This work presents the finite element analysis of partially stiffened cylinders subjected to axial compression at elevated temperatures. The compressive strength is calculated for self-weight conditions and the influence of the temperature on the material response is also investigated. In the oil industry, pressure vessels are commonly used operating at complex design conditions such as high-pressure profiles and/or elevated temperature gradients which affect considerably the structural response of inner components. Among them, risers become sensitive steel elements withstanding heavy compressive loading due to self-weight, as well as, insulation elements added toprotect them from the elevated temperature gradient. Most risers structurally fail at the bottom end due to buckling caused by self-weight and temperature effects. To remediate this situation and to guarantee the integrity of the riser, longitudinal stiffeners are welded at the bottom end. Hence, a proper determination of the compressive strength of the cylinder, taking into account the influence of the longitudinal stiffening and the corresponding temperature, is required.Results indicate that the use oflongitudinal stiffeners in deformed cylinders increases the strength to buckling in percentages that vary according to the cross-section of the profiles.

Several studies recognized that cracks in delayed coke drums resulted from low cycle fatigue induced by cyclic thermal stress [1], [2], [3]. According to a coke drum survey coordinated by API in 1996 [1], there are two different areas where cracks are produced. The first zone is located at the shell to skirt weld, and the second at the bulging areas found in the cylindrical section. In the second case, from 145 coke drums 57% reported that had shell bulging problems. Of the drums that bulged, 87% also showed cracks. In order to estimate the level of stress, finite element analysis of a bulged cylinder was performed. The study was carried out running a sequentially-coupled thermo-mechanical analysis. In the first step, a thermal analysis is used to estimate the variation of temperature with time. These temperatures serve as input to a mechanical stress analysis that was made in a second step. The assessment shows the influence of thermal gradients obtained for a period of 26 months for several bulging patterns that were identified from the analysis of 28 laser mappings of coke drums. The results indicated that the level of stress could reach the plastic deformation considering that the material has yield strength equal to the minimum specified in ASME Section II, part D. The zones where maximum axial stresses were found are consistent with areas where cracking has been reported in many coke drums.

Delayed coke drums are vertical thin-walled pressure vessels that operate under severe conditions by cyclic heating and quenching operations. The high stresses can lead to premature drum failure in the form of through wall cracking resulted from low cycle fatigue. One of the ways that has been used to reduce the growth of cracks localized near the shell to skirt weld is by the addition of different slots configurations in the skirt. In order to estimate fatigue life, a finite element analysis (FEA) was carried out using a decoupled thermo-mechanical assessment. In the first step, a thermal analysis was done to estimate the variation of temperature throughout the coking cycle. In the second step, this thermal response was combined with the mechanical loads to estimate the total stress and strain. The validity of the models has been demonstrated by several FEA results and by some field measured data. It is proposed a set of mathematical relations that allows predicting the behavior of fatigue lifetime as a function of slot size and hole radius.

In recent years the understanding of the relationship between drum damage and bulge sharpness has improved significantly. The authors of this paper developed a new parameter called bulge sharpness and have previously shown the relationship between sharpness and observed damage. Further to this study, the authors have exhaustively studied the evolution of stress cracking (elephant skin) on mid-course bulges and have estimated the likelihood of finding a particular type of surface damage based on the observed sharpness levels. This correlation has led to a proposed scale to categorize stress cracking into three levels: minor, intermediate, and significant. In addition, the progression of bulge sharpness over time was analyzed and it was determined through statistical modeling that bulge sharpness can have a range of rates of change or sharpness growth rates: low, medium, and high. These sharpness growth rates were subsequently studied and their relationship with overall cycle times analyzed. The study also shows that individual coke drums can experience different sharpness growth rates and there can be a distribution of these rates. To determine when repairs should be conducted, coke drum operators must consider the expected operational run. While the random nature of coke drum damage can defy such targets, bulge sharpness growth assessments can be used to better define when repairs should be conducted. Understanding current bulge sharpness levels, year-over-year sharpness growth rates and their distribution, can significantly assist in targeting areas of concern for optimized repair strategies and can also be used to avoid unnecessary repairs.

Laser mapping is a well-accepted technique for obtaining surface profiles of coke drum walls to identify bulges. The resulting data is used to track and trend vessel distortions and mapped to illustrate the shape of the vessel from a base radius and or previous inspection. Monitoring the development and evolution of these distortions over time in an accurate and consistent manner has been demonstrated to be an effective tool for predicting bulging induced crack in a coke drum. In this paper, the authors discuss several aspects of the laser mapping technique such as scanner positioning, data noise, laser range accuracy, missing data and the repercussions in the assessment of bulging. The effects on second derivative-based analysis are covered in detail and the use of techniques to reduce the effects of data noise and sensor motion are discussed.