UTPA STEM/CBI Courses/Introduction to Mechanical Engineering/Process Improvement

Course Title: Introduction to Mechanical Engineering

Lecture Topic: Process Improvement Challenge

Instructor: Timmer

Institution:UTPA

Backwards Design

Course Objectives

Primary Objectives- By the next class period students will be able to:
- Utilize statistical techniques to compare data
- Utilize design of experiments to improve process performance
- Introduce mathematical modeling
- Incorporate statistical analysis software in engineering

Sub Objectives- The objectives will require that students be able to:
- Use Minitab statistical analysis software
- Communicate results in oral and written formats

Difficulties- Students may have difficulty:
- Accessing Minitab (available at UTPA)

Real-World Contexts- There are many ways that students can use this material in the real-world, such as:
- Improving the performance of any process using experimental design
- Use statistical techniques to demonstrate process improvement

Model of Knowledge

Concept Map
- Understand and document yield at starting point using statistical techniques
- Find a better operating point
- Understand and document yield at new and better operating point using statistical techniques
- Compare starting and new operating point using statistical techniques

Content Priorities
- Enduring Understanding
  - Using statistical techniques to understand performance measure
  - Using design of experiments to find improved operating point
  - Using statistical techniques to compare data
- Important to Do and Know
  - Perform graphical analysis of data using appropriate statistical techniques
  - Perform numerical analysis of data using appropriate statistical techniques
  - Perform a response surface methodology design of experiments
- Worth Being Familiar with
  - Linear regression and mathematical modeling
  - Central composite design
  - Statistical test of hypothesis

Assessment of Learning

Formative Assessment
- In Class (groups)
  - Identify location and type of stationary point from 3-D graphs
  - Use regression equation to predict response variable values
- Homework (individual)
  - Pre-lecture quiz after student completes reading assignment
Summative Assessment
- In-class quiz over lesson

Legacy Cycle

OBJECTIVE

By the next class period, students will be able to:

Predict response variable values using regression output
Locate stationary plots by interpreting contour plots and 3-D plots
Determine type of stationary plots using contour plots and 3-D plots

The objectives will require that students be able to:

Utilize second-order polynomial to predict response values
Understand stationary points

THE CHALLENGE

Your first job assignment as a new process engineer in a chemical plant is to improve the process yield. A higher yield generates higher profits. The chemical process is very simple and has two controllable factors: reactant flow (flow) and reactor temperature (temp). Improving the process yield is accomplishing by finding a new setting for reactant flow and reactor temperature that provides a higher yield.

Your success in the company, as well as the company’s future success, is dependent upon your ability to improve the process yield and document the improvements that you have made. The use of statistics and statistical analysis will be important in completing this challenge.

GENERATE IDEAS

What method would you propose to find new values for flow and temperature that improve yield? How would your method work if your process had 6 input variables or 8 input variables or 10 input variables?

How can you be sure that you found the best possible new setting?

How can you prove to management that you improved the process?

You can request instructor handouts via mailto:timmer@utpa.edu Instructor handout request.

MULTIPLE PERSPECTIVES

The goal of improving a process that contains a single output and multiple inputs is challenging. The functional relationship between the inputs and the output must be discovered. This typically involves designing, conducting and analyzing an experiment. An engineer must decide on how experiments to conduct and the settings for each experiment. Once the data is collected it must be analyzed to determine functional relationship between the inputs and outputs and improved settings must be found.

RESEARCH & REVISE

OVERVIEW. A series of short lectures will be presented to provide the details necessary to accomplish the conceptual map given above. Students should be reminded that design of experiments is an advanced analytical technique. In the Manufacturing Engineering curriculum students receive this material in MANE 2332 – Engineering Statistics and MANE 4311 – Quality Control. This problem has been reduced and simplified for a freshman level introduction to engineering course. Students desiring to perform a design of experiments project should seek professional help. Students will be provided a handout containing the lecture slides.

Lecture 1. Descriptive statistics. A recorded lecture will review descriptive statistics. Two key ideas is the use of numerical summaries (e.g. mean, median, mode, etc.) and graphical analysis (histograms, boxplots and dotplots). Include a discussion of sample sizes. Students will use the software package Minitab to calculate numerical summaries and create graphs.

Lecture 2. Data analysis using Minitab. This lecture will introduce students to the web-based application to generate data for their process improvement project. Students will be shown how to use Minitab to perform the analysis and create reports.

Lecture 3. Introduction to Design of Experiments. This lecture will introduce students to design of experiments. The emphasis is on the creation of a response surface methodology model (CCD). This lecture includes a demonstration of using Mintab to create a central composite design.

Lecture 4. Analyzing a CCD. This lecture will introduce students to regression analysis and graphical analysis of a CCD. Since this is a two-factor design the graphical approach is superior. Unfortunately due to time limitations, we will discuss model building and refining regression models that is usually required in problem with more factors. The use of Minitab to analyze a CCD will be included in this lecture.

Lecture 5. Comparing set points. This lecture will present suggestions for comparing the old and new setpoints. The analysis will be limited to descriptive statistics and graphical approaches. The use of Minitab to compare set points will be included in this lecture.

A student website is available at http://quality.engr.utpa.edu/cbi/ProcessImprovementChallenge/

TEST YOUR METTLE

The web-based simulation of the chemical plant can be found at http://quality.engr.utpa.edu/cbiProcessImprovement. Use the simulated chemical plant and your knowledge gained from the Research and Revise section to complete the attached worksheet.

GO PUBLIC

Your instructor will provide instructions for preparing a written report based upon your Test Your Mettle worksheet. Suggested formats are technical reports or Power Point Presentations. Time permitting oral presentations can be given.

Pre-Lesson Quiz

What type of stationary point is displayed in the graph below?

What type of stationary point is displayed in the graph below?

What type of stationary point is displayed in the graph below?

What is the predicted value for y when x=5 using the equation shown below?

Test Your Mettle Quiz

From the contour plot and surface plot shown above, what type of stationary point is present?
From the contour plot and surface plot shown above, what are the values of A and B at the stationary point?
What is the predicted value of y from the equation below when A=1 and B=-1?