Ensemble of Hybrid Neural Networks to Compensate for Epistemic Uncertainties: A Case Study in System Prognosis

In this contribution, a case study considering an unexpected corrosion-fatigue crack propagation issue in an aircraft fleet is used to discuss how to compensate for incomplete knowledge in time dependent responses integration and extrapolation. For the considered application, degradation resulting from mechanical fatigue is well understood and accounted in the damage models. However, the unexpected corrosion effects are not accounted in damage integration, yielding a large discrepancy between predicted and observed crack lengths. To address this epistemic uncertainty in the fleet damage accumulation model, hybrid neural networks cells are formulated; where physics-informed layers address well-understood aspects of the degradation, and data-driven layers are trained to act as correction terms. The considered case study encompasses highly imbalanced data sets with uncertainties acting asynchronously. To improve overall accuracy, ensemble learning techniques are adapted to merge the resulting hybrid neural network cells predictions. Lastly, a heuristic based on optimal ensemble weights is presented to help in the decision-making task of defining safe operation of the fleet. Results show that our proposed approach was capable of compensating for the epistemic uncertainties, and that the proposed heuristic can be used to rank aircraft damage severity, allowing to prioritize aircraft for inspection and/or route reassignment.

Study on corrosion fatigue behavior and mechanism of 6005A aluminum alloy and welded joint

Purpose The purpose of this paper is to determine the corrosion fatigue behavior and mechanism of 6005A aluminum alloy and welded joint. Design/methodology/approach Electron back-scattered diffraction (EBSD) were adopted to characterize the microstructure of 6005A aluminum alloy and welded joint. Through potentiodynamic polarization, electrochemical impedance spectroscopy (EIS) and corrosion fatigue experiments, the corrosion fatigue behavior and mechanism of 6005A aluminum alloy base metal and welded joint were studied. Findings The results show that the corrosion fatigue crack initiation of 6005A aluminum alloy base metal and welded joint is mainly caused by the preferential anodic dissolution and hydrogen concentration in the areas with inclusions and welding defects. Originality/value The research is an originality study on the corrosion fatigue behavior and mechanism of 6005A aluminum alloy and welded joint.

Corrosion-Fatigue Crack Growth Performance of Titanium Grade 29 Welds in Tapered Stress Joints

Design of a steel catenary riser requires the use of connection hardware to decouple the large bending moments induced by the host floater at the hang-off location. Reliability of this connection hardware is essential, particularly in applications involving high pressure and high temperature fluids. One option for this connection hardware is the metallic tapered stress joint. Titanium (Ti) Grade 29 has been identified as an attractive material candidate for demanding stress joint applications due to its “high strength, low weight, superior fatigue performance and innate corrosion resistance”.2 Titanium stress joints for deepwater applications are typically not fabricated as a single piece due to titanium ingot volume limitations, thus making an intermediate girth weld necessary to satisfy length requirements. As with steel, the potential effect of hydrogen embrittlement induced by cathodic and galvanic potentials must be assessed to ensure long-term weld integrity. This paper describes testing from a joint industry project (JIP) conducted to qualify titanium stress joint (TSJ) welds for ultra-deepwater applications under harsh service and environmental conditions. Corrosion-fatigue crack growth rate (CFCGR) results for Ti Grade 29 flat welding-groove weld (1G/PA) gas tungsten arc welding (GTAW) specimens in seawater under cathodic potential and sour brine under galvanic potential are presented and compared to vendor recommended design curves.

EVALUATION OF CHARACTERISTICS OF RESISTANCE OF PROPAGATION OF CORROSION- FATIGUE CRACKS OF LONG-TERM OPERATED DRILL PIPES

The aim of the proposed article is to determine the patterns of the corrosion-fatigue cracks in long-term operational metal elements of drill strings (steel 45, 36G2S, 40HN) when drilling wells in corrosive environments (drilling fluid "Biocar", potassium polymer mud and on the air). Based on the obtained results of experimental studies, diagrams of cyclic corrosion crack resistance for the studied systems "metal - medium" were drawn. The cyclic crack resistance characteristics of long-used metals of drill string elements are determined - the values of constants (C and n) in power dependence of Paris, as well as the values of threshold and critical stress intensity factory. The influence of drilling mud of the corrosion crack resistance of the investigated metals of drill string elements was estimated quantitatively. In particular, in the environment of drilling mud "Biocar" the process of corrosion-fatigue crack development in the studied metals of the elements of drill strings is not significantly, in the range of 1-5%, differs from the development of fatigue crack on the air. Whereas in the potassium polymer drilling mud media, the rate of corrosion-fatigue cracks development in the investigated metals of the drill strings elements significantly, in the range of 15-35%, differs from the development of fatigue cracks on the air. It is established that under the influence of working loads the process of the corrosion-fatigue crack development in long-operated elements of drilling string is accelerated by more than 10% in the media of the potassium polymeric drilling mud than in the environment of drilling mud "Biocar".

The Fatigue Behavior and Mechanism of Large FV520B-I Specimens in a Very High Cycle Regime

We examined the fatigue properties in very high cycle regime of large FV520B-I specimens in an ultrasonic fatigue test. The fatigue mechanism in very high cycle regime didn’t change and the fatigue properties obviously degraded as the specimen size enlarged. The fatigue life decreased and the S-N curve moved downward due to the increase of inclusion size in large specimens. The maximum inclusion sizes in specimens were predicted by the method of statistics of extreme value. The prediction of fatigue strength using the modified Murakami model was closer to the test result, and the fitting of fatigue life using the corrosion fatigue crack initiation life model was less effective compared with the fitting of small specimen test results