Reliability estimation in random environment:: Different approaches

For most of the products, a real time depending variation of stress study implies the random environment approach. Several possible treatments are available to the users and each method offers different results for the same input data. The use of the equivalent constant stress, for example, gives a strong; pessimism while the use of cumulative damage model requires strong computation resources. Whatever method one might choose, approximations and/or estimations are to be done in order to extend the field observed stresses values over the whole life of the product. The present paper proposes a comparison of the possible methods. A field validation is proposed and may be the subject of incoming work.