No description available.

No description available.

No description available.

No description available.

No description available.

The micromechanisms by which ductile fracture extended from a pre-existing crack was experimentally observed for two classes of forged SA 508 pressure vessel steel. The micromechanisms were related to the measured values of fracture toughness characterised by the resistance to crack initiation and growth. This comparison was only possible with the aid of an accurate fracture resistance test technique which could determine the crack growth toughness from a single specimen. An unloading compliance test system was developed and was used for the construction of crack growth resistance curves. Microstructural parameters determined from a specimen were related to the toughness measured on that specimen and this proved invaluable in isolating the controlling parameters. The effect of orientation and location on the toughness of the materials was assessed. The crack growth resistance was sensitive to the orientation of the crack with respect to the maximum hot working direction and the bands of segregation associated with elongated manganese sulphide inclusions. The toughness was high when the crack plane was perpendicular to the segregation bands and low when the crack plane was parallel with the bands. The location of the crack-tip through the thickness of the forging had a minor effect on the crack growth resistance. A limited study of test temperature, strength level and isothermal ageing was undertaken. Testing within the dynamic strain ageing regime of temperature had a marked effect and reduced the crack growth resistance to below the value at room temperature. Increasing the strength level of one steel by re-heat treating had no effect on the crack growth resistance. Subsequent isothermal ageing treatments also had no effect on the resistance curves. The magnitude and extent of void formation around growing cracks was studied and related to the applied loading. The size, shape and distribution of inclusions was characterised for the materials and orientations used in the fracture tests. Correlations between inclusion parameters and toughness revealed the important microstructural parameters controlling initiation and crack growth. Simple models for initiation and crack growth resistance were developed which take the controlling parameters into account. These models are shown to agree reasonably well with some experimental data.

The R6 procedure is used for the prevention and prediction of crack behaviour and other defects in the reactor pressure vessel(RPV). The RPV material is an upper-bainitic, low alloy steel structure, which deforms inhomogeneously when yielding. The current codes that are used to design and calculate the fracture, within an RPV, assume that the material yields continuously as the size of the L¨uders strain is less than 2%. However, the work of Wenman et al[1] has shown that the inclusion of a L¨uders band during calculations can reduce the residual stress in a material, when compared to standard work-hardening models and, consequently, reduces the amount of conservatism. The objective of the research was to determine whether Wenman’s finding could be generalised and therefore initiate a re-evaluation of R6 procedure, when looking into materials that yield discontinuously. This required further investigation into L¨uders bands, such as using failure assessment diagrams (FADs). The findings from FADs showed that at the temperature range for an RPV steel at -155±C for different micro-structures (assuming that the material deforms homogeneously), this reduced the amount of conservatism. However, at fracture toughness values more representative of room temperature behaviour, the converse was true. That is, assuming a discontinuous yield point reduced the amount of conservatism. It was also shown that the tempered martensite structure could be used as an alternative to the current upper bainitic, low alloy steel that is used in RPVs. Further insight is gained into the nature of a L¨uders band, by developing a theoretical model that showed explicit relations between L¨uders strain and the mean free-path(ferrite path), dislocation density and the grain-size. It was also shown that an explicit relation between the L¨uders strain and carbon content was possible from known data, which a new parameter Á was derived, and is the derivative of the work-hardening exponent with respect to the lower yield stress.

This study used computer aided design software Solid Work to draw four models of pressure vessel, and to analyze the displacement and the stress by the finite element analysis software ANSYS. To carry on the main body of the pressure vessel and find the highest stress of the pressure vessels by finite element analysis. The stress analysis of the pressure vessel main body contains main nozzle, the skirt of the main body ban and the connected control line. And the stress analysis factor includes: the stress distribution situation by seismic force and the displacement change factor of the wind power and the stress distribution condition of the thermal load by expand with heat and contract with cold (normal temperature climb to high temperature). The researcher also discussed the difference of the stress distribution between individual analysis and the overall analysis. The present study used finite element analysis (contain main body, spray nozzle, skirt in view of the overall analysis ban) to carry on the shell individual analysis first, then using the boundary condition of the result displacements regarding connected spray nozzle, the pipeline by the shell analysis again carries on stress analysis of the spray nozzle and the pipeline. Based on the results of stress analysis by the finite element method, the researcher discussed the differences of stresses between overall analysis and the individual analysis results.

The R6 procedure is used for the prevention and prediction of crack behaviour and other defects in the reactor pressure vessel(RPV). The RPV material is an upper-bainitic, low alloy steel structure, which deforms inhomogeneously when yielding. The current codes that are used to design and calculate the fracture, within an RPV, assume that the material yields continuously as the size of the L¨uders strain is less than 2%. However, the work of Wenman et al[1] has shown that the inclusion of a L¨uders band during calculations can reduce the residual stress in a material, when compared to standard work-hardening models and, consequently, reduces the amount of conservatism. The objective of the research was to determine whether Wenman’s finding could be generalised and therefore initiate a re-evaluation of R6 procedure, when looking into materials that yield discontinuously. This required further investigation into L¨uders bands, such as using failure assessment diagrams (FADs). The findings from FADs showed that at the temperature range for an RPV steel at -155±C for different micro-structures (assuming that the material deforms homogeneously), this reduced the amount of conservatism. However, at fracture toughness values more representative of room temperature behaviour, the converse was true. That is, assuming a discontinuous yield point reduced the amount of conservatism. It was also shown that the tempered martensite structure could be used as an alternative to the current upper bainitic, low alloy steel that is used in RPVs. Further insight is gained into the nature of a L¨uders band, by developing a theoretical model that showed explicit relations between L¨uders strain and the mean free-path(ferrite path), dislocation density and the grain-size. It was also shown that an explicit relation between the L¨uders strain and carbon content was possible from known data, which a new parameter Á was derived, and is the derivative of the work-hardening exponent with respect to the lower yield stress. / © Crown Copyright

The storage of pressure liquefied gas in vessels is subject to various regulations and codes of practice. For example, Liquefied Petroleum Gas (LPG), a commercially relevant product, is subject to Health and Safety Executive Guidelines regarding cylinder/tank arrangements and spacing. In the event of an incident involving fire, the internal pressure and shell temperature of an LPG vessel will rise, and the weakening of steel at elevated temperatures can result in the structural failure of the shell. This can be avoided by the fitting of pressure relief valves, which vent material at a pre-set pressure. However, an ignited release can create a high velocity jet flame which, because of significant radiative and convective components, can generate intense, localised heat loads on neighbouring vessels or pipe-work. However, existing codes of practice have no special provision for the possibility of jet fire incidents. Owing to a lack of detailed information on the thermal response of a LPG vessel exposed to jet flame impingement, a series of laboratory scale tests with simulated, localised jet fire impingement on the exterior shell of a pressure vessel was required. The thermal response and the effects of key parameters, Le. fill level, magnitude of heated zone (Le. size and intensity) and position of simulated impingement, could then be examined for the part-validation of a suitable computer model. In addition, these studies could be used to interpret the results from concurrent full scale jet fire impingement trials. An appropriate pressure vessel was constructed to standard design codes, which incorporated a vent line and dump tank. A suitable LPG substitute was selected. Results from the studies indicated that mixing, and therefore thermal stratification, was highly dependent on the size of the heated zone and its position in relation to the liquid/vapour interface. High Speed Micro-Cinematography was successfully employed to film individual bubble streams within the vessel and to measure individual bubble sizes and velocities for various experimental configurations. Studies were also made on the venting characteristics. Sudden pressure relief caused severe agitation of the liquid phase and the breakdown of thermal stratification. In addition, swelling and aerosol generation through homogenous boiling within the liquid phase was observed. Comparisons with the nodal computer model revealed that the use of only single vapour and liquid nodes was a poor approximation to the detail observed in the small scale studies, where the incident heat flux was relatively low and the simulated region of impingement was highly localised. However, the bulk liquid and vapour temperatures and the pressure response up to the time of venting was generally well predicted. As the degree of engulfment increased the model became a better approximation. Although the full scale trials employed an almost fully engulfing jet flame rather than point source impingement, comparisons have allowed understanding of the liquid and vapour thermal gradients, and the subsequent breakdown of these during venting.