Process Reliability Training


Process Reliability Training (PR) is a 2-day basic training course for engineers in using Weibull analysis of production data to answer these questions. 
            1.  What is the reliability of my process? 
            2.  How can I use the information to improve margins and financial performance in my plant?
Traditional six-sigma efforts use statistical tools (most often the bell shaped Gaussian distribution) to help find root causes of problems and the six-sigma efforts provide mile-marker metrics.  Since most production output data is not bell shaped data, we need a more advanced tool to help define the problem solving metrics.

Weibull analysis of daily production output (using skewed Weibull distributions) helps quantify problems in four categories:
            1.  Define and show the reliability of the production process,
            2.  Identify and quantify losses due to reliability issues (special cause losses),
            3.  Identify and quantify losses from efficiency and utilization problems (common cause losses); and
            4.  Quantify the size of the hidden factory with suggestions of what to look for as corrective action
These metrics all go on one side of one sheet of paper in graphical format and they also form mile-marker metrics for the business.  Refer to the May 1997 Problem Of The Month, and subsequent hyperlinks (see below), to see examples of the Weibull production data analysis along with the descriptive mile-marker metrics.

Students solve problems during the class using their notebook or desktop computers, with the current version of SuperSMITHTM Weibull software to reinforce the lectures.   Authentic data files will be provided for use with the no-cost demonstration version of the SuperSMITH software so no randomization of data will occur as would be the situation if the students input their own data and this minimizes the software cost for the students for training purposes.

Each student is expected to bring an Excel TM spreadsheet of production data from one or two processes within his operations for work in class.  One dataset should be from a good process and the second dataset should be from a process that performs badly.

Subjects discussed in the course will usually be connected to money issues.  Thus the class will combine engineering tools and business tool for a no-nonsense approach to solving problems of a financial nature. The reliability tools discussed are:

  • How to find the reliability of a process
  • What data is typically required
  • Availability concepts and analysis
  • Data for use in solving reliability problems
  • Preparing reliability data for analysis
  • Normal probability plots
  • Log-normal probability plots
  • Weibull probability plots
  • Reliability block diagram models for cost of unreliability
  • What are the reliability losses and what causes them?
  • What are the efficiency and utilization losses and what causes them?
  • What is the size of the hidden factory and how many days of equivalent production are lost?
  • Monte Carlo simulations
  • Pareto distributions for vital problems
  • Critical items list
  • Reliability growth plots and forecast models based on previous experience
  • Role in achieving reliability by-
    • Production
    • Engineering
    • Maintenance
    • Management
  • How to measure process improvement
  • How Weibull process reliability techniques differ from six-sigma approaches

The PR course is 2-days in length (out by 3 PM on day two).  For training sessions conducted in-house, a 1-hour management overview is available for educating the mangers on issues of process reliability. The agenda for the 2-day course is:

Day 1--

Introductions and general information
What is reliability and what data is typically required
What is the difference between availability and reliability
Probability plots: Normal, Log-Normal, and Weibull
How do your prepare the data for analysis
How to make probability plots
Use of students data to illustrate how to make plots
Use of students data to find hidden factory losses
Over commitment and under production
Monte Carlo simulations for multi-plant output
Pareto distributions
Critical items list
Demonstration Project Introduced
Homework for Day 2 projects

Day 2--

Review of Day 1 Homework
Reliability Models for cost of unreliability
Computer demonstration of Monte Carlo models
Reliability growth plots and forecast models
Student work on their computers with their problems
Class discussion of students problems/opportunities
Roles for achieving process improvement
How to measure process improvement
Demonstration Project Preliminary Commitment
Management Overview (if required--1 hour duration)

Note:  Students must bring 1 or 2 data sets of production from their facility to analyze in class and present for discussion to the group to demonstrate they have grasped the concepts.

Note:
Class commences at 8:00 AM and ends at 5:00 PM with 45 minutes for lunch. A ten-minute break occurs in the morning and one in the afternoon.

Student demonstration projects must provide a minimum cost reduction of at least US$4,000 to pay for the cost of the student and instructors time.

During the PR, each student will be encouraged to study process reliability on his own as a homework/demonstration project.  The demonstration project will show the student can put both the art and science of engineering to work for reducing costs and improving output from the process.

 The PR course may usually require a one additional day of follow-up class concentrating on the demonstration projects. The follow-up class is a coaching effort conducted 6-8 weeks after the 3-day training. The students use the elapsed time for solving practical problems concerning their process. Coaching is usually required on demonstrations projects selected by the students to show they have mastered the skills of using process reliability tools for solving cost problems. This one-day follow-up class concerning the demonstration project is a coaching and tutoring effort to help the students across small problems delaying action on important cost issues.

Who should attend the Process Reliability training class?

          Production supervisors will find new tools for understanding how operations can improve reliability of their processes. They will learn how to influence improvements in availability, how they can assist in reducing process failures, and how they can calculate the cost of unreliability for making business decisions to attack problems of unreliability.

         Engineering personnel will find new modeling techniques for predicting process reliability based on how equipment is installed, operated, and maintained for making life cycle cost decisions in justifying new equipment and new processes.

         Maintenance engineers will find reliability tools helpful for providing supporting evidence during root cause analysis failure investigations. They will find reliability tools and techniques helpful for understanding failure data in their CMSS systems, and how failure data is used to justify making equipment more reliable as a business decision.

         Managers will find business aspects of process reliability helpful for measuring and motivating improvements in processes, procedures, people, and equipment to reduce the cost of unreliability through use of non-traditional tools as they ferret-out hidden factories wasting time and money. They will learn how to predict future failures as a selling point for improvement projects.

The Reliability Engineering Principles class is recommended prior to taking the Process Reliability Training class—however it is not a mandatory prerequisite.  The REP class will familiarize students with traditional reliability issues and background for the cost models.

You can download other articles from this site concerning process reliability:

·         Use Periods Of Low Production Output to Improve Process Reliability And Consistency, February 2009

·         Special Cause Variations, Common Cause Variations, and Process Reliability Plots, October 2008

·         Summary of Process Reliability, June 2008

·         Process Reliability Punch List March 2005

·         Process Reliability Line Segments April 2004

·         Process Reliability Plots With Flat Line Slopes May 2001

·         Key Performance Indicators From Weibull Production Plots May 1998

·         Production Reliability Example With Nameplate Ratings April 1998

·         Nameplate Capacity March 1998

·         Coefficient of Variation February 1998

·         Six Sigma January 1998

·         Production Output/Problems May 1997

·         Papers On Process Reliability As PDF Files For No-charge Downloads
   --Process and Equipment Reliability  May 2004
   --Process Reliability: Do You Have It?—What’s It Worth To Your Plant To Get It?  March 2002
   --Process Reliability  December 2001
   --New Reliability Tool for the Millennium: Weibull Analysis of Production Data  October 2000
   --Process Reliability and Six-Sigma March 2000

Refer to price lists for information about training costs.  Check the schedule for available times. 

Download a copy of a short Process Reliability Training brochure as a (35K) PDF file—if you need a PDF reader, download the free Adobe software.

Below is a Barringer process reliability plot showing production regimes with common cause problems and special causes:

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Below is a Barringer process reliability plot with losses described on one sheet of paper:

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Return to Barringer & Associates, Inc. homepage

Last revised 1/28/2010