Fatigue Life Enhancement of Welded Steel-Steel Composite during Crack Growth from Weak to Strong Steel: An Experimental Validation

Strength mismatch effect across weld interfaces, generated by welding weak and strong steels, influences fatigue and fracture properties of a welded bimetallic composite. Advancing fatigue crack tip in weak parent steel is shielded from the remote load when it reaches near the interface of ultra strong weld steel. Entry of crack tip plasticity into weld steel induces load transfer towards weld which dips crack growth rates thereby enhancing the fatigue life of the composite. A computational model for fatigue life prediction of strength mismatched welded composite under K dominant conditions is validated by experimental work in this paper. Notched bimetallic compact tension specimens, prepared by electron beam welding of weak alloy and strong maraging steels, are subjected to fatigue testing in high cycle regime.


Introduction
Bimetallic composites have shown considerable promise in life enhancement and material and weight optimization of engineering structures with the advent of a) welding processes like explosive cladding and diffusion bonding that are capable of joining dissimilar metals with good bond strength and b) efficient coating technologies. Their concepts, culminating in functionally graded materials, have been successfully tested in aerospace propulsion engines, pressure vessels, gas turbine disks, thin films and coatings, etc. In view of their high utility potential, their fatigue and fracture aspects assume lot of importance especially in load bearing applications. Although literature in this field exists, work still needs to be done towards establishment of fatigue life prediction methodologies of these composites which calls for validation of computational models by experiments. This paper is a step forward in that direction.
Literature survey indicates that the fatigue crack tip in weak steel is shielded from the remote load when it reaches near a thin interface of strong steel and that the effective near tip J integral is smaller than the remote J integral value [1]. These J integrals have hitherto been obtained numerically by finite element analysis. A computational model based upon the concept of crack tip stress intensity parameter is employed in the present work for theoretical treatment of the fatigue crack in a welded composite under K dominant conditions. The model is validated by experiments. A thick weld composite prepared by electron beam welding of weak alloy and strong maraging steels is employed in experimental work.

Theoretical review
High stresses develop ahead of the crack tip resulting in material plastic strain. Refer Fig. 1. The crack tip zone, comprising process and plastic regions, is commonly referred to as the yield zone. Dugdale [2] modeled the yield zone as a cohesive zone subjected to constant closing cohesive stresses, σ , equal to material yield strength, Y, in plane stress condition. These stresses are assumed as 3 Y in plane strain condition.
Refer Fig. 2. Ultra strong weld, W, obtained by welding weak parent steel, A, and strong interface steel, B, results in weld interfaces I and II. Effect of weld interface I is not felt by fatigue crack tip in parent steel till its distance, a, from the interface fulfills the condition, r a ∆ > Dugdale's cyclic cohesive zone length, r ∆ , under cyclic applied stress intensity parameter, applied K ∆ , is given by As the crack tip grows and reaches near weld interface I such that r a ∆ < , spread of cohesive zone across the interface results in load transfer towards weld steel because of its higher yield strength. This induces a diminishing or shielding effect of load over the crack tip. As the result, the magnitude of stress field and cyclic stress intensity parameter at the crack tip, tip K ∆ , drops resulting in dip in crack growth rates when compared with the values at crack tip in uni-material body made of parent steel alone.
The computational model [3] based upon Dugdale's cohesive zone criterion to obtain tip K ∆ at the tip of a Mode I fatigue crack in linear elastic regime near elastically matched weld interface I with the cohesive zone having extended into weld by distance, l, is as follows: Eqs. (1) and (2) are solved numerically for unknowns l and The model is valid when the crack tip crosses the weld and reaches near weld interface II, A σ and W σ in that case being replaced by W σ and B σ respectively resulting in

Experimental work
Bimetallic compact tension specimens were prepared by electron beam welding of 10 mm thick plates of hot rolled, low strength, ASTM 4340 alloy steel, A, and hot rolled, solution annealed, high strength, MDN 250 maraging steel, B. 1.5 mm wide ultra strong weld, W, was made between the steels such that the weld interface I was at the distance of 48.5 mm from the right end of the specimen. A notch, 30.5 mm long and 6.25 mm wide at the mouth, was machined perpendicular to weld in alloy steel side of the specimens. Fine tip at the notch was obtained by cutting with wire of 0.3 mm size. A bimetallic specimen is shown in Fig. 3. The specimens were subjected to high cycle tension-tension fatigue load of constant amplitude at ambient environment in ± 250 kN capacity material test rig. Details of load cycle were: Maximum load, max F = 14.7 kN, Minimum load, min F = 0.98 kN and Frequency = 20 cycles per second. The crack, generated at the notch tip by fatigue precracking, was made to grow till it became critical resulting in specimen fracture. Crack length, c, was measured at different positions from the right end of the specimen. Existence of linear elastic regime and non-occurrence of limit load conditions in alloy steel were ensured at each crack position with selected load values under the influence of load transfer. Number of cycles, , required for incremental crack growth were recorded. Crack growth rate, dc/d , was computed. . Constants C and m were found from initial data when the effect of weld interface I had not started.
Plain specimens of alloy steel with geometry and crack configuration identical with that of bimetallic specimens were fabricated and subjected to similar fatigue load for comparison of crack growth characteristics in them with those in the bimetallic specimens. Fractured specimens are shown in Fig. 4. Micrographs of crack surfaces at fracture locations in bimetallic and plain specimens are displayed in Fig. 5. values, helping the crack to grow longer before becoming critical. Micrograph of alloy steel near weld interface I in bimetallic specimen, when the specimen had not yet fractured, was different from the micrograph of fractured surface of plain alloy steel specimen which confirmed delayed fracture of bimetallic specimen. Effect of weld interface II was not experienced by the crack since it had already become critical before reaching that interface.

Results
Value of r ∆ for commencement of the effect of weld interface I was estimated as 4.07 mm at crack length at 44.43 mm. But it was found during experiments that the influence of weld commenced earlier at crack length of 38 mm. This was due to additional load transfer towards stronger materials on other side of weld interface I by eccentric loading over the crack. Correction  at different crack positions. Plane stress and plane strain conditions existed in alloy and in weld respectively for 10 mm thickness. Computation confirmed that the extent of spread of cohesive zone was restricted to the weld without crossing over into maraging steel since the maximum value of l at crack length of 48.5 mm was 0.12 mm which was less than the weld width of 1.5 mm. Plots of applied K ∆ and tip K ∆ during crack growth in a typical bimetallic specimen are presented in Fig. 6