Failure Time Point Distributions considering nonhomogenous effects and interaction
- We discuss failure time point processes with respect to maintenance and reliability of technical components or machines (failure-repair models).
For a given sequence of failure time points we assume that the corresponding counting process can be described as a nonhomogenous Poisson process.
Our approach is based on nonstationary intensity functions depending on the starting time of the corresponding machine. In the talk we present two different approaches:
an intensity function based on the lognormal distribution and a monotonously decreasing intensity function based on the exponential function.
For practical applications in manufacturing plants cluster processes based on several machines and their corresponding failure time points are discussed. Especially, we present an approach using fuzzy measures to model interaction among the machines. For example, let I be a time period. Obviously, we are interested in the total number of failures of all considered machines during this period. The fuzzy measure approach enables consideringWe discuss failure time point processes with respect to maintenance and reliability of technical components or machines (failure-repair models).
For a given sequence of failure time points we assume that the corresponding counting process can be described as a nonhomogenous Poisson process.
Our approach is based on nonstationary intensity functions depending on the starting time of the corresponding machine. In the talk we present two different approaches:
an intensity function based on the lognormal distribution and a monotonously decreasing intensity function based on the exponential function.
For practical applications in manufacturing plants cluster processes based on several machines and their corresponding failure time points are discussed. Especially, we present an approach using fuzzy measures to model interaction among the machines. For example, let I be a time period. Obviously, we are interested in the total number of failures of all considered machines during this period. The fuzzy measure approach enables considering interaction effects like synergy and suppression: Failures of a certain machine may cause an increasing probability of another machine to fail. On the other hand a higher number of failures of one machine can lead to less failures of other machines.…