Degradation process and failure estimation of drilling system based on real data and diffusion process supported by state space models

摘要: Abstract Technical systems used in adverse environments are subject to very intense degradation and their parts deterioration. Due the problematic placement of some parts, it is sometimes difficult, indicate level possible failure occurrence. Therefore, useful work with available field operation data. Since we possess such data apply progressive methods model degradation, able predict occurrence forecast residual life. At first, spectral analysis approaches. The capture extreme values structure. later filtered out avoid future estimations which might be affected by deformed inputs. In next step, use non-parametric smoothing state space models acquire trend, variance related statistics These characteristics as input parameters for specific new forms diffusion processes. With these processes would like evolvement represented one first passage time (FPT). FPT a moment when modelled trajectory hits predefined threshold – represents critical limit our observation. outcomes (i) modelling, deterioration prediction condition assessment, (ii) maintenance planning rationalisation, (iii) life cycle cost optimisation safety improvement.