Example of rainflow counting



In [1]:

    
import numpy as np
import pandas as pd
from metal_fatigue import rainfl
import matplotlib.pyplot as plt
import copy

%matplotlib notebook

Load data

Lets generate some load data and plot it.



In [3]:

    
n = 15
series = pd.DataFrame(data={"time":np.linspace(0,5,n),
                            "load":np.random.normal(loc = 5, scale= 3, size = n)})
fig,ax = plt.subplots()
series.plot.line(x = "time", y = "load", ax = ax)
ax.grid()

Rainflow counting with binned data

Fast way



In [4]:

    
rfm,cache = rainfl.rainflow(series=series.load,numbins=15)
rfm.plot2d()









    














    











    Out[4]:





(<matplotlib.figure.Figure at 0x209e9cc7358>,
 <matplotlib.axes._subplots.AxesSubplot at 0x209ea139ac8>)

Step by step

Classification (binning) and turning points



In [5]:

    
series.load.to_csv("test.txt",index=False)



In [6]:

    
# bin data. Result is a binned object 
binned_series = rainfl.bin_series(series=series.load,
                  minedge=series.load.min(),
                  maxedge=series.load.max()*1.01,
                  numbins=15)
# index of turning points
turn_p_index = rainfl.turning_points(binned_series.values)
turn_p = copy.deepcopy(binned_series)
turn_p.values = binned_series.values[turn_p_index]
    
    
fig1,ax1 = plt.subplots()
series.plot.line(x = "time", y = "load", ax = ax1)
ax1.plot(series.time.values, 
         binned_series.bins[binned_series.values]+0.5*binned_series.binsize, 
         label = "binned data")
ax1.plot(series.time.values[turn_p_index], 
         binned_series.bins[binned_series.values[turn_p_index]]+0.5*binned_series.binsize, 
         "-o",
         label = "turning points")
ax1.grid()
ax1.set_yticks(binned_series.bins)
ax1.legend()









    














    











    Out[6]:





<matplotlib.legend.Legend at 0x209ea13e198>

Counting and plotting



In [7]:

    
rfm, residuum = rainfl.binned_rainflow(turn_p)
fig2, ax2 = rfm.plot2d()

Rainflow counting with continuous data

Turnign points



In [8]:

    
turn_p_index = rainfl.turning_points(series.load.values)
fig3,ax3 = plt.subplots()
series.plot.line(x = "time", y = "load", ax = ax3)
ax3.plot(series.time.values[turn_p_index], 
         series.load.values[turn_p_index], 
         "-o",
         label = "turning points")
ax3.grid()
ax3.legend()









    














    











    Out[8]:





<matplotlib.legend.Legend at 0x209eaa1b828>

Cycle counting



In [9]:

    
range_, mean_, cycles_, cache = rainfl.continuous_rainflow(series.load.values[turn_p_index])
rf_res = pd.DataFrame(data={"range":range_,"mean":mean_,"cycles":cycles_})
rf_res

	cycles	mean	range
0	1.0	4.609917	0.649733
1	1.0	4.755280	6.749655
2	1.0	4.275223	7.837822
3	0.5	6.823543	6.349441
4	0.5	3.963968	12.068591