MAE486_Fall11_L22_S

MAE486_Fall11_L22_S - MAE 486 Design of Mechanical Systems...

Info iconThis preview shows page 1. Sign up to view the full content.

View Full Document Right Arrow Icon
This is the end of the preview. Sign up to access the rest of the document.

Unformatted text preview: MAE 486 Design of Mechanical Systems Lecture 22 Fall 2011 Today s topics: Manufacturing consideration of design (Chapter 13); Robust design; Achieving Part Standardiza7on •  With the importance of standardiza7on in mind, what we should avoid? –  Not free to make arbitrary decisions when sizing parts. •  A common misconcep7on: –  Minimum ­cost design minimum ­weight design. Why is it not always right? •  How to be aware of the existence of an iden7cal part? –  One way is to use Group Technology (GT); see Sec7on 13.7.3. Mistake ­Proofing •  Everybody makes mistakes—How to account for that in design? •  Technique: Poka ­yoke (Japanese: Mistake ­ proofing) •  Basic tenet: –  Human errors: individual operators •  Goal:  A state of zero defects (varia7on from design or manufacturing specs) Shigeo Shingo What are the frequent Mistakes •  Design mistakes: Where do they come from? •  Defec7ve material mistakes –  Poorly chosen: –  Failed to meet –  Poorly designed dies or molds Frequent Mistakes •  Manufacturing mistakes –  –  –  –  –  –  –  –  OmiXed opera7on OmiXed parts Wong orienta7on of part; misaligned part; wrong loca7on of part Selec7on of wrong parts Mis ­adjustments Commit a prohibited opera7on (e.g., accident) Added material or parts (e.g., dropping a screw into the assembly) Misread, mis ­measure, or mis ­interpret •  Human mistakes Metal ­Cu[ng Processes Metal ­Cu[ng Processes (cont’d) •  Primary mo9on •  Feed mo9on Machinability •  What does it mean? •  What cons7tutes to good machinability? •  What is the general trend? –  Hardness of the workpiece material  The machinability. –  Example? DFM Guideline for Machining •  Rule 1: How to achieve economy in machining? •  Rule 2: Sequence of machining maXers: •  Rule 3: DFM Guideline for Machining (cont’d) •  Rule 4: Avoid reclamping •  Rule 5: Use whenever possible; •  Rule 6: Avoid on unexposed surfaces of the workpiece (when the part is gripped in the work ­holding device); •  Rule 7: No interference •  Rule 8: Adjust  minimize the forma7on of sharp burrs. Process Capability •  What is process capability? –  The sta7s7cal informa7on about the parts produced by the a machine or process  The percentage of parts that fall outside of a specified tolerance band. •  Process capability index Cp Control charts Process Capability (cont’d) •  Usual limits on machine varia7on: ,  •  For mass produc7on, the percentage of defects is cri7cal Process Capability  ­ ­ Examples •  A spindle has a spec. on its diameter of 1.5+/ ­0.009 inch. If Cp=1. what is δ ?. ˆ •  What would be std to achieve Cp=1.33? •  If Cp=1.33 and mean is centered within the range, how many oversize parts? Process Capability •  If the mean is not centered within the tolerance range, process capability index Cpk •  The distance of the process mean differs from the midpoint of the tolerance region is give by •  How does Cpk and Cp relate to each other? Which one is bigger? Taguchi Methods •  What is it? –  The quality level of a product = The total loss from The failure of the product to deliver the expected performance + Harmful side effects of the product (including the opera7ng cost). Example for each case? Noise Factors and S/N Ra7os •  The input parameters affec7ng the quality of the product/process  •  Noise factors: •  What are they? –  Varia1onal noise: –  Inner noise: –  Design noise: –  External noise (outer noise): Signal ­to ­Noise Ra7os •  Ra7o of the mean (signal) and the varia7on (noise)  The response or output of experiment. •  Three forms of the S/N ra7o corresponding to three loss func7ons –  Nominal ­is ­best type –  Smaller ­the ­beXer type –  Larger ­the ­beXer ­type Understanding, why do these eqns. Make sense? Robust Design •  What is it? •  Parameter design: •  How to do parameter design? –  Step 1: –  Step 2: Parameter Design •  What is the approach? –  Extensive planned experiments. •  How to be efficient in experiments? –  Sta7s7cally designed –  Based on frac7onal factorial designs (a)  Three parameters P tested at two levels. (b) Frac7onal factorial design. All test combina7ons considered. Parameter Design (cont’d) •  What is the trade ­off? –  Min (# of tests) vs. Loss (Detailed informa7on of interac7ons). •  Two commonly used orthogonal arrays. (a)  L4 deals with three control factors at two levels. A total of 4 runs of DOE is needed. (b) L9 array considers four control factors each at three levels. A total of 9 runs of DOE is needed. Parameter Design (cont’d) •  Which to use? Depends on –  The number of control factors and noise factors. –  Whether you are seeking more resolu7on in the results, especially if you feel the responses will be nonlinear. •  Design of experiments usually consists of two parts. –  Part I: A design parameter matrix (inner array) from which the effects of the control parameters are determined through the use of a suitable orthogonal array. –  Part II: A smaller noise matrix (outer array) is formed for noise parameters. Taguchi Robust Design Method What are the steps? •  Step 1: Problem defini7on. –  Select the parameter to be op7mized and the objec7ve func7on. •  Step 2: Selec7on of design parameters; –  The control parameters (controlled under the designer) and the noise parameters (contribute to the varia7on caused by the environment. •  Step 3: Experiment design –  Select the appropriate frac7onal factorial array, the number of levels to be used, and the range of the parameters that correspond to these levels. •  Step 4: Do the experiments –  Follow DOE; •  Step 5: Results analysis; –  Calculate the S/N. •  Step 5: Repeat steps 1 to 4 if no clear op7mum value. OR: •  Step 5: Validate the results; –  Perform a confirming experiment when the method gives a set of op7mal parameter values. Robust Design Example •  Problem: –  A new prototype of new game box has indicator light failure •  Causes: –  Poor solder joints. •  Root cause: –  Use of improper solder paste (i.e., solder balls and flux). •  Objec7ve: –  Design the best condi7ons for making strong solder joints by using the Taguchi method. Robust Design Example (cont’d) •  Step 2: Select control Parameters: –  Four control parameters (with L9 orthogonal array, i.e., three levels) •  Step 2: Select noise parameters •  Three noise parameters (with L4 orthogonal array, i.e., two levels) Robust Design Example (cont’d) •  Step 3: Experiment design –  For each run in L9, we conduct four experiments. For example, run 2 in L9 is executed four 7mes to include the noise matrix. –  The first trial condi7on: a new can of paste, water rinse, and horizontal spray –  The last trial condi7on: a can of paste opened one year ago, a chlorocarbin cleaning agent, and horizontal spray for cleaning. •  Step 4: Experiment run –  For each of the four trials of run 2, measure a response that represents the objec7ve func7on that we are aXemp7ng to op7mize: shear strength of the solder joint measured at room temperature. Robust Design Example (cont’d) •  Step 5: Results analysis –  For four trials, find the average strength measurements and the standard devia7on. •  Step 5: Results analysis •  The appropriate response parameter: the S/N ra7o. Which one to use here? Robust Design Example (cont’d) •  Step 5: Obtain the parameter matrix. Robust Design Example (cont’d) •  Step 5: Determine the average response for each of the four control parameters at each of its three levels. •  Step 5: Plot the average S/N ra7os against the test level Robust Design Example (cont’d) •  Step 5: Observe the plots: Which factor is not important? S/N larger the beXer or the other way around? •  Step 6: Obtain the op7mum se[ng of the control parameters ...
View Full Document

Ask a homework question - tutors are online