I think the idea of the API test is really that it helps to rank the packings to give some confidence before the valve manufacturer conducts an expensive ISO 15848 test. Note of course that 500 ppm methane is also far easier to achieve than the ISO A, B, and C levels which are a lot tighter as far as I can see. (Compare the volume leak rates with those in the API 453 equations for example.)
Also, most people can understand figures like 500 ppm methane, far easier than something like 10E-06 mbar.l/sec/m or whatever of helium, which is impossible to measure and relate to compared with typical on-site refinery LDAR programme measurements. (These would be done with a flame ionisation detector calibrated with methane of course…)
I personally suspect that many valve buyers may already be a little confused about how to compare the ISO test results from one manufacturer to another. The piece in the API about re-testing in a real valve is because the test rig is never the same as your valve – packing diameter, stem material, packing section etc. will all influence sealing performance, especially in thermal cycling, which has a very significant effect. I am not sure how much of the ISO test is actually achievable – especially if you have a large diameter stainless stem in a butterfly valve going to low temperatures for example.
Maybe another good thread for this forum would be for people to say what ISO levels they have already achieved (albeit anonymously). This would be in terms of classes (block or control valve endurance 1, 2 or 3, and tightness A, B or C) with whatever kind of valve (ball, butterfly, gate etc.) and packing type (graphite, PTFE and so on…)
I have seen a lot of ISO testing and believe that some things may not be reasonably achievable. The API is more practical than the ISO, though some of the tests may still be superfluous to practical requirements in the real world.