Improving the quality of field data reliability analysis

Reliability engineers often face the field data with a small number of failures and a large number of suspensions (random censors). Problems arise frequently in engineering practice regarding the quality of field data reliability analysis. The existing methods may result in poor fitting to sample data for the studied cases. Two approaches are proposed and studied in this paper with intention to improve the quality of reliability data analysis. They are simple to implement and appear to have good statistical and convergence properties. Numerical examples are provided to illustrate the proposed methods. Further studies also are discussed and needed.