High-cyclic failure analysis of welded cast iron plates

R. Stonkus*, M. Leonavi*ius**, G. Petraitis***, S. Stupak**** *Vilnius Gediminas Technical University, Saul tekio al. 11, 10223 Vilnius, Lithuania, E-mail: rs@fm.vgtu.lt **Vilnius Gediminas Technical University, Saul tekio al. 11, 10223 Vilnius, Lithuania, E-mail: minleo@fm.vgtu.lt ***Vilnius Gediminas Technical University, Saul tekio al. 11, 10223 Vilnius, Lithuania, E-mail: gedp@vgtu.lt ****Vilnius Gediminas Technical University, Saul tekio al. 11, 10223 Vilnius, Lithuania, E-mail: stupakas@vgtu.lt


Introduction
None of any known techniques of welding guaranties a welded joint without a defect.The regulated sizes of defects are conditional.If a defect is found during a control or service exceeds the normal size, often a doubt arises whether to reject it or after performing some calculations to make clear that the crack appeared during operation does not cause the structural element critical state -failure [1][2][3].
The weld material of the joint should be of a uniform or increased strength than the base metal.However, during operation under the influence of cyclic stresses, the concept of uniform or larger strength condition changes and origins of cracks may be developed at the surface irregularities or of welding defects in weld material or in the heat affected zone.
In mining industry, some large elements are cast of cast iron and welded together then.The composition of base metal and weld material are different [8].

Testing procedures
Two cast iron plates are prepared in such a way that they should make a butt joint with double -V shape weld.The experimental part in the vicinity of the weld material is treated by an abrasive disc in the same way, as in a real structure.When doing in this way, some scratching remains, namely, concentrations.In joining places and supporting point fields, these places are grinded.The sam-ple is pushed lengthwise up to 5 mm in order to diminish the contact influence between the plate and supporting rollers, as shown in Fig. 1.
The static mechanical properties were determined from cylindrical tensile specimens (d 0 = 8.00 mm; L 0 = 30.0mm) from base metal (cast iron): conditional yield stress R p0.
Fig. 2 Loading history diagram Fig. 3 presents the microstructure of welded joint (with no etching) in heat affected zone when crossing from weld metal to the main one.Spheroid graphite can be distinguished there.When approaching the weld metal, it can be seen that the graphite shape and quantity change: beside a spheroid graphite there appears the field of small graphite and the remained graphite acquires the shape of flakes.Fig. 4 shows the structure of weld metal (carbon steel) and the base metal (cast iron) after etching.Here we can also see different derivatives in the heat affected zone, the graphite of spherical and flake-shapes in a ferrite-pearlite matrix and even graphite chains with a small dross layer.The irregularity of the graphite inclusions and the change in their shape also change the mechanical properties, and all this influence the resistance to cyclic loading.The metallic base of the basic metal consists of ferritic-pearlitic structure.By the chemical composition, mechanical properties and the structure it is possible to establish, that the plates are made of a strong cast iron.Such an iron is quite plastic, tensile, resistant to impact loading and easy welded after selecting a proper technology and a material for electrodes.Testing of such a connection for cyclic loading according to the programme coordinated with real structural element operational loading, when the number of cycles exceeds 10 8 , is necessary for ensuring the longevity for the exploited and being designed structures [14][15][16][17][18][19].From the point of view of the fracture mechanics the fatigue diagram is important -the dependence of crack growth rate on stress intensity factor K, allowing to estimate the material cyclic strength from the beginning of crack formation up to complete failure.The parameters of the curve (kinetic fatigue diagram) define the resistance of the material to a cyclic loading.According to the experimental data obtained (crack depth -number of cycles) the kinetic diagrams are designed applying 3 different formulas for K calculation: 1) Anderson [13] 1 where 2) Tada, Paris, Irwin [14]: where M is bending moment, t is width, W is thickness, a is crack size, f (a/W) is geometry function where V is maximal stresses, a is crack size, Į = a/W, f (Į) is correction function according to a formula According to some calculations, kinetic diagrams are made (crack growth rate -stress intensity factor) in Fig. 5.It is known that for determination of fracture toughness K C , the crack can be grown before it under low-cycle or high-cycle loading.By making a variety of materials kinetic fatigue diagrams, the specimens with cracks can be used for the determination of factors 'K th and K C .In our case, the threshold stress intensity factors 'K th and K C chosen according to the kinetic fatigue diagram can be used to calculate the structural elements.In the figure we can see that all the formula applied give similar results.Also characteristic stages of fracture are seen at it.For on its structure a large influence has been produced on the maximal meaning of loading value.An important diagram characteristics is the threshold stress intensity factor range, because it shows a crack nongrowing conditions in the specimen.Cracks developing below the cracking threshold are called short and more or less correspond to the spreading through a coarse-grained structure.In Fig. 5 a short cracking field is approximate; it is calculated by the Anderson's formula (1).For developing short cracks the microstructure has of the greatest importance.The middle part of the diagram is the macroscopic crack development; and it takes place, when the crack growth rate is from 10 -9 m/cycle to 2•10 -8 m/cycle.It is the stable crack development stage, limited from the right side by fracture toughness K C , which determines the end of crack developing stadium and the further sudden failure.
The fracture analysis enriches understanding of the conditions of crack formation, growing and total failure.The crack has appeared in the surface layer of weld metal (Fig. 6).
Fig. 6 The place of crack origin

Crack origin
It has been found that the crack is originated at the defect (1 mm depth) in a transitional zone from a weld material towards the base metal (transitional zone is made clearer in Fig. 7).During a defectoscope control the crack has been observed, when it has appeared on the surface.Fig. 7 Fracture of the specimen Further the crack is growing through the weld metal and moves to the base metal (Fig. 8).

Fig. 8 Crack development in depth
When intruding into the main metal (cast iron) in the transitional zone (the heat affect zone) the crack meets inside defects: pores, nonmetallic insertions (Fig. 9), which increase the crack growth rate.After making the crack growth rates versus stress intensity factor range diagram, additional tests have been performed for establishing compact specimens with a crack static fracture tests, during which were found the indices describing the failure -the critical stress intensity factor K C , as well as when all required conditions have been satisfied as well as the fracture toghness K 1C .This index is defined by the standard ASTM E 399-83.During the experiment of a static failure it is fixed the changeable force F amount and crack opening v and the diagram of their interdependence.For establishing the fracture toughness K 1C or the critical stress intensity factor K C it is necessary the value of force ing diagram (one of these diagrams is presented in Fig. 12) has been obtained.In this case the force F Q corresponds to 0-b straight line interaction.This force suits for calculating the fracture toughness K 1C value, when it satisfies the condition: F max /F Q d 1.1.When this and conditions of the standard are not satisfied, then only the value describing the critical stress intensity factor K C is calculated.Only a part of the conditions were met from tested specimens 1, 2, 3, 4.These conditions are checked after the test, i.e., was the biaxial strain state at the top of the crack or not, is the plastic deformation on the top of the crack excessive or not, is the crack front line not too curved and whether the ratio F max /F Q is satisfied.The obtained values of limiting stress factors 'K th and K C are applied for designing in mining industry and may be evaluated by the risk according to methods [20].

Conclusions
1.The experimental investigation has shown that under the stresses close to the operating conditions, no microcracks have been found at 100 mln.cycles.
2. Under increasing stresses the crack originates under a surface defect and it reaches surface.The surface roughness and base metal structure influences for further propagation.
3. It was found that a cracking development in surface layers in weld material (in steel) is slower (it include the field of short cracking) and when it pass to the main metal (cast iron), the crack growth rate increases.The article presents experimental analytical research results of seminatural welded cast iron specimen.It has been determined: the chemical structure of the base and weld material, static mechanical properties, hardness and microstructure.The static loading during the experiment is started from a close to structural elements loading.The investigation starts from a given stress changing programme.The experiment was performed according to the given programme on stress changing.The kinetic fatigue diagram is performed according to the experimental data.For determining the failure conditions an additional fracture analysis has been performed.

Fig
Fig.1The loading scheme of the specimen

Fig. 3
Fig.3The microstructure (without etching) in the heat affected zone from weld metal to the main one

Fig. 9
Fig. 9 The increased area of sample break In Fig. 10 a large accumulation of defects is seen, formed during welding.

Fig. 10
Fig. 10 Defects in the heat affected zone From analogical cast iron (as cast and normalised) plate CT specimen have been made (BxHxW = 25x60x50 mm).By methodology [6], 2 specimens of cast iron and 2 normalized specimens were tested.The dependence between the crack growth rate and stress intensity factor are presented in Fig. 11.We can see that 'K th changes from 7 to 10 MPa•m 1/2 , when the crack growth rate are v = 5•10 -11 -10 -10 m/cycle, meanwhile of the plate -'K th § 6.3 MPa•m 1/2 has been found, when the crack growth rate are v = 1•10 -9 -5•10 -10 m/cycle.

FGS 9. 1 FGSFig. 11
Fig. 11 Crack growth rate versus stress intensity factor range of cast and normalized cast iron (1, 2 -normalized; 3, 4 -as cast) F Q , found by performing the crack opening diagram analysis.In our investigations the I type of open-

Table Critical
4. The obtained kinetic fatigue diagram of welded joint differs from a cast and normalised cast iron crack growth rates versus stress intensity factor range diagram. 5.The obtained threshold stress intensity factor range values of the welded plate 'K th § 6.3 MPa•m 1/2 , when the crack growth rate v = 5•10 -10 -1•10 -9 m/cycle and CT specimens 'K th = 7-10 MPa•m 1/2 , when the crack growth rate v = 5•10 -11 -10 -10 m/cycle.6.The critical stress intensity factors of the investigated cast and normalized cast iron CT specimens vary between large limits: K C = 38.5 -54.6 MPa•m 1/2 .