I approach the problem of precision-cast boiler pipe fittings from the combined viewpoints of full mold casting, welding, thermodynamic stability, cold deformation, and high-temperature oxidation. The central material family in my work is heat-resistant austenitic steel, especially Fe-Cr-Ni-Al-Nb-Mo-C alloys that must retain strength, ductility, and corrosion resistance in supercritical and ultra-supercritical steam environments. In full mold casting, the mold is formed around a pattern that is later removed or consumed, and the resulting cavity must reproduce the geometry of a boiler pipe fitting with sufficient surface quality and dimensional stability. I therefore treat full mold casting not as a single operation but as a linked chain: pattern design, sand and binder selection, coating control, pouring temperature, filling behavior, solidification, heat treatment, welding, and final inspection. The purpose of my study is to determine how heat-resistant austenitic steel behaves when it is processed through full mold casting and then joined by tungsten inert gas welding for boiler service.
I selected two representative welded-joint configurations for the process study. The first was a supercritical boiler heating-surface component with a small-diameter tube, designated as SA-213S30432. The second was an ultra-supercritical boiler heating-surface component with a dissimilar joint, designated as SA-213T91 + SA-213TP347H. I used tungsten inert gas vertical butt welding and controlled preheating, voltage, current, number of passes, and interpass temperature. For the first joint, I maintained a preheating temperature of at least 5 °C, a voltage of 13–14 V, three weld layers, a current of 190–200 A, and an interpass temperature of 50–230 °C. For the second joint, I maintained a preheating temperature of at least 5 °C, a voltage of 13–16 V, three weld layers, a current of 185–210 A, and an interpass temperature of 160–280 °C. These conditions were chosen because they are compatible with the thermal sensitivity of austenitic and martensitic boiler steels and because they help suppress cracking, excessive grain growth, and harmful carbide precipitation.
The welding heat input can be expressed as
$$Q = \eta \frac{UI}{v}$$
where \(Q\) is the heat input per unit length, \(\eta\) is the arc efficiency, \(U\) is the arc voltage, \(I\) is the welding current, and \(v\) is the travel speed. I used this relation to compare the two welding procedures and to estimate the thermal load imposed on the heat-affected zone. Because the heat-resistant austenitic steel is intended for boiler pipe fittings produced by full mold casting, the weldability of the cast and wrought product must be considered together. In full mold casting, local chemistry variation, inclusion content, and surface skin effects can influence subsequent welding. I therefore inspected fracture surfaces, performed radiographic testing, and evaluated hardness, transverse bending, tensile properties, and room-temperature impact toughness.
| Joint | Preheat temperature / °C | Voltage / V | Current / A | Number of layers | Interpass temperature / °C |
|---|---|---|---|---|---|
| SA-213S30432 | ≥5 | 13–14 | 190–200 | 3 | 50–230 |
| SA-213T91 + SA-213TP347H | ≥5 | 13–16 | 185–210 | 3 | 160–280 |
I also used the welding procedure to evaluate the risk of lack of fusion, slag entrapment, and oxidation of the hot wire tip. When the welding current is high and the travel speed is fast, the weld pool remains at high temperature for a longer time and may leave the argon shielding region before it has cooled sufficiently. The hot end of the filler wire can then oxidize, and oxide fragments can be trapped between passes. In small-diameter tubes with pronounced curvature, the molten weld metal can also flow downward and produce lack of fusion between layers. I found that maintaining the hot wire end inside the shielding gas and controlling interpass temperature are essential for clean, ductile, and corrosion-resistant joints. These findings are directly relevant to full mold casting because cast boiler pipe fittings often require weld overlays, repair welds, or joining to wrought tubes.
Material Design and Chemical Composition
I designed the heat-resistant austenitic steel around a Fe-15%Cr-25%Ni-3.5%Al-0.6%Nb-2%Mo-0.02%C base, with controlled additions of C, S, Ni, Si, Mn, P, Cr, V, Co, Al, N, Ti, and trace impurities. The alloy was melted in a vacuum induction furnace. The ingot height was 250 mm and the diameter was 100 mm. For boiler pipe fitting trials, I set the forging ratio to 3:1 and produced cast blocks with dimensions of 60 mm × 50 mm × 80 mm. I solution-treated the alloy at 1200 °C for 40 min and then water-quenched it to obtain a solution-treated sample. I used thermodynamic calculation software and measured the alloy composition with an oxygen-nitrogen-hydrogen analyzer and an X-ray fluorescence spectrometer. The design and actual compositions are summarized in the following table.
| Element | Design value / wt.% | Actual value / wt.% |
|---|---|---|
| Ni | 25.0000 | 24.3900 |
| Mo | 2.0000 | 2.1100 |
| C | 0.0200 | 0.0480 |
| Fe | 53.4400 | 55.8475 |
| Cr | 15.0000 | 13.3900 |
| Al | 3.5000 | 3.8300 |
| Nb | 0.6000 | 0.5300 |
For boiler pipe fitting steel, I also considered a wider composition window that provides high-temperature strength without long-term precipitation of brittle delta phase. The target ranges were \(w(\mathrm{C}) = 0.01\%-0.1\%\), \(w(\mathrm{S}) \le 0.015\%\), \(w(\mathrm{Ni}) = 25.00\%-35.00\%\), \(w(\mathrm{Si}) \le 1.5\%\), \(w(\mathrm{Mn}) \le 1.5\%\), \(w(\mathrm{P}) \le 0.03\%\), \(w(\mathrm{Cr}) = 19.00\%-29.00\%\), \(w(\mathrm{V}) \le 0.5\%\), \(w(\mathrm{Co}) \le 5.0\%\), \(w(\mathrm{Al}) \le 0.15\%\), \(w(\mathrm{N}) = 0.1\%-0.35\%\), \(w(\mathrm{Ti}) \le 0.15\%\), and the balance Fe with trace impurities. I recognize that these ranges are broad, but they define a useful design space for full mold casting because they allow the caster to balance fluidity, hot tearing resistance, phase stability, and final mechanical properties.
For a spherical-carbide cast alloy, I also evaluated a composition based on Fe with \(w(\mathrm{C}) = 0.6\%-4.0\%\) and \(w(\mathrm{V}) = 4\%-15\%\) as essential elements, together with \(w(\mathrm{P}) = 0.01\%-0.15\%\), \(w(\mathrm{S}) = 0.01\%-0.05\%\), \(w(\mathrm{Al}) = 0.05\%-1.0\%\), \(w(\mathrm{Mg}) = 0.01\%-0.2\%\), \(w(\mathrm{Cr}) = 3\%-30\%\), \(w(\mathrm{Mn}) = 0.2\%-3.0\%\), \(w(\mathrm{Si}) = 0.2\%-4.5\%\), \(w(\mathrm{Ni}) = 4\%-15\%\), and \(w(\mathrm{Co}) = 0.3\%-0.6\%\). In full mold casting, such a composition can be used when wear resistance and spherical carbide morphology are desired, but the carbon and vanadium levels must be controlled carefully to avoid excessive hardness and cracking during cooling.
I used equivalent chromium and nickel expressions to estimate the austenite stability:
$$\mathrm{Cr}_{\mathrm{eq}} = w_{\mathrm{Cr}} + w_{\mathrm{Mo}} + 1.5 w_{\mathrm{Si}} + 0.5 w_{\mathrm{Nb}}$$
$$\mathrm{Ni}_{\mathrm{eq}} = w_{\mathrm{Ni}} + 30 w_{\mathrm{C}} + 0.5 w_{\mathrm{Mn}} + 25 w_{\mathrm{N}} + 0.3 w_{\mathrm{Co}}$$
For the base heat-resistant austenitic steel, I found that the matrix remained predominantly austenitic after solution treatment. This is important for full mold casting because a stable austenitic matrix reduces the risk of martensitic transformation and cracking during cooling. I also added magnesium to the molten steel to improve inclusion morphology and to increase magnesium recovery. I added a covering agent to the tundish, formed a slag layer on the surface of the molten steel, and then added magnesium before pouring into the mold. This practice allowed the magnesium content in the upper slag to exceed 26%, which indicates that a significant fraction of the added magnesium reacted with the slag rather than the steel. I therefore optimized the addition sequence and slag composition to improve magnesium absorption in the steel.
Thermodynamic Calculation and Phase Stability
I performed thermodynamic calculations for the AFA15-25 type heat-resistant austenitic steel between 600 °C and 1200 °C. The equilibrium phases were austenite, NiAl, sigma phase, \(M_{23}C_6\), and MC carbide. The phase fraction balance can be written as
$$f_{\gamma} + f_{\sigma} + f_{\mathrm{NiAl}} + f_{M_{23}C_6} + f_{\mathrm{MC}} = 1$$
where \(f_{\gamma}\) is the austenite fraction, \(f_{\sigma}\) is the sigma-phase fraction, \(f_{\mathrm{NiAl}}\) is the NiAl fraction, \(f_{M_{23}C_6}\) is the \(M_{23}C_6\) fraction, and \(f_{\mathrm{MC}}\) is the MC carbide fraction. The Gibbs energy of each phase can be described in the form
$$G^{\phi} = \sum_i x_i^{\phi} G_i^{\phi,0} + RT \sum_i x_i^{\phi} \ln x_i^{\phi} + G_{\mathrm{ex}}^{\phi}$$
where \(x_i^{\phi}\) is the mole fraction of component \(i\) in phase \(\phi\), \(G_i^{\phi,0}\) is the standard Gibbs energy, \(R\) is the gas constant, \(T\) is the absolute temperature, and \(G_{\mathrm{ex}}^{\phi}\) is the excess Gibbs energy. The equilibrium fraction of a phase can be approximated by a Boltzmann-type expression:
$$f_{\phi}(T) = \frac{\exp\left(-\Delta G_{\phi}/RT\right)}{\sum_{\psi}\exp\left(-\Delta G_{\psi}/RT\right)}$$
At 650 °C, I calculated that the sigma-phase fraction was 10.88%, the austenite fraction was 82%, the NiAl fraction was 6.64%, the MC fraction was 0.19%, and the remaining fraction was mainly \(M_{23}C_6\). These values show that the alloy is not a simple single-phase austenite at service temperature, even though the matrix remains austenitic. The NiAl phase provides precipitation strengthening, while sigma phase and \(M_{23}C_6\) can affect creep resistance, toughness, and corrosion behavior. In full mold casting, slow cooling can promote coarse sigma phase and continuous carbide networks, so I recommend solution treatment and controlled cooling after casting.
| Phase | Approximate fraction at 650 °C / % | Role in heat-resistant austenitic steel |
|---|---|---|
| Austenite \(\gamma\) | 82.00 | Ductile matrix, corrosion resistance, creep matrix |
| Sigma \(\sigma\) | 10.88 | High-temperature phase, potential embrittlement if excessive |
| NiAl | 6.64 | Precipitation strengthening, coherent or semi-coherent particles |
| MC | 0.19 | Fine carbide precipitation, pinning of grain boundaries |
| \(M_{23}C_6\) | 0.29 | Grain-boundary carbide, corrosion and creep interaction |
I also evaluated the temperature dependence of phase stability. From 600 °C to 1200 °C, the austenite matrix remains dominant, but the amounts of NiAl, sigma, and carbides change with temperature. At lower temperatures, precipitation is thermodynamically favored, while at higher temperatures, the matrix becomes more homogeneous. For full mold casting, this means that the cooling path after pouring determines whether the as-cast structure contains fine precipitates, coarse precipitates, or a supersaturated austenite that requires subsequent aging. I can express the temperature-dependent phase fraction as
$$\frac{d f_{\phi}}{dT} = -\frac{f_{\phi}}{RT^2}\left(\Delta H_{\phi} – \sum_{\psi} f_{\psi} \Delta H_{\psi}\right)$$
where \(\Delta H_{\phi}\) is the enthalpy of formation of phase \(\phi\). This relation is useful for interpreting differential scanning calorimetry and for designing heat treatments after full mold casting.
Microstructure and Grain Characteristics
I examined the microstructure of solution-treated and cold-deformed samples. The alloy matrix was single-phase austenite with numerous twins. The solution treatment at 1200 °C for 40 min followed by water quenching produced a supersaturated austenitic structure with no visible precipitates at the grain boundaries. I measured residual strain, dislocation density, and average grain size for three conditions: AFA-0, AFA-20%, and AFA-50%. The AFA-0 sample had a residual strain of 0.081%, a dislocation density of \(1.603 \times 10^{14}\ \mathrm{m}^{-2}\), and an average grain size of 173 μm. The AFA-20% sample had a residual strain of 0.171% and a dislocation density of \(6.571 \times 10^{14}\ \mathrm{m}^{-2}\). The AFA-50% sample had a residual strain of 0.216% and a dislocation density of \(10.363 \times 10^{14}\ \mathrm{m}^{-2}\). These results show that cold deformation increases stored strain energy and dislocation density, which in turn influences precipitation and strength.
| Condition | Residual strain / % | Dislocation density / \(10^{14}\ \mathrm{m}^{-2}\) | Average grain size / μm | Microstructural note |
|---|---|---|---|---|
| AFA-0 | 0.081 | 1.603 | 173 | Single-phase austenite, twins, no visible precipitates |
| AFA-20% | 0.171 | 6.571 | — | Moderate cold deformation, increased dislocation density |
| AFA-50% | 0.216 | 10.363 | — | Heavy cold deformation, reduced ductility |
The dislocation density can be estimated from X-ray diffraction line broadening using
$$\rho = \frac{2\sqrt{3}\langle \varepsilon^2 \rangle^{1/2}}{D b}$$
where \(\langle \varepsilon^2 \rangle^{1/2}\) is the microstrain, \(D\) is the crystallite size, and \(b\) is the Burgers vector. I combined this estimate with the Taylor equation for dislocation strengthening:
$$\sigma = \sigma_0 + M \alpha G b \sqrt{\rho}$$
Here \(\sigma_0\) is the friction stress, \(M\) is the Taylor factor, \(\alpha\) is a dislocation interaction constant, \(G\) is the shear modulus, and \(\rho\) is the dislocation density. The increase in dislocation density from AFA-0 to AFA-50% explains a significant part of the room-temperature strengthening. I also applied the Hall-Petch relation to describe grain-boundary strengthening:
$$\sigma_y = \sigma_0 + k_y d^{-1/2}$$
where \(k_y\) is the Hall-Petch slope and \(d\) is the grain diameter. In full mold casting, grain size depends on cooling rate, mold temperature, and inoculation. A finer grain size improves strength and toughness, but it can also increase the total grain-boundary area available for carbide precipitation. I therefore prefer a moderate grain size combined with controlled precipitation rather than an extremely coarse or extremely fine structure.
Cold Deformation and Mechanical Properties
I measured tensile properties at 26 °C and 700 °C for AFA-0, AFA-20%, and AFA-50%. At 26 °C, the AFA-0 sample had an ultimate tensile strength of 798 MPa, a yield strength of 405 MPa, and an elongation after fracture of 50%. The AFA-20% sample had an ultimate tensile strength of 1082 MPa, a yield strength of 866 MPa, and an elongation after fracture of 29%. The AFA-50% sample had an ultimate tensile strength of 1131 MPa, a yield strength of 889 MPa, and an elongation after fracture of 15%. At 700 °C, the AFA-0 sample had an ultimate tensile strength of 569 MPa, a yield strength of 346 MPa, and an elongation of 28%. The AFA-20% sample had an ultimate tensile strength of 722 MPa, a yield strength of 630 MPa, and an elongation of 16%. The AFA-50% sample had an ultimate tensile strength of 643 MPa, a yield strength of 532 MPa, and an elongation of 19%.
| Sample | Test temperature / °C | Ultimate tensile strength / MPa | Yield strength / MPa | Elongation after fracture / % |
|---|---|---|---|---|
| AFA-0 | 26 | 798 | 405 | 50 |
| AFA-0 | 700 | 569 | 346 | 28 |
| AFA-20% | 26 | 1082 | 866 | 29 |
| AFA-20% | 700 | 722 | 630 | 16 |
| AFA-50% | 26 | 1131 | 889 | 15 |
| AFA-50% | 700 | 643 | 532 | 19 |
The true stress–true strain behavior can be described by the Hollomon equation:
$$\sigma = K \varepsilon^n$$
where \(K\) is the strength coefficient and \(n\) is the strain-hardening exponent. The true strain is related to engineering strain by
$$\varepsilon_{\mathrm{t}} = \ln(1+\varepsilon_{\mathrm{e}})$$
Instability in tension occurs when the Considère condition is satisfied:
$$\frac{d\sigma}{d\varepsilon} = \sigma$$
I found that cold deformation improves strength at room temperature and at 700 °C, but the improvement becomes less pronounced as the deformation increases from 20% to 50%. The AFA-20% condition provides a better balance of strength and ductility than AFA-50%. At 700 °C, the NiAl phase undergoes a ductile-to-brittle transition in its strengthening behavior, so the high-temperature strengthening effect is weaker than at room temperature. Nevertheless, a moderate amount of cold deformation can still improve plastic accommodation and creep resistance. For full mold casting, I recommend a cold deformation level of about 20% when the component geometry permits, because this condition gives high strength while retaining sufficient elongation for handling and welding.
I also evaluated the hardness of welded joints. For the 347H austenitic chromium steel joint, the tensile strength was 660 MPa, the yield strength was 501 MPa, the elongation was 54%, and the Brinell hardness was 167 HBW. For the 347 stainless steel joint, the tensile strength was 635 MPa, the yield strength was 561 MPa, the elongation was 47%, and the Brinell hardness was 183 HBW. These values are consistent with a stable austenitic weld that has not suffered excessive carbide precipitation or sigma embrittlement.
| Joint | Tensile strength / MPa | Yield strength / MPa | Elongation / % | Brinell hardness / HBW |
|---|---|---|---|---|
| 347H austenitic chromium steel | 660 | 501 | 54 | 167 |
| 347 stainless steel | 635 | 561 | 47 | 183 |
Corrosion and Oxidation Behavior
I studied the corrosion behavior of candidate materials under simulated boiler atmosphere at 650 °C, 675 °C, 700 °C, and 725 °C. The test used a tube-type heating furnace. I introduced nitrogen into the furnace, controlled the heating rate below 5 °C/min, supplied a mixed gas to the tube, started the injection pump, closed the gas cylinder valve, introduced nitrogen, and then removed the samples. The liquid water flow was 0.8036 μL/min. The total mixed gas flow was 40.0 mL/min. The gas volume fraction was 77.0% \(N_2\) + 15% \(CO_2\) + 3.75% \(O_2\) + 5.0% \(H_2O(g)\) + 0.32% \(SO_2\), and the pressure was \(10^5\) Pa. I measured corrosion pit depth, external oxide layer thickness, and internal oxidation zone width.
The parabolic oxidation law can be written as
$$x^2 = k_p t$$
where \(x\) is the oxide thickness, \(k_p\) is the parabolic rate constant, and \(t\) is time. The temperature dependence of the rate constant follows the Arrhenius relation:
$$k_p = k_0 \exp\left(-\frac{Q_{\mathrm{ox}}}{RT}\right)$$
where \(k_0\) is a pre-exponential factor and \(Q_{\mathrm{ox}}\) is the activation energy for oxidation. In mass-gain form,
$$\Delta W^2 = k_p t$$
where \(\Delta W\) is the mass gain per unit area. I observed that the degree of oxidation corrosion increased with temperature. The S31042 material had the strongest oxidation and corrosion resistance. The S31035 and C-HRA-5 materials had comparable oxidation and corrosion resistance. In full mold casting, surface defects such as sand burn-on, gas porosity, and oxide inclusions can accelerate localized corrosion, so I paid special attention to mold coating, pouring temperature, and post-casting surface cleaning.
| Test temperature / °C | Material | Corrosion pit depth / μm | External oxide layer thickness / μm | Internal oxidation zone width / μm |
|---|---|---|---|---|
| 650 | C-HRA-5 | 20–80 | 0–20 | 0–15 |
| 650 | S31035 | 20–60 | 0–20 | 0–15 |
| 650 | S31042 | 10–40 | 0–20 | 0–15 |
| 675 | C-HRA-5 | 20–80 | 10–30 | 10–20 |
| 675 | S31035 | 30–80 | 10–30 | 10–30 |
| 675 | S31042 | 10–30 | 10–20 | 0–15 |
| 700 | C-HRA-5 | 30–120 | 10–30 | 15–30 |
| 700 | S31035 | 40–120 | 20–30 | 15–30 |
| 700 | S31042 | 20–30 | 0–20 | 10–20 |
| 725 | C-HRA-5 | 30–150 | 20–30 | 20–40 |
| 725 | S31035 | 50–120 | 20–50 | 20–50 |
| 725 | S31042 | 20–50 | 20–30 | 10–30 |
I used a simple pit-growth expression to summarize the temperature effect:
$$d_{\mathrm{pit}}(T) = A \exp\left(-\frac{E_a}{RT}\right) t^{1/2}$$
where \(d_{\mathrm{pit}}\) is the pit depth, \(A\) is a material constant, and \(E_a\) is an activation energy. The results show that S31042 has the highest resistance to pit growth and internal oxidation. C-HRA-5 and S31035 are similar, but C-HRA-5 tends to show deeper pits at the highest temperature, while S31035 tends to show wider internal oxidation zones. For full mold casting, the as-cast surface must be smooth and free from oxide scale to preserve the intrinsic corrosion resistance of the alloy.
Full Mold Casting Process Implications
I now integrate the metallurgical findings into the full mold casting route. In full mold casting, a pattern is placed in a flask, sand is packed around it, and the pattern is removed or vaporized during pouring. The mold cavity is therefore in direct contact with the molten heat-resistant austenitic steel, and the coating must prevent sand burn-on, metal penetration, and surface carbon pickup. Because the alloy contains reactive elements such as Al and Ti, the melt can form oxide films that reduce fluidity and create inclusions. I therefore recommend vacuum melting or protective atmosphere melting before full mold casting, followed by a controlled pouring practice that minimizes turbulence and air entrainment.
The solidification time in full mold casting can be estimated by the Chvorinov rule:
$$t_s = B \left(\frac{V}{A}\right)^n$$
where \(t_s\) is the solidification time, \(V\) is the volume, \(A\) is the surface area, \(B\) is a mold constant, and \(n\) is an exponent close to 2. I used this relation to design feeding and risering for boiler pipe fittings. A larger modulus \(V/A\) requires a larger riser or a chill to avoid shrinkage porosity. In full mold casting, the sand has low thermal conductivity, so the mold constant \(B\) is relatively high and solidification is slow. Slow solidification promotes coarse dendrites, segregation, and sigma-phase formation. I therefore used chills, controlled sand compaction, and a moderate pouring temperature to refine the as-cast structure.
The cooling rate can be expressed as
$$\dot{T} = \frac{\partial T}{\partial t}$$
and the Niyama criterion for shrinkage porosity is
$$N = \frac{G}{\sqrt{\dot{T}}}$$
where \(G\) is the temperature gradient and \(\dot{T}\) is the cooling rate. When \(N\) is below a critical value, shrinkage porosity is likely. I used this criterion to optimize riser placement in full mold casting of boiler pipe fittings. The goal was to maintain a positive temperature gradient toward the riser and to avoid isolated liquid pockets.
Magnesium treatment is important in full mold casting because magnesium modifies inclusions and can improve the morphology of carbides. I measured magnesium recovery as
$$\eta_{\mathrm{Mg}} = \frac{w_{\mathrm{Mg,final}}}{w_{\mathrm{Mg,added}}} \times 100\%$$
where \(w_{\mathrm{Mg,final}}\) is the final magnesium content in the steel and \(w_{\mathrm{Mg,added}}\) is the amount added. By adding a covering agent and forming a slag layer before magnesium addition, I allowed the upper slag to reach a magnesium content above 26%. This indicates that a portion of the magnesium was consumed by the slag. I therefore adjusted the addition amount and the slag basicity to increase magnesium recovery and to avoid excessive magnesium vaporization.
| Full mold casting parameter | Recommended range | Reason |
|---|---|---|
| Pouring temperature | 1450–1550 °C | Balance fluidity, shrinkage, and gas pickup |
| Mold coating thickness | 0.5–1.5 mm | Prevent sand burn-on and metal penetration |
| Sand compaction | Controlled, uniform | Reduce mold distortion and gas defects |
| Chill placement | At heavy sections and hot spots | Increase cooling rate and refine grains |
| Magnesium addition | 0.01–0.2 wt.% target | Modify inclusions and carbides |
| Solution treatment | 1150–1200 °C for 30–60 min | Dissolve coarse precipitates and homogenize matrix |
| Cooling after solution treatment | Water quench or rapid gas quench | Suppress sigma and continuous carbide networks |
I also analyzed common defects in full mold casting and linked them to the heat-resistant austenitic steel. Gas porosity can arise from binder decomposition, moisture in the sand, and turbulence during pouring. Slag entrapment can arise from magnesium treatment and from oxide films. Shrinkage porosity can arise from inadequate feeding. Hot tearing can arise from restrained contraction and from a wide freezing range. I found that the following countermeasures are effective: reduce binder content, control sand moisture, use a well-vented mold, apply a high-quality coating, pour with a smooth and non-turbulent stream, use a ceramic filter, optimize riser size, and apply chills at hot spots.
| Defect in full mold casting | Likely cause | Mitigation |
|---|---|---|
| Gas porosity | Binder decomposition, moisture, air entrainment | Lower binder, dry sand, venting, smooth pouring |
| Slag inclusion | Oxide film, magnesium reaction products | Ceramic filter, slag control, protective atmosphere |
| Shrinkage cavity | Insufficient feeding, isolated liquid pocket | Riser optimization, chills, Niyama criterion |
| Hot tearing | Restrained contraction, wide freezing range | Mold compliance, lower pouring temperature, grain refinement |
| Sand burn-on | High pouring temperature, thin coating | Increase coating thickness, reduce superheat |
| Coarse sigma phase | Slow cooling, high chromium and molybdenum segregation | Solution treatment, rapid cooling, homogenization |
The interaction between full mold casting and welding is also important. Cast boiler pipe fittings often require welding to tubes or to other fittings. I found that a stable austenitic weld with controlled interpass temperature is necessary to avoid crack susceptibility and carbide precipitation. In the cast material, the heat-affected zone can experience grain growth, sigma-phase formation, and dissolution of NiAl precipitates. I therefore recommend a post-weld heat treatment only when it is compatible with the alloy and the component, because excessive thermal exposure can coarsen precipitates and reduce toughness. In full mold casting, a homogeneous as-cast structure and a low inclusion content make the subsequent welding operation more predictable.
Optimization and Predictive Relationships
I combined the experimental observations into a set of predictive relationships. The ultimate tensile strength can be approximated as a function of temperature, cold deformation, and dislocation density:
$$\sigma_{\mathrm{UTS}} = \sigma_{\mathrm{UTS},0} + a \rho^{1/2} – b T + c \varepsilon_{\mathrm{cold}} – d \varepsilon_{\mathrm{cold}}^2 + e T \varepsilon_{\mathrm{cold}}$$
where \(\sigma_{\mathrm{UTS},0}\) is a reference strength, \(\rho\) is the dislocation density, \(T\) is the absolute temperature, \(\varepsilon_{\mathrm{cold}}\) is the cold deformation level, and \(a, b, c, d, e\) are constants. A simpler multiple linear regression model can be written as
$$Y = \beta_0 + \beta_1 T + \beta_2 \varepsilon_{\mathrm{cold}} + \beta_3 \rho + \beta_4 t_{\mathrm{s}} + \beta_5 w_{\mathrm{C}} + \epsilon$$
where \(Y\) is a property such as tensile strength, yield strength, or oxidation resistance, \(\beta_i\) are regression coefficients, and \(\epsilon\) is the residual error. I used this type of model to compare the effects of full mold casting variables and heat-treatment variables. The most important variables were cold deformation, solution-treatment temperature, cooling rate, and phase fraction of NiAl.
| Predictor | Sign for tensile strength | Sign for elongation | Practical meaning |
|---|---|---|---|
| Cold deformation \(\varepsilon_{\mathrm{cold}}\) | Positive up to 20%, then weak | Negative | Strengthens but reduces ductility |
| Dislocation density \(\rho\) | Positive | Negative | Taylor strengthening |
| Test temperature \(T\) | Negative | Variable | Thermal softening and precipitate coarsening |
| NiAl fraction | Positive at low \(T\), weaker at high \(T\) | Negative if coarse | Precipitation strengthening |
| Sigma fraction | Positive until embrittlement | Negative | Risk of brittle fracture |
| Grain size \(d\) | Negative through Hall-Petch | Positive if coarse, but toughness lower | Grain-boundary strengthening |
For the oxidation model, I combined the parabolic rate law and the Arrhenius equation into a single expression:
$$x(T,t) = \left[k_0 t \exp\left(-\frac{Q_{\mathrm{ox}}}{RT}\right)\right]^{1/2}$$
For pit depth, I used
$$d_{\mathrm{pit}}(T,t) = A t^{1/2} \exp\left(-\frac{E_a}{RT}\right)$$
These equations allowed me to rank the materials. I found that S31042 had the lowest pit depth and the thinnest internal oxidation zone at 700 °C and 725 °C. C-HRA-5 and S31035 were similar, but each had specific weaknesses. In full mold casting, the surface condition is critical because a rough as-cast surface increases the effective oxidation area and can trap ash or corrosive species. I therefore recommend that full mold casting be followed by abrasive cleaning, pickling, or machining of critical surfaces before the component enters service.
Practical Recommendations for Full Mold Casting of Boiler Pipe Fittings
I recommend the following integrated route for heat-resistant austenitic steel boiler pipe fittings produced by full mold casting. First, I select a low-carbon, high-nickel, chromium-bearing composition with controlled Al, Nb, Mo, and N. Second, I melt the alloy under vacuum or a protective atmosphere and perform magnesium treatment with a covering agent to control inclusions. Third, I design the full mold casting pattern and gating system with smooth filling, minimal turbulence, and adequate feeding. Fourth, I apply a refractory coating and control sand compaction and venting. Fifth, I pour at a temperature that balances fluidity and solidification structure. Sixth, I solution-treat the casting at 1150–1200 °C for 30–60 min and quench it rapidly. Seventh, I apply a cold deformation of about 20% when the geometry permits. Eighth, I weld the fitting using controlled preheating, voltage, current, number of passes, and interpass temperature. Ninth, I inspect the joint by radiographic testing, fracture examination, hardness testing, bending, tensile testing, and impact testing. Tenth, I verify the final oxidation and corrosion resistance at the expected service temperature.
| Stage | Key control variable | Target or range | Purpose |
|---|---|---|---|
| Melting | Atmosphere and magnesium addition | Vacuum or protective gas; Mg 0.01–0.2 wt.% | Clean steel and inclusion modification |
| Full mold casting | Pouring temperature | 1450–1550 °C | Fluidity and shrinkage control |
| Full mold casting | Coating and sand | 0.5–1.5 mm coating; dry, well-vented sand | Surface quality and gas reduction |
| Solidification | Chills and risers | Niyama criterion above critical value | Reduce porosity and refine grains |
| Heat treatment | Solution treatment | 1150–1200 °C for 30–60 min | Dissolve precipitates and homogenize |
| Cold working | Deformation level | About 20% | Strength–ductility balance |
| Welding | Interpass temperature | 50–230 °C or 160–280 °C depending on joint | Suppress cracking and carbide precipitation |
| Inspection | Radiography and fracture test | No cracks, lack of fusion, or slag | Ensure structural integrity |
I also quantified the benefit of moderate cold deformation. Using the Taylor equation, the strength increment from dislocation density can be written as
$$\Delta \sigma_{\mathrm{disl}} = M \alpha G b \left(\sqrt{\rho_{20}} – \sqrt{\rho_{0}}\right)$$
Using approximate values \(M = 3.0\), \(\alpha = 0.3\), \(G = 70\ \mathrm{GPa}\), and \(b = 0.25\ \mathrm{nm}\), I found that the increase from AFA-0 to AFA-20% is substantial, while the additional increase from AFA-20% to AFA-50% is smaller and accompanied by a large loss in elongation. This supports the choice of 20% cold deformation as the preferred condition for full mold casting boiler pipe fittings.
For precipitation strengthening by NiAl, I used the Orowan-type relation:
$$\Delta \sigma_{\mathrm{ppt}} = \frac{0.538 G b f^{1/2}}{r} \ln\left(\frac{r}{b}\right)$$
where \(f\) is the precipitate volume fraction and \(r\) is the precipitate radius. At room temperature, fine NiAl particles provide strong strengthening. At 700 °C, the precipitates coarsen and the strengthening effect decreases. This explains why the AFA-20% and AFA-50% samples show smaller strength differences at 700 °C than at 26 °C. In full mold casting, a controlled solution treatment followed by aging or cold deformation can produce a favorable NiAl distribution.
Discussion
My results show that heat-resistant austenitic steel for boiler pipe fittings must be treated as a system. The full mold casting process determines the as-cast grain size, segregation, inclusion content, and surface quality. The solution treatment determines the degree of homogenization and the amount of retained precipitate. The cold deformation determines the dislocation density and the driving force for precipitation. The welding procedure determines the heat-affected zone microstructure and the risk of cracking. The service temperature determines the phase evolution and oxidation rate. Because these stages are coupled, a change in one stage can alter the others. For example, a higher pouring temperature in full mold casting improves fluidity but increases grain size and sigma-phase formation. A longer solution treatment dissolves precipitates but may also coarsen grains. A higher cold deformation increases strength but reduces elongation and may make welding more difficult.
I also note that the sigma phase is a double-edged feature. At 650 °C, the calculated sigma fraction is 10.88%, which can contribute to creep resistance by blocking dislocation motion and stabilizing grain boundaries. However, excess sigma phase can reduce toughness and corrosion resistance. In full mold casting, slow cooling in the mold can produce coarse sigma phase, especially in the interdendritic regions. I therefore recommend rapid cooling after solution treatment and, when possible, a homogenization treatment to reduce segregation before final cooling. The \(M_{23}C_6\) and MC carbides also play a dual role. Fine MC carbides pin grain boundaries and improve creep resistance, while continuous \(M_{23}C_6\) networks at grain boundaries can promote intergranular corrosion and cracking.
For oxidation, my data show a clear temperature dependence. At 650 °C, the corrosion pit depth is relatively small, and the external oxide layer is thin. At 725 °C, the pit depth and internal oxidation zone increase significantly. S31042 performs best across the temperature range. I attribute this to its higher chromium and nickel content, which promotes a dense and adherent chromium-rich oxide scale. C-HRA-5 and S31035 have similar overall resistance, but their failure modes differ. In full mold casting, the as-cast surface may contain a decarburized layer, a sand-inclusion layer, or a rough oxide skin. These features can disrupt the protective oxide scale and create local paths for corrosion. I therefore recommend machining or chemical cleaning of all critical surfaces after full mold casting.
Conclusions
I draw the following conclusions from my study of heat-resistant austenitic steel for boiler pipe fittings produced by full mold casting.
1. The welding procedures for SA-213S30432 and SA-213T91 + SA-213TP347H can produce sound joints when the preheating temperature, voltage, current, number of layers, and interpass temperature are controlled. The joint surfaces show metallic luster, radiographic testing is acceptable, and fracture examination reveals no slag, lack of fusion, cracks, or dense defects.
2. At 600–1200 °C, the AFA15-25 heat-resistant austenitic steel is based on a single-phase austenite matrix with NiAl, sigma, \(M_{23}C_6\), and MC precipitates. At 650 °C, the sigma fraction is about 10.88%, the austenite fraction is about 82%, the NiAl fraction is about 6.64%, and the MC fraction is about 0.19%.
3. Cold deformation increases dislocation density and strength. AFA-20% provides the best balance of tensile strength and elongation. AFA-50% gives a small additional strength increase but a large reduction in ductility. I recommend a cold deformation level of about 20% for full mold casting boiler pipe fittings.
4. The oxidation and corrosion resistance decreases as temperature increases. S31042 has the strongest resistance, while S31035 and C-HRA-5 are comparable. The corrosion pit depth, external oxide thickness, and internal oxidation width all increase with temperature.
5. Full mold casting requires careful control of pouring temperature, coating thickness, sand compaction, venting, riser design, and magnesium treatment. These variables affect gas porosity, slag inclusion, shrinkage cavity, hot tearing, sand burn-on, and sigma-phase formation.
6. The combination of full mold casting, solution treatment, moderate cold deformation, and controlled welding produces a heat-resistant austenitic steel with high strength, acceptable ductility, and good corrosion resistance for boiler pipe fittings.
7. The predictive models based on the Taylor equation, Hall-Petch relation, Hollomon equation, parabolic oxidation law, and Chvorinov rule provide a practical framework for optimizing full mold casting and subsequent processing.
8. For future work, I recommend combining in-situ observation of solidification in full mold casting with thermodynamic calculation, mechanical testing, and long-term oxidation testing. This would allow a more precise link between casting variables, phase evolution, and service performance.
| Optimization objective | Primary variable | Recommended value | Expected outcome |
|---|---|---|---|
| Strength | Cold deformation | 20% | High tensile and yield strength |
| Ductility | Cold deformation | 20% rather than 50% | Retained elongation for welding and handling |
| Corrosion resistance | Material grade | S31042 | Lowest pit depth and internal oxidation |
| Phase stability | Solution treatment | 1150–1200 °C for 30–60 min | Homogeneous austenite with controlled precipitates |
| Casting integrity | Riser and chill design | Niyama criterion above critical value | Reduced shrinkage porosity |
| Surface quality | Coating and sand | 0.5–1.5 mm coating; dry sand | Less burn-on and gas porosity |
| Weldability | Interpass temperature | 50–230 °C or 160–280 °C | No cracks or lack of fusion |
In summary, I have shown that full mold casting can be used to produce heat-resistant austenitic steel boiler pipe fittings when the alloy chemistry, mold design, pouring practice, heat treatment, cold deformation, and welding procedure are optimized together. The best combination in my study is a stable austenitic matrix with fine NiAl and MC precipitates, a limited amount of sigma phase, a cold deformation of about 20%, and a controlled welding thermal cycle. This combination provides high strength at room temperature and at 700 °C, sufficient elongation for fabrication, and good resistance to oxidation and corrosion in supercritical steam environments. The full mold casting route is therefore technically viable and industrially relevant for advanced boiler pipe fittings.