I treated the repair of large turbine-casing casting defects as a coupled thermal, metallurgical, and mechanical problem. In my experience, casting defects such as porosity, sand inclusions, slag inclusions, hot tears, cold cracks, shrinkage cavities, and incomplete fusion zones are common in large steel castings because of complex geometry, long solidification times, and uneven cooling. When these casting defects remain unrepaired, they reduce load-bearing area, generate stress concentrations, and shorten service life. For turbine casings, which operate under high temperature and pressure, the repair of casting defects must restore not only geometry but also mechanical integrity, creep resistance, fatigue resistance, and dimensional stability.
I focused on a low-alloy pearlitic heat-resistant cast steel, ZG20CrMoV, because it represents many large turbine-casing casting defects encountered in power-generation equipment. My aim was to use welding simulation to understand how repair welding of casting defects changes the temperature field, stress field, strain field, and final quality. I combined numerical simulation with actual repair welding trials. I selected process parameters, compared candidate repair schemes, evaluated residual stress, optimized the welding procedure, applied post-weld heat treatment, and validated the final repair by non-destructive inspection.
I began by classifying the casting defects that are most relevant to large turbine casings. The following table summarizes how I related each type of casting defect to repair-welding difficulty and inspection requirements.
| Casting Defect Type | Typical Origin | Repair-Welding Concern | Inspection Focus |
|---|---|---|---|
| Porosity | Gas entrapment during solidification | Gas expansion, incomplete fusion, residual porosity | Radiography, ultrasonic testing |
| Sand inclusion | Mold or core erosion | Slag entrapment, lack of fusion | Visual, penetrant, radiographic |
| Slag inclusion | Non-metallic melt reaction products | Incomplete cleaning, weld contamination | Radiography, ultrasonic testing |
| Hot tear | Constrained contraction during solidification | Crack propagation into heat-affected zone | Penetrant, magnetic particle, ultrasonic |
| Cold crack | Hydrogen and residual stress after cooling | Delayed cracking, brittle heat-affected zone | Ultrasonic, magnetic particle |
| Shrinkage cavity | Insufficient feeding during solidification | Deep excavation, high heat input, distortion | Radiography, ultrasonic testing |
| Incomplete fusion defect | Poor welding or casting flow | Lack of sidewall fusion during repair | Ultrasonic testing, radiographic testing |
I found that casting defects in large turbine casings cannot be treated as simple surface flaws. Many casting defects are irregular after excavation, and the repair cavity often has a complex three-dimensional shape. I therefore modeled the excavated casting defect as a realistic groove, not as a perfect rectangular notch. In my simulation I used a base block with dimensions of 600 mm × 300 mm × 250 mm. The excavated casting defect was approximated as a groove 180 mm long, 60 mm wide, and 40 mm deep, with a sidewall angle of 12° and a root radius of 8 mm. I used reciprocating weld passes, multiple layers, and multiple beads to represent the manual arc repair of casting defects.
I selected ZG20CrMoV because its chemical composition and service conditions make it a representative material for high-pressure and intermediate-pressure turbine casings. The material is a pearlitic heat-resistant cast steel with good high-temperature strength and structural stability, but it also has significant welding challenges. I summarized the nominal chemical composition in the following table.
| Element | Range or Limit, wt.% | Role in Repair Welding of Casting Defects |
|---|---|---|
| C | 0.18–0.25 | Controls hardenability and strength |
| Si | 0.20–0.60 | Deoxidation and solid-solution strengthening |
| Mn | 0.40–0.70 | Hardenability and sulfide shape control |
| Cr | 0.90–1.20 | Oxidation resistance and creep strength |
| Mo | 0.50–0.70 | Creep strength and hardenability |
| V | 0.20–0.30 | Precipitation strengthening and reheat-cracking sensitivity |
| Cu | ≤0.30 | Residual element |
| S | ≤0.030 | Hot-cracking sensitivity |
| P | ≤0.030 | Temper embrittlement sensitivity |
I calculated the carbon equivalent to evaluate the weldability of this material when repairing casting defects. The International Institute of Welding carbon equivalent is expressed as:
$$CE_{IIW}=C+\frac{Mn}{6}+\frac{Cr+Mo+V}{5}+\frac{Ni+Cu+Si}{15}$$
For ZG20CrMoV, I obtained a carbon equivalent of approximately 0.81%. This value is well above the commonly used threshold of 0.45%, which means that the material has poor weldability under rapid cooling. When I repair casting defects in this steel, the heat-affected zone can transform to martensite, and the combination of hydrogen and restraint can produce cold cracks. The large size and high rigidity of turbine casings further increase restraint, so the repair of casting defects must be supported by preheating, interpass temperature control, and post-weld heat treatment.
I also evaluated reheat-cracking sensitivity because the repaired casting defects must survive post-weld heat treatment and long-term high-temperature service. I used two commonly cited parameters. The first is:
$$\Delta G=Cr+3.3Mo+8.1V+10C-2$$
The second is:
$$\Delta P_{SR}=Cr+Cu+2Mo+10V+7Nb+5Ti-2$$
For the material chemistry I considered, the calculated values were positive and exceeded the usual threshold, which indicated a high sensitivity to reheat cracking. I therefore concluded that the repair of casting defects in ZG20CrMoV must minimize residual stress and avoid excessive heat input, while still providing enough preheat to prevent cold cracking. This balance became the central problem of my process optimization.
I used temperature-dependent material properties in the simulation because the repair welding of casting defects involves rapid heating, melting, solidification, and cooling. The following table summarizes the trends I applied for specific heat, thermal conductivity, elastic modulus, thermal expansion coefficient, and yield strength.
| Temperature, °C | Specific Heat, J kg⁻¹ °C⁻¹ | Thermal Conductivity, W m⁻¹ °C⁻¹ | Elastic Modulus, GPa | Thermal Expansion, 10⁻⁶ °C⁻¹ | Yield Strength, MPa |
|---|---|---|---|---|---|
| 20 | 420 | 41.2 | 216.7 | 11.86 | 490 |
| 100 | 448 | 43.1 | 212.5 | 12.18 | 441 |
| 200 | 502 | 46.2 | 206.3 | 12.58 | 415 |
| 300 | 515 | 44.0 | 198.2 | 13.18 | 419 |
| 400 | 527 | 41.8 | 190.5 | 13.06 | 419 |
| 500 | 552 | 39.0 | 178.8 | 13.96 | 392 |
| 550 | 560 | 37.4 | 174.3 | 14.10 | 304 |
| 600 | 572 | 36.1 | 169.5 | 14.20 | 251 |
| 700 | 584 | 34.8 | 165.4 | 14.32 | — |
| 750 | 602 | — | — | 14.50 | — |
I modeled the repair welding of casting defects using a finite-element welding simulation approach. I chose a volumetric heat source because a simplified uniform heat source cannot accurately capture the molten pool, the heat-affected zone, and the thermal cycles that control residual stress. A double-ellipsoidal heat source is widely used for arc welding. For the front part of the heat source, I used the following form:
$$q_f(x,y,z,t)=\frac{6\sqrt{3}f_f Q}{a_f b c \pi\sqrt{\pi}}\exp\left(-\frac{3x^2}{a_f^2}-\frac{3y^2}{b^2}-\frac{3z^2}{c^2}\right)$$
For the rear part, I used a similar expression:
$$q_r(x,y,z,t)=\frac{6\sqrt{3}f_r Q}{a_r b c \pi\sqrt{\pi}}\exp\left(-\frac{3x^2}{a_r^2}-\frac{3y^2}{b^2}-\frac{3z^2}{c^2}\right)$$
The total arc power and net heat input are given by:
$$Q=\eta U I$$
$$H_{net}=\frac{\eta U I}{v}$$
Here, η is the arc efficiency, U is the voltage, I is the current, and v is the welding speed. I used these equations to compare two candidate repair-welding schemes for casting defects. I kept the heat-source type, the reciprocating path, and the base geometry identical so that the comparison would isolate the effects of preheat, interpass temperature, electrode type, current, and voltage.
The two candidate schemes I evaluated are summarized in the following table.
| Parameter | Scheme I: Matching Heat-Resistant Electrode | Scheme II: Austenitic Stainless Electrode |
|---|---|---|
| Electrode | R317, diameter 5.0 mm | A307, diameter 5.0 mm |
| Preheat temperature | 250 °C | 150 °C |
| Interpass temperature control | About 200 °C, not above 350 °C | About 100 °C, not above 300 °C |
| Current | 200 A | 200 A |
| Voltage | 28 V | 28 V |
| Travel speed | 5 mm/s | 5 mm/s |
| Pass strategy | Reciprocating multi-layer multi-bead | Reciprocating multi-layer multi-bead |
| Post-weld heat treatment | 700 °C × 6 h | Stress relief considered separately |
I selected the matching heat-resistant electrode scheme as the superior route for repairing casting defects in ZG20CrMoV. The main reason was material compatibility. A matching ferritic or pearlitic heat-resistant weld metal reduces the mismatch in thermal expansion, creep behavior, and long-term microstructural stability. Although an austenitic stainless electrode can improve ductility and simplify preheating in some cases, the dissimilar joint can develop carbon migration, interfacial stress, and long-term creep mismatch when the casing operates at high temperature. For casting defects in a large turbine casing, I judged that long-term stable behavior was more important than short-term welding convenience.
I then examined the transient temperature field during the repair of casting defects. The temperature field is not only a thermal problem; it drives thermal expansion, plastic strain, residual stress, microstructure, and cracking. I used a moving heat source and a reciprocating path. The molten pool formed a quasi-steady temperature field. Ahead of the heat source, the isotherms were dense, which indicated a steep temperature gradient. Behind the heat source, the isotherms were more widely spaced, which indicated slower cooling and heat accumulation. As the repair proceeded from layer to layer, the maximum temperature increased and then stabilized. In my simulation, the maximum molten-pool temperature reached approximately 1517 °C.
The following table summarizes the thermal characteristics I observed at selected times and layers when repairing the casting defects.
| Stage | Maximum Temperature, °C | Thermal Gradient, K/m | Heat-Affected Zone Behavior | Repair-Welding Implication |
|---|---|---|---|---|
| First layer completion | About 1450–1500 | 3775 | Local steep gradient near stop position | Risk of thermal stress and local hardening |
| Second layer completion | About 1500 | 3883 | Heat accumulation begins | Interpass control becomes critical |
| Fifth layer completion | About 1510 | 4685 | Broader hot zone, repeated tempering | Residual stress may redistribute |
| Eighth layer completion | About 1517 | 4909 | High thermal load near final passes | Cracking risk if cooling is too fast |
I inserted characteristic points near the repair region to track temperature histories. The temperature at these points rose extremely quickly when the heat source approached, then fell rapidly as the heat source moved away. During the first pass, the peak temperature reached about 1648.7 °C. Within a short time, the temperature dropped below 400 °C. This rapid thermal cycle can refine the microstructure, but it can also create hard phases and high local stress if the preheat and interpass temperature are not controlled. In the first three passes, the peak temperature at the characteristic point exceeded the melting range of ZG20CrMoV, which means that the point remelted and resolidified during the repair of casting defects. This repeated melting and solidification contributed to a more uniform final deposit, but it also increased the complexity of the thermal history.
I observed that the trough temperature between passes remained near 200 °C. This indicated that the base material retained a stable thermal condition during the multi-pass repair of casting defects. The material did not show uncontrolled heat accumulation at the characteristic point. Nevertheless, the temperature gradients increased as the repair progressed. The maximum temperature gradient rose from about 3775 K/m in the first layer to about 4909 K/m in the eighth layer. These gradients are important because thermal stress scales with the constrained thermal strain:
$$\epsilon_{th}=\alpha(T)(T-T_0)$$
$$\sigma_{th}\approx \frac{E\alpha\Delta T}{1-\nu}$$
When the thermal stress exceeds the temperature-dependent yield strength, plastic deformation occurs. The residual stress after cooling is the accumulated consequence of incompatible thermal strain, plastic strain, phase transformation, and external restraint. For casting defects repaired in a massive turbine casing, external restraint is high because the surrounding base metal is cold and rigid relative to the small repair cavity.
After selecting the matching heat-resistant electrode scheme, I analyzed the stress field. I used the von Mises equivalent stress as the main comparison parameter because it condenses the multiaxial stress state into a scalar that can be compared with the yield strength. The von Mises equivalent stress is:
$$\sigma_v=\sqrt{\frac{1}{2}\left[(\sigma_x-\sigma_y)^2+(\sigma_y-\sigma_z)^2+(\sigma_z-\sigma_x)^2+6(\tau_{xy}^2+\tau_{yz}^2+\tau_{zx}^2)\right]}$$
I also tracked equivalent elastic strain and equivalent plastic strain. Equivalent elastic strain indicates recoverable deformation under the local stress state. Equivalent plastic strain indicates irreversible deformation that remains after cooling. High plastic strain is especially dangerous because it can be associated with strain aging, microcracking, and reduced fatigue life.
The following table summarizes the stress and strain results I obtained for the original Scheme I before optimization.
| Result Quantity | Original Scheme I Value | Location | Interpretation for Casting Defects Repair |
|---|---|---|---|
| Maximum transient equivalent stress | About 310 MPa | Completed bead near repair edge | Approaching yield limit; local stress concentration |
| Maximum X-direction normal stress after cooling | About 273 MPa | Near repair surface | Tensile residual stress on outer repair zone |
| Maximum Y-direction normal stress after cooling | About 342 MPa | Center of heat-source path | Compressive-to-tensile transition and constraint effect |
| Maximum Z-direction normal stress after cooling | About 383 MPa | Repair zone beside base metal | High through-thickness stress; possible crack initiation |
| Maximum equivalent elastic strain | About 0.097 | Repair/base interface | Large elastic accommodation; assembly accuracy risk |
| Maximum equivalent plastic strain | About 0.24 | Bottom of repair zone | Irreversible deformation; residual stress risk |
| Maximum yield stress after cooling | About 297 MPa | Bottom of repair zone | Local plastic deformation likely |
The original Scheme I produced acceptable temperature control but excessive residual stress in some locations. I found that the bottom of the repair zone and the interface between the repair weld and the base metal were the most critical regions. The stress distribution was not uniform. The outer surface of the repair zone showed tensile stress, while the side near the base metal showed compressive stress. This pattern resulted from the expansion of the heated repair deposit against the colder surrounding base metal, followed by constrained contraction during cooling. The reciprocating path also produced local stress concentrations at the start and stop positions of each bead.
I decided to validate the original scheme by performing an actual repair of casting defects under the same nominal conditions. I prepared the surface, removed the defect by mechanical excavation, and produced a groove with an angle close to 12°. I preheated the part, maintained the interpass temperature, used the specified electrode, and completed the multi-layer repair. After cooling, I performed X-ray inspection. The X-ray image showed clear indications of porosity and crack-like linear indications in the repaired region. The porosity appeared as dark spots or rounded indications, while the crack indication appeared as an irregular bright band or line. This confirmed that the original process was not reliable for the repair of casting defects in this material.
The defects I found after the original repair trial are summarized in the following table.
| Inspection Result | Observed Indication | Likely Cause | Process Implication |
|---|---|---|---|
| Porosity | Dark rounded spots in radiographic image | Insufficient gas escape, excessive heat input, moisture, or poor cleaning | Adjust current and voltage; improve drying and cleaning |
| Crack | Irregular bright linear band | High restraint, rapid cooling, residual stress, hydrogen | Reduce heat input; optimize preheat and interpass control |
| Local stress concentration | Predicted by simulation | Constrained contraction in thick casting | Use stress-relief heat treatment and better pass sequence |
I concluded that the original process had to be optimized. The optimization targets were clear. I needed to reduce heat input, reduce the peak temperature gradient, lower the residual stress after cooling, and maintain sufficient fusion and preheat to avoid cold cracking. I therefore adjusted the welding parameters while keeping the same overall strategy: matching heat-resistant electrode, reciprocating multi-layer deposition, and post-weld heat treatment. The optimized parameters are summarized below.
| Parameter | Original Scheme I | Optimized Scheme I | Reason for Change |
|---|---|---|---|
| Initial base temperature | 20 °C | 25 °C | Small increase in initial thermal stability |
| Ambient temperature | 25 °C | 25 °C | Kept constant |
| Preheat temperature | 250 °C | 200 °C | Reduce excessive thermal expansion while still preventing cold cracking |
| Heating rate | 150 °C/h | 100 °C/h | More uniform temperature distribution |
| Interpass temperature | 200 °C | 200 °C | Maintain stable thermal condition |
| Current | 200 A | 180 A | Lower heat input |
| Voltage | 28 V | 26 V | Lower arc energy |
| Travel speed | 5 mm/s | 5 mm/s | Maintain productivity and stable deposition |
| Path | Reciprocating | Reciprocating | Keep balanced heat distribution |
| Post-weld heat treatment | 700 °C × 6 h | 700 °C × 6 h | Retain stress relief and microstructural improvement |
I recalculated the net heat input using the optimized parameters. The original net heat input was:
$$H_{net,original}=\frac{\eta U I}{v}=\frac{0.65\times 28\times 200}{5}=728\ \text{J/mm}$$
The optimized net heat input was:
$$H_{net,optimized}=\frac{\eta U I}{v}=\frac{0.65\times 26\times 180}{5}=608.4\ \text{J/mm}$$
This reduction of about 16.4% in net heat input was important. Lower heat input reduces the peak temperature, narrows the heat-affected zone, and reduces the thermal strain that drives residual stress. However, if the heat input is reduced too much, lack of fusion and cold cracking can occur. Therefore, I did not reduce the heat input aggressively. I also increased the preheat uniformity by slowing the heating rate. The optimized process therefore balanced crack prevention and stress reduction.
After optimization, I re-ran the simulation for the repair of casting defects. The temperature field became more uniform. The molten pool still formed properly, and the fusion boundary was continuous. The maximum temperature remained high enough for metallurgical bonding, but the surrounding temperature distribution was less extreme. The high-temperature zone expanded more gradually from layer to layer. I did not observe unstable overheating or cold regions that could cause lack of fusion.
I summarized the optimized thermal results in the following table.
| Thermal Metric | Original Scheme I | Optimized Scheme I | Improvement |
|---|---|---|---|
| Maximum molten-pool temperature | About 1517 °C | About 1480–1510 °C | More controlled peak |
| Maximum thermal gradient | About 4909 K/m | Reduced and more uniform | Lower thermal stress driving force |
| Heat-affected zone spread | Progressive and locally steep | Progressive but smoother | Better interpass stability |
| Characteristic-point peak | About 1648.7 °C | Lower peak with stable trough | Less severe thermal shock |
| Trough temperature | Near 200 °C | Near 200 °C | Maintained stable base condition |
I also analyzed the optimized stress field. The maximum transient equivalent stress decreased significantly. The stress concentration at the start and stop positions became smaller. The stress at the bottom of the repair zone decreased, and the plastic strain was much lower. The optimized process still produced residual stress, but the residual stress was below the level that would cause immediate cracking. The following table compares the original and optimized stress and strain results.
| Quantity | Original Scheme I | Optimized Scheme I | Effect of Optimization |
|---|---|---|---|
| Maximum transient equivalent stress | About 310 MPa | About 267–306 MPa | Lower and more stable |
| Maximum X-direction normal stress | About 273 MPa | About 227 MPa | Lower tensile residual stress |
| Maximum Y-direction normal stress | About 342 MPa | About 213 MPa | Reduced constraint stress |
| Maximum Z-direction normal stress | About 383 MPa | About 329 MPa | Lower through-thickness stress |
| Maximum equivalent elastic strain | About 0.097 | About 0.065 | Less elastic accommodation |
| Maximum equivalent plastic strain | About 0.24 | About 0.09 | Much less permanent deformation |
| Maximum plastic strain rate | Not controlled | About 0.10662 s⁻¹ | Limited local plastic flow |
The reduction in equivalent plastic strain from about 0.24 to about 0.09 was one of the most important results. Plastic strain is directly linked to residual stress and dimensional instability. When repairing casting defects in a large casing, even small plastic strains can accumulate over multiple passes and produce distortion or cracking. The optimized process therefore improved both the thermal and mechanical responses of the repair.
I then evaluated post-weld heat treatment. For ZG20CrMoV, a stress-relief treatment at 700 °C for 6 hours is appropriate because it allows creep relaxation of residual stresses and promotes tempering of hard microstructures. I used the Larson–Miller parameter to relate the heat treatment to long-term service exposure:
$$P=T(C+\log_{10}t)$$
Here, T is the absolute temperature, t is time in hours, and C is a material-related constant. I used C ≈ 26 for this class of steel. The 700 °C × 6 h treatment corresponded to a significant thermal exposure, equivalent to long-term service at a lower temperature. This treatment helped to relax residual stress without causing excessive grain growth or softening.
I summarized the post-weld heat treatment effect in the following table.
| Condition | Maximum Equivalent Stress | Maximum Plastic Strain Rate | Microstructural Effect |
|---|---|---|---|
| After optimized repair, before heat treatment | Up to about 306 MPa locally | About 0.10662 s⁻¹ locally | Hard, strained repair zone |
| After 700 °C × 6 h heat treatment | About 100 MPa | About 0.06 | Relaxed residual stress and tempered microstructure |
After post-weld heat treatment, the maximum equivalent stress dropped to about 100 MPa, which is well below the yield strength of the material. The plastic strain rate was also low. This indicated that the heat treatment successfully reduced the driving force for cracking and improved the stability of the repair. I considered this essential for the long-term performance of casting defects repairs in turbine casings.
I performed an actual repair welding trial using the optimized parameters. I prepared the casting defect cavity, cleaned the surface, preheated the part to 200 °C with a controlled heating rate, maintained the interpass temperature near 200 °C, and deposited the repair using the optimized current and voltage. I used the same reciprocating strategy and controlled the layer thickness. After welding, I applied the 700 °C × 6 h post-weld heat treatment and cooled the part under controlled conditions. I then performed X-ray inspection.
The X-ray image of the optimized repair showed a uniform and continuous repaired region. I did not observe porosity, slag inclusions, lack of fusion, or crack-like indications. The repair of casting defects therefore met the quality requirement. This validated my simulation-based optimization. The comparison between the original and optimized trials is summarized below.
| Trial | Process Condition | X-ray Result | Conclusion |
|---|---|---|---|
| Original repair of casting defects | R317, 250 °C preheat, 200 A, 28 V, 5 mm/s | Porosity and crack indications | Process not acceptable |
| Optimized repair of casting defects | R317, 200 °C preheat, 180 A, 26 V, 5 mm/s | No observable defects | Process acceptable |
I believe the main reasons for the improved result were lower heat input, more uniform preheating, and better control of the thermal cycle. Lower heat input reduced the maximum temperature gradient and the amount of plastic strain. More uniform preheating reduced the risk of cold cracking and helped the repair zone accommodate thermal contraction. The post-weld heat treatment then relaxed the remaining residual stress. This combination was more effective than simply increasing preheat or changing the electrode.
I also considered the practical constraints of repairing casting defects in large turbine casings. The casing is massive, so full furnace preheating may not be possible. Local preheating must be controlled carefully to avoid excessive temperature gradients. The repair cavity may be irregular, which makes manual deposition difficult. The operator must maintain the correct interpass temperature, clean each pass, and avoid excessive heat input. Simulation is useful because it provides a window into the thermal and mechanical history that cannot be measured easily in production. I used simulation not as a replacement for experience but as a guide for selecting parameters and interpreting inspection results.
I further summarized the key process variables and their effects on casting defects repair quality in the following table.
| Process Variable | Effect on Temperature Field | Effect on Stress and Strain | Effect on Casting Defects Repair Quality |
|---|---|---|---|
| Preheat temperature | Raises base temperature and reduces cooling rate | Reduces thermal gradient and cold-cracking risk | Too low risks cracking; too high increases distortion |
| Interpass temperature | Controls heat accumulation between passes | Stabilizes thermal cycles and plastic strain | Must be maintained within a narrow range |
| Current | Increases heat input and molten-pool size | Increases thermal stress and residual stress | Excessive current promotes porosity and distortion |
| Voltage | Changes arc length and heat distribution | Affects bead shape and stress concentration | Must match electrode and joint geometry |
| Travel speed | Controls local heating and cooling rate | Slower speed increases heat input | Must ensure fusion without overheating |
| Pass sequence | Controls heat accumulation and tempering | Reciprocating paths balance stress | Poor sequence increases distortion and cracks |
| Post-weld heat treatment | Provides controlled heating and cooling | Relaxes residual stress | Essential for creep-resistant steel repairs |
I also found that the repeated thermal cycles during multi-pass repair of casting defects had a beneficial tempering effect on previously deposited layers. Each new pass reheats the underlying material, which can temper hard phases and reduce stress. However, if the heat input is too high, this benefit is offset by excessive thermal expansion and plastic deformation. The optimized process therefore used enough passes to refine and temper the deposit but avoided excessive heat input. This balance is critical for casting defects that require deep excavation and multi-layer filling.
The temperature history at characteristic points also showed that the repair of casting defects is not a single thermal event. It is a sequence of heating and cooling cycles. Each cycle can cause phase transformation if the temperature exceeds the transformation range. The transformation temperature range for this steel is approximately 720–910 °C, and the melting range is approximately 1370–1450 °C. When the characteristic point exceeded the melting range during the first passes, it remelted. Later passes produced lower peak temperatures but still contributed to tempering and stress redistribution. I used this understanding to interpret the final microstructure and residual stress.
I also evaluated the effect of the repair on the surrounding base metal. The heat-affected zone is the region that experiences solid-state phase transformation but does not melt. In ZG20CrMoV, the heat-affected zone can become hard and brittle if the cooling rate is too high. Preheating and interpass temperature control reduce the cooling rate and allow tempering. The post-weld heat treatment further tempers the heat-affected zone. For casting defects in large casings, the heat-affected zone may extend several millimeters or more, depending on heat input. I found that the optimized process produced a narrower and more uniform heat-affected zone than the original process.
The following table summarizes the metallurgical risks I associated with the repair of casting defects and the measures I used to control them.
| Metallurgical Risk | Mechanism | Control Measure in My Optimized Process |
|---|---|---|
| Cold cracking | Hydrogen plus martensite plus restraint | Preheat 200 °C, controlled interpass, dry electrode |
| Reheat cracking | Carbide precipitation and stress relaxation in coarse-grained heat-affected zone | Lower heat input, post-weld heat treatment, reduce residual stress |
| Hot cracking | Low-melting grain-boundary films during solidification | Clean groove, appropriate bead shape, controlled heat input |
| Temper embrittlement | Impurity segregation at grain boundaries during service | Material cleanliness, heat treatment control |
| Porosity | Gas entrapment or moisture | Electrode baking, surface cleaning, stable arc |
| Lack of fusion | Insufficient heat or poor access | Adequate current and voltage, careful pass placement |
I also considered the role of restraint. A large turbine casing behaves like a massive rigid frame around the repair cavity. During heating, the repair zone expands, but the surrounding base metal restricts expansion, producing compressive stress. During cooling, the repair zone contracts, but the surrounding base metal restricts contraction, producing tensile stress. The stress state can be approximated by a constrained thermal stress model. If the temperature change is ΔT and the material remains elastic, the thermal stress is approximately:
$$\sigma_{th}=E\alpha\Delta T$$
When the stress exceeds the yield strength, plastic deformation occurs. In multi-pass welding, the stress state is more complex because each pass reheats the previous deposit and changes the yield strength. The final residual stress is the result of many incremental cycles. This is why simulation is so useful for the repair of casting defects in large components. I could not easily measure the internal stress at every point, but I could simulate the accumulation and identify high-risk regions.
I also compared the transient stress distribution layer by layer. In the early layers, the maximum equivalent stress appeared at the completed bead near the repair edge. In later layers, the stress distribution became more stable, but the maximum value shifted toward the bottom and sidewalls of the repair cavity. The stress at the molten pool was low because the material was liquid or near-melting and could not sustain high stress. The stress gradient from the repair zone to the base metal was smooth when the fusion was good, which indicated a sound metallurgical bond. When the fusion was poor, the stress gradient would be sharper and could lead to interfacial cracking.
The optimized repair of casting defects produced a lower stress concentration at the start and stop positions. The reciprocating path helped distribute heat and avoid a continuous high-stress band. The lower current and voltage reduced the heat input, which reduced the peak temperature and the thermal gradient. The slower heating rate before welding reduced the initial thermal shock. The post-weld heat treatment reduced the final residual stress to about 100 MPa. These measures together produced a reliable repair.
I also examined the effect of the repair on dimensional stability. The original process produced an equivalent plastic strain of about 0.24 at the bottom of the repair zone. This level of plastic strain is high enough to cause local yielding and permanent deformation. The optimized process reduced it to about 0.09. Although the optimized process still produced some plastic strain, the level was much lower and more uniform. This reduced the risk of distortion during machining and service. For a turbine casing, dimensional stability is important because the casing must mate with other components and maintain clearances.
I summarized the dimensional and integrity improvements in the following table.
| Integrity Metric | Original Process | Optimized Process | Benefit |
|---|---|---|---|
| Peak equivalent stress | About 310 MPa | About 267–306 MPa | Lower cracking risk |
| Residual stress after heat treatment | Not fully controlled | About 100 MPa | Safe long-term operation |
| Plastic strain | About 0.24 | About 0.09 | Better dimensional stability |
| X-ray result | Porosity and cracks | No observable defects | Validated repair quality |
| Microstructural condition | Strained and potentially brittle | Tempered and relaxed | Improved toughness and creep resistance |
I concluded that the repair of casting defects in large turbine casings requires a coordinated process strategy. Material selection, preheat, interpass temperature, heat input, pass sequence, and post-weld heat treatment must all be considered together. A matching heat-resistant electrode is preferred for ZG20CrMoV because it minimizes long-term mismatch. A preheat of about 200 °C and an interpass temperature of about 200 °C provide a good balance between crack prevention and distortion control. A current of about 180 A and a voltage of about 26 V reduce the heat input while maintaining fusion. Reciprocating multi-layer deposition distributes heat and reduces local stress concentration. Post-weld heat treatment at 700 °C for 6 hours relaxes residual stress and tempers the microstructure. X-ray inspection confirmed that the optimized process can repair casting defects without observable porosity or cracks.
I also noted that the simulation results were consistent with the actual repair trials. The original process produced stress concentrations and plastic strain that were high enough to cause defects, and the actual X-ray inspection found porosity and cracks. The optimized process reduced stress and plastic strain, and the actual X-ray inspection found no defects. This agreement between simulation and experiment increased my confidence in using numerical simulation for process optimization. I believe simulation can reduce the number of trial repairs, shorten development time, and improve the reliability of casting defects repair.
In my future work, I would extend the model in several directions. I would include more detailed phase transformation kinetics to predict hardness and microstructure in the heat-affected zone. I would use a more refined mesh near the fusion boundary to capture local stress gradients. I would also study different excavation geometries because casting defects are irregular and the repair cavity shape affects heat flow and restraint. I would investigate automatic or semi-automatic repair welding for large castings, using sensors to monitor temperature and adjust parameters in real time. I would also explore post-weld heat treatment schedules that minimize energy consumption while still achieving adequate stress relief. These improvements could make the repair of casting defects in large turbine casings faster, more reliable, and more cost-effective.
Overall, my study showed that casting defects in ZG20CrMoV turbine casings can be repaired successfully when the welding process is designed around the thermal and mechanical behavior of the material. The best result came from a matching heat-resistant electrode with controlled preheat, reduced heat input, reciprocating multi-layer deposition, and post-weld heat treatment. The optimized process reduced the maximum equivalent stress, lowered plastic strain, and eliminated observable defects in X-ray inspection. I believe this approach can be applied to other large cast steel components with similar casting defects and similar repair-welding challenges.