Hello Guys
We’re considering using PMUs with different sampling rates within the same openXDA environment.
Do you believe it’s possible to process COMTRADE files from PMUs with different sampling rates in openXDA without issues?
I ask this because each COMTRADE file specifies its own sampling rate, but openXDA appears to enforce a single frequency setting in its global configuration.
Thanks for the help,
Regards
Hello Renan_Pires,
To be clear, since you mentioned PMUs, are you aware that openXDA is primarily intended for processing point-on-wave data, not synchrophasor data?
As for the system frequency setting, that value specifies the number of cycles per second on the power system you are monitoring. This value should be 60 for US power systems, 50 for European power systems, etc. The sampling rates of your devices are handled independently.
Thanks,
Stephen
Thanks for the quick response
Apologies for the mix-up.
Since COMTRADE files and synchrophasors come from the same equipment, I ended up conflating the two in my question.
Locally, we have two servers with independent openXDA installations—one for COMTRADE files at 50Hz and another at 60Hz.
In an effort to unify the environments, we started wondering whether it would be possible to have a single setup handling both sampling rates.
From your response, I understand that processing two different sampling rates within the same environment shouldn’t be an issue, since openXDA handles this information individually for each device.
In other words, the system-wide configured frequency (as shown in the settings) does not affect the processing of the files. That was my main concern in cases where the values differ.
Did I understand that correctly? Apologies if I didn’t.
Thanks again
Perhaps I am the one who is not understanding. Please clarify whether 60Hz and 50Hz are sampling frequencies or system frequencies. I would think that point-on-wave data collected at 1 sample per cycle wouldn’t be very useful when processed using openXDA. Can you give me a better idea of how you’re currently making use of openXDA?
It does affect the processing of files, but it has nothing to do with your device’s sampling frequency, which for point-on-wave data is typically much higher than 60Hz or 50Hz.
First of all my apologies.
Both 50Hz (used in Europe and some South American countries) and 60Hz (used in Brazil) are system frequencies. We use openXDA + PQ Dashboard + openSEE to process and visualize oscillography data triggered by threshold violations in voltage, frequency and power quality.
50Hz enviroment:
60Hz enviroment:
Currently, we have a separate server for each system frequency.
The idea is to explore whether it would be possible to handle both system frequencies within a single server.
Okay, I understand now.
The way openXDA was designed, it’s not possible to merge these servers into one, and it looks like a redesign might be a little bit tricky. The most obvious approach would be to move the SystemFrequency setting into the Meter table. Many usages of the SystemFrequency setting would be trivial to fix up, but there are two usages you’d have to watch out for.
The first is in the data readers which load data from data files into a MeterDataSet. This step comes first, before the ConfigurationOperation has had a chance to find the corresponding Meter record in the database. Both the PQDIFDataReader and the SELEVEDataReader make use of the SystemFrequency setting to parse data from these file formats. Currently, it sounds like you’re only processing COMTRADE files so perhaps you can simply break these two readers. However, if you think you might use these formats in the future, you can probably get away with simply removing the code from PQDIFDataReader that depends on system frequency. You won’t be able to do that with SELEVEReader so you’d need to come up with something else there.
The other problem is going to be on the visualization side. The openXDA engine, for the most part, processes event files in isolation without correlating it with other data. But many of the visualizations are designed to look at the system in aggregate, and the assumption has always been that the system frequency is the same between meters. So you may encounter situations where the code loads multiple waveforms from different meters and assumes their system frequencies are the same. You’d have to decide whether it even makes sense to compare waveforms from the 60Hz system to the 50Hz system and, if so, what adjustments you’d have to make to the algorithms that depend on system frequency.
Thanks,
Stephen
