In the realm of metal casting, the investment casting process stands out for its ability to produce components with exceptional dimensional accuracy and surface finish. My research is fundamentally driven by the pursuit of further enhancing the quality of castings produced through this sophisticated investment casting process. A critical challenge in any casting operation, including the investment casting process, is the control of solidification to mitigate defects such as shrinkage porosity and hot tears. While techniques like chills are common in sand casting, their application within the ceramic shell of an investment casting is often impractical. An alternative method, which forms the core of this investigation, is the post-pouring application of directed air jets onto the shell surface to locally accelerate cooling. This technique, though sometimes employed empirically, lacks comprehensive quantitative analysis. Therefore, in this study, I systematically explore the air jet cooling method within the investment casting process, aiming to establish quantitative relationships between key operational parameters—specifically nozzle diameter and jet-to-surface distance—and the resulting cooling rate and thermal gradient.
The foundational principle of this cooling technique is rooted in the physics of impinging jet flow. When compressed air is expelled through a nozzle, it forms a free jet stream that subsequently strikes the shell surface. The flow field is classically divided into three regions: the free jet region, the stagnation region, and the wall jet region. The heat extraction occurs most intensively within the stagnation region where the fluid momentum is converted, leading to a thin hydrodynamic and thermal boundary layer and consequently high convective heat transfer coefficients. The characteristics of this flow, and thus the cooling efficacy, are predominantly governed by the nozzle geometry and its positioning relative to the shell. For a circular nozzle, the initial core of the jet, where velocity remains constant, has a length $L_0$ empirically related to the nozzle diameter $D$ by $L_0 \approx 6.2D$. The relative distance $H/D$ (where $H$ is the nozzle-to-surface distance) is therefore a critical dimensionless parameter determining whether the potential core impinges on the surface, which significantly affects the heat transfer distribution. The local Nusselt number ($Nu$), a dimensionless measure of convective heat transfer, for a stationary circular jet impinging on a flat surface can be correlated with Reynolds number ($Re$) and Prandtl number ($Pr$) through relationships of the form:
$$Nu = C Re^m Pr^n f\left(\frac{r}{D}, \frac{H}{D}\right)$$
where $C$, $m$, $n$ are constants, $r$ is the radial distance from the stagnation point, and $f$ is a function describing the radial decay of the heat transfer coefficient. This theoretical framework guides the experimental design to quantify the effects on the investment casting process.
To simulate the thermal conditions of a freshly poured investment casting, I designed and constructed a specialized test platform. The core of this setup was a disk-shaped ceramic shell, representative of a typical shell system in the investment casting process. Its composition, detailed in Table 1, mirrored standard practice with a zircon flour primary layer and a molochite-based backup coat. The shell’s internal cavity was instrumented with an array of K-type thermocouples arranged radially from the intended impingement center, allowing for spatial and temporal temperature mapping.
| Layer | Binder | Refractory Flour | Viscosity (s) |
|---|---|---|---|
| Primary | Silica Sol (GS-30) | Zircon | 40 ± 1 |
| Backup | Silica Sol (GS-30) | Molochite | 35 ± 1 |
The shell was heated uniformly by placing it on a bed of pre-heated zircon sand maintained at 700°C, insulated to minimize lateral heat loss. Once the entire shell reached a stable, uniform initial temperature of 450°C—simulating the thermal state of a shell shortly after metal pouring in the investment casting process—the insulation over the top surface was removed. The air jet system, comprising a compressor, regulator, flow meter, and a nozzle mounted on an adjustable stand, was then activated. The volumetric flow rate was held constant at 6 m³/h for all experiments to isolate the effects of nozzle geometry and distance. The nozzle was aligned perpendicularly to the shell surface, targeting the geometric center (the location of Thermocouple 1). Temperature data from all thermocouples were recorded at a high frequency until the entire shell cooled below 150°C. The experimental matrix, as outlined in Table 2, was devised to methodically investigate the variables.
| Test Group | Nozzle Diameter, $D$ (mm) | Jet-to-Surface Distance, $H$ (mm) | $H/D$ Ratio | Constant Parameters |
|---|---|---|---|---|
| 1 | 4 | 150 | 37.5 | Flow Rate = 6 m³/h Initial Shell Temp = 450°C Shell Geometry = Constant |
| 2 | 4 | 100 | 25.0 | |
| 3 | 4 | 50 | 12.5 | |
| 4 | 4 | 30 | 7.5 | |
| 5 | 4 | 20 | 5.0 | |
| 6 | 4 | 10 | 2.5 | |
| 7 | 4 | 5 | 1.25 | |
| 8 | 6 | 20 | 3.33 | |
| 9 | 8 | 20 | 2.5 | |
| 10 | 10 | 20 | 2.0 |
The primary metrics for analysis were the cooling rate at the impingement center (stagnation point) and the radial temperature gradient developed in the shell. The cooling rate was derived from the inverse of the time required for the center point to cool through a specified temperature interval (e.g., from 400°C to 300°C). The temperature gradient, $\nabla T$, at a given time and radial position $r$ was approximated using the discrete temperature measurements:
$$\nabla T(r) \approx \frac{T(r) – T(r+\Delta r)}{\Delta r}$$
where $\Delta r$ is the distance between adjacent thermocouples. Special focus was placed on the gradient at the stagnation point ($r=0$) at specific temperature milestones.
The analysis of Test Groups 1-7, where nozzle diameter was constant at 4 mm, revealed a pronounced nonlinear relationship between jet distance $H$ and cooling effectiveness. Figure 1 (represented conceptually below via a formula-derived trend) summarizes the cooling time for the 400°C to 300°C interval. The cooling time $t_c$ as a function of $H/D$ can be modeled with a quadratic relation, indicating an optimum:
$$t_c \approx \alpha \left(\frac{H}{D} – \beta\right)^2 + \gamma$$
where $\alpha$, $\beta$, and $\gamma$ are fitting constants. The minimum cooling time, corresponding to the maximum cooling rate, was consistently observed at $H/D \approx 5$ across multiple temperature intervals. This optimal point for the investment casting process aligns with the jet core length theory. When $H/D < 6.2$, the potential core impacts the surface, delivering a region of uniformly high velocity and thus high heat transfer. However, at very small $H/D$ values (e.g., 1.25 or 2.5), the confined space can lead to flow restriction and increased static pressure, potentially causing the wall jet to deflect prematurely or reducing the entrainment of cooler ambient air, thereby diminishing heat extraction efficiency. Conversely, for $H/D >> 6.2$, the core dissipates before impact, and the jet spreads significantly, reducing the centerline velocity and stagnation point heat transfer coefficient, which is proportional to $Re^m$. The velocity $V$ at the nozzle exit for a given flow rate $Q$ is $V = 4Q/(\pi D^2)$, and the Reynolds number is $Re = \rho V D / \mu$. Therefore, the local heat transfer coefficient $h$ scales approximately as:
$$h \propto \frac{k}{D} Re^{0.5} \quad \text{(for turbulent jet impingement)}$$
where $k$ is thermal conductivity of air. As $H$ increases beyond the core length, the effective $Re$ at impact decreases due to momentum decay, leading to lower $h$.
| $H/D$ Ratio | Cooling Time, 400°C→300°C (s) | Normalized Cooling Rate (1/s) | Temperature Gradient at $T_c$=300°C (°C/mm) | Flow Regime at Surface |
|---|---|---|---|---|
| 37.5 | 182 | 0.0055 | 1.8 | Fully developed, low velocity |
| 25.0 | 167 | 0.0060 | 2.1 | Post-core, spreading jet |
| 12.5 | 142 | 0.0070 | 2.9 | Transition region |
| 7.5 | 125 | 0.0080 | 3.5 | Core boundary impact |
| 5.0 | 118 | 0.0085 | 3.1 | Core impingement |
| 2.5 | 130 | 0.0077 | 2.7 | Confined impingement |
| 1.25 | 155 | 0.0065 | 2.3 | Highly confined flow |
The temperature gradient, crucial for directing solidification and feeding in the investment casting process, exhibited a different optimal point. As shown in Table 3, the maximum radial temperature gradient at the stagnation point, measured at the instant the center reached 300°C, occurred at $H/D \approx 7.5$. This can be explained by the radial profile of the heat transfer coefficient. At the optimal cooling rate distance ($H/D=5$), the heat extraction is very intense but also relatively concentrated. At a slightly larger distance ($H/D=7.5$), the jet spreads more before impact, creating a broader stagnation zone with a slightly less peaked but more radially extensive heat transfer profile. This results in a larger temperature difference between the rapidly cooled center and the adjacent zones within a fixed small radial increment, thereby maximizing $\nabla T$. The radial distribution of Nusselt number often follows a correlation like:
$$\frac{Nu}{Nu_0} = 1 – b \left(\frac{r}{D}\right)^c$$
for $r/D$ up to a certain limit, where $Nu_0$ is the stagnation point Nusselt number, and $b$, $c$ are constants that depend on $H/D$. The gradient in heat flux, and consequently temperature, is directly influenced by the derivative of this profile.
Investigating the effect of nozzle diameter while maintaining $H=20$ mm and constant flow rate (Test Groups 5, 8, 9, 10) yielded clear trends. Despite all configurations satisfying $H < L_0$ (i.e., core impingement), increasing the diameter $D$ significantly reduced the cooling rate at the stagnation point. This is a direct consequence of fluid dynamics. With constant volumetric flow $Q$, the exit velocity $V$ is inversely proportional to $D^2$: $V = \frac{4Q}{\pi D^2}$. Since Reynolds number $Re = \frac{\rho V D}{\mu}$, substituting gives $Re = \frac{4\rho Q}{\pi \mu D}$. Therefore, $Re \propto 1/D$. A lower $Re$ signifies less turbulent intensity and momentum, leading to a lower convective heat transfer coefficient $h$, as $h \propto Re^{0.5-0.8}$ for such jets. The cooling time $t_c$ for the same temperature interval increased markedly with $D$, as quantified in Table 4. This inverse relationship is critical for process design in the investment casting process, where selecting an appropriately small nozzle is key for intense local cooling, albeit with consideration for potential flow supply limitations.
| Nozzle Diameter, $D$ (mm) | $H/D$ | Exit Velocity, $V$ (m/s) | Reynolds Number, $Re$ (x10⁴) | Cooling Time, 400°C→300°C (s) | Normalized Cooling Rate (1/s) |
|---|---|---|---|---|---|
| 4 | 5.0 | 132.6 | ~3.5 | 118 | 0.0085 |
| 6 | 3.33 | 58.9 | ~2.3 | 145 | 0.0069 |
| 8 | 2.5 | 33.2 | ~1.7 | 185 | 0.0054 |
| 10 | 2.0 | 21.2 | ~1.3 | 238 | 0.0042 |
$$V = \frac{4Q}{\pi D^2}, \quad Re = \frac{\rho V D}{\mu} = \frac{4\rho Q}{\pi \mu D}$$
$$t_c \propto \frac{1}{h} \propto Re^{-m} \propto D^m \quad \text{with } m \approx 0.5-0.8$$
Contrastingly, the variation in nozzle diameter, under these specific test conditions, did not induce a systematic or significant change in the magnitude of the temperature gradient at the stagnation point. While the absolute temperatures changed at different rates, the spatial shape of the temperature field—dictated by the relative distribution of $h(r)$—remained similar for different $D$ at the same $H/D$ when $H$ was held constant. The primary driver for the gradient was the $H/D$ ratio governing the jet spread, not the absolute velocity scale for this constant $H$ comparison. However, it must be noted that in a broader parameter space, if $H$ were adjusted to maintain a constant $H/D$ ratio while changing $D$, both velocity scale and geometric scale would change, potentially affecting the gradient.
The implications of these findings for the investment casting process are substantial. The ability to quantitatively tailor local cooling allows for precise control over solidification sequences in complex castings. For instance, to eliminate a hot spot, the operator should target a $H/D$ ratio near 5 for the most rapid heat extraction. If the goal is to establish a strong directional temperature gradient to promote progressive solidification towards a feeder, a setting near $H/D=7.5$ might be more effective. Furthermore, for a given air supply capacity, using a smaller nozzle diameter will generate a higher velocity jet and faster cooling, which is advantageous for the investment casting process when addressing isolated heavy sections. However, practical constraints like shell fragility under extreme thermal shock and the potential for jet detachment at very high velocities must also be considered. The heat flux $q”$ imposed on the shell surface can be estimated by:
$$q” = h (T_{shell} – T_{air})$$
where $h$ is derived from the empirical correlations developed through this study. Integrating this boundary condition into a numerical simulation of the investment casting process would significantly improve the predictive accuracy for solidification modeling.
In conclusion, this experimental investigation successfully decouples and quantifies the effects of two pivotal parameters in the air jet cooling technique as applied to the investment casting process. The study demonstrates that the nozzle-to-surface distance, expressed as the $H/D$ ratio, exhibits a dual optima: one for maximizing cooling rate ($H/D \approx 5$) and another for maximizing thermal gradient ($H/D \approx 7.5$). These optima are directly linked to the hydrodynamics of jet impingement, specifically the interaction of the potential core with the surface and the subsequent radial spread of the wall jet. Secondly, under constant flow rate conditions typical of many foundry setups, increasing the nozzle diameter severely diminishes the cooling rate at the stagnation point due to the inverse square relationship with jet velocity and the consequent reduction in Reynolds number, a critical factor for the investment casting process where intense local cooling is often required. However, the nozzle diameter showed a less pronounced effect on the developed temperature gradient for a fixed absolute distance $H$. These results provide a scientific foundation for moving beyond empirical practice, enabling process engineers to strategically select and adjust air jet cooling parameters to achieve desired solidification patterns, thereby enhancing casting integrity and yield in the sophisticated investment casting process. Future work could expand this model to include the effects of jet inclination, multiple jets for larger areas, and integration with real-time thermal monitoring for closed-loop control within the investment casting process.