How to draw an OC Curve (operating characteristic curve)

[PafflesWancakes]

Good Evening!

I just wanted to ask if anyone could assist me with some problems? I'm supposed to draw an OC Curve and have been looking online. I can't seem to find anything that is helping.

This was the question:

Disetronic Medical Systems and the United States Food and Drug Association (FDA) notified healthcare professionals and patients about a nationwide recall of D-TRON adapters, used with the D-TRONplus insulin pump. The adaptors can potentially over-deliver a maximum amount of up to 1.8 units of insulin. Use of these recalled adapters may pose a potentially life-threatening situation to certain children using the pump. Other users who are insulin sensitive may also be at increased risk. The root cause of this failure mode is a sporadic failure of a valve inside the D-TRON adapter to close completely, which can result in a potential over-delivery of a maximum amount of up to 1.8 I.U. of insulin. As an immediate correction, Disetronic has notified the caregivers and physicians of the pump users 13 years of age and under to immediately discontinue the use of the affected adapters. The corrective action will include modification of the valve assembly process, which must be re-validated. You have been selected as the senior process validation engineer to participate in this important initiative.

The severity, occurrence and detectability for the valve assembly process are as follows: S= 4, O=2, D=2. This gives an RPN of 16 and a risk index of 2. A risk index of 2 for attribute data requires an LTPD of 1% and a confidence interval of 95%. The AQL of the valve assembly process is 0.07%.

Using this information as your basis, select an appropriate sampling plan from the attached table that ensures consumer’s risk (patient risk) is within your selected LTPD, and also ensures at least a 95% chance that a process meeting this LTPD will pass one lot of the validation. Draw the OC curve.

Table:
Sample Size AQL LTPD 0.05 LTPD0.10
n = 300; a = 0 0.02% 1% 0.76
n = 470;a = 1 0.08% 1% 0.90
n = 100; a = 0 0.05% 3% 2.51
n = 315; a = 1 0.11% 1.5% 1.15

 

[reynald]

Since it seems that this is an assignment, I would only give you a clue.
You have to figure it out yourself so as to maximize learning retention.
In Excel
If you have finite population size, use hypergeometric distribution:
=HYPGEOM.DIST(sample_s,number_sample, population_s,number_pop,cumulative = TRUE)


If you have relatively infinite population size use:
=BINOM.DIST(number_s,trials,probability_s,cumulative = true)

go to Excel's internal documentation for more details on the formula. You have to give several values of the probability of defect and compute the probability of seeing the acceptance number. Then plot the series.