In our extensive practice within heavy industrial maintenance, we have frequently confronted the challenge of repairing critical large-scale steel castings that operate under extreme conditions. One of the most demanding scenarios involves the steel casting archways of wide and heavy plate finishing mills. These monumental steel castings, weighing hundreds of tons, form the backbone of the rolling mill’s load-bearing system. Their failure, often in the form of deep fatigue cracks, poses a severe threat to continuous production. Traditional repair methods, such as milling away the cracked section, significantly reduce the cross-sectional area and load-bearing capacity of these vital steel castings, forcing operation at derated loads. Complete replacement is prohibitively expensive and time-consuming, with lead times exceeding nine months. This narrative details our first-person experience and methodology in developing and executing a successful in-situ welding repair procedure for such massive, cracked steel castings, ensuring their full operational integrity without disassembly.
The subject steel castings are the four vertical archways of a 10,000-tonne capacity finishing mill. Two of these archways developed significant horizontal fatigue cracks originating at the inner transition radius (R150mm) on the top underside, a critical stress zone. Non-destructive testing (NDT) revealed one crack with a surface length of 360mm and a maximum depth of 95mm, and another approximately 400mm long and 50mm deep. Alarmingly, the deeper crack was propagating at a rate of 1-2mm per day under continuous operation. The material, specified as GS-20Mn5V low-alloy cast steel according to DIN 17182, presents specific challenges. The chemical composition and mechanical properties of this grade of steel castings are summarized below.
| Element | C | Si | Mn | S | P | Ni | V | Cr | Mo |
|---|---|---|---|---|---|---|---|---|---|
| Wt. % | 0.18-0.22 | 0.60 | 1.00-1.50 | ≤0.015 | ≤0.020 | 0.30 | 0.05 | 0.35 | 0.12 |
| Property | Tensile Strength | Yield Strength | Elongation | Impact Energy |
|---|---|---|---|---|
| Value | ≥500 MPa | ≥280 MPa | ≥22% | ≥22 J |
The root cause analysis for the cracking in these massive steel castings converged on several interrelated factors. Primarily, the location is subjected to immense alternating loads. During rolling, each pass imposes a force cycle on the archway. With a rolling schedule of 20-30 plates per hour and 6-10 passes per plate, the inner radius zone endures 120 to 300 significant stress cycles hourly. The stress concentration factor (Kt) at such a geometric discontinuity can be approximated for preliminary assessment. For a shoulder fillet in a rectangular member under bending, the stress concentration is significantly higher than the nominal stress. The nominal bending stress (σnom) is given by:
$$ \sigma_{nom} = \frac{M \cdot y}{I} $$
where M is the bending moment, y is the distance from the neutral axis, and I is the area moment of inertia. The localized peak stress (σpeak) becomes:
$$ \sigma_{peak} = K_t \cdot \sigma_{nom} $$
In these steel castings, the combination of a sharp 90° corner (in the horizontal plane) with a 150mm fillet (in the vertical plane) creates a complex triaxial stress state, elevating Kt and making the site prone to fatigue initiation. Furthermore, metallurgical examination of samples extracted from the crack region revealed inherent casting defects. These included macroscopic shrinkage cavities and microstructural banding, as shown in the micrographs. Banded structures, consisting of alternating layers of ferrite and pearlite with differing hardness, act as planes of weakness. The hardness variation across these bands can be described by the local carbon concentration, influencing the phase fraction. The volume fraction of pearlite (Vp) can be estimated using the lever rule approximation for hypereutectoid steels under non-equilibrium cooling:
$$ V_p \approx \frac{C_0 – C_{\alpha}}{C_{cementite} – C_{\alpha}} $$
where C0 is the local carbon content, Cα is the carbon in ferrite (~0.022%), and Ccementite is 6.7%. The localized chemical segregation during solidification of these thick-section steel castings leads to such inhomogeneity, reducing ductility and toughness locally. The synergy of high cyclic stress, severe stress concentration, and inherent material discontinuities typical of large steel castings culminated in progressive fatigue crack initiation and growth.
Assessing the weldability of these GS-20Mn5V steel castings was our next critical step. The primary concern was the high risk of hydrogen-induced cold cracking and the formation of brittle microstructures in the heat-affected zone (HAZ). We employed the International Institute of Welding (IIW) carbon equivalent (CE) formula to quantify the hardenability and cracking tendency:
$$ CE = C + \frac{Mn}{6} + \frac{Cr + Mo + V}{5} + \frac{Ni + Cu}{15} $$
Substituting the mid-range values from the chemical composition table:
$$ CE = 0.20 + \frac{1.25}{6} + \frac{0.35 + 0.12 + 0.05}{5} + \frac{0.30 + 0.00}{15} $$
$$ CE = 0.20 + 0.208 + 0.104 + 0.02 = 0.532 $$
For conservative assessment, using the maximum values yields a CE exceeding 0.55. A CE value above 0.40 generally indicates poor weldability requiring preheat. Our calculated value confirmed a significant淬硬倾向, necessitating strict thermal control. Furthermore, the immense structural restraint posed by the 185-tonne steel casting itself and the 2000+ tonnes of attached mill components creates a restrained condition that inhibits shrinkage, leading to high residual stresses. The three-dimensional heat flow during welding in such a massive steel casting can be modeled simplistically using the Rosenthal equation for a thick plate, but the practical implication is rapid heat dissipation, increasing the cooling rate (Δt8/5). The cooling rate is critical for HAZ microstructure; excessive cooling promotes martensite formation. The need for controlled preheat, interpass temperature, and post-weld heat treatment (PWHT) was unequivocal.
The repair of these critical steel castings was conducted entirely in-situ, with the mill framework under load. The first operation was defect removal. For the shallow crack (50mm deep), we used mechanical grinding to open a U-shaped groove with a root radius of 15mm, followed by penetrant testing (PT). For the deep crack (95mm+, propagating to 115mm), we selected controlled carbon arc gouging. To prevent thermal shock from exacerbating the crack, the entire region 300mm around the defect was preheated to 200±20°C and soaked for 6 hours. Gouging was then performed under this sustained preheat using a 10mm carbon electrode at 500A with 0.5MPa compressed air. The final groove was also PT and ultrasonically tested (UT) to ensure complete defect removal, revealing a sound but geometrically complex cavity in the steel casting.
We selected a hybrid welding approach to balance control, productivity, and mechanical properties. The root and hot pass layers were deposited using Shielded Metal Arc Welding (SMAW) for better penetration control in the horizontal position. The filler and cap passes were made with Gas Metal Arc Welding (GMAW) using a pulsed spray transfer mode to increase deposition rate and improve operator comfort for the extensive volume required. The chosen consumables were:
| Process | Consumable | Specification | Key Characteristics |
|---|---|---|---|
| SMAW | E5015-N1P (CHE507RH) φ3.2mm | GB/T 5117 | Low-hydrogen, high toughness (-40°C impact ≥54J) |
| GMAW | G49A3 C1/M21 S6 (CHW-50C6) φ1.2mm | GB/T 8110 | Mn-Ni alloyed wire, CO2/Ar mix gas, good strength & toughness |
The detailed welding parameters and procedure were meticulously controlled:
| Stage | Process | Current (A) | Voltage (V) | Heat Input (kJ/cm) | Travel Speed (cm/min) | Gas Flow (L/min) |
|---|---|---|---|---|---|---|
| Preheat | N/A | N/A | N/A | N/A | N/A | N/A |
| Target | — | — | — | — | — | — |
| Root/Pass 1-2 | SMAW | 110-130 (DC+) | 22-24 | ~10-12 | 10-15 | N/A |
| Filler/Cap | GMAW (Pulse) | 240-260 (Avg) | 25-28 | 12-15 | 15-25 | 20 (82%Ar/18%CO2) |
Thermal management was paramount. The preheat of 200°C was maintained throughout the welding operation. We monitored interpass temperature rigorously, ensuring it never fell below 180°C. The heat input (Q) for each pass was calculated and controlled:
$$ Q = \frac{\eta \cdot V \cdot I}{v \cdot 1000} \quad \text{[kJ/mm]} $$
where η is the arc efficiency (0.8 for SMAW, 0.9 for GMAW), V is voltage, I is current, and v is travel speed in mm/s. We aimed for a range of 1.0-1.5 kJ/mm to avoid excessive grain growth while preventing too rapid a cooling rate. To mitigate residual stresses during deposition, we employed intermediate stress relief. After filling approximately 60mm of the 115mm deep groove, we performed a localized intermediate post-weld heat treatment (IPWHT). The steel casting region was heated to 560°C, held for 8 hours, and cooled slowly under insulation. This step reduced the accumulated stress from the initial weld layers. Welding then resumed to complete the joint.
The welding technique itself was critical. We used a multi-layer, multi-pass, staggered sequence to distribute heat evenly. Each pass was peened immediately after deposition using a round-nose pneumatic tool (R≥6mm), except for the root and cap passes. Peening induces beneficial compressive stresses, partially counteracting tensile shrinkage stresses. The relationship between peening intensity and induced strain can be complex, but the practical goal was a slight plastic deformation of the weld metal surface. The entire operation was conducted by a relay team of welders to ensure continuity, preventing any unnecessary cooling cycles that could jeopardize the integrity of the repair on these large steel castings.
Upon completion of welding, a final and more extensive post-weld heat treatment (PWHT) was executed in-situ. The repaired section of the steel casting was heated uniformly to 560°C, soaked for 16 hours to ensure through-thickness temperature equalization, and then furnace-cooled (within the insulation) to below 250°C before exposure to ambient air. The PWHT cycle serves multiple functions: it tempers any martensite formed in the HAZ, promotes diffusion of residual hydrogen, and most importantly, reduces residual stresses through creep relaxation. The stress relief efficiency can be conceptually related to the Larson-Miller parameter (PLM) for creep, though for practical purposes, the hold time and temperature were selected based on standard guidelines for low-alloy steel castings. The final step was contour grinding to blend the weld reinforcement smoothly into the parent metal and to restore a generous radius (R16mm) at the previously sharp corner, effectively reducing the future stress concentration factor for this steel casting.
The efficacy of our repair procedure for these monumental steel castings was validated through comprehensive non-destructive and operational testing. Magnetic particle testing (MT) over the entire weld and adjacent heat-affected zone showed no indications of surface cracks. Ultrasonic testing (UT), using a calibrated distance-gain-size (DGS) method, confirmed the soundness of the volumetric weld metal and full fusion with the base steel casting, with no reportable planar or volumetric defects. The repaired archways were returned to service at the full 10,000-tonne rated load. We monitored the repair zones periodically using acoustic emission sensors and periodic UT scans during planned maintenance outages for over a year. No crack initiation or propagation was detected. The mill has operated at full capacity and design parameters since the repair, demonstrating that the restored steel castings perform identically to their original condition.
This successful intervention underscores several key principles for repairing large, critically loaded steel castings. Firstly, a thorough root cause analysis integrating metallurgy, stress analysis, and operational history is non-negotiable. Secondly, weldability assessment must go beyond standard carbon equivalent calculations; the immense restraint and heat-sink effect of multi-ton steel castings demand tailored thermal procedures. Our use of a hybrid welding process combined with intermediate and final PWHT proved highly effective. The strategic application of intermediate stress relief during the welding of thick sections is a technique we find invaluable for massive steel castings, as it prevents the buildup of critical stress levels during deposition. Furthermore, the selection of high-toughness, low-hydrogen consumables is essential to match and often exceed the base metal’s properties, particularly in fracture-critical zones of large steel castings.
The economic and operational impact of developing this in-situ repair capability for steel castings is profound. It eliminates the need for multi-month shutdowns and the multi-million-dollar cost of casting and installing new archways. More importantly, it preserves the original design strength and fatigue life of the component, unlike mechanical removal methods. The knowledge gained extends beyond rolling mill archways to other massive steel castings in heavy industry, such as those used in forging presses, shipyard gantries, and mining shovel bases. The repair of such steel castings in-situ represents a pinnacle of advanced welding engineering, requiring a deep synthesis of materials science, structural mechanics, and practical field execution. As the global industrial infrastructure ages, the ability to reliably and robustly repair these foundational steel castings will become increasingly vital, ensuring safety, productivity, and sustainability.