PSI - Issue 22

5

Yaorong Feng et al. / Procedia Structural Integrity 22 (2019) 219–228 Author name / Structural Integrity Procedia 00 (2019) 000 – 000

223

160

G1 G2

140

120

100

80

60 12MPa&25℃ 11MPa&10℃ 10MPa&3℃ 9MPa&2℃ 8.5MPa&1℃ BTC predicted CVN/J

pressure-temperature combination along pipeline

Fig. 2. Requirement for crack arrest toughness under different pressure-temperature combinations along pipeline

3.4. Correction of prediction results of Battelle hyperbolic model When the BTC method is used to predict the crack arrest toughness of high strength and toughness steel pipe, the calculation results need to be corrected. There are 3 kinds of correction methods usually adopted: first, the Battelle prediction results are modified according to the results of full-scale burst test with the same or similar specifications, and this method takes into account various influencing factors (including the rising shelf factor CVP/CV100, here CVP is the Charpy impact energy at upper shelf, and CV100 is Charpy impact energy of 100% sear fracture surface), which is relatively comprehensive and reliable. The second is based on the fracture separation of steel pipe, through statistical analysis of the actual steel pipe fracture separation situation, the Battelle prediction results are modified, this is only to consider the impact of the separation of the fracture surface and targeted correction, has a great limitation. The third is the use of safety factor correction, taking into account the dispersion of the data, the correction coefficient (that is, the safety factor) 1.4 is taken on the calculation results of the hyperbolic model. This is a mathematical way of data processing. Tubular Goods Research Institute (TGRI) of China National Petroleum Corporation (CNPC) collected the X80 steel pipe full-scale gas burst test database carried out in Europe, America and Japan, and the correction coefficient obtained from the analysis is shown in Figure 3. The modified coefficient 1.71 includes all the crack propagation points, which is safer, but the required crack arrest impact energy is higher, and the correction coefficient 1.43 does not include a gas burst test of Japanese Sumitomo Metal using air as medium, which is measured by the high crack arrest impact energy, which may be related to the obvious fracture separation in steel pipe.

X80,1219,NG,P X80,1219,NG,A X80,1422,NG,P X80,1422,NG,A X80,1422,Air,P X80,1422,Air,A

300

1.71:1

1.43:1

250

1:1

200

100 Tested C V ,J 150

50

0

0

50

100

150

200

250

300

BTC Predicted C V , J

Fig. 3 Comparison of measured values of crack arrest toughness with Battelle prediction results