Prediction of Thermal Fatigue in Tooling for Die‐casting Copper via Finite Element Analysis

Amit Sakhuja and Jerald R. Brevick

Citation: AIP Conference Proceedings 712, 1881 (2004); doi: 10.1063/1.1766807
View online: http://dx.doi.org/10.1063/1.1766807
View Table of Contents: http://scitation.aip.org/content/aip/proceeding/aipcp/712?ver=pdfcov
Published by the AIP Publishing

Abstract

Recent research by the Copper Development Association (CDA) has demonstrated the feasibility of diecasting electric motor rotors using copper [1].

Electric motors using copper rotors are significantly more energy
Recent research by the Copper Development Association (CDA) has demonstrated the feasibility of die-casting electric motor rotors using copper [1].

Electric motors using copper rotors are significantly more energy efficient relative to motors using aluminum rotors. However, one of the challenges in copper rotor die-casting is low tool life.

Experiments have shown that the higher molten metal temperature of copper (1085 °C), as compared to aluminum (660 °C) accelerates the onset of thermal fatigue or heat checking in traditional H-13 tool steel. This happens primarily because the mechanical properties of H-13 tool steel decrease significantly above 650 °C.

Potential approaches to mitigate the heat checking problem include: 1) identification of potential tool materials having better high temperature mechanical properties than H-13, and 2) reduction of the magnitude of cyclic thermal excursions experienced by the tooling by increasing the bulk die temperature.

A preliminary assessment of alternative tool materials has led to the selection of nickel-based alloys Haynes 230 and Inconel 617 as potential candidates. These alloys were selected based on their elevated temperature physical and mechanical properties. Therefore, the overall objective of this research work was to predict the number of copper rotor die-casting cycles to the onset of heat checking (tool life) as a function of bulk die temperature (up to 650 °C) for Haynes 230 and Inconel 617 alloys.

To achieve these goals, a 2D thermo-mechanical FEA was performed to evaluate strain ranges on selected die surfaces. The method of Universal Slopes (Strain Life Method) was then employed for thermal fatigue life predictions.

INTRODUCTION

Major factors affecting the fatigue life of the die surface in high pressure die-casting applications are: 1) the initial bulk die temperature, 2) the temperature, heat of fusion and specific heat of the casting alloy, and 3) the elevated temperature physical and mechanical properties of the die material [2]. Comparing the die-casting of aluminum versus copper, experience shows that the thermally induced problems with die materials are more serious with copper alloys. Aluminum alloys are die-cast at relatively low temperatures (675 °C), compared to copper alloys that are cast at temperatures in the 1200 °C range.
The higher temperatures in the copper die-casting process induce high compressive strains at the die cavity surface during molten metal injection. The magnitude of the thermal gradient and strains on the die surface are directly correlated to the temperature of the die surface immediately prior to molten metal injection (initial bulk die temperature). Once the casting begins to cool, it contracts, resulting in a reduction of heat transfer to the cavity surface due to less contact pressure.
Subsequent to ejection of the casting from the die cavity, the cooling of the die surface via exposure to cool air or spray lubricants may also result in tensile stresses. Cyclic heating and cooling during the die-casting operation results in alternating compressive and tensile strains (strain range) on die surfaces. The cyclic strains eventually initiate cracks, and as the number of casting cycles accumulates, the cracks may grow deeper into the die. Eventually this results in the castings mechanically adhering to the die, or small areas of the die surface actually breaking out. In either case, thermal fatigue cracking creates an undesirable surface finish on the casting.
The formation of cracks on the die surface due to thermal fatigue is commonly called heat checking.

Previous CDA copper die-casting experiments using H13 tool steel dies at an initial bulk die temperature of 200 °C has yielded as few as 20 die-casting cycles to the onset of heat checking in the metal delivery system of the die (Figure 1, areas 3 and 4). Nickel-based alloys Haynes 230 and Inconel 617 may offer improved resistance to heat checking as die materials because of their excellent elevated temperature mechanical properties. Also, increasing the initial bulk die temperature may reduce the thermal gradient and strain range induced on the surface of the die material.

However, conducting experimental copper die-casting campaigns to assess the heat checking resistance of various die materials at various initial bulk die temperatures is expensive. A lower cost alternative is computer modeling using the Finite Element Analysis (FEA) technique. Using FEA the strain range on die surfaces can be evaluated, and subsequently the method of universal slopes can be employed to predict cycles to failure.

Therefore, the objective of this research was to predict the number of copper die-casting cycles to the onset of heat checking in the metal delivery system of a rotor die using FEA. Haynes 230 and Inconel 617 were evaluated at initial bulk die temperatures of 200 °C, 350 °C, and 650 °C. Initial bulk die temperatures above 650 °C are impractical to maintain in a production environment.

FIGURE 1. 2D cross-section of the copper rotor die:  1. Steel laminations, 2. End ring (Cu), 3. Gate area  (Cu), 4. Runner (Cu), and 5. Steel arbor [1].
FIGURE 1. 2D cross-section of the copper rotor die:
1. Steel laminations, 2. End ring (Cu), 3. Gate area
(Cu), 4. Runner (Cu), and 5. Steel arbor [1].
FIGURE 4. Strain versus distance plot for surface 1 at  t = 20 s [6].
FIGURE 4. Strain versus distance plot for surface 1 at
t = 20 s [6].

CONCLUSIONS

As shown in Table 6, of the three temperatures simulated (200, 350, and 650 °C) an initial bulk die temperature of 650 °C resulted in the best thermal fatigue life performance for both Haynes 230 and Inconel 617 alloys. The predicted thermal fatigue life of the tooling was increased by a factor of approximately three for the nodes evaluated when the bulk die temperature was increased from 200 °C to 650 °C.

When considering Haynes 230 versus Inconel 617, the Inconel 617 was predicted to perform slightly better in thermal fatigue resistance at 650 °C at four out of the five locations analyzed in Fig. 3 (A, B, C, and E).

The results of this research suggest that replacing H-13 with either Inconel 617 or Haynes 230 nickel based alloys, along with increasing the bulk die operating temperature to 650 °C, would significantly delay the onset of heat checking in the metal delivery system of copper rotor casting tooling. Finally, the predicted number of cycles to the onset of heat checking in this research should be considered somewhat conservative for two reasons.

First, the analyses were performed in the metal delivery area of the die – known from experiments to demonstrate the earliest heat checking.

Second, due to lack of available high temperature mechanical property data, the RA values used in the model were most likely conservative.

REFERENCES

[1] Cowie, J.G., Peters, D.T., Brush, Jr., E.F.,
Madison, S.P., “Materials and Modifications to Die Cast the Copper Conductors of the Induction Motor Rotor”, Die-Casting Engineer, September 2001.
[2] Manson, S. S., Thermal Stress and Low Cycle Fatigue, McGraw-Hill, New York, 1966, ISBN #65-25918.
[3] Sirinterlikci, A., “Thermal Management and Prediction of Heat Checking in H13 Die-casting Dies”, PhD Dissertation, The Ohio State University, Columbus, OH, 2000.
[4] Shi, Q., “Prediction of thermal distortion and thermal fatigue in shot sleeves”, PhD Dissertation, The Ohio State University, Columbus, OH, 2002.
[5] Haynes International, “Haynes 230 Databook”, www.haynesintl.com.
[6] Sakhuja, A., “Evaluation of die-casting tooling using FEA modeling”, Master’s Thesis, The Ohio State University, Columbus, OH, 2004.