COMPUTATIONAL WELDING MECHANICS

Preheat prediction

Several investigators have developed relationships for predicting preheat levels to avoid the HAZ hydrogen cracking problem. Bibby et al [28] present a review of the predictive methods to assist in managing hydrogen in welding applications.

The following method may be used to determine the necessary preheat temperature for steel welding. This method is based on Stout H Slit or Tekken Test experimental results, [29]:

1- Obtain the carbon equivalent of the steel to be welded using equation (6-2 8c)

2- Obtain the estimate of the hydrogen content of the process.

3- Determine Kt using the chart in Figure 6-19, and the weld stress <JW using with the help of the restraint intensity in Figure 6-20.

4- Calculate Cl using the values of CE, HMS, Kt, aw and equation (6-28b)

5- Calculate (/100)cr from Cl using equation (6-28a) or obtain from Figure 6-21.

6- Finally select the preheating temperature, taking into account h, 2b and Q/v, so that the^100 > (tm)cr, Figures 6-22 and 6-23.

The chart method for determining HAZ preheat to avoid НІС described above, adopted from Bibby et al [28], is convenient but less general than a computer managed system.

The Slit Test is a self-restrained cracking test that has been applied to pipeline girth welds. In the Slit Test, welds are performed at a sequence of preheat temperatures. The lowest preheat temperature at which hydrogen cracking does not occur is the
critical preheat temperature. To use this method both the critical and actual weld cooling times must be calculated.

(6-28)

(ґ,00)с,. =exp(67.6C/3 -182.0C/2 +163.8C7-41.0) (6-28a)

Cl = CEn + 0.151ogtfyB +0.301og(0.017^(7w) (6-28b)

The carbon equivalent is easily determined from the chemical composition of the base material. The weld pool diffusible hydrogen level HHIJ can be estimated as follows [28 and 34]:

1.0 ml/lOOg for low hydrogen GMA situations;

3.5 ml/1 OOg for SA W and low hydrogen SMA W;

15.0 ml/1 OOg for rutile/acid electrodes and;

30-40 ті/1 OOg for cellulosic field electrodes;

The joint configuration stress concentration factor varies from a low of about 1.5 for a full V root pass to about 7 for a single bevel mid-thickness root pass, Figure 6-19.

The local stress at the toe of a root weld, cr,, is K, sw where , is

7 W 7 l W Vv

the stress across a weld (MPa) . The mean stress across a weld is given by sw = 0.057?f where Rf < 20sy ; sy is the yield stress (MPa);

sw = sy + 0.0025(7?, — sv), where Rf > 20sy. Restraint Rf(N/mm

mm) is by definition the force per unit length necessary to expand or contract the joint gap by a unit length. It is evaluated by the expression Rf =71r/ [arctan(0.017/i)-(/;/400)2]. The restraint coefficient rf (MPa/mm) is a fundamental factor that assumes a two dimensional stress rate. The restraint Rf is thickness dependent, which accounts for three-dimensional effects. rf can be estimated by

knowing that it varies from a low of about 400 MPa/mm for normal restraint situations to 700 MPa/mm for high restraint situations. By combining the carbon equivalent, the diffusible hydrogen level and the local stress level, a cracking index Cl can be calculated and from that the critical time t.....

It is then necessary to calculate the actual weld cooling time. The cooling time can calculated from the following time-temperature relationship, adopted from Yurioka et al [35] and Bibby et al [28]:

T = T~ + {--t! V)A,/. + (T0-TJ)-exp(-^• 0 (6-29)

)di(cm) лІАтіїл ° °° kV

where Q/v (J/mm) is the heat input per unit length of weld, the thermal conductivity к =0.06 J/mm°C is recommended, the thermal diffusivity X=I4.6 mm2 /°С, convection coefficient a = l,5xl(T5 J / mm2s2°C), S is the surface area of the plate (mm2), V is the volume of the plate (mm3), h is plate thickness (mm), t is time (s), b is the half width of the electrical strip preheater (mm), Г is the ambient temperature (°С), To is the preheat temperature

(°С)-

cooling time can be calculated

Г=Г”+{(^М V^+(r°_rjF^(i)^)}'expf^'0 (6'31)

Д =b/j4A(t + tph) (6-3 la)

(32=ь і (6-3 ib)

The preheat is adjusted in equation (6-30) until the cooling time calculated in equation (6-31) satisfies the inequality of equation (6­28).

Figure 6-22: Relationship between cooling time to 100°C and preheating temperature (Q/v=170() J/mm), adopted from Bibby et al [28].