nikhilgupta10
commented
7 years ago nikhilgupta10 imported these comments from Sourceforge: "dchassin":Let me take a look at the real-time aspect of this and if it looks like it more a powerflow issue, then I'll send to Frank.,
"dchassin":The players go from 5/19 to 5/22 but today is 7/3. For a real-time time simulation the player files have no data to provide that matches the current time. That's just not how they work. When it loads them there is nothing to do and the system probably is settling into a steady state.
As far as I'm concerned the players are working as designed. You're just trying to use them in a way (albeit an interesting way) that is far outside the specifications. Sorry about that but not much can be done for you using players and I'm not going to recommend changing players for this situation.
One alternative approach that might work for you is the set up the player as a loop so that it always has something to read and then just append that to a single line that has the current time and initial condition. Otherwise you're going to have to provide the data some other way, i.e., push it in using the server mode or create a runtime class that does all this for you.
If you choose to do the latter, you might want to post the code in the forums or as an example in the repo because that is a use case that is likely to occur again. It's just not one players were ever meant to address.,
"dchassin":- status: new --> assigned
- assigned_to: Jason Fuller
- Resolution: none --> wontfix
- Component: none --> tape
- Version: --> trunk
-
Type: defect --> documentation ,
"dchassin":I'm recommending we document this issue and not attempt to change player for this use case.Hopefully power_engr will post a solution that will help other in the future. ,
"tesswilliams":A few clarifying notes after a conversation with power_engr:
- This question concerned code in Ticket 716, not a released version.
- In 716, there is a new feature, realtime rate mode, where one can run at realtime rate but at an arbitrary timestamp, making it possible to use players while in realtime mode. Dave's response only applies to the traditional version of realtime mode, where simulation time is tied directly to system clock time.
-
It is possible that the issue described is due to a platform compatibility problem. Try the system on a windows machine and see if that helps. Please let us know if it does solve the problem, as it is something we'll want to address before we push these features to a release version. ,
"dchassin":Ok. I'll run the test on code/ticket/716:4711 and post my feedback on ticket:716.,
"dchassin":Can I get a confirmation that this is really for ticket 716 and not ticket 797?,
"dchassin":Question: do you look for the output while gridlabd is running or after it has exited? The reason I ask is I think recorder doesn't flush its output buffers very often. I think there is a flush feature in collector but as far as I know none such in recorder. This may be needed for you to see output before exit (or a lot of data is generated).,
"dchassin":FYI, ticket 850 tracks the status of adding flush to recorders.,
"dchassin":While testing ticket 850 code I noticed that the code posted to exemplify this problem uses group_recorder, which does support flushing. See http://gridlabd.me.uvic.ca/gridlabd/wiki/index.php/Flush_interval for details.,
"dchassin":- status: assigned --> closed ,
"dchassin":At this point this report is considered a user error unless further information is reporting suggesting otherwise. The group_recorder has a flush option that remedies the reported problem on the test system at UVic.,
I am trying to get GridLAB-D to run in real-time mode and have run into a set of issues. I am working with the IEEE123 model in which I replaced the static loads with player files, with values that update every 1 second. I also added PV inverters with batteries at all the nodes that have loads, but the problem described below persists even if I remove all the inverters. When I switch to real-time mode (#set run_realtime=1), the group recorders set up to capture the real and reactive power measurements respectively from the triplex meters at each node that has load, stops recording any data. I am attaching a simplified version of the files I am trying to run, but that still exhibits the problem. I am running GridLAB-D 3.0.0-4711 (Hatwai) 64-bit MACOSX RELEASE
The main file in the attached zip file is IEEE123_ts_Ab_RT.glm. Currently real-time mode is off, i.e., line 37 is commented out. If you run this, you should get values in all output files. If you uncomment line 37 and run real-time mode, it runs, but there is no output in some of the csv files. This version reads only two player files, which should provide a variable load at node 1. I have set the start time of the simulation to the first time step in the player files. My player files have a date range of May 19 to June 12th, 2012. My clock settings are clock { timezone PST+8PDT; starttime '2012-05-26 13:45:01'; stoptime '2012-05-26 14:59:59';}
Thank you. ,