ASME A112.18.3:2002 pdf download.PERFORMANCE REQUIREMENTS FOR BACKFLOW PROTECTION DEVICES AND SYSTEMS IN PLUMBING FIXTURE FTTINGS.
The upper confidence limits are based on measured failure rates for devices subjected to the durability tests defined in ASME A112.18.3-2002.
The upper confidence limits, for a Pa of 97.5% unilateral, being usual for technical products of the type under examination, are given in Table Al for selected sample sizes and failure rates.
If, for example, 2 out of n = 10 specimens (= 20%) fail to meet the requirements in the life test, the upper confidence limit for all of these products in the field is 55.6%. This means the maximum probable failure rate may account for 55.6%.
For devices of the type under examination, it is proposed not to permit an upper confidence limit, i.e., the maximum probable failure rate in the field, in excess of 56%.
It shall also be mathematically possible to add the results of several tests performed to the same specification on differing numbers of specimens. This technique is referred to as a multiple sampling scheme and is widely recognized in quality control work.
Therefore, adding a second test result from 5 test spedmens with I failure to the above described result produces an upper confidence limit for the final test result of <48.1% and means, with respect to the above assumed limit of 56%, that the product tested in this way has passed the test. It is proposed that the tested device, or the entire product in the case of integrated solutions, be certified.
The described method of computing the reliability of safety systems is mathematically exact, noninterpretable, extremely simple for test engineers and manufacturers to use, and extremely flexible in application. The latter two criteria are of particular significance to pragmatic approval testing.
The upper confidence limits, for a Pa of 97.5% unilateral, being usual for technical products of the type under examination, are given in Table Al for selected sample sizes and failure rates.
If, for example, 2 out of n = 10 specimens (= 20%) fail to meet the requirements in the life test, the upper confidence limit for all of these products in the field is 55.6%. This means the maximum probable failure rate may account for 55.6%.
For devices of the type under examination, it is proposed not to permit an upper confidence limit, i.e., the maximum probable failure rate in the field, in excess of 56%.
It shall also be mathematically possible to add the results of several tests performed to the same specification on differing numbers of specimens. This technique is referred to as a multiple sampling scheme and is widely recognized in quality control work.
Therefore, adding a second test result from 5 test spedmens with I failure to the above described result produces an upper confidence limit for the final test result of <48.1% and means, with respect to the above assumed limit of 56%, that the product tested in this way has passed the test. It is proposed that the tested device, or the entire product in the case of integrated solutions, be certified.
The described method of computing the reliability of safety systems is mathematically exact, noninterpretable, extremely simple for test engineers and manufacturers to use, and extremely flexible in application. The latter two criteria are of particular significance to pragmatic approval testing.
