Version 1 of PPgNN for Fatigue Life of Superalloys

This model is uses the PPgNN architecture to predict Creep-Fatigue and Thermomechanical Fatigue Life of Ni-base Superalloys. The model will plot data points that have similar parameters to the one inputted into the model including runout cases. Tests that did not runout are displayed as a blue dot, while tests that did runout are displayed as a red triangle. This model is explained in greater detail in our paper.

Reference Statement

This is neural network model uses data that was collected from multiple sources. The complete data set used to train the model is stored here. Listed in the about section are where all the sources from where data was used to train the model shown on the home page.

Parameters

Max Temperature \((T_{max})\): The maximum temperature of the specimin in Celcius \((^{\large\circ}C)\)

Min Temperature \((T_{min})\): The minimum temperature of the specimin in Celcius \((^{\large\circ}C)\)

Alternative Temperature Strain Phase Factor \((ATSPF)\): A variable dependent on the difference of temperature and if the test was Isothermal Fatigue \(\text{(IF)}\), In-Phase Thermomechanical Fatigue \(\text{(IP TMF)}\), or Out-of-Phase Themromechanical Fatigue \(\text{(OP TMF)}\). If the test was:
 \(\text{IF}\) then \(\text{TSPF} = 0\) \(\text{IP TMF}\) then \(\text{TSPF} = 1\) \(\text{OP TMF}\) then \(\text{TSPF} = -1\)
and ATSPF is caluclated by
 \(ATSPF = \frac{T_{max} - T_{min}}{500}\;*\;TSPF\)

Cyclic Frequency \((Freq)\): The cyclic frequency of the test, not including dwell times, on the specimin in \((Hz)\).
Total frequency displayed in the title of the plot is the cyclic frequency including dwell times calculated using the equation \(\text{Total Freq} = \frac{1}{(\frac{1}{Freq} + 60(Dwell_C + Dwell_T))}\).

Compression Dwell \((Dwell_C)\): Compression dwell applied to the specimin during the test in minutes \((min)\)

Tensile Dwell \((Dwell_T)\): Tensile dwell applied to the specimin during the test in minutes \((min)\)