Abstract
The numerical simulation is similar to the experimental research, which uses the computer. 3D printing is an experimental study, which is compared with simulation. 3D printing and precision casting of impellers are frequently used in industrial scenarios. To address the problem of low casting yield of aluminium alloy impellers in gypsum-type casting, three-dimensional modelling by magics, casting simulation, and parameter optimization method by applying ProCAST software are used to improve the casting yield of impellers. The results show that the optimized structure of the casting system is more reasonable, so that the shrinkage and shrinkage defects are concentrated in the location of the riser; the maximum stress value of the casting and the casing during the solidification stage is 2.71 MPa, which is much smaller than the stress limit of the casing; the maximum horizontal displacement of the casting along the X and Y axes after solidification is completed at 0.025 mm and 0.022 mm, respectively, and the maximum displacement along the Z axis due to the presence of the riser is 0.009 mm, which greatly improves casting accuracy and molding quality.
1. Introduction
As a typical part of mechanical equipment, impeller is widely used in automotive and other fields. Because of its structural complexity, traditional machining production time is long and the cost is high. At present, a company has produced a batch of impeller castings by gypsum casting, which are made of ZAlSi7MgA and code named ZL101A aluminum alloy. In order to verify the casting rationality of the actual production of the existing impeller model, the original gating system was simulated by ProCAST to judge the casting defects, and the casting qualification rate was improved by optimizing the structure and process parameters.
Saravana Kumar et al, based on the vortex height of the fluid and the settling time of the reinforced particles, optimized the main process parameters such as the blade geometry, mixing speed, and impeller position of the stirring casting [1]. TorabiParizi et al. have attempted to evaluate the effect of stir and rheo-casting techniques on microstructure and mechanical properties of AZ80-0.5wt%Ca/1.5 vol%Al2O3 nanocomposite; type and amount of the eutectic phase and particles distribution were influenced remarkably by casting techniques [2]. Das et al. used micron-sized activated carbon powder to manufacture titanium carbide-reinforced MMC, reacted with titanium materials through in-situ technology and generated TiC intermetallic compound particles in al-4.5% Cu matrix, thus generating al-4.5-Cu-5% TiC MMC, aiming to find the best processing parameters [3]. Xu et al. introduced γ-The phase composition and precipitation of TiAl alloy are introduced γ Texture evolution of TiAl alloy, summarized the influence γ-Factors of mechanical properties of TiAl alloy in compression, tension, creep, fatigue and tribological tests [4]. Shabani and Mazahery proposed a hybrid genetic algorithm/particle swarm optimization algorithm to effectively estimate the optimal process conditions for preparing nanocomposites by casting [5]. Rana et al. tested the influence of each parameter on the responsiveness and sufficiency of the hardness model by variance analysis and Fisher f test, and the mathematical model affecting the hardness of composite materials was established [6]. Sridhar et al., through this numerical investigation, estimated the temperature distribution profiles by finite element program (ANSYS) for different composition of composites [7]. Barot and Ayar obtained the cooling curves of plates with different geometric shapes by simulating the process of gravity casting with permanent mold and placing virtual thermocouples at different positions [8]. Chandra Kandpal et al. effectively reduced the sand casting defects by adopting correct methods in the sand casting process to reduce the occurrence of casting defects [9]. Singh et al. discussed the influence of particle embedment and the main factors of mechanical stirring during the processing of aluminum matrix composites, which improved the mechanical properties of pure aluminum alloy [10]. Singh et al. emphatically introduced the research achievements and the latest progress of several researchers in the field of stirring casting technology, and the stirring casting technology of composite materials that needs to be improved in the future is proposed [11]. Kandpal et al. studied the relationship between microstructure and related mechanical properties of sand-casting samples. With the increase of binder content, the hardness of castings was improved [12]. Mckay and Anbalagan, through two finite element simulation methods of static structure and explicit dynamics, analyzed the impeller with different materials, and the impeller failure is successfully identified and predicted [13]. Wallace et al. described the technology used to produce the semisolid impellers in detail, and the influence of modified processing parameters on the tendency of impeller surface cracks is given [14]. Adithiyaa et al. prepared aluminum hybrid metal matrix composites with nano reinforced particles, and mechanical properties were tested to evaluate the maximum deformation, hardness, and impact energy [15]. Lei et al. simulated the mold filling and solidification process by using the magmatic casting simulation software and found and analyzed the root causes of defects [16]. Nimbulkar and Dalu studied the design of the existing pouring and feeding system and optimized the pouring and feeding system with Auto CAST X1 casting simulation software to reduce the scrap rate [17]. Nyemba et al. optimized using computational fluid dynamics and sizing of the gating system, and waste is expected to be reduced from 37% to 24% [18]. Xue et al. proposed an optimized salt metal reaction method and multistage heat treatment for Mg Ag microalloyed TiB2/Al-4.5Cu composite, which improved its strength plasticity balance [19]. Choudhari et al. tried to carry out the whole method, simulation, and optimization in AutoCAST X software based on VGM and made significant improvement in casting quality [20].
The existing literature has carried out a lot of research on various precision casting methods and the actual production of a variety of parts, but there are still some deficiencies in the production of aluminum alloy impeller parts using the gypsum casting method, unlike the production using investment casting, the gypsum shell has poor air permeability and lower thermal conductivity, and the impeller, a thin-walled complex curved surface part, is more prone to appear porosity, shrinkage, and other defects during production. Based on this, the simulation of impeller mold filling and solidification process is taken as a reference, and the prediction of defects is analyzed and studied. The defects are reduced by optimizing the process parameters to guide the actual casting production.
2. Model Processing
In this paper, the structural features and dimensional values of the initial casting system are measured, and the plaster casing 3D model adapted to the actual production of the impeller casting is designed using Magics modelling software, outputted in wireframe IGES format to maximise the retention of features, and then closed using the “Create plane by closed curve or edge” command in the “Geometry tool.” The impeller model is then reconstructed in reverse using the “Create plane by closed curve or edge” command in the “Geometry Tools” to close the casing. The mathematical model and reconstruction model are shown in Figure 1.
3. Numerical Simulation of Gypsum Casting
3.1. Pretreatment and Parameter Setting
After the model reconstruction is completed, the grid is divided. The line areas at different structures can be assigned different grid sizes according to the simulation needs by observing the model with wireframe. The number of initial gating system surface grids is 20764, and the number of volume grids is 246938. The grid division is shown in Figure 2.
Preheating the first shell was performed in the pouring casting process, metal liquid fluidity cannot be guaranteed in the first part into the shell, and pouring at too low preheating temperature can result in the plaster shell and aluminum alloy liquid temperature difference. It is especially easy in the impeller class thin-walled castings molten mold casting because of defects such as insufficient pouring and cold partition, but too high plaster mold preheating temperature will easily prolong the cooling time of the cavity, resulting in coarse grains in the finished casting and reduced mechanical properties of the casting. In this paper, the preheating temperature of the plaster mould before casting is selected at 250°C based on production experience.
After completing the determination of the casting-related environmental parameters, the CAST module is applied to the casting and shell material selection. The TYPE command is used to select the corresponding material type, setting the alloy as the casting material type because the software comes with the material library missing when selecting the material, and by consulting the data for ZL101A custom creation and selection, the initial filling rate is set as 0.. Combined with the previous analysis results, the initial pouring temperature is selected as 680°C. The initial pouring temperature was chosen to be 680°C. The initial filling rate was set to 100, the initial temperature was set to 250°C, and the type of heat transfer between the shell and the metal liquid was set to COINC. 45 mm/s was calculated as the pouring speed, which is 0.3831 kg/s. The parameters were set as shown in Table 1.
(a)
(b)
(c)
(a)
(b)
3.2. Simulation of the Filling and Solidification Process
Viewing the solidification and simulation results in the temperature field display mode is a common observation method in the casting process. The original simulation results are shown in Figure 3. At 0 s, the aluminum alloy melt starts to enter the mold cavity, contacts the sprue cup, and at 0.66 s, the metal liquid drops to the bottom. At this time, the mold filling in the horizontal direction of the mold cavity starts from bottom to top. At 2.9 s, the impeller part is filled, and the sprue part continues to fill. Finally, the filling of the whole mold cavity was completed in 8.06 s.
After the mold filling is completed, the solidification process can be analyzed by observing the subsequent temperature gradient changes. The part below the liquidus is in the completed state of solidification, and the part around the liquidus is the solid-liquid coexistence part. With the heat exchange between the liquid metal and the shell, the shell and the environment, and the liquid metal and the environment, the overall temperature continues to decrease. The simulation results of the solidification process are shown in Figure 4. At 261.55 s, the blade part started to solidify first due to thin wall. In 1961.55 s, the impeller blade part was completely solidified, and the middle support part was not conducive to heat dissipation due to the thick wall thickness, resulting in heat accumulation in the middle part and slower solidification. At 2351.55 s, the impeller part has been solidified. With the increase of time, the gating system part begins to solidify, and the overall temperature of the model continues to drop.
(a)
(b)
(c)
(d)
3.3. Defect Prediction Results
Throughout the casting process of the impeller, it is observed that the aluminium alloy liquid undergoes a change from a high temperature liquid state cooling into a low temperature solid state, and the alloy model shrinks accordingly as the temperature decreases after cooling. At the beginning of the solidification stage, the part of the liquid metal first cooled in the cavity cannot be subsequently replenished from the riser, and the prematurely cooled part with a hole-like defect after solidification is the shrinkage. Later in the solidification stage, the branch crystal framework in the complementary shrinkage channel obstructs the flow of metal liquid, reducing the area of the complementary shrinkage channel, resulting in a large number of tiny, tiny holes that can easily be formed in the parts that cannot be replenished by metal liquid, which is shrinkage. The use of computer software to analyse the simulation results before the actual casting can be observed for defects with a greater tendency to produce, greatly reducing the probability of casting defects and reducing the corresponding casting cycle. In this paper, the observation of the tendency of casting defects is viewed through the Shrinkage Porosity criterion, and the relevant results are shown in Figure 5.
According to the analysis of observation criteria, the overall casting of the impeller is in good condition, but the middle part will form a shrinkage cavity defect with a diameter of about 4-5 mm at the final stage due to the heat accumulation generated during the solidification process. Through testing, the process of placing cold iron at the bottom can reduce the size of the defect to about 3 mm at the lowest. After adjusting various environmental parameters for multiple sets of simulation, the defect can never disappear.
(a)
(b)
(a)
(b)
4. Process Optimization Scheme
4.1. Structural Optimisation
Compared to other casting methods, the thermal conductivity and exhaust of the plaster casing is poor, so the design of the plaster casting pouring system needs to ensure excellent exhaust and sequential solidification of the casting. Based on the analysis of the initial casting results of the impeller, the original structure was improved to a side injection pouring system, taking into account the need to ensure the smooth exhaust of the plaster casing and the sequential solidification of the castings during the solidification phase and to advance the pouring time of the central solid part of the impeller in order to reduce the time for heat to gather during solidification and to reduce defects.
In the improved gating system, the position where the aluminum alloy liquid is injected into the gypsum cavity of the impeller is selected to be placed at the position where the lower plane size of the impeller is thick and flat, and the bottom surface is not used as a working plane in subsequent applications. After casting, it is not easy to damage the key parts when the casting is separated from the runner. The corresponding structure is changed to side pouring to ensure the sequential solidification and advance the filling and solidification time of the middle part of the impeller compared with most parts. At the same time, multiple runners can ensure the flow stability and avoid turbulence.
The improved three-dimensional model of the pouring system and the shell is shown in Figure 6.
The number of meshes is inversely proportional to the cell size; the more meshes are generated in the selected area, the smaller the unit mesh size and the more precise the results of the selected area, which helps to improve the accuracy of the simulation. However, the greater the accuracy of the calculation, the better. An excessive number of meshes can ensure the accuracy of the simulation but will increase the difficulty of the solution or even lead to invalid calculation solution. Therefore, for the key parts of the model, the denseness of the meshing should be increased to obtain more accurate simulation results, while for the structure of the riser and sprue, the density of the meshing should be reduced to improve the simulation efficiency and at the same time to ensure the accuracy of the simulation results.
For the overall structure of the casting system, as the impeller is the main casting, the grid cell size of the impeller part is set to 2 to ensure the denseness of the division; the calculation of the connection between the sprue and the casting does not need to be too strict, so the grid size can be set to 3; the upper part of the sprue is not used as the main target part of the simulation, so the grid size is set to 5. No complex calculations are required for the shell; the mesh size is generated using the default value of 10, as shown in Figure 7.
After the surface mesh has been generated, the results are checked by first setting some relevant parameters in the Check Surface Mesh dialogue box, usually setting the Minimum Elm value to 0.45 mm, adding 10 to the Overlap selection and setting Fill Holes to 0–10. As a result, the repair stage was selected as Auto Correct for automatic computer grid repair, and the repair results will be displayed in the dialogue prompt box. If there is still part of the automatic repair which cannot solve the division results, select manual completion of the corresponding cell re-division, deletion, and other operations. When the division results displayed that the mesh is ok, selecting the module automatically generates the body mesh function key. As shown in Figure 8 for the mesh division results, this model successfully completed the mesh division behind the mesh for 70301 and the body mesh for 1596201.
4.2. Parameter Optimisation
In the casting process, the three data of pouring temperature, pouring speed, and shell preheating temperature are the most controllable and at the same time the most influential environmental parameters on the casting results, and a certain defect may arise under the combined effect of one or more parameters. Therefore, an orthogonal test design was carried out to compare and analyse the results of multiple tests, to find out the influence of different factors on the test results while reducing the time required for the test, to rank the different variables in order of priority, to analyse the influence law, and to find the best test parameters. The three parameters were arranged and planned as shown in Table 2.
Among the groups of parameters, the combination test, combined with the simulation results of orthogonal test data was further analyzed, by comparing different combinations of simulation results to find the better process parameters. The casting simulation results of the total shrinkage porosity parameters, the resulting stress value as a measure of standard, and the combination of the smallest impact value corresponding to the test parameters were identified as the optimal process parameters. The overall score for the impact capability is expressed in S to express the results as shown in Table3.where F—along the horizontal yz is the maximum stress value in the horizontal direction. —Total shrinkage porosity of the casting.
In order to compare the impact of different test parameters on the quality of castings, the extreme difference is introduced. K represents the ability of each parameter to influence the quality of the casting; the greater the influence of the test group, the greater the influence, the greater the corresponding range K value of the test group.The results are shown in Table 4.where H—sum of the impact values correspondsto each set of parameter factors. N—The number of trials for each group of parameters.
The results of the orthogonal experiments were calculated by designing the extreme differences, and the impact on the quality of the finished casting was greatest at the pouring temperature, followed by the shell preheating temperature, and lowest at the pouring speed. Comparing the impact values of several groups of tests, it was found that the sum of the three factors was smallest when the pouring temperature was taken as the second group, the preheating temperature as the second group, and the pouring speed as the third group. Theoretically, the best quality castings were produced with this combination of parameters, which corresponded to a pouring temperature of 700°C, a preheating temperature of 250°C and a pouring speed of 55 mm/s, corresponding to a mass of 0.3322 kg/s. The final optimum process parameters were determined as follows, as there were no simulations of the corresponding parameters in the 9 groups of orthogonal tests designed, it was necessary to carry out simulations with this parameter condition, further investigating whether this test condition is reasonable. (Table 5)
(a)
(b)
(a)
(b)
(c)
(d)
(a)
(b)
4.3. Form-Filling and Solidification Process
The casting simulation process of the optimized gating system is shown in Figure 9. The whole process can be summarized as follows: 0 s is defined as the time when the aluminum alloy liquid initially contacts the upper surface of the sprue cup in the mold shell, indicating that the alloy liquid begins to enter the gating system. At 0.8427 s, the alloy liquid begins to enter the impeller cavity, and the gas in the cavity is gradually discharged as the metal liquid slowly fills the cavity in 4.3156 s; partial filling of integral cavity impeller is completed. It takes 5.7684 s to complete the overall filling of the gating system. During the complete filling process, it is observed that the process of metal liquid filling the cavity is relatively stable. According to the simulation results of the corresponding temperature field, the temperature of aluminum alloy liquid in the cavity is very high after the completion of filling. Due to the characteristics of thin-walled surfaces, the temperature drop gradient of the blade area is greater than that of other parts under the effect of a good heat dissipation environment, and the overall temperature drop gradient shows a sector decreasing trend from the edge of the blade to the center of the impeller.
The solidification state of the molten liquid is accurately judged by the solid phase colour difference diagram, where the colour depth is higher than the solid phase line temperature corresponding to the colour of the casting part for the liquid state, and the opposite low area indicates a change from the liquid state to the solidified state, distributed between the casting area, which indicates the current moment for the solid-liquid coexistence state. Common defects in the casting of various parts, such as shrinkage are mainly generated in the solidification phase of the metal liquid. As can be observed in Figure 10, heat build-up is inevitable at the root of the impeller blade due to the wall thickness, but the diameter of the thermal joints generated by pouring with the side-injection optimized pouring system is significantly reduced, and the heat build-up generated during solidification is significantly improved.
As can be seen from the temperature difference distribution Figure 11, the impeller blade section is the first to enter a fully solidified state, where solidification is most rapid due to the small wall thickness. At 547.32 s, the temperature at the lower end of the impeller blade starts to drop below the liquid phase line temperature; at 1627.32 s, the entire impeller section has been solidified.
4.4. Defect Prediction
Figure 12 shows the simulation results of the impeller part after improving the structure of the pouring system and optimizing the process parameters. The Shrinkage Porosity criterion shows that under the current conditions, there are zero casting defects in the casting part and that the overall distribution of defects in the pouring system are concentrated at the final solidification of the sprue cup, which is a good casting situation. The original pouring system has a large pouring gate and a large contact area with the casting, which makes the heat dissipation process of the high temperature metal liquid longer, and the relatively delayed pouring time at the central entity makes the heat easily accumulate; the optimized pouring system can effectively reduce the heat accumulation problem and reduce the defect rate of the impeller.
4.5. Stress Field and Displacement Prediction
As the metal liquid gradually solidifies, casting stresses are generated in the solid-liquidtwo-phase area of the pouring system. As the metal liquid is poured into the cavity along the negative direction of the X-axis in this simulation of the pouring system, the longitudinal stresses along the X-axis are effectively released and can be ignored, and the horizontal YZ-axis integrated stresses are shown below. (Figures13 and 14).
From the previous graph, it can be seen that the maximum value of the integrated stress along the horizontal direction of the YZ axis is only 2.71 MPa, which is much less than the material stress limit and will not cause thermal fracture deformation. Also, along the Y and Z direction displacement, as shown in the figure below, it can be found that the corresponding displacement is small and will not affect the casting accuracy.
(a)
(b)
5. Conclusion
To address the problem of common shrinkage and loosening defects of aluminium alloy impellers in gypsum casting, the redesigned model was simulated and the process parameters were optimized using ProCAST software, and the following conclusions were obtained from the analysis of the test results:(1)By analysing the simulation results of the original pouring system and the redesigned pouring system, it was determined that the use of the side-injection-distribution internal sprue pouring system can effectively reduce the probability of shrinkage defects at the internal wall thickness of the impeller, which is conducive to improving the actual production of the gypsum type casting ZL101A impeller.(2)Through orthogonal tests on the design of pouring parameters and optimization of process parameters, it was concluded that the order of influence on the quality of the castings was the pouring temperature which was the largest; the casing preheating temperature being the second largest, and the pouring speed being the lowest. After the adoption of the new pouring system, the process parameters were set at a pouring temperature of 700°C; a casing preheating temperature of 250°C and a pouring speed of 0.3322 kg/s are used to minimise the defect rate of impeller castings in actual production and to ensure that stress and displacement do not affect the quality.
Data Availability
The data used to support the findings of this study are available from the corresponding author upon request.
Conflicts of Interest
The authors declared that they have no conflicts of interest.
Acknowledgments
The research was supported by the Program for the Introduction of Foreign Intellects in Hebei Province (2022) “Research on Key Technologies and Equipment of Efficient and Clean Power Generation for Carbon Neutralization.”