1757-899x_33_1_012001_squeeze Pins 233i3q

Home

Search

Collections

Journals

About

us

My IOPscience

A Method to set process parameters of local squeeze in HPDC

This content has been ed from IOPscience. Please scroll down to see the full text. 2012 IOP Conf. Ser.: Mater. Sci. Eng. 33 012001 (http://iopscience.iop.org/1757-899X/33/1/012001) View the table of contents for this issue, or go to the journal homepage for more

details: IP Address: 213.201.34.20 This content was ed on 04/02/2015 at 14:34

Please note that and conditions apply.

MCWASP XIII IOP Conf. Series: Materials Science and Engineering 33 (2012) 012001

IOP Publishing doi:10.1088/1757-899X/33/1/012001

A Method to set process parameters of local squeeze in HPDC I Ohnaka1, JD Zhu2, A Sugiyama1 and F Kinoshita3 1

ieSol Co./Osaka University, Toyonaka-shi, Higashitoyonaka-cho 1-32-12, Osaka 560-0003, Japan 2 Multi-Flow Software CO.LTD, Canada 3 QUALICA Inc. , Suita-shi, Japan E-mail:[email protected] Abstract. In HPDC processes, pressurization of the mushy regions via local squeeze pins and plunger during the intensification stage is very important to decrease porosity defects. In order to better understand the process and to be able to set the process parameters more properly, we developed a simulation code to solve the flow field of mushy regions and porosity formation based on the following main assumptions: 1) Local squeeze pins and plunger tip can be treated as pressure boundaries, while the squeeze pins are treated as moving boundaries. 2)The intensification and local squeeze pressure can propagate inside regions where the solid fraction is less than critical values. 3) In the pressurized regions, the cast metal is treated as a Newtonian fluid and both the solid and liquid flow together when the solid fraction is low, and the D'Arcy flow only exists when the solid fraction goes higher. 3) Only the liquid flows in the unpressurized regions, following the D'Arcy's law. 4) Porosities grow in elements when the pressure decreases below critical values depending on the element condition. The simulation was applied to HPDC castings and showed that the pressure-duration was similar to the measured one though the simulated pressure drop in the pressurized region was much smaller than the measured one. The pressure drop, pressure-duration and pin-penetration depth were discussed and the simulation seems to be helpful to determine local squeeze parameters. Future challenges are also discussed.

1. Introduction In high pressure die-casting(HPDC), pressurization of the mushy region is very important to decrease the porosity defects, because many gas bubbles are entrained during mold filling and risers to feed solidification contraction are not available. In particular, local squeeze, namely pressurization with local squeeze pins is often applied on thick-wall regions of the casting prone to porosity defects[1]. Because better propagation of applied pressure decreases porosity more significantly, several works have been reported on pressure propagation and the effect of pressure on the quality of castings[2-8]. However, very few can be found on porosity simulation in HPDC[9-11], especially the simulation of effect of local squeeze pins [12,13]. This paper presents a practical simulation method of the pressure propagation and porosity formation to properly set process parameters of local squeeze. 2. Simulation method Because exact simulation of pressure propagation in the HPDC is very difficult, following assumptions and equations have been used in the simulation:

Published under licence by IOP Publishing Ltd

1

MCWASP XIII IOP Conf. Series: Materials Science and Engineering 33 (2012) 012001

1) Boundaries between casting and the plunger tip or local squeeze pins are treated as pressure boundaries(figure 1). 2) The intensification and local squeeze pressures propagate inside regions where the solid fraction is less than a critical value, f gate

IOP Publishing doi:10.1088/1757-899X/33/1/012001

Pin 2

for intensification and f sqz for the local squeeze, respectively. Namely, the pressures from the plunger tip continue to propagate into the casting as long as the gate does not "freeze off" by reaching the critical value f gate across the cross-section. The

Pin 1

Gate regions where external pressure can and cannot propagate are called pressurized and unpressurized regions, respectively. 3) The pressurized regions are assumed as a Newtonian fluid, where Figure 1. Pressure boundary both the solid and liquid flow together in the regions where solid fraction is less than g SC ( see equation 5) and only the liquid flows in the regions where solid fraction is higher than g SC , following the D'Arcy law. The following discrete equations developed with the Direct Finite Difference Method[14-16] have been used, while the gravity and convection are neglected. Note that these equations are used only for flow elements without porosities. The elements with porosities are treated as pressure boundary. Mass balance (for elements);

( ρ S − ρ L )i

∂VSi ( ρ SU )ij = ∂t ∑

Momentum balance (for staggered elements); u j − ui ut +∆t − uit (α − 1) µ0 f L Vi V j = S ( Pi − Pj ) + ∑ (µα S )− ( + )ui ρVi i ∆t dij Ki K j 2

(1)

(2)

where ρ: density[kg/m3], μ: effective viscosity[ Pa ⋅ s ], µ 0 :viscosity of pure liquid[ Pa ⋅ s ], Δt: time step[s], d: distance between nodal elements[m], u: velocity[m/s], S: surface area[m2] , U:velocity on element surface[m/s], V: volume[m3], K: permeability[m2], f L : liquid fraction[-], α : 1 for the low-fraction-solid regions and 0 for high-solid-fraction regions. Subscript i,j : element number, s : solid, L: liquid 4) In the unpressurized regions, the liquid flow follows the D'Arcy's law. 1 (3) = U ij {Pj − Pi + ρ L g ( h j − hi )} di d j µ( + ) Ki K j The permeability, K, is calculated wih equztion 4 that is a simplified Ergan's equation[17].

K=

(1 − f S )3 d s2 100 f S2

(4)

where g : gravitational acceleration, f S : solid fraction, P : pressure, h : melt head, d S : grain diameter of solid phase. 5) The effective viscosity in the pressurized region is calculated with Mori-Ototake's equation[18], assuming equiaxed solidification structure. 0.5d S S a for (5) = µ µ0 [1 + ] f S < f SC 1/ f S − 1/ f SC where S a :specific surface area of grain, f SC :critical fraction solid. Although S a varies with time, it is assumed as a constant here (see table 1). It was found that the pressure drop is not sensitive to S a via several case studies.

2

MCWASP XIII IOP Conf. Series: Materials Science and Engineering 33 (2012) 012001

IOP Publishing doi:10.1088/1757-899X/33/1/012001

6) The temperature field is solved with the thermal energy balance equation described in Ref[14-16], neglecting the convection term. The latent heat of fusion is considered by the temperature recovery method. 7) The volume change of gas entrapped during mold filling follows the equation of state for ideal gas: (6) PV = m0 RT 3 where P : pressure of the gas[Pa], V : volume of the gas[m ], R : gas constant[J/kgK], T : temperature[K], m0 : mass of the gas[kg]. 8) The penetration depth of squeeze pins is calculated from the volume shrink during solidification of their territories bounded by the critical solid fraction. The compression of porosities is also considered. The pin penetration terminates when the solid fraction ahead of the pins, say 5mm from the pin tip, becomes greater than a critical value. 9) The thermal properties of squeeze pins are the same with those of the casting and there is no thermal resistance between the casting and the squeeze pins, while the region where the pins exist is treated as a solid region. Solidification simulation for the case where a pin was inserted by 8mm showed that the fraction solid ahead of the pin tip rapidly increased at the very early stage, but soon the solidification rate decreased approaching rather similar value to the case where no pin exists.

Pressure / MPa

Simulated

40 ADC12

30

Measured

20 AC4C 10 0 0

Figure 2. Casting specimen used by Iwata et.al[4]

0.5

1.0

1.5 2.0 Time / s

2.5

3.0

Figure 3. Comparison of measured and simulated pressure change

Pressure drop/ kPa

Solid Fraction

-0 AC4C-2 ADC12 Ref. 21

AC4C-1 Ref. 22

AC4C-1

-4

AC4C-2 -8 ADC12

-12

0 Temperature /0C

Figure 4. Fraction solid and temperature relationship used in the simulation 3

0.5 1.0

1.5 2.0 Time / s

2.5 3.0

Figure 5. Simulated pressure drops at the measuring point

MCWASP XIII IOP Conf. Series: Materials Science and Engineering 33 (2012) 012001

IOP Publishing doi:10.1088/1757-899X/33/1/012001

Table 1. Physical properties used in the simulation Physical parameters Density [kg/m3] Specific Heat[kJ/(kgK)] Thermal cond.[W/mK] Latent heat [kJ/kg] Liquidus Temp.[℃] Solidus Temp.[℃] Ejection Temp.[[℃] Shrinkage coef. [%]

AC4C 2400 1.19 100 425 615 562 620 6

ADC12 2480 1.19 100 495 580 520 600 4

f gate , f sqz

0.9

0.99

d S in Eq.4 and 5 [μm] µ 0 Sa in equation 5 f SC in equation 5 [-]

20

20

200

200

0.45

0.45

P

Figure 6. Simulated flow pattern at 1.2s for AC4C alloy (Cross-section AA' in figure 2) This is because the pin temperature rapidly increased and kept due to its low thermal conductivity (SKD61) compared to that of the casting. Therefore, the assumption is reasonable for this work. 10) Porosity is calculated from the flow field assuming pores start to grow at elements when the pressure there becomes lower than a critical value. Once the critical pressure is reached, the pressure in the elements is set to zero and the net-loss of melt volume in these elements is calculated and used as the increase of porosity volume (porosity growth). 3. Simulation of pressure propagation The measured data of Iwata et al.[4] were selected to evaluate the simulation accuracy of high pressure propagation, because the casting dimensions are clearly described as in figure 2. They measured the Figure 3 compares the measured and simulated pressure changes with using solid fraction temperature relationship shown in figure 4. In the simulations the gate pressure was set as follows; 0MPa at 0s, 40MPas at 0.3s, 40MPa between 0.3 and 3s, and 0MPa at 3.1s. Other data used in the simulation can be found in table 1. Figure 5 shows simulated pressure drops (pressure at the measuring point - pressure at the biscuit [40MPa]) and figure 6 is the simulated flow field for the AC4C casting at 1.2s, showing a reasonable flow pattern. Although the simulated pressure drop is much smaller than the measured one, the pressure-duration is rather similar between the simulation and measurement as shown in figure 3. Further, the pressure drop strongly depends on the solid fraction - temperature relationship of casting alloys as shown in figure 3 and figure 5, though the ADC12 casting showed a larger drop unlike the experimental result. The reasons for the difference in the pressure drop may be as follows: 1) Boundary condition Actually the boundary at the plunger tip is not a uniform pressure boundary. This is because the tip movement is controlled by the surrounding of the biscuit (P in figure 6) when the surrounding solidifies. If its strength is high enough, the force of the tip is balanced with the reaction force of the surrounding, resulting in pressure drop in the central part with lower strength. Solving such problem requires a structural analysis with considering the thermo-mechanical behaviour of the solidified part and the of the casting with the mold.

4

MCWASP XIII IOP Conf. Series: Materials Science and Engineering 33 (2012) 012001

IOP Publishing doi:10.1088/1757-899X/33/1/012001

2) Basic equation and viscosity The basic equation (equation 2) considers only the shear resistance which strongly depends on the flow velocity. Because the velocity of the flow is usually low, resultant pressure drop is also low. Although a higher effective viscosity causes a higher pressure drop, even a viscosity 1000 times larger than coal-tar did not cause such high pressure drops as measured. 3) Measurement error Although the pressure sensor was set at the boundary between the casting and mold with a heatinsulating-sheet insert, the solidification near the sensor may decrease the pressure on the sensor when the solidified region has a strength that varies with alloy composition. The displacement of the sensor due to the mold expansion may also contributes to the pressure drop. Currently it is not clear yet which is the main reason caused the difference in the pressure drops. However, the estimation of the pressure-duration may be more important than that of the pressure drop, because the former affects the casting quality more significantly [4]. The reason why the pressure-duration affects the porosity defects may be explained as follows. From assumption 6, equation 7 can be used to calculate the gas volume change if the temperature change is small; (7) V = PV 0 0 /P where P0 is the initial gas pressure, namely, the pressure at the end of mold filling, and is usually much lower than the intensification pressure, P , for example P0 ≈ 1MPa and P ≈ 40MPa. Therefore the entrapped gas becomes very small in a very short period. When the pressure decreases, however, the gas in the porosities expands. With a longer pressure-duration, the progress of solidification during this period will be larger and the gas will expand less. This may be ed by the fact that a kind of extrusion of cast metal occurred when a squeeze pin was pulled out too early( figure 7), while the effect of pressure decrease induced by the pulling-out cannot be excluded. Another hypothesis for the effect of pressure-duration might be gas absorption into the liquid phase during the period. However this effect may be small, because the melt surface of the gas bubbles may be covered with an oxide film preventing the gas diffusion, and the pressure-duration was in the order of a few seconds in this particular experiment. Based on the fact that complete blocking of the pressure propagation requires a certain thickness of solidified shell, the pressure-duration in simulation was determined by the time when the solid fraction at 1.7mm from the mold, where the sensor is attached reached the critical solid fraction. Sensitivity was also checked by changing the distance from 1.7mm to 0.8mm, and it was found that the pressureduration becomes 20% shorter. As can be seen in figure 3, both simulated and measured results show a same trend, the pressure-duration of AC4C is shorter than ADC12. This is mainly due to the smaller latent heat and higher solidus temperature of AC4C as compared with those of ADC12(table 1). Because usual castings produced by the HPDC have many thin and thick parts unlike the casting shown in figure 2, the pressure-duration may be determined mainly by the solidification of thin parts and the starting time to push the pins. Therefore, the proposed method using the critical solid fraction, f gate and f sqz may be useful as shown below. 4. Local squeeze casting Simulation was carried out on a local squeeze casting of ADC12 where two squeeze pins were set on the thick parts as shown in figures 8 and 10. Although the gas entrapment during mold filling was simulated (the method is reported in [11]), it's effect didn't appear in the porosity simulation. This is because the entrapped gas was compressed to very small volume during the intensification stage. It is reported that the local squeeze drastically decreased the porosity defects as shown in figure 8[1]. In the actual casting, the gate pressure of 52MPa was applied until 3s and squeeze pin pressure of 195MPa from 1s to various pulling-out times. The pin-diameter was 25mm(total pin cross-sectional area was 980mm2).

5

MCWASP XIII IOP Conf. Series: Materials Science and Engineering 33 (2012) 012001

IOP Publishing doi:10.1088/1757-899X/33/1/012001

Figure 9 shows the simulation result for the case where the local squeeze is not applied, showing many porosities formed in the product. When the local squeeze is applied from 3.5s to the end of the solidification, most of the defects are eliminated(figure 10) just as the observation(figure 8). The squeeze-pin-penetration depth was about 14 mm as shown in figure 11( when the solidification shrinkage coefficient is 5%, it increases to 17.8mm). It is greatly affected by the solidification at the head of the pins, because the pin movement is terminated when the solid fraction becomes greater than the critical value.

Figure 8 Section through the center of squeeze pins in Fig.10, showing good quality (squeeze starting time: 3s , pulling-out time:20s)

Figure 7 Section through the center of squeeze pins in figure 10, showing kind of extrusion of the metal (Pull-out time: 12s after mold filling [1]).

Pin 1

Pin 2

Gate Figure 9 Simulated porosity without local squeeze

Figure 10 Simulated porosity with local squeeze

Distance/mm

16 12 8 4 0 0

5

10

15

20

25

30

35

Time / s Figure 11 Simulated travel curve of the squeeze pins 6

MCWASP XIII IOP Conf. Series: Materials Science and Engineering 33 (2012) 012001

IOP Publishing doi:10.1088/1757-899X/33/1/012001

The pin-penetration depth measured for the casting ranged from 20 to 27 mm depending on the squeeze starting time[1], for example, 27mm for 2s, 22mm for 5s. If the penetration depth is calculated from the volume of solidification shrinkage of the product (475x103mm3), it is about 19x103mm3, hence the penetration depth is 19mm. However, it should be less than 19mm as the simulation shows, because solidification proceeds before starting the squeeze in the product. The large value of 27mm that was obtained when the squeeze starting time was 2 seconds after the mold filling may be due to the pushing back of the plunger tip, because the intensification time was 3 seconds and the squeeze pressure was higher than the plunger tip pressure. However, the value of over 20mm for the starting time after 4 seconds cannot be explained only by the solidification shrinkage. The reasons for the larger penetration depth may be as follows; 1) Mold expansion Usually the mold temperature increases until 10 or 20 second after the mold filling depending on the mass of casting and mold cooling conditions. Therefore, the mold should expand after the mold filling, resulting in increase of casting volume and hence the penetration depth. This also could affect the porosity defects. 2) Compression of porosities If it took time to compress the porosities for some reasons, the porosities could remain even after the intensification stage and increase the penetration depth.

Although we cannot reject the latter, the former might be more possible. 5. Concluding remarks A simulation method for setting process parameters after mold filling in HPDC has been developed and the comparison of simulation with observations showed the followings: 1) The pressure-duration was similar to the measured one and determined by solidification at the measuring point. 2) In the pressurized region, no significant pressure drop was obtained from the simulation. 3) The pressure propagation varies with the solid fraction-temperature relationship, though simulated pressure drops were much smaller than measured ones. 4) Why the longer pressure-duration causes smaller porosity defects may be explained by the suppression of the expansion of the compressed porosity gas due to the progress of solidification during the pressure-duration. Although the proposed simulation method should use proper critical solid fractions to determine the pressurized regions and termination of squeeze pins, and cannot consider the mold expansion, it can roughly estimate the interaction between the plunger and squeeze pin, time variation of the pin penetration depth and degree of porosities. Therefore, it seems the simulation is helpful to set the process parameters. However, there are still many challenges to improve the simulation accuracy such as visco-plastic analysis as a problem between the casting and mold, which enable the simulation without using the critical solid fractions. The consideration of the pin material properties and thermal resistance between the pin and casting are also required. References [1] Zhu J, Yokoyama H, Ohnaka I, Murakami T, and Cockcroft S, Prediction of Casting Defects in High-pressure Die Casting Processes by Using JSCAST, 2009 Modeling of Casting, Welding and Advanced Solidification Processes XI I, (Warrendale, TMS ), 361-368 [2] Nishi N, Sasaki H, Hirahara T and Takahashi Y , 1988 Imono 60 777-783 [3] Iwamoto N and I.Kuboki 1994 Trans. Japan Die Casting Conference, JD94-11 p 94 [4] Iwata Y,Sugiyama Y, Iwahori H and Awano Y, 2000 J.Japan Foundry Engineering 72 263-267 [5] Ikeda S, Matsumoto Y, Murakami M, Xiong S, and Hu B 2008 Relation btween Casting Conditions and the Molten Metal Pressure in a Die Mold, Trans. 2008 Japan Die Casting Congress, JD08-18, 119-123 [6] Kato E, Nishiyama N, Nomura H, Asai K and Tanigawa S, 2002 J. JFES 74 370-376 [7] Tanigawa S, Asai K, Yang Y, Nomura H and Kato E, 2003 J.JFES 75 525-531

7

MCWASP XIII IOP Conf. Series: Materials Science and Engineering 33 (2012) 012001

[8] [9] [10] [11]

[12] [13] [14] [15] [16]

[17] [18] [19] [20] [21] [22]

IOP Publishing doi:10.1088/1757-899X/33/1/012001

Ueki T,Tashiro M,Tajima A,Kato H,Yageta K and Kambe H, 2010 Measurement of melt and die behavior in die-casting process, Trans. 2010 Japan Die Casting Congress, JD10-25, 157-162 Kubo K 1997 Method for Numerically Predicting Casting Defects, NADCA, T97-015 Kubo K, Asao H, and Deki N, 2008 Progress in Computer Simulation of Die Casting, Japan Die Casting Congress, JD08-30 197-203 Kimatsuka A, Ohnaka I, Zhu J, Sugiyama A and Kuroki Y,Mold Filling Simulation for Predicting Gas Porosity 2006 Modeling of Casting, Welding and Advanced Solidification Processes XI I, (Warrendale, TMS ), 603-610 Li S, Sannakanishi S, Tsumori M, Akase M and Anzai K, 2003 Method for Numerical Prediction of Porosity Defects in Squeeze Casting, Trans. NADCA, T03-14 Li S, Mine K,Sannakanishi S and Anzai K 2006 Qunatitative Simulation of Shrinkage Porosity by Pressure in Squeeze Casting, Japan Die Casting Congress, JD06-30, 199-204 Ohnaka I, 1979 Classification of Numerical Methods for Transient Heat Transfer Problems and Improved Inner Nodal Point Method, ISIJ, 65 1737-1746 Zhu J and Ohnaka I, 1996 Three Dimensional Computer Simulation of Mold Filling of Casting by Direct Finite Difference Method”, J. of JFS, 68 668-673 J.D.Zhu and I.Ohnaka, 1995 Treatment of Free Surface Boundary Condition for the Simulation of Mold Filling, Modeling of Casting, Welding and Advanced Solidification Processes VI???, (Warrendale, TMS ) 359-372 Ergan S, Fluid flow through packed columns, Chem.Eng.Prog., 1952 48 89-93 Mori Y and Ototake N, On the Viscosity of Suspensions, J. Chemical Engineering of Japan, 1956 20 488-494 Ohnaka I, Zhu J, Sugiyama A and Kinoshita F, 2011 Proc. 11th Asian Foundry Congress, Foundry Inst. Chinese Mechanical Eng. Soc. Guanzhou China, 152-157 Sugiyama Y, Iwahori H, Yonekura K and Oguchi Y, 1994 IMONO 66 412-417 Iimai H, Kameyama Y, Suzuki Y, Nagayama Y, Kato E and Nomura H, 2001 Denso Technical Review, 6(2) 100-106 Sim J, Jang Y, Moon J, Kim J, Min K and H C, 2009 ISIJ International , 49 1700-1709

8