You can easily set up InterPSS grid computing solution where there are network-connected servers or PC/laptops, as shown in the above diagram. All computers, participating in the grid computing, should be located in the same subnetwork. Different operating systems, such as Windows, Linux, McOS, could be used, as far as there is a Java Runtime (Java 6.0) installed on each machine.
When install InterPSS, please make sure that the Grid Computing Library is included in the installation. After the installation, use the <InterPSS Install>/runGrid.bat batch file to run InterPSS in Grid Computing mode. You may need to make changes to reflect your Gridgain installation location in the runGrid.bat file, as follows:
REM
REM set Gridgain installation location
REM
set GRIDGAIN_HOME=<Graingain installation location>
You need first to start one or more Gridgain agents, then start InterPSS by running the runGrid.bat. After starting InterPSS, you can check if your grid computing environment is set up properly by selecting the Tools/Debug/Test Grid Environment.
Please note: The guide is based on the following versions:
InterPSS : version 1.4, available from SourceForge
GridGain : version 1.6.1 (please contact us support@interpss.com if you cannot find this version)
After the installation, use the <InterPSS Install>/runGrid.bat batch file to start GEditor in Grid Computing mode. You may need to make changes to reflect your Java and/or Gridgain installation location in the runGrid.bat file, as follows:
set GRIDGAIN_HOME=<GridGain installation location>
"C:\Program Files\Java\jre1.6.0\bin\java" -> your Java installation location
You need first to start one or more Gridgain agents, then start GEditor by running the runGrid.bat. After starting GEditor, you can check if your grid computing environment is set up properly by selecting the Tools/Debug/Test Grid Environment.
You may get the following message, if the environment is not set up properly or there is no grid node running in the environment.
You have to run the same version GridGain across all nodes (Master and Remote). Otherwise you might get the following error message:
Remote node has different: build version
Local node: 1.6.0-16102007
Remote node: 1.5.1-13082007
InterPSS provides a GUI to allow you to select a remote node and distribute a simulation case to run remotely. Please note, you can only run one simulation case one time. Therefore this approach is more for testing purpose.
Select Run Loadflow and launch the Loadflow Analysis dialog. You will then see the Enable Grid Computing checkbox enabled, if your grid computing environment is set properly. After selecting the checkbox, the available Grid nodes in your Grid Computing environment will be discovered and listed in the Select Grid Node dropdown list. At this time you can select any available grid node to distribute your Loadflow job.
Click the Run button to run InterPSS grid computing. If you are running the grid job on a particular grid node the first time, after starting the environment, it may take a while to get Loadflow calculation results back. The reason is that the compiled InterPSS simulation engine code, a set of Java class files, has to be deployed through the network to the remote grid node. Click the Refresh button to discover new grid node(s) and refresh the available grid node list.
Please note:
Loadflow Reuslts
The Loadflow result indicates that the computation is performed at the “Remote Logical Node – 1”. Also, you will see Loadflow logging information printed out in the remote Gridgain agent output console, as follows:
You can distribute a typical InterPSS transient stability simulation case to a remote grid node and perform the simulation remotely. The simulation result will be sent back to the master node through messages for display purpose.
Please Note: Currently you are limited to use only InterPSS built-in controller models for the grid-based distributed simulation. We will provide detailed documentation in the future to guide user to package user-defined controller models to make them ready for the grid computing purpose.
Run Simulation Case Remotely
Assume you are working on a transient stability simulation project, for example, the samples/DStab-5Bus.ipss project in the sample workspace. One or more grid agent(s) are running in your computer sub-network. . Check the Enable Grid Computing and select an available remote grid node to distribute the simulation.
Transient stability simulation result will be sent from the remote grid node to the master node real-time through messaging while running the simulation. This may increase your computer network traffic and slow down the simulation process if too much simulation information is sending back, and the simulation power network is quite large. By selecting the Output State/Variable Filter checkbox and clicking the Output Filter tab, you can select a sub-set information to be sent back for display. In the following example, only machine states at Bus 0004 and bus states at Bus 0001 are sent back to the master node for display.
While running transient stability simulation at a remote grid node, you should see the log information printed in the remote grid node console as follows:
Display Simulation Results
Simulation result at a remote grid node are sent through messaging back to the master node real-time while running the simulation. The result is saved into InterPSS database at the master node. After completing a remote simulation run, right click a bus to plot the simulation result.
The following is a sample plot of a bus voltage response. You should see no difference between running the simulation case locally node or at a remote grid node.
To evaluate the transient stability of a real-world power network requires multiple simulation runs. InterPSS grid computing allows you to distribute a set of power system simulation cases to remote grid notes and perform simulation in parallel. Simulation results are sent back to the master grid node through network messages for display purpose. Currently, you have to use InterPSS Custom Scripting Run option and XML as the scripting language to perform multiple case run.
Follow the following steps to run the IEEE-14Bus sample case:
Follow the following steps to run the DStab-5Bus sample case: