Resumen de Simulation on reliability analysisof linear consecutive K-out-of-n: G systemsfor Weibull parameter estimation with incomplete failure data
The area of Reliability, as well as the new area of Digital Twin (DT), is faced with the needto bring its theoretical or virtual model closer to real models. For this, it needs to determine thebest statistical distribution and its parameters to approximate and tune the real model as bestas possible using real data, which is often censored or incomplete data. In this way, it becomespossible to simulate the estimated useful life of equipment, systems and components virtually.This article develops an algorithm for simulating the estimation of Weibull parameters forconsecutive K-out-of-n systems that exist in many modern systems, such as oil pipeline systems,computer ring networks, telecommunications, etc.The article develops the complex and coherent systems theory and the respective reliabilitymodels. With a new and original approach, algorithms were designed with a simulation thatgenerated random and censored data (right-censored and type I data). The case studies oflinear consecutive K-out-of-n systems were developed to validate the algorithms K-out-of-n:G.A set of simulations was designed with the variation of the different parameters of the reli-ability models to compare, tune and optimize the estimation of parameters and the simulationof these complex systems.One relevant result shows that the more censored data in the sample, the more significantthe bias and the error about the true value. Increasing the parameter ß (shape factor) propor-tionally increases the bias.One of the relevant results shows that the more censored data in the sample, the greaterthe bias and error about the true value. Increasing the parameter ß (shape factor) proportion-ally increases the bias.
