Somat eDAQlite - 17 Data Processing Algorithms

I2773-4.4 en
221
Somat eDAQ lite
17 Data Processing Algorithms
17.1 Peak Valley Processing Algorithm
There are three states in the peak valley processing algorithm. The algorithm starts in
the initialize state. Thereafter it toggles between the peak search and valley search
states. These states are detailed as follows.
Initialize
The algorithm tracks the maximum and minimum input values until the difference
between the maximum and minimum exceeds the specified hysteresis level. If the
minimum value preceded the maximum value, then the minimum value is the first
valley and the algorithm state switches to peak search. If the maximum value
preceded the minimum value, then the maximum value is the first peak and the
algorithm state switches to valley search.
Peak Search
The algorithm searches for a peak, tracking the maximum and minimum values since
the last stored valley until their difference exceeds the hysteresis level. At this point,
the maximum value is output as the next peak and the algorithm state switches to
valley search.
Valley Search
The algorithm searches for a valley, tracking the maximum and minimum values since
the last stored peak until their difference exceeds the hysteresis level. At this point, the
minimum value is output as the next valley and the algorithm state switches to peak
search.
17.2 Rainflow Cycle Counting Algorithm
Rainflow counted cycles are typically used in low cycle fatigue damage analyses. The
rainflow counting algorithm is based on the “one-pass” algorithm described in the
paper Simple Rainflow Counting Algorithms, International Journal of Fatigue, January
1982, by D. Socie and S. Downing. The algorithm described in this paper generates
the set of closed cycles for the input peak valley sequence, assuming that the
sequence repeats itself.
However, to support proper “rainflow histogram addition” (i.e., generating a composite
rainflow histogram from multiple rainflow histograms defined in a specific sequence),
the eDAQ
lite stores both the sequence of unclosed reversals and the histogrammed
set of closed cycles in the Rainflow DataMode.
NOTE
The algorithm requires an allocated memory stack to store the reversals that have not
yet closed. The eDAQ
lite uses a fixed stack size of 1024, which should suffice for the
vast majority of applications. In the rare event that this allocation is insufficient, the
eDAQ
lite aborts the Rainflow DataMode processing and sets an error flag in the
output data file.
HBM: public