Probability theory of degradation of thermoelectrtic energy converters and its use to determine the reliability of thermoelectric materials

Abstract

A probabilistic theory of degradation of thermoelectric energy converters has been developed, which describes such stages of their life cycle as the run-in period, which is described by the Weibull function, the period of stable operation, during which the failure rate is not necessarily constant, and the period of mass failures. To correctly describe these stages, the theory introduced the limit resource of a thermoelectric energy converter, which is consumed over time according to a nonlinear law and is precisely determined by the limit capabilities of thermoelectric and other materials, as well as contact structures. But this limit resource is a generalized integral characteristic of the limit resource capabilities of all materials used to create a thermoelectric energy converter and cannot be unambiguously “distributed” between them. After its complete consumption, there is a complete loss of performance of the thermoelectric energy converter. The parameters of the theory are determined by the least squares method based on processing the results of life tests.  By combining the results obtained with a diffusion-nonmonotonic distribution of failure times, the parameters of which are also determined by the least squares method, it is possible to obtain such integral indicators of the reliability of materials as the average limit resource, 95% limit resource, the minimum achievable equivalent failure rate and relative errors, in this case – with a confidence probability of 0.95).