Lifetime Analytic Prognostic for Petrochemical Pipes Subject to Fatigue Abdo Abou Jaoude*, Hassan Noura**, Khaled El-Tawil***, Seifedine Kadry****, and Mustapha Ouladsine***** * Paul Cézanne University Aix-Marseille, France, and the Lebanese University (EDST), Lebanon, (Tel: 961-3-303034; email: abdoaj@idm.net.lb) ** United Arab Emirates University, Department of Electrical Engineering, (email: hnoura@uaeu.ac.ae) *** Lebanese University, Faculty of Engineering & EDST, (email: khaled_tawil@ul.edu.lb) **** American University of the Middle East, Kuwait, (e-mail: skadry@gmail.com) ***** Laboratoire des Sciences de l'Information et des Systèmes (LSIS, UMR CNRS 7296), Paul Cézanne University, Aix-Marseille, France, (email: mustapha.ouladsine@lsis.org) 1. INTRODUCTION Prognostic is a process encompassing a capacity of prediction. It is the ability to estimate the remaining useful lifetime (RUL) of equipment in terms of its functioning history and its future usage. Predicting the RUL of industrial systems becomes currently an important aim for industrialists knowing that the failure, whose consequences are generally very expensive, can occur suddenly. The classical strategies of maintenance [1] based on a static threshold of alarm are no more efficient and practical because they do not take into consideration the instantaneous product functioning state. The introduction of a prognostic approach as an "intelligent" maintenance consists of the analysis, the health follow up and monitoring, based on physical measurements using sensors. Previous specialized prognostic studies belong generally to three types of models: the first type is the "Data-based models" relying on the statistics of large measured data (as examples we can cite the works based on degradation behavior described by abaci and using expert description of system: Process-Mission-Environment [2], the works based on artificial intelligence, machine learning [3], neural network [4], fuzzy logic [5], etc.). Their methodologies are described as not very precise and efficient. The second type is the "Experience-based models" [4] (based on measurements taken from health monitoring of machine like for example those based on expert judgment, stochastic model, Markovian process, Bayesian approach, Reliability analysis, Optimization of preventive maintenance, etc.). Their prognostic methodology proves to be simple but inflexible toward changes in system behavior and environment. The third type is the "Physical models" based on mathematical description of degradation process and its level evolution using NDI monitoring (Non-Destructive Inspection). It is described to be more flexible and precise than the two first types. This paper presents an analytical prognostic methodology based on analytic laws of damage, such as Paris' and Miner's laws. It belongs to the third type of models. Whenever the damage law of the studied system is available analytically, the advantage of this approach is therefore its realistic and precise features in determining the system remaining useful lifetime (RUL). Pipelines are petrochemical systems transporting oil and natural gas over long distances and in huge quantities. Their life prognostic is vital in this industry since their availability has crucial consequences. The DNV 2000 rules propose for pipelines a target probability of failure about 10 -5 . Their main failures are due to seismic ground waves, soil settlements, buckling, deformations, internal and external corrosion, stress concentration in welding and fitting, vibration and resonance, pressure fluctuation over long period. The fatigue failures by cracks propagation are detected by cracks detection tools. Three case studies of pipes are considered here: unburied, buried and subsea (offshore pipes). Each one of these situations requires different physical parameters like: corrosion, soil pressure and friction, water and atmospheric pressure. This paper is organized as follows: first the mechanical fatigue model is presented, followed by a proposed Abstract: The high availability of technological systems like aerospace, defense, petro-chemistry and automobile, is an important major goal of earlier recent developments in system design technology. In petrochemical industries, pipelines are the principal element of hydrocarbon transport systems. They are crucial for human activities since they serve to transport oil, water, and natural gases from sources to all sites of consumers. A new analytic prognostic model is developed in this paper and applied to three cases of pipes subject to internal pressure, to the effects of corrosion, and to soil loading. This will initiate micro-cracks in the tubes body that can propagate suddenly and lead to failure. The increase of pipes performance, availability, and the reduction of their global mission cost, need to develop a suitable prognostic process. Consequently, a new approach based on analytic laws of degradation is proposed in this paper. From a predefined threshold of degradation D, the Remaining Useful Lifetime (RUL) is estimated from this prognostic model. Keywords: Prognostic, Pipelines systems, Fatigue, Paris' law, Miner's law, Degradation trajectories, Analytic model, RUL. 8th IFAC Symposium on Fault Detection, Supervision and Safety of Technical Processes (SAFEPROCESS) August 29-31, 2012. Mexico City, Mexico 978-3-902823-09-0/12/$20.00 © 2012 IFAC 707 10.3182/20120829-3-MX-2028.00143