Reliability model is the foundation of reliability analysis. Conventional reliability modeling takes the entire system as the research object

and uses all fault data to fit the distribution function of the system

estimate parameters

and optimize the model

thus to determine the distribution type and distribution function of the system

and then to calculate the reliability indicators of the system. Wind turbine is a typical complex electromechanical system

with different functions

structures

and fault forms of each subsystem. It is obviously inappropriate to use one distribution function to determine the fault distribution of the entire system. Therefore

based on the collected and sorted fault data of wind turbines

the distribution function optimization-based reliability modeling and analysis technology of wind turbines is proposed. By applying the commonly used exponential distribution

normal distribution

log-normal distribution

Weibull distribution and gamma distribution

the distribution function fitting

parameter estimation and goodness of fit analysis of the fault interval of each subsystem of wind turbines are carried out

and the distribution function and subsystem reliability function of the fault interval time of each subsystem are determined. On this basis

a Copula connection function is used to establish the reliability function model for wind turbines

taking into account the fault correlation between subsystems. Moreover

an example analysis is conducted on the fault data of an offshore wind turbine

which verifies the feasibility of the proposed method.