This is a repository copy of Deformation of thin plates subjected to impulsive load : Part III

– an update 25 years on.

White Rose Research Online URL for this paper:

http://eprints.whiterose.ac.uk/159056/

Version: Accepted Version

Article:
Chung Kim Yuen, S., Nurick, G.N., Langdon, G.S. orcid.org/0000-0002-0396-9787 et al. (1
more author) (2017) Deformation of thin plates subjected to impulsive load : Part III – an
update 25 years on. International Journal of Impact Engineering, 107. pp. 108-117. ISSN
0734-743X

https://doi.org/10.1016/j.ijimpeng.2016.06.010

Article available under the terms of the CC-BY-NC-ND licence

(https://creativecommons.org/licenses/by-nc-nd/4.0/).

Reuse
This article is distributed under the terms of the Creative Commons Attribution-NonCommercial-NoDerivs
(CC BY-NC-ND) licence. This licence only allows you to download this work and share it with others as long
as you credit the authors, but you can’t change the article in any way or use it commercially. More
information and the full terms of the licence here: https://creativecommons.org/licenses/

Takedown
If you consider content in White Rose Research Online to be in breach of UK law, please notify us by
emailing eprints@whiterose.ac.uk including the URL of the record and the reason for the withdrawal request.

eprints@whiterose.ac.uk
https://eprints.whiterose.ac.uk/
Accepted Manuscript

Title: Deformation of thin plates subjected to impulsive load: part III – an

update 25 years on

Author: S. Chung Kim Yuen, G.N. Nurick, G.S. Langdon, Y. Iyer

PII: S0734-743X(16)30395-5
DOI: http://dx.doi.org/doi: 10.1016/j.ijimpeng.2016.06.010
Reference: IE 2720

To appear in: International Journal of Impact Engineering

Received date: 17-2-2016

Revised date: 15-6-2016
Accepted date: 22-6-2016

Please cite this article as: S. Chung Kim Yuen, G.N. Nurick, G.S. Langdon, Y. Iyer, Deformation
of thin plates subjected to impulsive load: part III – an update 25 years on, International Journal
of Impact Engineering (2016), http://dx.doi.org/doi: 10.1016/j.ijimpeng.2016.06.010.

This is a PDF file of an unedited manuscript that has been accepted for publication. As a service
to our customers we are providing this early version of the manuscript. The manuscript will
undergo copyediting, typesetting, and review of the resulting proof before it is published in its
final form. Please note that during the production process errors may be discovered which could
affect the content, and all legal disclaimers that apply to the journal pertain.
 1 Deformation of Thin Plates Subjected to Impulsive Load: Part III – An Update 25 years

2 on

3 S. Chung Kim Yuen, G. N. Nurick, G. S. Langdon and Y. Iyer

5 Blast Impact and Survivability Research Unit (BISRU), Department of Mechanical

6 Engineering, University of Cape Town, Private Bag, Rondebosch 7701

7 *Corresponding author email: Genevieve.langdon@uct.ac.za

9 Highlights
10  Air blast experiments on flat steel plates
11  Review of 25 years of blast experimentation
12  Update of dimensionless analysis
13  Check to see that post 1989 test data correlates with dimensionless analysis
14 approaches

15 Abstract

16 In 1989, Nurick and Martin published two review papers on the deformation of thin steel

17 plates subjected to impulsive air-blast loading. The state of the art has progressed

18 significantly in the following 25 years, and this review paper restricts itself to experimental

19 studies that investigate the response of monolithic metal plates subjected to air-blast loading

20 generated by detonating plastic explosive. From the large numbers of experiments reported, it

21 is shown that the failure progressions in circular and quadrangular plates are similar and can

22 be adequately described by three “failure modes” – namely large plastic deformation (mode

23 I), tensile tearing (mode II) and shearing (mode III) although the severity and location of

24 these failures on the plates is primarily determined by spatial distribution of the blast loading

25 across the plate surface, and that boundary conditions significantly influence the onset of

26 shearing and tearing failures due to variation in the in-plane movement of the plate material.

Page 1 of 33
27 The non-dimensional analysis approaches used by Nurick and Martin have been expanded to

28 include the effects of load localisation and stand-off distance, and show good correlation with

29 the expanded sets of test data published since 1989. It is concluded that these approaches still

30 hold merit as simple tools for evaluating the likely effect of a close proximity air blast load

31 on a flat metal plate.

32 Keywords

33 Blast loading, plates, failure modes, deformation, non-dimensional analysis

34

35 Notation

36 B plate width

37 I impulse

38 L plate length

39 R exposed plate radius

40 Ro charge radius

41 S stand-off distance

42 t plate thickness

43 W TNT equivalent mass

44 Z Hopkinson scaled distance

45  permanent mid-point displacement

46 c non-dimensional impulse for circular plates

47 cS non-dimensional impulse for circular plates, incorporating stand-off distance

48 q non-dimensional impulse for quadrangular plates

49  material density

50  characteristic stress = quasi-static yield stress

51

Page 2 of 33
52 1. Introduction

53 The study of structural response to impulsive blast loading has been performed for many

54 years. In 1989, Nurick and Martin[1, 2] reviewed data from previous experimental and

55 theoretical investigations into the response of flat, monolithic metal plates subjected to blast

56 loading. In addition, new experimental data was reported and non-dimensional analysis was

57 utilised with the aim of developing a simple, empirical prediction of the displacement of a

58 blast loaded plate with fully clamped boundary condition[2]. Since 1989, there have been

59 many experimental, analytical and numerical modelling investigations into the response of

60 structures to blast loading. These have expanded the types of structures examined to include

61 different plate geometries, stiffened and welded structures, sandwich panels, composite

62 materials and monolithic metal plates with different boundary conditions. Tensile tearing and

63 shear failures have also been investigated alongside large plastic deformation responses.

64 Different forms of loading, including localised loading and stand-off distance effects have

65 also been studied.

66

67 This paper reviews the literature that has published since 1989 in an attempt to update the

68 work presented by Nurick and Martin[1, 2]. Using the post 1989 developments in non-

69 dimensional analysis, much of the more recent experimental data is then translated into non-

70 dimensional impulse and displacement parameters and plotted alongside the data presented in

71 ref[2] to determine if the empirical relationships proposed in 1989 are still valid. Due to the

72 myriad of papers in this area, this review paper restricts itself to experimental studies that

73 investigate the response of monolithic metal plates subjected to air-blast loading generated by

74 detonating plastic explosive. Unless otherwise stated, the reviewed results are concerned with

75 flat mild steel plates.

76

Page 3 of 33
 77 2. Experimental studies since 1989

78 In order to summarise the experimental work performed in the past twenty five years, the

79 studies presented in the literature are summarised in Table 1 according to the following

80 classifications:

81  Loading type: uniform, localised, varying stand-off distance,

82  Plate geometry: circular or quadrangular; stiffened or flat,

83  Boundary conditions, and

84  Failure mode: large plastic deformation, tensile tearing, shear failure.

85

86 2.1 Uniform loading conditions

87 Teeling-Smith and Nurick[3] investigated the progression in failure of clamped circular

88 plates subjected to uniform blast loading. Photographs of plate failures at different impulses

89 are shown in Figure 1. At low impulse levels, large plastic deformation (known as Mode I

90 failure) was observed, with mid-point displacement increasing with increasing impulse. As

91 impulse was increased, thinning occurred at the plate boundary (known as Mode Ia or Mode

92 Ib, depending upon the proportion of the circumference that exhibited necking) which was a

93 precursor to tensile tearing along the boundary edge[3]. As impulse reached a threshold

94 value, partial tearing (known as Mode II*) of the plate edge occurred. This Mode II* failure

95 is the transition between Mode I and Mode II (tensile tearing of the boundary edge) failure.

96 As impulse increased further, plates exhibited tensile tearing of the boundary edge. If the

97 impulse was increased beyond the threshold required for complete boundary tearing, the mid-

98 point displacement of the plates decreased with increasing impulse and the failure mode

99 tended towards transverse boundary shear (known as Mode III)[3]. The observed failure

100 modes were similar to those reported by Menkes and Opat[4] for explosively loaded beams

101 and the definitions of the failure modes are summarised in Table 2 for blast-loaded plates.

Page 4 of 33
102

103 Nurick and co-workers[5, 6] investigated the influence of boundary conditions on the failure

104 of uniformly loaded circular plates in follow-on studies from the work reported in ref[3].

105 Thomas and Nurick[5] compared the fully clamped and built-in boundary conditions and

106 showed that the boundary condition has little influence upon plate response during Mode I

107 (large plastic deformation) failures. However, the onset of boundary thinning and

108 subsequently boundary tearing were significantly influenced by the type of boundary

109 condition. The built-in plates exhibited the onset of thinning and tearing at lower impulses

110 since the boundary was more rigid and the clamped boundary was unable to fully prevent in-

111 plane movement of the plate[5]. There was also a difference in the curvature of the plates at

112 the boundary – the clamped plates exhibited curvature within the clamped region whereas the

113 built-in plate curvatures began at the built-in edge[5]. Nurick et al.[6] investigated the effect

114 of a sharp or filleted edge on the response of fully clamped circular plates. The filleted edge

115 clamps delayed the onset of thinning and tearing failures to higher plate displacements,

116 whereas the sharp edged clamp failed at the lowest impulse levels. The sharp edge caused an

117 indentation in the plate which initiates necking[6]. Plates with curved edges exhibited

118 necking without any indentations, similar to that observed during tensile test[6].

119

120 Cloete et al.[7] reported results from tests involving uniformly loaded clamped circular

121 plates. In additional to the boundary clamping, the plates either had a centrally located hole

122 (referred to as annular) or were prevented from deforming with a central support. The central

123 support was a Hopkinson bar, enabling transient force measurements to be taken during the

124 structural response phase. During the annular plate tests, the duration of the blast load was

125 typically in the order of 50 ms and showed the classical blast wave characteristics of a sudden

126 rise followed by an exponential decay. Three distinct modes of failure for fully clamped

Page 5 of 33
127 circular plates were observed from the tests on plates with a central support. The failure mode

128 definitions were similar but not identical to those reported for circular plates[3]. A distinction

129 was made between tensile and shear failures that occur prior to and subsequent to large

130 plastic deformation, something which was only possible due to the transient measurements

131 from the Hopkinson bar. The maximum transverse shear stress that the plates could sustain

132 along the inner boundary was relatively consistent and marked the transition from failure

133 subsequent to large deformation and failure prior to large deformation[7].

134

135 Olson et al.[8] and Nurick and Shave[9] investigated the progression in failure of clamped

136 quadrangular plates subjected to uniform blast loading. Photographs of plate failures reported

137 in ref[9] are shown in Figure 2. Similar failures modes were observed for quadrangular plates

138 as for circular plates, when the work in refs[8, 9] is compared to the work of Teeling-Smith

139 and Nurick[3]. At low impulse levels, large plastic deformation was again observed, with

140 mid-point displacement increasing with increasing impulse. As impulse was increased,

141 thinning occurred at the centre of the plate boundary which progressed to tearing failure with

142 further increases in impulse[8, 9]. The tensile tearing also initiated at the centre of one of the

143 boundary edges and progressed towards the corners at higher impulses[8]. In-plane

144 movement (often referred to as “pulling-in”) was observed as displacement increased and

145 tearing occurred[8]. If the impulse was increased beyond that required for complete boundary

146 tearing, the mid-point displacement of the plates decreased with increasing impulse and the

147 failure mode tended towards transverse boundary shear (known as Mode III).

148

149 Bonorchis and Nurick[10] extended the work on boundary conditions reported in refs [5, 6]

150 to study the influence of boundary conditions on the response of quadrangular plates

151 subjected to uniform loading. Plates with clamped, built-in and welded (both TIG and MIG)

Page 6 of 33
152 boundaries were tested under the same loading conditions. Similarly to the results obtained

153 by Thomas and Nurick[5] for circular plates, large plastic deformation response was

154 unaffected by boundary condition, but the initiation of tensile failure was significantly

155 influenced by the imposed boundary condition. The impulse required for boundary tearing

156 was lowest for the built-in plates and next lowest for the welded plates. The clamped plates

157 required the highest impulse to initiate tearing, again due to the in-plane movement of the

158 clamped plates.

159

160

161 2.2 Localised blast loading

162 Nurick and co-workers[11, 12] reported results from localised blast tests on fully clamped

163 circular plates. The loading was generated by detonating centrally located disks of plastic

164 explosive (PE4) in close proximity to the plates. The dimensions of the PE4 disk (diameter

165 and charge height) were varied, along with charge mass, to produce a range of responses in

166 the plates from low levels of plastic deformation through to tearing of the plates in the central

167 region. Photographs of failed plates are shown in Figures 3 and 4[13]. The localisation of the

168 loading caused a change in plate profile when compared to uniformly loaded plates. Instead

169 of a single dome, the locally loaded plates exhibited a global dome with a secondary inner

170 dome superimposed on top in the central region. The inner dome diameter increased with

171 increased charge diameter, as might be expected. At higher impulses, thinning and tearing

172 occurred in the central region, followed by capping failure. After capping, the remaining plate

173 has a centrally located hole, the diameter of which is also proportional to the charge diameter.

174 As impulse increased beyond that required for capping, radial cracks propagated in the plate

175 away from the central cap and caused petalling failure.

176

Page 7 of 33
177 Yuen and Nurick[13] investigated the effects of load-plate diameter ratio and plate thickness

178 on the response of locally blast-loaded circular plates with built-in boundaries. In many tests,

179 the same capping failures observed in refs[11, 12] were observed, but it was also shown that

180 thicker plates with small load diameters exhibited petalling failures rather than capping

181 failure. Once the load-plate diameter was greater than or equal to 0.4, the loading was less

182 localised and tearing failure occurred at the plate boundary rather than in the central region.

183

184 Jacob et al.[14] reported a study on the influence of plate and load geometry on the response

185 of clamped quadrangular plates subjected to localised blast loading. Plate thickness was

186 varied from 1.6 mm to 4 mm. The plate aspect ratio was varied from 1:1 (square) to 2.4:1

187 (rectangular). Various charge diameters and charge masses were used. It was found, that over

188 the whole range of plate thicknesses, charge diameters and aspect ratios, the response of the

189 quadrangular plates was similar in form to that of circular plates[14]. Photographs of some

190 typical plate profiles are shown in Figure 5, where the inner dome atop a global dome is again

191 observed for Mode I failure (followed by capping failure at higher impulses) as it was for

192 circular plates in refs[11-13]. Similar results were reported by Langdon et al.[15] for built-in

193 square plates subjected to localised blast loads.

194

195 2.3 Plates subjected to blast loading at different stand-off distances

196 Jacob et al.[16] performed an experimental investigation into the influence of stand-off

197 distance on the response of clamped circular plates. The blast loading was generated by

198 detonating small disks of PE4 at the open end of a tube. The tube directed the blast loading

199 towards the clamped target plate. The plate had a diameter of 106 mm and a thickness of 1.6

200 mm. Various tube lengths were used to provide stand-off distances ranging from 25 to 300

201 mm. Large plastic deformation occurred in most of the experiments. At low stand-off

Page 8 of 33
202 distances (up to 40 mm), the plate response was typical of those observed for locally loaded

203 plates refs [11-13]. At large stand-off distances (above 100 mm), the plate response was

204 typical of uniformly loaded plates, with stand-off distances between 40 and 100 mm, being

205 described as a transition between the two loading regimes. Increasing stand-off distance

206 therefore decreases load localisation. Jacob et al.[16] concluded that loading could be

207 considered to be uniform when the stand-off distance exceeded the largest plate dimension

208

209 Neuberger et al.[17, 18] performed scaled experiments where the blast loading was

210 generated by detonating spherical TNT charges (varying mass from 0.468 kg to 8.75 kg) at

211 different stand-off distances (65 mm to 200 mm). The clamped circular test plates had

212 diameters of either 0.5 m or 1 m and were manufactured from RHA steel. The main foci of

213 the investigations were the achievement of geometrically similar scaling during the blast

214 tests[17] and the influence of spring-back on response[18]. Neuberger et al. [17, 18]

215 observed large plastic deformation of the plates that was similar to that observed by Jacob et

216 al. [16].

217

218 Jacinto et al.[19] reported results from large scale experiments on quadrangular plates. The

219 loading was generated by detonating charges with equivalent TNT masses varying from 0.8

220 kg to 10 kg at stand-off distances of 30 m and 60 m. Two sets of plates were tested – one with

221 dimensions 1.5 m by 1m by 2 mm (with three free edges and one edge clamped in a concrete

222 base) and one with dimensions 0.95 m by 0.95 m by 0.9 mm (clamped in a frame on all four

223 sides). Due to the large stand-off distance relative to the plate size, the loading was assumed

224 to be uniformly distributed. Pressure transducers recorded transient pressure histories that

225 were typical of far field loading, as shown in Figure 6. Most of the paper was focussed upon

226 numerical modelling. Jacinto et al.[19] commented that the plates clamped in frames showed

Page 9 of 33
227 greater vibrational damping than the plates with one edge clamped in a concrete base. This

228 observation reinforces the findings reported by Nurick et al.[5, 6, 10, 20] on the importance

229 of boundary conditions.

230

231 Large scale field tests involving the detonation of up to 26300 kg of ordnance per test were

232 reported by Yuen et al.[21, 22]. Square plates (exposed area of 0.5 by 0.5 m) were clamped in

233 test frames and situated at various stand-off distances from the explosive detonation site. The

234 ordnance was laid out on the ground in carpet-like form (rather than the typical cylindrical or

235 spherical shaped charges used in laboratories). The location of the charge on the ground,

236 unusual explosive layout and casing of the explosive due to the ordnance types led to

237 complex loading conditions (with ground effects, geometry effects and fragmentation damage

238 respectively). Despite the complex loading, the resulting deformation was similar to that

239 observed in lab-scale tests involving uniformly loaded clamped square plates[8, 9]. At the

240 closer stand-off distances the ordnance casings became projectiles as a result of the

241 explosions and perforated some of the plates. The perforation damage, however, seemed to

242 have little influence on the global response of the panels. Hopkinson-Cranz scaling was used

243 (and modified) to evaluate the loading parameters and allow empirical predictions of mid-

244 point displacement. The empirical predictions gave reasonable agreement with the

245 experimental results.

246

247 2.4 Stiffened plates

248 Investigations into the response of stiffened plates to blast loading have mainly focussed on

249 uniformly distributed loading[23-35]. As might have been anticipated, the addition of

250 stiffeners, whether they are clamped into position[29, 31], welded[32], integral with the

251 plate[30, 33], or riveted[34, 35] resulted in lower global displacements for a given impulse.

Page 10 of 33
252 Schleyer et al.[32] showed that the direction of loading and the amount of in-plane restraint

253 influenced the response of stiffened plates subjected to pulse pressure loading. Loading the

254 plates such that the stiffeners were in tension resulted in slightly increased mid-point

255 displacement compared with loading the plates with the stiffeners in compression. Plates with

256 no in-plane restraint exhibited displacements that were much larger than those with in-plane

257 restraint along the boundary edges[32]. The addition of stiffeners to the unrestrained plates

258 had no discernable effect upon displacement[32]. The geometry, number and position of the

259 stiffeners are of importance in the response of clamped plates. Yuen and Nurick[33] showed

260 that uniformly loaded plates with more stiffeners are more effective at reducing permanent

261 displacements, and that locating the stiffener along the mid-lines of the plates is more

262 effective at reducing displacements than placing stiffeners at other positions. Stiffeners were

263 shown to reduce the plate deformations but this led to the initiation of tearing failures at

264 lower impulses when compared to unstiffened flat plate responses[33]. Tearing initiated

265 along the plate boundary, usually along the side parallel to the stiffeners.

266

267 Localised loading of the same stiffened plates used by Yuen and Nurick[33] was reported by

268 Langdon et al.[15]. A general failure progression of the stiffened plates emerged from the

269 experimental results[15]: localised central dome, limited by the stiffener geometry; thinning

270 along the stiffener-plate edges; tearing away from the stiffener edge and towards the

271 boundary; petalling and in one case, stiffener rupture. The locations of tearing were strongly

272 dependent upon the stiffener location[15], as shown in the photographs in Figure 7.

273

274 3. Non-dimensional analysis

Page 11 of 33
275 Non-dimensional analysis can be used to compare the results from experiments performed on

276 panels of different scale, different material properties and tested under different loading

277 conditions.

278

279

280

281 3.1 Circular plates

282 In 1989, Nurick and Martin[1, 2] derived an expression, shown in Eq. (1) for circular plates,

283 that provided a non-dimensional impulse parameter for comparing blast-loaded panels of

284 different geometries (exposed radii and thicknesses) and different materials (densities and

285 characteristic stresses), to be treated similarly. The expression for non-dimensional impulse

286 given in Eq. (1) was then modified by Nurick and Martin[2] for localised blast loading of

287 circular plates as shown in Eq. (2). The experimental studies conducted since 1989 have

288 provided additional insight into the behaviour of plates under different loading conditions and

289 greater understanding of the factors affecting plate performance. In parallel with the

290 experimental studies reported above, Nurick and co-workers[14, 16] have further developed

291 the non-dimensional analysis of blast loaded plates to account for some of these additional

292 loading conditions.

293

294 (1)

295 Where R = exposed plate radius, t = plate thickness, I = impulse,  = material density,  =

296 characteristic material stress (quasi-static yield stress for monolithic metals) and is known

297 as the non-dimensional impulse parameter for circular plates.

298

299 (2)

Page 12 of 33
300 Where R0 = load radius of a centrally located disc of explosive.

301

302 A modification to account for stand-off distance in the response of circular plates was

303 proposed by Jacob et al.[16], shown in Eq. (3), based upon tests conducted over a range of

304 stand-off distances (from 13 to 300 mm) where the blast was directed along a tube towards a

305 106 mm diameter metal target plate, as described previously.

306 (3)

307
308 Where S = stand off distance (SOD) between charge surface and target plate surface.

309

310 Nurick and Martin[2] found an empirical relationship between permanent deflection-

311 thickness ratio and non-dimensional impulse using data from over 100 tests. This

312 relationship, expressing in Eq. (4), is used for large inelastic displacement only, and not for

313 failures involving tensile tearing of the plate (capping in the plate centre or tearing at the

314 boundary) or shear failure (either in the central region or boundary of the plate).
315
316 (4)

317 Where  = permanent midpoint deflection of the plate.

318

319 The empirical relationship in Eq. (4) was determined from the slope of the non-dimensional

320 displacement versus non-dimensional impulse curve and had a correlation coefficient of

321 0.974. Since 1990, hundreds of additional tests have been performed on circular plates

322 subjected to various loading conditions. Data from these tests is combined with data obtained

323 prior to 1990 and is plotted in Figure 8 in terms of non-dimensional impulse and non-

324 dimensional displacement. The pre-1990 data went up to a non-dimensional impulse value of

Page 13 of 33
325 approximately 25; 140 of the new data-points are in the range where non-dimensional

326 impulse is between 25 and 50. In total, 699 data-points are plotted in Figure 8, along with a

327 line described by Eq. (4) and a new regression line, given by the expression in Eq. (5). A

328 correlation coefficient of 0.928 is obtained for Eq. (5). As indicated in Table 3, 74% of the

329 data is predicted within a plate thickness and 90% of the data within 2 plate thicknesses of the

330 expression in Eq. (5). There is little difference between the new empirical expression in Eq.

331 (5) and the 1989 proposal[2], validating the original prediction in Eq. (4).
332
333 (5)

334
335 From an inspection of Table 3 and Figure 8, it is evident that the additional data agrees

336 particularly well with the data prior to 1990 for non-dimensional impulses below 25. Above a

337 non-dimensional impulse of 25, there is an upward trend in the displacement data away from

338 the lines described by Eqs (4) or (5). At higher non-dimensional impulses, thinning or

339 necking of the plates is likely to be occurring at the boundary, reducing the levels of in-plane

340 restraint and hence increasing the displacement exhibited by those plates as also observed by

341 Nurick and Teeling-Smith[36] for the range of deflection-thickness ratio higher than 10.

342
343 It should be noted that the empirical equations presented herein are restricted to mostly mild

344 steel plates which are known to be highly strain rate sensitive. There are a few data points on

345 aluminium plates. The slight variation in the various figures could be related, partly, to the

346 different strain rate sensitive properties of the various mild steels. Jones[37] showed how

347 strain rate effects can be considered for plating under explosive loadings. Yao et al.[38] very

348 recently presented some experimental results on locally blast loaded mild steel plates that were very

349 similar to those presented by Nurick and co-workers[14, 16]. Yao et al. [38] showed that there was a

350 linear trend of increasing displacement-thickness ratio with increasing dimensionless damage number

351 Dex, where Dex used the Hopkinson scaled distance (shown in Eq. (6)) instead of impulse (I) as the

Page 14 of 33
352 loading parameter in the dimensionless analysis. This is similar to the approach taken by Yuen et

353 al.[22] for large scale explosions outdoors which employed empirical formulae to estimate the

354 impulse using the peak over-pressure and duration calculations from the Hopkinson scaled distance.

355 However, the stand-off distance range was small (12.9-17.6 mm) and the effects of the charge

356 geometry and plate stiffness were not considered. Hence, this work uses impulse with the appropriate

357 correction factors for stand-off distance, charge shape and plate geometry shown in Eq. (3).

358 (6)

359 Where Z = Hopkinson scaled distance and W = equivalent TNT mass of charge

360

361 3.2 Quadrangular plates

362 A similar approach to non-dimensional analysis has been followed for quadrangular plates.

363 The non-dimensional impulse q (or damage number) for quadrangular plates proposed by

364 Nurick and Martin[2] in 1989 is given in Eq. (7). This was modified by Jacob et al.[14] to

365 incorporate a loading parameter for localised loading, as shown in Eq. (8).

I
366 q  (7)
 BL 
1

2
2
2t

367
368 Where B = plate breadth; L = plate length

369

  LB 
I  1  ln  
  2 
  R0 
370  ql  1
(8)
2 t ( LB  )
2 2

371
372 In 1989, Nurick and Martin[2] plotted the deflection-thickness ratios obtained from uniform

373 loading tests against the non-dimensional impulse parameter given in Eq. (7). A regression

374 analysis was performed to obtain a line of best fit. The equation of the best fit line was

375 proposed as an empirical prediction for the large deformation response of blast-loaded

Page 15 of 33
376 quadrangular plates and is shown in Eq. (9). A probability of 90% is obtained to predict mid-

377 point deflection within a plate thickness.

378
 
379    0 . 471  q  0 . 001 (9)
 t q

380 Eq. (8) had a correlation coefficient of 0.984, where the number of data-points was 156.

381 The deflection-thickness ratios of quadrangular plates subjected to uniform and localised

382 loads for all experiments pre-1989 and post-1989 were plotted against the appropriate non-

383 dimensional impulse parameters from Eqs. (7) and (8). The graph is shown in Figure 9 and

384 comprises results from 356 experiments. A new empirical relationship, given by Eq. (10),

385 with a correlation coefficient of 0.94 was obtained from a regression analysis of all 356 data-

386 points.
387
 
388    0 . 506  q  0 . 158 (10)
 t q

389
390 A statistical analysis showed that both Eqs. (9) and (10) would predict the mid-point

391 deflection of a quadrangular plate subjected to either localised or uniform blast load with

392 similar probability within one or two plate thickness (72% and 92% respectively), as

393 indicated in Table 4. At higher dimensionless numbers, it appears that Eq. (10) would provide

394 better prediction.

395
396

397 3.3 Combining the circular and quadrangular plate data

398 Nurick and Martin[2] hypothesised that Eqs. (5) and (9) could be used to determine the

399 deflection-thickness ratio of any circular or quadrangular plates subjected to impulsive

400 loading, provided that the plates do not suffer tearing or shear failure. If a circular and

Page 16 of 33
401 quadrangular plate of equal area, thickness and material properties are subjected to impulsive

402 loading over the entire area (that is, uniformly loaded) then the ratio of Eqs. (1) and (7) is:

403
c 2
404   1 . 128 (11)
q 
0 .5

405
406
If Eq. (11) is substituted into Eq. (5) then a new empirical expression is formulated for plates
407
of either circular or quadrangular geometry, Eq. (12):
408


409  0 . 480  q  0 . 277 (12)
t

410
411
Eq. (11) was used to combine all the experimental data for circular and quadrangular plates,
412
including the locally loaded tests and those performed at various stand-off distances, with the
413
appropriate corrections to given in Eqs. (2), (8) and (3).The deflection-thickness ratios of
414
all plates subjected to blast loads (pre-1989 and post 1989) were plotted against the non-
415
dimensional impulse parameter as shown in Figure 10. A regression analysis was performed
416
and yielded the empirical prediction given by Eq. (13), which has a correlation coefficient of
417
0.92 for 1054 data-points.
418


419  0 . 446  q  0 . 261 (13)
t

420
421 A statistical analysis, summarised in Table 5, showed that Eq. (13) provides a better

422 correlation for plate prediction than Eq. (12).

423

424

Page 17 of 33
425 4. Concluding comments

426 From the large numbers of experiments conducted since 1989, it has been found that the

427 failure mode progression of circular and quadrangular plates is similar[3, 9]. The Mode I

428 deformation profile of uniformly was characterised by a single global dome, whereas locally

429 loaded plates exhibited a local inner dome atop a global dome profile. Capping and petalling

430 failures occurred in the centre of locally loaded plates, as opposed to tensile and shear failures

431 at the boundaries of uniformly loaded plates[3, 9]. The locally loaded plates have been

432 observed to exhibit tearing at the boundary when the load-plate diameter ratio is increased

433 above 0.4[13]. Boundary conditions have a significant influence on the tearing and shear

434 failures of plates due to the difference in in-plane movement achieved with each boundary

435 condition. Rigid boundaries, such as built-in and welded boundaries, will initiate necking and

436 tearing failures at lower impulses compared to clamped boundaries[5, 10]. The effect of

437 “edge sharpness” of the tensile boundary failure of clamped plates showed that filleting the

438 clamp delayed the onset of tearing by preventing indentation of the plate boundary[6].

439 Increasing stand-off distance resulted in less localised loading of plates and it was found that

440 when the stand-off distance exceeds the largest plate dimension, loading could be considered

441 to be uniform[16].

442

443 The addition of stiffeners reduced the displacements of clamped quadrangular plates

444 subjected to uniformly distributed blast loading, particularly if they were placed along the

445 mid-lines of plates[33]. However, the initiation of tearing along the boundary can occur at

446 lower impulses[33]. In locally loaded stiffened plates, the stiffener placement also affected

447 the tearing location, with tearing initiating along the stiffener-plate edges[15].

448

Page 18 of 33
449 In parallel with experimental studies, the non-dimensional analysis used by Nurick and

450 Martin[2] has also been expanded to include the effects of load localisation for quadrangular

451 plates[14] and the influence of stand-off distance for circular plates[16]. When the new post-

452 1989 experimental data is plotted in non-dimensional form, the original empirical predictions

453 proposed in ref[2] are very close to the best fits for the data, with particularly good

454 correlation observed for circular plates. The data for circular and quadrangular plates has

455 been combined into one set and a new empirical prediction is proposed in Eq. (12), which is

456 very similar in form to those previously used. Thus, 25 years on from the original work by

457 Nurick and Martin[2], it is concluded that this analysis technique is still useful and the

458 original empirical predictions give good correlation with experiments. It should be noted that

459 the effects of both strain hardening and strain rate sensitivity have not been specifically

460 considered and that both effects are embedded in the empirical equations.

461

462 Acknowledgements

463 The authors wish to thank the staff and students at BISRU for their help, and the Mechanical

464 Engineering workshop at the University of Cape Town for their technical assistance. The

465 authors are grateful to the UCT University Research Committee and the National Research

466 Foundation (NRF) of South Africa for their financial support. Opinions expressed and

467 conclusions arrived at, are those of the authors and are not necessarily to be attributed to the

468 NRF.

469
470 References

471 [1] G.N. Nurick, J.B. Martin, Deformation of thin plates subjected to impulsive loading - a
472 review. Part I: Theoretical considerations, International Journal of Impact Engineering, 8
473 (1989) 159-169.
474 [2] G.N. Nurick, J.B. Martin, Deformation of thin plates subjected to impulsive loading - a
475 review. Part II: Experimental Studies, International Journal of Impact Engineering, 8 (1989)
476 171-186.

Page 19 of 33
477 [3] R.G. Teeling-Smith, G.N. Nurick, The deformation and tearing of thin circular plates
478 subjected to impulsive loads, International Journal of Impact Engineering, 11 (1991) 77-91.
479 [4] S.B. Menkes, H.J. Opat, Tearing and shear failures in explosively loaded clamped beams,
480 Experimental Mechanics, 13 (1973) 480-486.
481 [5] B.M. Thomas, G.N. Nurick, The effect of boundary conditions on thin plates subjected to
482 impulsive loads, The 5th International Symposium on Plasticity and its Current Application,
483 (1995) 85-88.
484 [6] G.N. Nurick, M.E. Gelman, N.S. Marshall, Tearing of blast loaded plates with clamped
485 boundary conditions, International Journal of Impact Engineering, 18 (1996) 803-827.
486 [7] T.J. Cloete, G.N. Nurick, R.N. Palmer, The deformation and shear failure of peripherally
487 clamped centrally supported blast loaded circular plates, International Journal of Impact
488 Engineering, 32 (2005) 92-117.
489 [8] M.D. Olson, J.R. Fagnan, G.N. Nurick, Deformation and rupture of blast loaded square
490 plates - Predictions and experiments, International Journal of Impact Engineering, 12 (1993)
491 279-291.
492 [9] G.N. Nurick, G.C. Shave, The deformation and tearing of thin square plates subjected to
493 impulsive loads-an experimental study, International Journal of Impact Engineering, 18
494 (1996) 99-116.
495 [10] D. Bonorchis, G.N. Nurick, The effect of welded boundaries on the response of
496 rectangular hot-rolled mild steel plates subjected to localised blast loading, International
497 Journal of Impact Engineering, 34 (2007) 1729-1738.
498 [11] T. Wierzbicki, G.N. Nurick, Large deformation of thin plates under localised impulsive
499 loading, International Journal of Impact Engineering, 18 (1996) 899-918.
500 [12] G.N. Nurick, A.M. Radford, Deformation and tearing of clamped circular plates
501 subjected to localised central blast loads, Recent developments in computational and
502 applied mechanics: a volume in honour of John B. Martin, (1997) 276-301.
503 [13] S. Chung Kim Yuen, G.N. Nurick, The significance of the thickness of a plate when
504 subjected to localised blast load, 16th International Symposium on Military Aspects of Blast
505 and Shock, (MABS 16), (2000) 491-499.
506 [14] N. Jacob, S. Chung Kim Yuen, G.N. Nurick, D. Bonorchis, S.A. Desai, D. Tait, Scaling
507 aspects of quadrangular plates subjected to localised blast loads - experiments and
508 predictions, International Journal of Impact Engineering, 30 (2004) 1179-1208.
509 [15] G.S. Langdon, S. Chung Kim Yuen, G.N. Nurick, Experimental and numerical studies on
510 the response of quadrangular stiffened plates -Part II- subjected to localised load,
511 International Journal of Impact Engineering, 31 (2005) 85-111.
512 [16] N. Jacob, G.N. Nurick, G.S. Langdon, The effect of stand-off distance on the failure of
513 fully clamped circular mild steel plates subjected to blast loads, Engineering Structures, 29
514 (2007) 2723-2736.
515 [17] A. Neuberger, S. Peles, D. Rittel, Scaling the response of circular plates subjected to
516 large and close-range spherical explosions. Part I: Air-blast loading, International Journal of
517 Impact Engineering, 34 (2007) 859-873.
518 [18] A. Neuberger, S. Peles, D. Rittel, Springback of circular clamped armor steel plates
519 subjected to spherical air-blast loading, International Journal of Impact Engineering, 36
520 (2009) 53-60.
521 [19] A.C. Jacinto, R.D. Ambrosini, R.F. Danesi, Experimental and computational analysis of
522 plates under air blast loading, International Journal of Impact Engineering, 25 (2001) 927-
523 947.

Page 20 of 33
524 [20] D. Bonorchis, G.N. Nurick, The influence of boundary conditions on the loading of
525 rectangular plates subjected to localised blast loading - Importance in numerical
526 simulations, International Journal of Impact Engineering, 36 (2009) 40-52.
527 [21] G.N. Nurick, S. Chung Kim Yuen, N. Jacob, W. Verster, D. Bwalya, A.R. Vara, Response of
528 Quadrangular Mild-Steel Plates to Large Explosive Load, 2nd International Conference on
529 Design and Analysis of Protective Structures (DAPS), (2006) 30-44.
530 [22] S. Chung Kim Yuen, G.N. Nurick, W. Verster, N. Jacob, A.R. Vara, V.H. Balden, D. Bwalya,
531 R.A. Govender, M. Pittermann, Deformation of mild steel plates subjected to large-scale
532 explosions, International Journal of Impact Engineering, 35 (2008) 684-703.
533 [23] R. Houlston, J.E. Slater, N. Pegg, C.G. DesRochers, On analysis of structural response of
534 ship panels subject to air blast loading, Computers and Structures, 21 (1985) 273-289.
535 [24] R. Houlston, C.G. DesRochers, Nonlinear structural response of ship panels subjected to
536 air blast loading, Computers and Structures, 26 (1987) 1-15.
537 [25] R.B. Schubak, M.D. Olson, D.L. Anderson, Nonlinear analysis of one-way stiffened plates
538 under blast loads, Structures Under Shock and Impact I, (SUSI I), (1989) 345-354.
539 [26] R.B. Schubak, M.D. Olson, D.L. Anderson, Non-linear rigid-plastic analysis of stiffened
540 plates under blast loads, in: P.S. Bulson (Ed.) Structures Under Shock and Impact II, (SUSI II),
541 Computational Mechanics Publications, Portsmouth, UK, 1992, pp. 521-532.
542 [27] R.B. Schubak, M.D. Olson, D.L. Anderson, Rigid-plastic modelling of blast loaded
543 stiffened plates - Part I: one way stiffened plates, Third International Symposium on
544 Structural Crashworthiness and Failure, (1993).
545 [28] R.B. Schubak, M.D. Olson, D.L. Anderson, Rigid-plastic modelling of blast loaded
546 stiffened plates - Part II: partial end fixity, rate effects and two-way stiffened plates, Third
547 International Symposium on Structural Crashworthiness and Failure, (1993).
548 [29] G.N. Nurick, A.G. Conolly, Response of clamped single and double stiffened rectangular
549 plates subjected to blast loads, Structures Under Shock and Impact III (SUSI III), (1994) 207-
550 220.
551 [30] G.N. Nurick, M.D. Olson, J.R. Fagnan, A. Levin, Deformation and tearing of blast-loaded
552 stiffened square plates, International Journal of Impact Engineering, 16 (1995) 273-291.
553 [31] G.N. Nurick, D.M. Lumpp, Deflection and tearing of clamped stiffened circular plates
554 subjected to uniform impulsive blast loads, Structures Under Shock and Impact IV (SUSI IV),
555 (1996) 393-402.
556 [32] G.K. Schleyer, S.S. Hsu, M.D. White, R.S. Birch, Pulse pressure loading of clamped mild
557 steel plates, International Journal of Impact Engineering, 28 (2003) 223-274.
558 [33] S. Chung Kim Yuen, G.N. Nurick, Experimental and numerical studies on the response of
559 quadrangular stiffened plates -Part I- subjected to uniform blast load, International Journal
560 of Impact Engineering, 31 (2005) 55-83.
561 [34] R.L. Veldman, J. Ari-Gur, C. Clum, A. DeYoung, J. Folkert, Effects of pre-pressurization on
562 blast response of clamped aluminum plates, International Journal of Impact Engineering, 32
563 (2006) 1678-1695.
564 [35] R.L. Veldman, J. Ari-Gur, C. Clum, Response of pre-pressurized reinforced plates under
565 blast loading, International Journal of Impact Engineering, 35 (2008) 240-250.
566 [36] G.N. Nurick, R.G. Teeling-Smith, Predicting the onset of necking and hence tearing of
567 thin plates subjected to impulsive loads - an experimental view, Structures Under Shock and
568 Impact II (SUSI II), (1994) 431-445.

Page 21 of 33
569 [37] N. Jones, Dynamic inelastic response of strain rate sensitive ductile plates due to large
570 impact, dynamic pressure and explosive loadings, International Journal of Impact
571 Engineering, 74 (2014) 3-15.
572 [38] S. Yao, D. Zhang, F. Lu, Dimensionless numbers for dynamic response analysis of
573 clamped square plates subjected to blast loading, Archive of Applied Mechanics, 85 (2015)
574 735-744.
575

Increasing Impulse

Increasing Impulse
576

577 Figure 1: Photographs showing the failure of uniformly loaded circular plates with increasing

578 impulse [3]

579

580

Increasing Impulse

Increasing Impulse
581
582

583 Figure 2: Photographs showing the failure of uniformly loaded square plates with increasing

584 impulse [9]

585

586

587

588

Page 22 of 33
589 (a) Capping failure (b) Petalling failure (c) Tensile tearing at the

590 boundary

591 Figure 3: Photographs of locally loaded circular plates [13]

592

Inner dome

Global dome

593

594 Figure 4: Photographs of cross-sections from two built-in circular plates subjected to

595 localised blast loading [13]

596

597

598

599 Increasing
impulse
600

601

602

603

604

605

606 Figure 5: Photographs of cross-sectional profiles from quadrangular plates subjected to

607 localised blast loading [14]

608

Page 23 of 33
609

610

611

612

613

614

615

616

617

618
619
620

621

622

623 Figure 6: Pressure-time history recorded from a far-field explosion, Jacinto et al. [19]

624

Page 24 of 33
625

626

627

628

629
Single stiffened, Double stiffened,
630
9.8 Ns 9.5 Ns
631

632

633

634

635
Cross stiffened, Double cross
636
9.9 Ns stiffened, 11.2 Ns
637 Figure 7: Photographs showing the tearing locations of various stiffened plates subjected to

638 localised blast loading [15]

639

Page 25 of 33
 25
/t = 0.427cs + 0.298
R² = 0.928
Mid-popint Deflection Thickness Ratio

20

/t= 0.425cs + 0.277

15

Pre 1989
Teeling Smith and Nurick [3]
10 Thomas and Nurick [5]
Nurick et al. [6]
Nurick and Radford [12]
Chung Kim Yuen and Nurick [13]
Jacob et al. [16]
5
Nurick and Lumpp [31]
Nurick and Teeling Smith [36]
New test data
New test data
0
0 10 20 30 40 50
640 Dimensionless Number c

641 Figure 8: Graph of displacement-thickness ratio versus non-dimensional impulse for circular

642 plates, both pre-1989 and post-1989

643

644

645

646

647

648

649

650

651

Page 26 of 33
 25

/t = 0.506q - 0.158

R² = 0.940
20
Mid-popint Deflection Thickness Ratio

/t = 0.471q

15
Pre 1989
Thomas and Nurick [5]
Olson et al. [8]
10 Nurick and Shave [9]
Jacob et al. [14]
Langdon et al. [15]
Bonorchis and Nurick [20]
5 Nurick et al. [30]
Nurick and Connolly [29]
Chung Kim Yuen and Nurick [33]
New test data
0
0 5 10 15 20 25 30 35 40 45 50
Dimensionless Number q
652

653

654

655

656

657

658

659

660

661

662 Figure 9: Graph of displacement-thickness ratio versus non-dimensional impulse for

663 quadrangular plates, both pre-1989 and post-1989

664

665

Page 27 of 33
666

667

668

669

670

671

25
Mid-popint Deflection Thickness Ratio

20

/t = 0.446 + 0.261

R² = 0.920
15

10

Quadragular
Circular
0
0 10 20 30 40 50
672 Dimensionless Number 

673 Figure 10: Graph of displacement-thickness ratio versus non-dimensional impulse for the

674 combined circular and quadrangular plate data, both pre-1989 and post-1989

675

676

677

678

679

Page 28 of 33
680 Table 1: Summary of experiments performed since 1989 on air-blast loaded metal plates
Load Plate Reference Notes Nominal Plate Year
1
type type dimensions (mm)
Teeling-Smith and Failure Modes I, II and III of  500 - 1000 x 10 - 1991
Nurick [3] circular plates 20
Thomas and Built-in versus clamped  100 x 1.6 1995
C Nurick [5] boundaries
Nurick et al [6] Influence of edge sharpness:  60 -120 x 1.6 1996
clamped plates
Uniform

Cloete et al [7] Annular and centrally support  100 x 1.6 2005

plates, failure
Olson et al [8] Failure progression, tearing 89 x 89 x 1.6 1993
initiation
Nurick and Shave Failure progression, tearing 89 x 89 x 1.6 1996
Q
[9] initiation
Bonorchis and Varied boundary conditions 188 – 200 x 108 - 2007
Nurick [10] 120 x 3
Wierzbicki and Failure modes  100 x 1.6 1996
Nurick [11]
Nurick and Failure modes, capping  100 x 1.6 1997
C
Radford [12]
Yuen and Nurick Influence of plate thickness,  100 x 1.6 - 3.6 2000
Localised

[13] load-to-plate diameter ratio

Jacob et al. [14] Varied load and plate geometries 160 – 290 x 160 - 2004
290 x 1.6 – 4.0
Q (different aspect
ratios)
Langdon et al. [15] Flat and stiffened plates 126 x 126 x 1.6 2005
Jacob et al. [16] Influence of stand-off distance on  106 x 1.9 2007
Mode I failure, use of tube to
Stand-off distance

direct blast
C Yao et al. [38] Mild steel, 12.9 – 17.6 mm 160 x 160 x 1.6 -4.0 2015
stand-off distance, dimensionless
analysis
Neuberger et al. Scaling  100 x 1.6 2007

Page 29 of 33
 [17]
Neuberger et al. Springback of armour steel  1000 x 20 2009
[18]
Jacinto et al. [19] Boundary conditions, model 950 – 1000 x 950 – 2001
Q validation 1500 x 0.9 – 2.1
Yuen et al. [21, 22] Large scale field tests, scaling 500 x 500 x 3-6 2006
Schubak et al [25- One and two way stiffeners, 4000 - 4572 x 2438 1992-
28] model validation – 4000 x 6 - 7 1993
Nurick et al.[30] Influence of integral stiffeners 89 x 89 x 1.6 1995
Schleyer et al[32] Effects of loading direction, in- 1000 x 1000 x 2 2003
Uniform

plane restraint on pulse loaded

QS
welded stiffened plates
Yuen and Nurick Influence of integral stiffeners 126 x 126 x 1.6 2005
[33]
Veldman [34] Riveted stiffened plates, pre- 508 x 610 x 1.6 2008
pressurisation
1
681 C = circular, Q = quadrangular, QS = quadrangular stiffened
1
682 C = circular, Q = quadrangular, QS = quadrangular stiffened

683

684

685

Page 30 of 33
686 Table 2: Descriptions of failure modes for plates subjected to uniformly and locally

687 distributed blast loading, according to Jacob et al [16]

Failure Uniform Localised

Description
Mode loading loading

Mode I large inelastic deformation  

large inelastic deformation with necking around part of

Mode Ia 
the boundary

large inelastic deformation with necking around the entire

Mode Ib  
boundary

large inelastic deformation with thinning in the central

Mode Itc 
area

large inelastic deformation with partial tearing around

Mode II* 
part of the boundary

Mode II*c partial tearing in the central area 

Mode II tensile tearing at the boundary  

increasing mid-point deflection with increasing impulse

Mode IIa 
with complete tearing at the boundary

decreasing mid-point deflection with increasing impulse

Mode IIb 
with complete tearing at the boundary

Mode IIc complete tearing in the central area – capping 

Mode III transverse shear failure at the boundary 

Petalling 

688

689

Page 31 of 33
690 Table 3: Statistical variation of deflection-thickness ratio for circular plates subjected to blast

691 load for all data

Number of data points ±1 plate thickness ±2 plate thickness

c , Eq. (4) 699 74% 90%

0  c  25 559 80% 93%

c > 25 140 49% 77%

692
693 Table 4: Statistical variation of deflection-thickness ratio for quadrangular plates subjected to

694 blast load for all data

Number of data points ±1 plate thickness ±2 plate thickness

q , Eq.(8) 354 72% 94%

0  q  25 320 77% 96%

q > 25 34 47% 65%

q , Eq.(9) 354 73% 92%

0  q  25 320 74% 97%

q > 25 34 44% 65%

695

696 Table 5: Statistical variation of deflection-thickness ratio for the combined circular and

697 quadrangular plate data

±1 plate thickness ±2 plate thickness

Eq. (12) 61% 88%

Eq. (13) 71% 89%

698

699

Page 32 of 33
700

Page 33 of 33