Indexes of the counted cycle range

[addarnr]

Hello,

I am in search of a rainflow counting algorithm in Python that provides the indexes of the counted ranges (similar to the rainflow counting in Matlab) also. Would this implementation rainflow() (rainflow 2.2.0 by PyPI) provide the indexes? If not, is there a way to modify this rainflow() code to obtain the indexes? Are there any other Python based rainflow implementations that provide indexes? Thank you!

[iamlikeme]

Hi! The current version does not provide indices, but it seems fairly straightforward to implement. Are you willing to implement this feature and submit a pull request? Or at least a test case? I am not sure about other Python based implementation of rainflow counting. You would have to research yourself.

[addarnr]

Hello, Yes, I am willing to test a case.
Thank you for your support!

[addarnr]

Hello, please let me know what to do at my end. Thank you!

[iamlikeme]

Hi again. It would be great if you could provide a sample time series and the expected result including indices. Some sample time series are already included in the unit tests, e.g. this one. Perhaps you could generate the expected result using Matlab?

[addarnr]

Hello, yes i will do that and get back to you.

Thanks!

On Thu, Apr 9, 2020, 1:07 PM Piotr Janiszewski notifications@github.com wrote:

Hi again. It would be great if you could provide a sample time series and the expected result including indices. Some sample time series are already included in the unit tests, e.g. this one https://github.com/iamlikeme/rainflow/blob/b1990b4e0c9189479df0b5f720c842e8a9bc66a1/tests/test_rainflow.py#L6. Perhaps you could generate the expected result using Matlab?

— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub https://github.com/iamlikeme/rainflow/issues/31#issuecomment-611672617, or unsubscribe https://github.com/notifications/unsubscribe-auth/AIFGUYD5YZ7533AZYUBRLFDRLYFGDANCNFSM4MCMKPXQ .

[addarnr]

Hello, I tested (2) series, one from the link you sent me and one my own. The resulting cycle counts, ranges, means, start indices, and end indices are attached.

Thank you!

SERIES 1

series1 = [0, -2, 1, -3, 5, -1, 3, -4, 4, -2, 0];

Count    Range    Mean    Start_Index    End_Index
_____    _____    ____    ___________    _________

 0.5       2        -1         1             2    
 0.5       3      -0.5         2             3    
 0.5       4        -1         3             4    
   1       4         1         6             7    
 0.5       8         1         4             5    
 0.5       9       0.5         5             8    
 0.5       8         0         8             9    
 0.5       6         1         9            10    
 0.5       2        -1        10            11   

SERIES 2

series2 = [-1.5 1 -3 10 -1 3 -8 4 -2 6 -1 -4 -8 2 1 -5 0 2.5 -4 1 0 2 -0.5]

Count    Range    Mean     Start_Index    End_Index
_____    _____    _____    ___________    _________

 0.5      2.5     -0.25         1             2    
 0.5        4        -1         2             3    
   1        4         1         5             6    
 0.5       13       3.5         3             4    
   1        6         1         8             9    
   1       14        -1         7            10    
   1        7      -1.5        14            16    
   1        1       0.5        20            21    
 0.5       18         1         4            13    
 0.5     10.5     -2.75        13            18    
 0.5      6.5     -0.75        18            19    
 0.5        6        -1        19            22    
 0.5      2.5      0.75        22            23    

[iamlikeme]

Thank you for the test cases. I will work on an implementation and get back to you.

[gsokoll]

What is the benefit of this functionality ? Other than for academic interest, it seems to have little to no real world use.

[addarnr]

No, that is not true.
There are cases where the rainflow count of one signal can be used as a representation of another signal that cannot be directly measured. Throttling of a car is an example. Every throttle change causes some damage to the car engine. You can estimate these damages by the rainflow count of the throttle if there is data already available to correlate the two. In this case you need the indices of the throttle rainflow counts to correlate the two.

[gsokoll]

Note that the Matlab rainflow implementation includes an option to return the indices of reversals identified in the data, not cycles. This is not the same as what you have asked for here. https://www.mathworks.com/help/signal/ref/rainflow.html

[addarnr]

Yes the linear indices are what I am after. Sorry if I was not clear. The test results I sent include these linear indices.

On Tue, Apr 14, 2020, 7:26 PM Geoff Sokoll notifications@github.com wrote:

Note that the Matlab rainflow implementation includes an option to return the indices of reversals identified in the data, not cycles. This is not the same as what you have asked for here. https://www.mathworks.com/help/signal/ref/rainflow.html

— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub https://github.com/iamlikeme/rainflow/issues/31#issuecomment-613747207, or unsubscribe https://github.com/notifications/unsubscribe-auth/AIFGUYBSKGFE65U6VORMQXLRMT5MVANCNFSM4MCMKPXQ .

[iamlikeme]

The feature you requested is now released in v3.0.0 on PyPI. Here is a usage example:

series = [0, -2, 1, -3, 5, -1, 3, -4, 4, -2, 0]
list(rainflow.extract_cycles(series))
# Output
[(2, -1.0, 0.5, 0, 1),
(3, -0.5, 0.5, 1, 2),
(4, -1.0, 0.5, 2, 3),
...
(2, -1.0, 0.5, 9, 10)]