Scale testing of profiled stainless steel blast walls

Transcription

1 Scale testing of profiled stainless steel blast walls G. S. Langdon & G. K. Schleyer Department of Engineering, University of Liverpool, UK Abstract A series of laboratory tests were carried out at the University of Liverpool Impact Research Centre on panels of ¼ scale stainless steel blast walls and on various panel/connection systems. The panel design was based on a deep trough trapezoidal profile with welded angle connections top and bottom and free sides. The loading applied to the test panel was a triangular pulse pressure representative of a gas explosion overpressure. The aim of this work was to investigate the influence of the connection detail on the overall performance of the panel/connection system under pulse pressure loading and to develop appropriate analytical and numerical models for correlation with the test results. Large permanent plastic deformations were produced in the panel without rupture. The work has shown that the connection detail can significantly influence the response of the panel to extreme pressure loading. Keywords: blast walls, corrugations, connections, buckling, pulse pressures. 1 Introduction An extensive programme of experimental work has been carried out at the University of Liverpool Impact Research Centre on ¼ scale stainless steel blast wall panels with connections subjected to pulse pressure loading to determine their various modes of failure and blast resistance beyond the design limit. The aim of the work was to investigate the influence of the connection detail on the overall performance of the panel/connection system under pulse pressure loading and to develop appropriate analytical and numerical models of the blast wall/connection system for correlation with the test results. This is the first time that an experimental study of the behaviour of blast walls made from profiled

2 112 Structures Under Shock and Impact VIII stainless steel sheet has considered a detailed investigation of the modes of failure and end effects of the support construction. Blast loading of various flat, stiffened and corrugated panels has been studied by previous researchers [1-4] and by the blast wall manufacturers themselves, with the aim of enhancing their blast resistance capacities. However, due to the design of the experimental facilities available to the blast wall manufacturers, the connection details used on offshore platforms have not been examined as part of the test programme. Experimental testing of the connection details would have required major structural modifications to the test rig and this was not commercially viable. The facilities available at the University of Liverpool were able to incorporate the connection details at a scale suitable for laboratory work. The most common design practice is to assume a single plastic hinge formation at the ends and mid-span of the blast wall (treated as a one-way member) and to ignore end effects. The capacity of the wall is therefore based on a simple bending resistance model and generally leads to conservative designs. The actual capacity of the wall to resist a dynamic event is considerably larger than the simple design estimate when considering the support construction and attachment to the primary framework. The experimental and modelling work at Liverpool has enabled a more accurate assessment of the ultimate capacity of the blast wall based on the influence of the connection detail. A reliable method of including the connection detail in the assessment of the blast response of the wall should result in improved safety through better understanding of the blast wall behaviour in an accidental gas explosion. The paper provides an overview of the work carried out on the scaled panels and includes typical test results. Some finite element (FE) modelling results of the blast wall are briefly described and compared with the experiments. 2 Blast wall design Typical blast walls consist of stainless steel profiled sheet, about 12 m wide and 4 m high as shown in Fig. 1. Blast walls are welded top and bottom and down both sides to the primary steelwork through angle connections. Due to the length to height ratio, the sides of blast walls are usually represented as free for analysis purposes. This allows them to be analysed as a one-way spanning member. The top and bottom connections usually consist of two angles welded together and to the structural steelwork of the platform. Blast walls are rated to a certain pressure, that is they are designed by the manufacturer to resist a certain dynamic pressure (usually in the range of 1 to 4 bar) with some permanent deformation (typically 1/40 th of the height of the blast wall). There is no agreed, common basis for determining this rating or an acceptable deformation. In general, industry standard guidance notes [5] as issued by FABIG and the Steel Construction Institute (SCI) are used to design blast walls for fire and blast loading. Technical Note 5 [6] provides further detailed guidance on the design of stainless steel blast walls to resist explosion loading.

3 Structures Under Shock and Impact VIII 113 Figure 1: Typical blast wall construction made from profiled stainless steel sheet with end connections top and bottom. Table 1: Deep trough test panel properties. Mass of beam section / unit length 4.2 kg/m Second moment of area 13.9 cm 4 Young s modulus MN/mm 2 Moment of resistance 2.3 knm Base yield stress N/mm 2 Design dynamic yield stress N/mm 2 Elastic stiffness 2.1 kn/mm Elastic-plastic stiffness 2.1 kn/mm Plastic resistance 18.6 kn Displacement to first yield 6.61 mm Displacement to full plasticity 8.81 mm Plastic displacement mm Maximum displacement (occurs after 32 msec) mm Ductility ratio 2.6 Natural elastic period 7.5 msec The blast wall test panel design was based on a non-symmetric trapezoidal deep trough profile, angle connections top and bottom and free sides. The blast wall was rated using the time-domain equivalent SDOF method and assumed to behave as a one-way span beam with cross-sectional properties of the deep trough profile. The properties are given in Table 1. End fixity was assumed to be pinned at both ends. The resistance of the beam to dynamic loading was due to inertia, elastic bending stiffness and plastic hinge formation at the mid-span of the beam. A strain-rate enhancement factor of was used to specify the

4 114 Structures Under Shock and Impact VIII dynamic yield stress of the material. A triangular pulse load of 950 mbar overpressure and load duration of 50 msec was applied to the SDOF model and produced a maximum permanent displacement of 14 mm, considered acceptable. The blast panel was therefore rated at 950 mbar for a triangular pulse load duration of 50 msec 3 Experimental investigation 3.1 Material characterisation Quasi-static and dynamic uni-axial tensile tests were performed on the materials used to manufacture the test panels and connections at strain rates ranging from s -1 to 118 s -1. All the test panels were manufactured from the same batch of material, namely 2, 3 and 4 mm thick AISI 316L austenitic stainless steel sheet and with the same orientation. The corrugated plate was made from the 2 mm thick sheet material and the angle connections were made from the 3 and 4 mm thick sheet material. Tensile test specimens were taken from this same batch of material, half of the specimens with their major axis in the rolling (longitudinal) direction and half with their major axis perpendicular to the rolling (transverse) direction. The entire material test results for the 2, 3 and 4 mm thick specimens in both rolling directions are presented in Table 2 as a range of values. Further details on the properties for each thickness and orientation are available in reference [7]. Table 2: Material properties for 2,3 and 4 mm thick stainless steel specimens. Static Yield (0.2% proof) MPa Variation in dynamic Yield (compared to static Yield) +18 to -7 % Cowper-Symonds constant D (at Yield) Cowper-Symonds constant q (at Yield) Static UTS MPa Variation in dynamic UTS (compared to static UTS) to -7.2 % 3.2 Panel tests Pulse pressure tests were performed on three types of panels in both the A and B loading directions as shown in Fig. 2. The difference between the three panels was in the length of 3 mm thick flexible angle. All other dimensions remained the same. The tests gathered data on the influence of the connection on the overall response of the panel and to helped to identify the modes of failure of the panels and connections. The data was also used to validate the modelling work later described. A summary of the main test results is given in Table 3. The test rig used in this work, shown schematically in Fig. 3, is capable of producing repeatable and uniform pulse pressure loading on a structural component by virtue of a dynamic pressure gradient generated across the test panel. This is achieved by the timed blow-down of two back-to-back pressureloading chambers. The test panel is clamped to a support plate that is sandwiched

5 Structures Under Shock and Impact VIII 115 between two halves of a pressure vessel. The two halves clamp together over the support plate to form a bolted joint. Thus, the support plate and test panel divides the pressure vessel into two pressure-loading chambers. A large flanged nozzle in each of the chambers allows the pressure in the chambers to be released quickly by means of a bursting diaphragm system. A controlled pressure pulse is applied to the test component essentially in four steps, namely (1) both chambers are pressurised at the same time keeping the pressure equal on both sides of the test plate, (2) the pressure is released quickly on one side, (3) a finite time later, the pressure on the other side is released quickly and (4) both sides return to atmospheric pressure conditions when the loading is over. Further details of the test facility can be found in [1] B 915 A 915 Dimensions in mm Figure 2: Short, intermediate and long angle connections on panels (side elevation view). Lifting frame Support plate Clamping frame Diaphragm clamping rings 0.5m 1m Test plate Diaphragm clamping rings Pressure loading chambers Figure 3: Schematic of the pulse pressure rig.

6 116 Structures Under Shock and Impact VIII Strain gauges were attached to most of the test panels. Other instrumentation employed included pressure sensors to measure the load and displacement transducers to measure transient response. Typical large deformation response of a test panel with long angle connections loaded in the A direction is given in Fig. 4. A similar panel loaded in the opposite direction (B) is shown in Fig. 5. Figure 4: Deformed blast wall panel C1803 subjected to a pulse pressure of 1.37 bar (A load direction). Figure 5: Deformed blast wall panel C1806 subjected to a pulse pressure of 1.22 bar (B load direction). The pulse pressure test results indicate that the blast walls have a lower elastic limit in the B direction. The tests with small and intermediate size connections gave a B-direction elastic limit of 7.3 mm whilst the results indicate an elastic limit of 11 mm in the A direction. The directionality of the elastic limit displacement is due to the asymmetry of the corrugated profile and appears independent of connection type (the connections do not move while the corrugations deform elastically). The yield pressure, defined here as the

7 Structures Under Shock and Impact VIII 117 pressure at which inelastic strains at the mid-point of the central corrugation occur or the pressure at which permanent displacement is produced in the panel (also at the panel centre as this is the point of maximum applied bending moment) varies with both load direction and connection type employed. The yield pressure is lower in the B load direction and is also reduced by increasing the blast wall flexibility, i.e. by increasing the flexible angle length. As support restraint decreases (flexible angle increases), the connections move inwards more easily and larger displacements are induced at the panel centre for a given pressure. Thus lower yield pressures were recorded for longer connection details. Table 3: Summary of pulse pressure tests on blast wall panels. Connection Type Short Intermediate Long Test Reference P peak (bar) t m (msec) t dur (msec) δ max (mm) δ perm (mm) PREBLA CORR C GCORR (B) CORR (B) 30.2 > C C C C (B) C (B) C C C C (B) C (B) P peak : peak magnitude of dynamic pressure loading in or (B) load direction; t m : time at peak magnitude of dynamic pressure loading; t dur : load duration; δ max : maximum transverse displacement at centre of panel; δ perm : permanent transverse displacement at centre of panel. 4 Finite element modelling A finite element numerical model was used to predict the large deformation behaviour of a 195 mm deep (short angle connection) panel, subjected to pulse pressures varying from 0.5 to 1.92 bar, incorporating the end connection detail as

8 118 Structures Under Shock and Impact VIII part of the structure analysed. The finite element software, ABAQUS/Standard v. 6.2, was used to perform the simulations. One half of the corrugation was modelled, due to symmetry, with connections at each end of the span. The profiled section was modelled using S4R shell elements and the connections were modelled using C3D8R solid elements. Static stress analyses were performed, and due to the anticipated unstable nature of the buckling process, a small damping factor (2 x 10-4 of the strain energy change during the step) was applied to the model. The pressure was applied as a uniformly distributed load to the profiled section, and its time varying nature defined using tables based on test data. Run-times for the dynamic models were of the order of a few hours. Corrugation stretching Buckled flange Figure 6: Residual longitudinal stress (S11) contour plot, showing the permanent shape for test ref C622, peak pressure = 1.92 bar. Plastic strain at the 3 mm thick connection angle Plastic strain at the rigid angle weld-line Compressive strain pattern (a) t = 14 msec (b) t = 32 msec Figure 7: Logarithmic strain (LE11) contour plots showing failure progression for panel C622, peak pressure = 1.92 bar.

9 Structures Under Shock and Impact VIII 119 The finite element model predicted a permanent displacement of 78 mm at the mid-span position for the 195 mm deep blast wall panel (test ref. C622) at a pressure of 1.92 bar. This compares favourably with a corresponding experimental result of 69 mm. Fig. 6 shows residual stresses (in the direction along the corrugation) contour plots after loading. The buckling behaviour is clearly visible. One advantage of FE modelling is that it can often demonstrate the failure process in more detail than that which can be observed experimentally. For example, Fig. 7 shows that the flexible angle in the connection detail yields partway through its thickness before the corrugated panel yields at the centre (14 and 32 msec, respectively), providing additional information on the failure progression of the panel. 5 Discussion and conclusions An extensive programme of experimental work has been carried out on ¼ scale stainless steel blast wall panels with connections subjected to pulse pressure loading to determine their various modes of failure and blast resistance beyond the design limit. The panel design was based on the deep trough trapezoidal profile with welded angle connections top and bottom and free sides. The panels were approximately 880 mm wide by 1000 mm high. The profile depth was 40 mm and the overall depth of the panels including end connections was 195, 255 and 315 mm, respectively. The work has shown that the connection detail can significantly influence the response of the panel to pulse pressure loading. While the blast panels could withstand pulse pressures well in excess of the rated design limit, the panels responded with large permanent plastic deformations up to 602 mm (ductility ratio 82.5 approximately) without rupture. The end restraint was a principal factor in the response of the panel. Consequently, the experimental work focused on the connection detail. The performance of the ¼ scale panels was fully investigated using a pulse pressure test facility to generate pulse pressure loading in the range bar peak pulse pressure ( msec load duration). Three types of panels were tested to compare the influence of the flexible angle length (60, 120, and 180 mm) on the overall performance of the panel/connection system. In general, the flexibility of the end connection and thus the panel increases as the angle length increases. This was observed in the yield pressure, defined as the pressure at which plastic strains at the mid-span of the panel occur or the pressure at which permanent displacement is produced in the panel. As support restraint decreases (flexible angle length increases) the connections bend more easily and larger displacements are produced in the panel for a given test pressure. For very large displacements, the panels adopt a membrane type mode in which high in-plane forces can be induced in the supports and consequently transferred to the primary framework with the risk of causing permanent damage to the framework. This will occur more readily with shorter angle connections. Thus it is important to optimise the design of the blast panel to absorb as much energy in bending and stretching and at the same time limit displacements that could affect other equipment and processes. The buckling capacity of the deep trough profile

10 120 Structures Under Shock and Impact VIII section is affected by loading direction. Thus tests were carried out in both transverse-loading directions. In the design direction, the panels exhibited high post-buckling strength and deformation resistance due to the dishing of the flange in tension and subsequent flattening of the profile. It was found that, due to the non-symmetry of the profile about the plane of the panel, buckling of the section was initiated at a lower pressure in the opposite direction. In this direction (B), the connection geometry allows the flexible angle to fold back on itself thus offering less restraint and greater flexibility of the panel. No postbuckling strengthening mechanisms were observed in the B-direction tests. The tests offered a unique opportunity to observe progressive failure of the panels at different pulse pressure loads in both the directions discussed in the previous statements. This data was used to refine the finite element numerical models of the panel/connection system. The ultimate strength of the panel, that is rupture, was never achieved although the panel tests produced deformations well beyond the rated design capacity of the panel. The design capacity of the panel was based on a simple bending mode of deformation and simple supports with no account for end restraint and large deformations. Consequently, the actual strength of the panel is substantially higher than the rated design capacity. Thus, future design of blast walls should consider a more appropriate assessment of the panel s resistance taking into account end restraint, connection flexibility and stretching modes of deformation as well as bending modes. Acknowledgements The authors gratefully acknowledge the Engineering and Physical Sciences Research Council, Mobil North Sea Ltd, the Health and Safety Executive and Mech-Tool Engineering Ltd for sponsoring this work. Thanks are also due to the following members of the Impact Research Centre who have been of assistance at various stages, namely, Mr. C. Eyre, Mr P. Smith, Dr. R. Birch, Dr. M. White and Mr. M. Simmons. References [1] Schleyer, G.K., Hsu, S.S., White, M.D. and Birch, R.S., Pulse pressure loading of clamped mild steel plates, International Journal of Impact Engineering, 28, pp , [2] Louca, L.A. and Pan, Y. Behaviour of Stiffened Plates and Corrugated Panels Subjected to Impulsive Loading, Proc. 7th Int. Offshore and Polar Engng Conf., pp , [3] Nurick, G.N. and Shave, G.C., Int. J. Impact Engng 18, pp , [4] Langdon, G.S. and Schleyer, G.K., Int. J. Impact Engng (in press). [5] Steel Construction Institute. Interim Guidance Notes for the Design and Protection of Topside Structures Against Explosion and Fire, Steel Construction Institute, [6] Steel Construction Institute. Design Guide for Stainless Steel Blast Walls FABIG Technical Note 5, Steel Construction Institute, [7] Langdon, G.S. & Schleyer, G.K. J. Strain Analysis for Engineering Design, (in press).