Artificial Intelligence Based Approach for Predicting Fatigue Strength Using Composition and Process Parameters

Accurate prediction of the fatigue strength of steels is vital, due to the extremely high cost (and time) of fatigue testing and the often fatal consequences of fatigue failures. The work presented in this paper is an extension of the previous paper submitted to OMAE 2019. The main objective of this manuscript is to utilize Artificial Intelligence (AI) to predict fatigue strength, based on composition and process parameters, using the fatigue dataset for carbon and low alloy steel available from the National Institute of Material Science (NIMS) database, MatNavi. A deep learning framework Keras is used to build a Neural Network (NN), which is trained and tested on the data set obtained from MatNavi. The fatigue strength values estimated using NN are compared to the values predicted by the gradient boosting algorithm, which was the most accurate model in the OMAE 2019 paper. The comparison is done using metrics such as root mean square error (RMSE), Mean Absolute Error (MAE), Coefficient of Determination (R2) and Explained Variance Score (EVS). Thereafter, the trained NN model is used to make predictions of fatigue strength for the simulated data (1 million samples) of input parameters, which is then used to generate conditional probability tables for the Bayesian Network (BN). The main advantage of using BN over previously used machine learning algorithms is that BN can be used to make both forward and backward propagation during the Bayesian inference. A case study illustrating the applicability of the proposed approach is also presented. Furthermore, a dashboard is developed using PowerBI, which can be used by practicing engineers to estimate fatigue strength based on composition and process parameters.

Additive Manufacturing in the Oil and Gas Industry Overview

The API 20S Standard is designed to play a crucial role in leveraging Additive Manufacturing (AM) to foster innovation in the oil and gas industry. The paper, in association with the standard, will facilitate the understanding of how AM will enable equipment design improvements, faster prototyping, and better inventory management. By way of discussing the progress, challenges, and lessons learned from the standardization process, the paper aims to encourage a safer, broader, and faster adoption of AM technologies in the mainstream oil and gas applications. The paper will summarize the streamlining process, feedback from the API 20S task group, and current status of the standardization efforts. Additionally, upcoming challenges and the potential for the oil and gas industry industries to contribute to the standard will be summarized. The paper will also showcase a novel tiered approach (Additive Manufacturing Specification Levels) to allow the users of the document to match different levels of criticality.

On the Effect of Various Heat Treatments on Microstructure of AISI 4130 Steel Used in Sour Service Pipes

Nowadays welding is the most common way to connect metal parts and structures. One of the challenges connected to welding it that heat output from the welding alters the microstructure of the metal creating the heat affected zone (HAZ) near the weld. In steel welds HAZ is often harder and more brittle than the base material due to formation of martensite. This might cause hydrogen induced cracking and speed up the fatigue of the weld. To mitigate the martensite formation in the HAZ different heat treatments, like preheat, interpass and PWHT are often applied. However, for 4130 steel, preheat and interpass temperatures are not expected to restrict martensite formation due to materials slow transformation rate. Preheat and interpass temperatures are still important for hydrogen diffusion and reduction of tension in the weld. This paper investigates the effect of different heat treatments on the microstructure of AISI 4130 steel used in sour service pipes. The welding and sample preparation were performed in accordance with ISO 15156 and ASME B31.3 standards. Two sample sets were produced: one with and one without preheating. The hardness tests of weld profiles were performed in accordance with ISO 15156-2 international standard. Comparison of hardness profiles indicated that preheat had virtually no effect on hardness of the steel in HAZ, although it affected hardness of fusion zone. Preheated samples were further heat treated in a furnace simulating PWHT effect. Three different PWHT condition were tested. The hardness profiles indicated that PWHT led to noticeable changes in steel microstructure. In order to understand those microstructure changes, the heat treatment of the steel during production process was reviewed and microscopic investigations of the weld profiles were performed.

Residual Stresses and Stress Intensity Factor Calculations for an Orthotropic Steel Deck

The welding residual stress in an orthotopic steel deck is investigated through experimental measurements and finite-element simulations. The simulated residual stress fields are in reasonable agreement with the measured results. The weight function method are used to investigate the stress intensity factors (SIFs) at the surface and deepest points of the semi-elliptical surface cracks, subjected to a combination of external load and through-thickness welding residual stress. Different crack aspect ratios and relative depths are analyzed. The results reveal that the transverse residual stress is always tensile through the plate thickness at the middle section, which makes the SIFs of the surface and deepest points larger than those without considering the residual stress. However, the non-linear reduction for transverse residual stress through the thickness causes the SIFs to decrease for aspect ratios of 0.4, 0.6 and 1.0.

Machine Learning Approach for Estimating Residual Stresses in Girth Welds of Topside Piping

Residual stresses are internal self-equilibrating stresses that remain in the component even after the removal of external load. The aforementioned stress when superimposed by the operating stresses on the offshore piping, enhance the chances of fracture failure of the components. Thus, it is vital to accurately estimate the residual stresses in topside piping while performing their fitness for service (FFS) evaluation. In the present work, residual stress profiles of girth welded topside sections of P91 pipes piping are estimate using a machine learning approach. The training and testing data for machine learning is acquired from experimental measurements database by Veqter, UK. Twelve different machine learning algorithms, namely, multi-linear regression (MLR), Random Forest (RF), Gaussian process regression (GPR), support vector regression (SVR), Gradient boosting (GB) etc. have been trained and tested. In order to compare the accuracy of the algorithms, four metrics, namely, Root Mean Square Error (RMSE), Estimated Variance Score (EVS), Maximum Absolute Error (AAE), and Coefficient of Determination (R^2) are used. Gradient boosting algorithm gives the best prediction of the residual stress, which is then used to estimate the residual stress for the simulated input parameter space. In the future work authors shall utilize the residual stress predictions from Gradient boosting algorithm to train the Bayesian Network, which can then be used for estimating less conservative through-thickness residual stresses distribution over a wide range of pipe geometries (radius to thickness ratio) and welding parameters (based on heat input). Furthermore, besides topside piping, the proposed approach finds its potential applications in structural integrity assessment of offshore structures, and pressure equipment’s girth welds.

Validation of the Master Curve Approach With Various Welding Conditions: Groove Shapes, Heat Inputs and Welding Processes

Engineering critical assessment (ECA) is a procedure for evaluating the soundness of structures with flaws and has been widely applied for assessing the structural integrity. ECA procedure requires reliable fracture toughness data to assess the effect of defects. Ideal data are typically obtained from samples taken during construction of an engineering structure or from the structure afterward, but there are cases in which removal of the test samples is impossible due to the continued operation of the structure. To this end, Appendix J of the BS 7910 provides a procedure for estimating fracture toughness values from appropriate Charpy impact test data. However, the correlation between Charpy impact energy and fracture toughness is known to be overly conservative with not sufficient theoretical background in fracture mechanics perspective. In this regard, the revised BS 7910:2019 provides an improved method for calculating the reference temperature by applying the yield strength and the Charpy upper shelf energy based on empirical data. The target of this study is to validate the master curve approach in the modified BS 7910 for two common offshore grade steels with explicit considerations for various groove shapes, heat inputs and welding processes. For the purpose, the master curves are compared in terms of the reference temperature calculated from Charpy impact test according to BS 7910:2013 and the newly revised 2019 version of BS 7910. The modified master curve resulted in less conservative fracture toughness values anticipated from the decreased reference temperature. The estimated fracture toughness values exhibited a good correlation with experimentally obtained toughness values. The influence of various groove shapes, heat inputs and welding processes in estimating fracture toughness based on the master curve approach is discussed. In addition, the effect of impact test sample locations within weld metals toward estimated fracture toughness values is evaluated.

Characterization of Magnetic Field in Crack Propagation Stage and Application to Non-Destructive Evaluation

Monitoring of fatigue crack propagation is very important in industrial fields. Stress-induced magnetic measurement is a newly developed non-destructive testing technique which can detect early failure of ferromagnetic materials in service. A lot of experiments demonstrate that magnetic measurement is more sensitive compared with other non-destructive testing technologies. In order to explore the correlation between crack propagation and magnetic hysteresis under cyclic stress in X70 pipeline steel, a series of stress-controlled tests were carried out and the changes in magnetic field around cracks during their propagation process were observed throughout the tension-tension fatigue tests. The variations of magnetic field and stress intensity factor K in the crack propagation stage were studied. The results obtained allowed the division of the magnetic behavior of the investigated steels into three stages corresponding to three distinct crack propagation stages. It was found that the magnetic field varies with crack propagation, and the stress intensity factor K increased with the increase of loading cycles. A strong correlation between the variation of the magnetic field and stress intensity factor was recognized, regardless of the loading conditions, maximum load or stress ratio. The results suggest that non-destructive evaluation of fatigue cracks would be possible using this relationship.

Numerical Evaluation of Crack in Polyvinylidene Fluoride (PVDF)

Polyvinylidene fluoride (PVDF) is an engineering thermoplastic having a high degree of sensibility to crack, which affects long-term mechanical behavior. This study evaluates the crack-sensitive of PVDF for one commercial-grade through the development of a numerical model. Firstly, tensile tests using DIC were performed on both uncrack and pre-crack specimens to get experimental tensile as DIC-displacement, displacement-control, and load data. For pre-crack specimens, it was proposed two values of depth: 1.0 and 1.5 mm, opened by razor blade. All specimens were uniaxial tests at 23°C under 5 mm/min. Secondly, tensile tests using extensometer were implemented for uncrack samples to determine material parameters for calibration of the numerical model and comparison with DIC-displacement. Finally, a numerical model based on the FE was implemented using ANSYS-student that inputs PVDF’s material properties, which considered the elastic-plastic behavior in simulation tests. The PVDF demonstrated significant crack sensitivity, as it can be seen in experimental and numerical data. And, the numerical model developed based on MKHP was successfully agreement against experimental data obtained by Blue Hill 3 software. Therefore, the results allowed us to observe that pre-crack acts as a stress concentration and the numerical model got well simulates this influence on the PVDF mechanical behavior.

Collapse Resistance Under Combined External Pressure and Bending Deformation of Coated Linepipe

As offshore pipeline projects have expanded to deeper water regions with depths of more than 2 000 m, higher resistance against collapse by external pressure is now required in linepipe. Collapse resistance is mainly controlled by the pipe geometry and compressive yield strength. In UOE pipe, the compressive yield strength along the circumferential direction changes dramatically due to tensile pre-strain that occurs in pipe forming processes such as the expansion process. In order to improve the compressive yield strength of pipes, it is important to consider the Bauschinger effect caused by pipe expansion. As the mechanism of this effect, it is understood that internal stress is generated by the accumulation of dislocations, and this reduces reverse flow stress. Compressive yield strength is also changed by the thermal cycle associated with application of fusion-bond epoxy in pipe anti-corrosion coating by induction heating. In the typical thermal heat cycle of this coating process, the maximum heating temperature is from 200 °C to 250 °C. In this case, compressive yield strength increases as an effect of the thermal cycle, resulting in increased collapse resistance. Thus, for deep water application of UEO linepipe, it is important to clarify the conflicting effects of the Bauschinger effect and the thermal heat cycle on compressive yield strength. During installation of deep water pipelines by a method such as J-lay, curvature is imposed on the pipe axis, but the circumferential bending that leads to ovalization is determined by the interaction of the curvature of bending deformation. This bending deformation decreases collapse resistance. The interaction of external pressure and bending is also important when evaluating collapse. Against this background, this study discusses the collapse criteria for coated linepipe and their bending interaction against collapse based on a full-scale collapse test under external pressure with and without bending loading. The effect of the thermal heat cycle on linepipe collapse criteria is also discussed based on the results of tensile pre-strain tests with simulation of the thermal cycle and a collapse calculation by FEA.

Development of 3D FE Models for Incremental Forming Process Analysis of UOE Linepipes

Three dimensional (3D) FE models have been developed for simulation of the incremental forming processes employed in UOE linepipe manufacturing at Tata Steel Hartlepool 42” Pipe Mill, namely edge crimping (C-press) and pipe expansion (expander). Transitional zone behaviour resulted from both forming operations as previously identified in practice has been revealed, for the first time, by the 3D FE simulations and preliminarily analysed. It was demonstrated that the transitional zone features observed in both edge crimping and pipe expansion were initiated in the plate/pipe feed-in side/area of the dies, and then formed within the working length/face of the forming dies. Detailed examination and analysis of the FE results, in terms of plastic strain and residual stress patterns as well as contact pressure distributions, have shown that the initiation of the transitional zone features was due to the redundant plastic deformation caused by (cantilever) bending in the material feed-in area of the dies mainly along the longitudinal direction. It is therefore believed that minimisation/elimination of such redundant plastic bending effect in the longitudinal direction would lead to minimisation/elimination of the unfavourable deformation features in the so-called transitional zones, which could be achieved through improved die designs and possibly forming process parameter settings.