate simulation outputs
  • Frequency, Pe, Pm, Q, and Voltage Magnitude and Angle compared to PSDS.
  • Branch power flow of P, Q, and I.
  • Accounts for loss of system inertia and change in system losses.