|
WeibPar is a stand-alone 32-bit Windows program designed for Windows XP.
INPUT:
Text file data input, and on-screen user select choices
- Data:
- Time Independent (Fast Fracture)
- Uncensored Failure (single failure mechanism)
- Censored Failure (up to 10 different failure mechanisms)
- J-Integral (JC)
- Fracture Toughness (K1C)
- Time Dependent Data
- Dynamic Fatigue (Monotonically Increasing Uniaxial Loading)
- Specimen Geometries:
- None (raw data option)
- Tensile (Circular or Rectangular)
- Bend Bar (3 pt, 4 pt, or pure bend)
- User Input Effective Volume or Area
- Finite Element Model (arbitrary specimen geometry)
- Flaw Types - Volume or Surface
- Pooled Data - different specimen geometry, uncensored data sets may be pooled
- Data ranking formula options
- Confidence Level options
- Fixed Weibull Modulus and/or Weibull Threshold options
OUTPUT:
Both Text results and Graphical results are displayed on-screen and in optional output files.
- Linear Regression - 2 and 3 parameter Weibull (Uncensored, Biased)
- Maximum Likelihood - 2 and 3 parameter Weibull
- Biased - Uncensored, Censored, or Pooled
- Unbiased - Uncensored (Bain, Antle, Thomain & Bootstrap)
- Confidence Bounds
- Monte-Carlo Bootstrap - Bounds on the failue data (Uncensored, Censored, Pooled)
- Bain, Antle, Thomain - Bounds on the parameter estimates (Uncensored)
- Likelihood Ratio - Bounds on the parameter estimates (Uncensored)
- Goodness of Fit - Log Likelihood, Sum of Squared Residuals, Correlation Coefficient, Anderson-Darling
- Probability Density Function Plot - includes PDF of parameter estimates and a Histogram of the data
- Parameter Estimates are Exported to the CARES Material Parameter (CMP) File for subsequent use in component reliability
OTHER:
- WeibPar interfaces with CARES to perform the following advanced features:
- Using a Finite Element Model for arbitrary specimen geometry parameter estimation
- Generating confidence bounds on component reliability via Monte-Carlo bootstrap techniques
- Generate Data - A Monte-Carlo routine is available which allows the user to randomly generate sample data sets according to user inputs for the number of data points and the 3 Weibull parameters. (Setting the threshold parameter to zero will simulate a 2 parameter data set.)
For further information please review the WeibPar User Guide.
|
