Block 6 - lecture 2 Flashcards
why will the need for acceptance sampling decrease?
TQM (total quality management) being used more - working with suppliers more so need less data for acceptance
AQL?
acceptable quality level
acceptable quality level?
quality desired by the customer
where is the AQL specified?
contract
ways that AQL can be specified?
- defective units per 10,000
- as a fraction eg. 0.0001
producer’s risk (alpha)?
risk that the sampling plan will fail to accept good parts (Type I)
typical producer’s risk?
5 percent
why are customers interested in low risk like the producer?
returning materials:
- disrupts
- wastes time
- impacts relations
LTPD?
Lot tolerance proportion defective
Lot tolerance proportion defective?
worse level of quality tolerable
customer risk (beta)?
risk of accepting bad parts (Type II)
typical customer’s risk?
10 percent
ANI?
Average Number of items Inspected
OC curve?
Operating Characteristics curve
Operating Characteristics curve?
a graph for performance of a sampling plan
probability of accepting the lot over the proportion defective
what should n and c (sample size and acceptance number) be based on?
AQL (acceptable quality level)
alpha (producer’s risk)
LTPD (lot tolerance proportion defective)
beta (customer risk)
what is the type of distribution for proportion defective?
binomial distribution
difference between binomial and poisson distributions?
Both for attribute data.
binomial distributions = discrete events (eg. proportion of separate parts)
Poisson distributions = continuous events (eg. count of parts)
when can a poisson distribution be used to approximate proportion defective (binomial)?
n>20
p<0.05
when does the producer’s risk = 1 - probability of acceptance?
at AQL
when does the customer’s risk = the probability of acceptance?
at LTPD
steps to make an OC curve?
- find p for AQL and LTPD (probability for 1 part)
- find np (probability for sample)
- use ‘cumulative poisson probabilities’ chart in bklet (probability under c)
- repeat for many p values
- plot on graph
how does the sample size affect the shape of the OC curve?
larger n = lower proportion defective, thus the graph is squashed steeper
how does the sample size affect the producer’s and customer’s risk? why?
- np increases
- probability of acceptance drops for given proportion defective
- producer risk increases
- customer risk decreases
how to decrease producer’s and customer’s risk with sampling?
- increase sample size (reduces customer’s risk)
- increase acceptance number (reduces producer’s risk)
AOQ?
Average Outgoing Quality
Average Outgoing Quality?
expected proportion of defects that the plan will allow to pass
Rectified inspection?
- rejected lots will be replaced
- defective accepted units will be replaced
AOQL?
Average Outgoing Quality Limit
Average Outgoing Quality Limit?
The value of proportion defective (there can’t be less defects than in the sample)