|
|
Weibull Log Normal Workshop
Instructors: Dr. Robert B. Abernethy, Wes Fulton
SuperSMITH™ Version 5.0 and The New Weibull Handbook, 5th Edition!
Workshop participants – Included with the registration fee is a license for Version 5.0 of Fulton Findings’ SuperSMITH™ package, a major new release of the world’s most widely used and accepted Weibull Analysis software, along with the new Fifth Edition of The New Weibull Handbook - a $960 value!
Workshop Schedule
- March 5, 2007
7:30am Registration
8:00am-4:00pm Workshop
- March 6, 2007
8:00am-4:30pm Workshop
- March 7, 2007
8:00am-4:00pm Workshop
RMS (Reliability-Maintainability-Safety-Supportability) engineering is perhaps the most important new discipline in product development due to new credible, accurate, quantitative methods. Weibull Analysis is foremost among these new tools. New advanced Weibull and Crow-AMSAA techniques provide accurate failure and warranty claims forecasts including cost analysis. They yield useful results even with extremely small samples and with data that has deficiencies. Special methods developed for these data problems, such as Weibayes, are presented with actual case studies. Inspection data will be treated with both interval and probit analysis. Test substantiation of new designs with minimum cost is presented. WinSMITH™ Weibull produces risk forecasts with renewal and optimal component replacement. Many case studies will be illustrated. Students are fully qualified to conduct Weibull Analysis at workshop completion.
Benefits of Attending
By attending this workshop, you will be able to use Weibull's unique capabilities to:
- Analyze and predict design, development, production, and service failures
- Model product lifetime and reliability
- Evaluate optimal parts replacement and maintainability plans
- Analyze inspection data
- Reduce test substantiation, time and costs
- Get the latest software and Handbook updates - a $960 value
Who Should Attend
Engineers responsible for reliability, safety, supportability, maintainability, materials, warranties, life cycle cost, design, structures, instrumentation and logistics will find these Weibull techniques extremely useful.
Prerequisites
An engineering undergraduate degree in any discipline would be beneficial.
Text & Software Included
Each attendee will receive the entire SuperSMITH™ package, a $960 value. SuperSMITH™ contains: the new Fifth Edition of The New Weibull Handbook written by Dr. Abernethy with many experts helping. The new edition has been completely rewritten with new sections and new methods for warranty analysis, goodness of fit, S/N curves, Crow-AMSAA modeling, Life Cycle Costs and Weibull Libraries.
The SuperSMITH™ package also includes the latest version, 5.0-32, of WinSMITH™ Weibull, WinSMITH™ Visual and YBATH software which is used in the workshop; and the Play-Time Tutorial Software Workbook. This state-of-the art, Windows™ software includes Normal & Log Normal Analysis and allows the student much more time for actual hands-on PC experience, interpreting data of all types. There are many new features. This software is very "user friendly" and provides: automatic iteration of t0, testing up to ten data sets to see if they are significantly different, MLE and rank regression solutions, labels, test substantiation designs with minimum cost, goodness of fit, inspection data analysis, confidence bounds with censored data, reliability prediction with confidence level, optimal component replacement intervals, future failure predictions with & without renewals, distribution analysis, accelerated testing, S/N curve production, etc. The software includes all the new methods in the Fifth Edition of The New Weibull Handbook. The students will exercise all the software systems to gain experience with all types of problems using a computer tutorial. They will also analyze data generated from class fatigue experiments and data they bring to class from their home organization.
Weibull Experiments
The students will rupture wire samples in low cycle fatigue and analyze the data using WinSMITH™Weibull software.
Crow-AMSAA
Reliability growth modeling (Crow-AMSAA) is also presented on the third day. It is the fastest growing application of life data analysis. It is useful for three purposes: (l) tracking repairable systems, (2) tracking component development, and (3) analyzing systems where early portions of the data is missing. An example of the last is electrical power cables buried decades ago by utilities that did not keep failure data in the early years. It is the best technique for analyzing "dirty" data. Failure and cost predictions are accurate even with dirty data. The newest application, renewal/repairable system analysis, has some advantages and disadvantages over Monte Carlo methods. WinSmithTMVisual performs the Crow-AMSAA calculations and provides plots of the results.
Workshop Content
The Weibull, the Poisson, Log Normal, Exponential, Gumbel and Binomial, the distributions that form the basis for RMS Engineering, are covered. The third day is devoted to confidence intervals, RMS models, and reliability growth analysis. Class work is used to reinforce key concepts. Lectures are based on actual case studies. The students use personal computers on the second and third day to analyze dozens of Weibull and Log Normal problems. The students will be fully capable of performing basic and advanced RMS Engineering analysis with their own software on completion of the workshop.
DAY 1 - Undergraduate Weibull Analysis
7:30am Registration
8:00am-4:30pm Workshop
- Introduction & Overview
- How to do Weibull Analysis
- Interpretation of Good & Bad Weibulls, and log normals
- Interpretation of Bad Weibulls - Cusps, Dog Legs, Curves, Outliers, Goodness of Fit
- Suspension and Sample Size Effects on Uncertainty
- Failure Predictions and Weibull Risk Analysis: How Many Failures Will There be in the Next year, Five Years? With and Without Parts Replacement? How many Warranty Claims? Cost Analysis. Comparisons with Crow-AMSAA predictions.
- Case Studies, Customer Usage, Maintenance Planning,
- Batch problems, detection and analysis
- Weibull Experiments, Wire Rupture, Torsion, LCF
- Classwork, Problems and Solutions
DAY 2 - Postgraduate Weibull Analysis
8:00am-4:30pm Workshop
- Review and Discussion of Class Work and Playtime Exercises
- Maximum Likelihood Weibull and Log Normal Theory and Application including Reduced Bias Adjustments (MLE-RBA) for extreme accuracy with small samples
- Rank Regression vs. Maximum Likelihood, When Should I Use MRR vs. MLE vs, MLE-RBA, Advantages & Disadvantages? Best Practices?
- What is the best plotting position - Mean or Median or Blom or Hazen
- Extremely Small Samples Analysis, What is Best Practice for small samples?
- Inspection Methods: Inspection Option, Probit and MLE Interval analysis, which method is most accurate?
- Goodness of Fit with the three best methods: Critical Correlation Coefficient, the likelihood ratio method and the Anderson-Darling. Which method is the best, most powerful?
- Weibayes, best practice for small samples with and without failures,
- Test Substantiation with minimum time and cost, Is the new design significantly better than the old design, zero failure, zero-one failure and Sudden Death tests
- Accelerated Testing, S/N curves, Step-Stress
- Weibull libraries with automatic input
- The Logic Diagram - A summary of Weibull Methods, now available in Spanish as well as English
- "Playtime With SuperSMITH™" Tutorial, Hands-On Computer Experience with All Kinds of Problems, Experimental Wire Data Distribution Analysis
DAY 3 - Confidence Intervals and System Models
8:00am-4:00pm
- Confidence Intervals, "The Good, the Bad and the Complicated," what is best practice for confidence intervals comparing the six different methods?
- Comparing Data Sets New Design versus old design, Vendor A versus B, Alloy A versus B. multiple data sets…are they significantly different?
- The Binomial & Poisson, Discrete Distributions Related to the Weibull and Useful for RMS & Quality
- Probabilistic Design: Convolutions of Life and Usage Distributions and crack and crack-to-rupture distributions
- Batch Problems: How can we detect batch problems and estimate the size of the batch?
- Crow-AMSAA Reliability Growth Modeling - New Useful Technology for Tracking: Component Development, components and systems in service, management problems, with missing portions of data, mixtures of failure modes, dirty data. Includes Cost analysis.1
- The Exponential Distribution Related to the Poisson and the Weibull, Assumptions and Applications
- Kaplan-Meier Survival Analysis
- Production Process control with Weibull Analysis for major cost savings
- System Degradation Models: performance degradation, crack growth, vibration data
- Hands-on Computer Problems with the Playtime Tutorial
- The “Ph.D” Oral Examination
| Organized by: |
 |
|
|
|
