IEEE Power & Energy Magazine - March/April 2020 - 65

The server and clients provide only channels
for communications, while it is left to the modules
to specify the data formats for data exchange.
connections and exchange data generically, or more specifically, for
✔✔ connecting to a data server at a specified address
✔✔ querying the names of modules connected to the same
server
✔✔ broadcasting or sending data to specified clients
✔✔ querying data availability and synchronizing data.
In particular, the exchanged data in the LTB have
been divided into parameter and time-stamped data, on
which different streaming approaches have been applied.
Parameter data include the device parameters in the
loaded test system as well as the parameters for algorithms in modules. Time-stamped data are the values for
the variables a module produces for a certain simulation
time. Typically, parameter data are requested a limited
number of times, whereas the time-stamped data are produced continuously. Therefore, the downstream module
(the data consumer) is allowed to initiate the query of
parameter data on demand, and the upstream module (the
data producer) is allowed to initiate time-stamped data
streaming when available.
The LTB provides a handshake mechanism for requesting time-stamped variables. Before the simulation starts,
downstream modules are required to identify themselves to
the upstream modules, which will then respond with the list
of variables it can provide during the runtime. The downstream modules will respond with the indices of the variables they need to complete the handshake. During the simulation, the upstream modules will send the requested data to
its downstream modules in an array that has the requested
order accompanied by a time stamp.
The workflow summarized in Table 1 is applied to integrate the decoupled modules for systematic simulation.
Compared with a typical power system simulator, the LTB
requires additional steps to initiate the data server and the
modules in sequence. The initialization step consists of
three substeps:
1) self-initialization of modules for internal states
2) querying parameters provided by the upstream module
3) handshaking with the upstream module for variables.
When the initialization is completed, the simulator runs the computation, outputs data, and progresses
the simulation time. Modules will also enter the stage of
active calculation and data exchange once the data arrives.
The program execution will automatically exit if the predefined simulation time is reached or an exception occurs
in any module.
march/april 2020

A Demonstration Case Study
This section showcases studies for wide-area, closed-loop
control on the LTB. Data streaming from the modules is
visualized in a web-based tool called LTB-Web, which renders plots and animations in the user's browser. LTB-Web
provides a systemic overview of the base electrical quantities, including voltage magnitude, voltage phasors, and
frequency, on top of a geographic view. It also provides
the reload and replay function for comparing recorded test
cases and observing the control differences simultaneously.
The controller developed and tested is a wide-area damping controller (WADC) for damping interarea oscillations.
The controller utilizes the measurements for computing additional reactive power control signals for actuators, which are
wind generators in this case study. In the event of actuator
failure or unavailability, the control allocator can dynamically reallocate the control signal to the remaining controllers. The block diagram for the feedback loop with measurement-based control allocation is depicted in Figure 5.

Effectiveness of the Wide-Area
Damping Controller
The WADC has been integrated and tested as a MATLAB
module in the Western Electricity Coordinating Council
(WECC) test system, using ePHASORsim and the custom
OpalAPIControl Python interface. The test system is a modified WECC 181-bus system to include a high wind penetration rate (20% by energy). The system consists of 31 conventional generators with a total capacity of 48.49 GW and
table 1. The execution workflow for module
integration in the LTB.
1: Start the data server process.
2: Start the module processes.
3: Initialization
a) Begin the module self-initialization.
b) Query parameters from the upstream modules.
c) Perform handshake with the upstream modules.
4: Start the simulation.
a) The simulator calculates numerical integration.
b) The simulator and modules exchange data for each step.
c) The simulator progresses the simulation time at wallclock speed.
5: Exit when
a) the end of simulation is reached or an exception occurs.
ieee power & energy magazine

IEEE Power & Energy Magazine - March/April 2020

Table of Contents for the Digital Edition of IEEE Power & Energy Magazine - March/April 2020