Hi Edi,
Unfortunately, there are no universal models. Every material, except for cement, is exposed to aging phenomena that increase failure probability with time. These phenomena are driven by interactions with the environment, which is specific to a field (and to individual wells, to a degree).
Some aging can be monitored: for metal loss in tubulars you can run wireline logs, inspection when pulling tubing, or the good old coupons. This way you can extrapolate a corrosion rate to the minimum acceptable thickness. If you're experiencing Stress Corrosion Cracking, logs are useless - and degradation behaves more or less randomly.
For elastomers you have no alternative to a statistical approach built on experience in your own fields. No averages, no looking at your neighbor.
This is even more true of safety valves, the WBE's we activate to close barriers across the well. Aging is impossible to predict, so you will rely on statistics. The advantage of a tolerable leak rate, as defined by API 14B, is that you can monitor degradation and predict incipient failure (that's why NOROG 117 has this puzzling "degraded" well category). You should really take advantage of it to detect trends, and deploy prevention (i.e., maintenance) and mitigation (i.e., planning intervention) measures to reduce the risk of failure.
So, in a nutshell, "models" have to come from your own field. Data you have been collecting. Other people's data are like other people's bank accounts: they may be uplifting, but it's not your money.
Best regards,
------------------------------
Matteo Loizzo
Well integrity consultant
matteo.loizzo@mac.comBerlin, Germany
------------------------------