EddyPro 4.1 Crashes on detrend with over-range data

Environmental Forum Forums EddyPro EddyPro 4.1 Crashes on detrend with over-range data

Tagged: , , ,

This topic contains 8 replies, has 3 voices, and was last updated by  gerardo 5 years, 10 months ago.

Viewing 9 posts - 1 through 9 (of 9 total)
  • Author
    Posts
  • #1431

    JonE
    Participant

    Hi,

    I persistently get EP4.1 to crash after processing one or a few files, which often (always?) have some bad over-range gas values. It crashes on ‘detrend’, both in Advanced and Express Mode, on two different WinXP machines. Typical Console Output up to crash is:

    file(s): ..TOA5_AngliaFarmFlux_fast20Hz_2013_02_20_1400.dat

    Total No. of samples in file(s): 35994

    2013-02-26 21:45:03.046: processing new flux averaging interval

    Number of samples of current period: 35994

    Samples range in raw file(s)..

    start= 1, end= 35994

    Part 1 of 1. Date/time: 2013-02-20 14:30

    Calculating statistics.. done.

    Raw level statistical screening..

    Spike detection/removal test..

    Spike detection/removal test.. done.

    Amplitude resolution and/or dropouts test..

    Amplitude resolution and/or dropouts test.. done.

    Absolute limits test.. done.

    Skewness & kurtosis test.. done.

    Done.

    Verifying time series integrity.. done.

    Performing stationarity test.. done.

    Re-calculating statistics (2)..

    Re-calculating statistics (2).. done.

    Cross-wind correction.. done.

    Re-calculating statistics (3).. done.

    Re-calculating statistics (4).. done.

    Performing tilt correction by: double_rotation

    Done.

    Re-calculating statistics (5).. done.

    Compensating time lags.. done.

    Re-calculating statistics (6)..

    Re-calculating statistics (6).. done.

    Detrending..

    –> Windows crash dialogue —> EddyPro Crash dialogue


    Before this file it successfully processed six 30 minute files. With different input files the crash occurs on the second file.

    I think the problem is to do with very high gas values when the LI7500 windows are wet – but I think earlier versions of EP coped OK with out of range gas values. Clearly EP should cope with these bad data and not crash.

    Crash behaviour is the same with all the different Advanced detrending options.

    Is there a work around or a fix for this please? Thanks for the nice software (when it doesn’t crash).

    Cheers

    Jon.

    #1721

    JonE
    Participant

    … test files for the above crash scenario attached.

    Jon

    #1722

    JonE
    Participant

    OK using the above setup files and these two raw data files attached – the first file processes Ok and then the second field crashes EP just before the detrend:

    file(s): ..TOA5_AngliaFarmFlux_fast20Hz_2013_02_20_1400.dat

    Total No. of samples in file(s): 35994

    2013-02-27 0:04:18.218: processing new flux averaging interval

    Number of samples of current period: 35994

    Samples range in raw file(s)..

    start= 1, end= 35994

    Part 1 of 1. Date/time: 2013-02-20 14:30

    Calculating statistics.. done.

    Raw level statistical screening..

    Spike detection/removal test..

    Spike detection/removal test.. done.

    Amplitude resolution and/or dropouts test..

    Amplitude resolution and/or dropouts test.. done.

    Absolute limits test.. done.

    Skewness & kurtosis test..

    Skewness & kurtosis test.. done.

    Done.

    Verifying time series integrity.. done.

    Performing stationarity test.. done.

    Re-calculating statistics (2).. done.

    Cross-wind correction.. done.

    Re-calculating statistics (3).. done.

    Re-calculating statistics (4).. done.

    Performing tilt correction by: double_rotation

    Done.

    Re-calculating statistics (5).. done.

    Compensating time lags.. done.

    Re-calculating statistics (6).. done.

    —> CRASHES HERE (1st file processes OK)

    Thanks

    Jon.

    #1723

    gerardo
    Participant

    Hi Jon,

    thanks for your inquiry. I downloaded the files and will post a feedback as soon as I can (probably in a few days).

    Thanks,

    Gerardo

    #1724

    gerardo
    Participant

    Hello Jon,

    are you using the latest release of EddyPro (ver.4.1)?

    I just processed your 2 raw files with 4.1, using your project file, only changing in/out paths and with various detrending options, and didn’t incur in the crash..

    Not sure how to proceed. Any suggestions?

    Gerardo

    #1726

    JonE
    Participant

    Hi thanks for the fast response. Yes I’m using the latest eddy pro 4.1, downloaded yesterday. did you test under Win XP? Eddy pro 4.1 seems a bit unstable for me on Win XP. today I opened a ep project that ran ok on 4.0 & ep crashed viewing the raw file column selection setup, when I scroll to the right, that reminded me it did the same editing the above setup yesterday, in the same way, scrolling to the right to see the input columns of the raw data file in the setup.

    is it possible to download version 4.0 as I don’t have the install file anymore, & it’s not on my current PC. this might be the best way forward to get data processed for next week.

    thanks

    Jon.

    #1727

    gerardo
    Participant

    Hello Jon, please contact me by email (gerardo.fratini@licor.com). We can continue the discussion by email and post here what we find at the end.

    Thanks,

    Gerardo

    #1837

    Marta
    Participant

    Hi Gerardo,

    I would like to know if the problem that Jon had with the absolute limits has been solved. I’m using EddyPro 4.1 as well, and with some set of data it crashes if the “filter outranged values” option is flagged while it processes all the data without problems if the option is not flagged. Which is is the risk if I do not flag this option? I presume I simply will have to filter outraged data by myself in a postprocessing, isn’t it?

    Thanks in advance,

    Marta

    #1838

    gerardo
    Participant

    Hi Marta,

    sorry for the late reply, I lost the RSS feed for the Forum for a while.

    Yes, the problem was fixed and the fix will be available in 4.2 (coming soon, although I have been saying that for a while now..).

    Of course, if you don’t filter out outranged values you may incur in spikes in the flux time series, which you would have to eliminate in post-processing. In some cases though, you may simply get wrong fluxes which look realistic: those cases would be hard to spot.

    Please bear with us as we finalize this long promised 4.2.

    Thanks,

    Gerardo

Viewing 9 posts - 1 through 9 (of 9 total)

You must be logged in to reply to this topic.

Comments are closed.