Lean Six Sigma Black Belt Sample Paper 2: Improve 1. The operators of a manufacturing cell work out a more orderly arrangement for tool storage and establish a schedule to maintain cleanliness daily. These improvements are best described by which approach to problem solving? 5S Visual aids Poka yoke SMED2. Power of the design can be increased by: Replicating the center points Replicating the corner points Both of the above None of the above3. The correct sequence in a DoE study is Design selection -> Brainstorm factors -> Decide response(s) -> Determine optimal settings of factors for desired response -> Conduct experiments Design selection -> Decide response(s) -> Brainstorm factors -> Conduct experiments -> Determine optimal settings of factors for desired response Decide response(s)-> Brainstorm factors -> Conduct experiments -> Fit data into design -> Determine optimal settings of factors for desired response Decide response(s) -> Finalize factors & their levels-> Design selection -> Conduct experiments - > Determine optimal settings of factors for desired response4. In a fractional factorial design, the 2 factor interactions (2FI) are confounded with other 2FI’s. What is the resolution of this design Resolution III Resolution IV Resolution V Resolution VI5. Two drugs A & B were evaluated for their effectiveness. However, drug A was tested only in individuals with <30 years age and drug B was tested only in individuals with >30 years age. In such a case the effect of the drugs and age of the individuals is said to be Blocked Confounded Folded Replicated6. Calculate the effect of factor “B” if 65 -65 5 -57. One of the following is not correctly matched to its purpose Replication: Understand sources of variability that are not seen in repeat measurements Center point: Understand the effect of hard-to-change factors Blocking: Technique to include other factors in our experiment which might contribute to undesirable variation. Randomization: Eliminate effects of unknown or uncontrolled variables8. If the effect of factor A is 10, it means When A changes from high to low, the response on an average increases by 10 When A changes from low to high, the response on an average decreases by 10 Both of the above None of the above9. With respect to DoE, “Block” for_________ factors and “Randomize” for__________factors known nuisance, unknown noise unknown noise, known nuisance known nuisance, known noise unknown nuisance, unknown noise10. The disadvantage of using Pugh Matrix is: A datum is required to create the matrix Each alternative is compared to the datum but not to other alternatives The extent of bettering or worsening is not factored in the calculations All the above11. Hindering CTQ’s from HoQ can be optimized to achieve the “Sweet spot” using: (Sweet Spot: Targets for multiple CTQ’s are achieved): Design of Experiments De Bono six thinking hats QFD Pugh Matrix12. To check if the relationship between X & Y exhibits curvature (non-linearity) we introduce_______in the design Center point(s) Block(s) Randomization Folding13. Determine the number of runs that would be required if we had 3 factors, 2 levels, 2 replicates of the corner points, 3 Center points and 1 block. 33 19 27 22 14. All the pairs below are incorrectly matched, EXCEPT Full factorial design: Changing one variable at a time (keeping others constant) to understand the impact on the output Trial and Error: Randomly changing multiple inputs to observe changes to output Fractional factorial design: Running all factors at all levels and developing a model One-factor at a time: Running a subset of the full factorial design15. One of the following DoE can be preferably used as a screening design? Completely randomized full factorial design with center points Replicated factorial Multilevel factorial Fractional factorial16. Which of the following is an example of a pilot? Prototypes developed for early evaluation by end users Implementation at one work area/region Test marketing All the above17. With respect to an interaction existing between two factors, one of the following is TRUE about the interaction plot: The plot will show two overlapping lines representing the low and high levels of one of the factors The plot will show two parallel line representing the low and high levels of one of the factors The plot will show two non-parallel lines representing the low and high levels of one of the factors, which may or may not be intersecting All the above scenarios are possible18. To assess the significance of factors in either a fractional or a full-factorial design, a black belt can use the_________in Minitab Half-normal plot of effects ANOVA table Pareto plot of effects All the above19. One factor at a time studies would not be able to identify interactions between the factors as compared to full factorial designs. The statement is: False True True, only if center points are included in the design Cannot comment until resolution is known20. Statements given below are true, EXCEPT: In Resolution III designs, the main effects are confounded with 2 FI’s In Resolution IV designs, the main effects are confounded with 3 FI’s The maximum resolution in fractional factorial designs is Resolution V Low resolution designs can serve the purpose of screening rather than optimization if we have many factors 21. Calculate the interaction effect (AB) for the following data and match with the appropriate statement 0; There is a strong interaction between A & B 0; There is no interaction between A & B 20; There is an interaction between A & B 80; There is an interaction between A & B22. What resolution will you choose to perform a two-level fractional factorial design; 6 factors, 2 replicates necessary; if the budget doesn’t allow more than 20 experiments IV VI II III23. All the following statements regarding reduction of a model for 3 factors (A,B & C) are true, EXCEPT: Insignificant C can be dropped even if BC is significant Insignificant AB can be dropped even if A and B are significant AC can be significant even if A and B are insignificant If ABC is significant, then no model reduction can be done. 24. One of the following statements is incorrect: a. DoE is done preferably using historical data b. Multiple Regression can be run on available historical data c. Interactions can be better identified from data generated through designed experiments d. DoE is a better approach than One-factor at a time (OFAT)25. Residual is an estimate of experimental error obtained by subtracting the ________ from the_________. Common cause, Special cause Sigma level (Long term), Sigma level (Short term) Observed value, Predicted value Mean, USL26. An experimental run is conducted several times for the same treatment combination to neutralize the effect of any uncontrollable, unknown factor. This is called___________. However, conducting repeated runs at the same experimental conditions (with a time gap in most cases) is called__________. Replication, Repetition Repetition, Replication Reproducibility, Repeatability Randomization, Blocking
