Bayesianmaximum entropy network-based damage precursor identification and reliability-based maintenance scheduling optimization Bayesian/maximum entropy network-based damage precursor identification and reliability-based maintenance scheduling optimization Damage diagnosis and remaining life prediction of pipeline infrastructure systems is still a challenging problem despite tremendous progress made during the past several decades, such as the damage accumulation in plastic gas distribution pipes. Historically, two entirely different approaches are used for structural system performance prediction (i.e. data-driven or physics-based predictive models). Data-driven approaches used nondestructive inspection technique (optical images, ultrasound, acoustic measurement, etc.) and experts justification (personal experience on trending function, normal range of operations, etc.) to extrapolate system future behaviors. Physics-based models use underlying mechanisms (crack initiation and propagation model, chemical diffusion functions, oxidization rate, etc.) to predict system future behaviors. Information fusion between two approaches will enable accurate risk assessment and mitigation planning. This is a collaborative project of GTI, ASU, and CU-Denver aiming to develop and implement this methodology for plastic pipeline systems. The specific goal of the ASUs research is: 1) Develop an automatic damage precursor identification methodology using Bayesian/maximum entropy network; 2) Develop a reliability-based maintenance scheduling optimization framework for plastic pipeline systems.
|Effective start/end date||10/1/15 → 9/30/17|
- DOT: Pipeline Hazardous Materials Safety Administration (PHMSA): $179,913.00
Explore the research topics touched on by this project. These labels are generated based on the underlying awards/grants. Together they form a unique fingerprint.