Manufacturability of Turbine Blade Die from Composite Material Using Rapid Tooling Techniques D. Safaeian 1, M. Shakeri 2 and A.L. Darvish 2 Faculty of Engineering, Mazandaran University, Shariati Ave. P.O. Box: 484, Babol, Iran shakeri@nit.ac.ir ABSTRACT Rapid tooling is a development behind of CAD/CAM and rapid prototyping technology. Nowadays, in most large industries such as automotive, aerospace, house equipments and medicine facilities, rapid tooling is used. In this paper manufacturability of turbine blade die from composite material using rapid tooling have been investigated and advantages of this procedure relative to traditional turbine blade wax injection die making have been considered and finally production with investment casting method are also concerned. As a case study, a second stage turbine blade Ruston TA1750 is selected to produce wax model utilizing rapid tooling techniques by composite material, silicon rubber, MJM rapid prototype techniques and traditional wax injection and their characteristics are also compared. Key word: rapid tooling, rapid prototype, CAD/CAM, turbine blade, investment casting, composite material, silicon rubber, MJM 1. Introduction Today, the competition of the industries in the overall the world are provided to fabricate a new product with high quality, low cost and low lead time of marketing, thereby, using of CAD/CAM systems, rapid prototyping and rapid tooling are not avoidable. At the present time, in most large industries such as automotive, aerospace, house equipments and medical facilities, rapid tooling is used. When solid modeling system was introduced in 1970, many researchers try to make physical model directly from geometric model data without using traditional tooling. From 1970 to 1980, companies had some achievement in manufacturing parts layer by layer by photopolymer material. In 1986, 3D System company was introduce SLA rapid prototyping method and first rapid prototype machine was manufactured [1]. In rapid prototyping against traditional machining method, workpiece is produce with additive method and layer by layer, result in this method is faster than from the subtraction of the material. This method hasn t the restriction of machining method 1 -M.S. of Mazandaran University 2 -Assistant Professor of Mazandaran University
and complex feature can produce with high accuracy and high quality [2]. Rapid tooling is extension of rapid prototyping, in reality it is the development of CAD/CAM and rapid prototype technologies. It uses from rapid prototype technology directly or indirectly and reduce the product cycle time [2]. Using of RP&RT allow the casting foundry to produce workpiece without using tooling in low volume production. That helps to approve design, process condition and runner parameters. Result in this method decreases time and cost for production of sample parts with best quality, rapid and low cost [3, 4]. While 20 or 50 casting parts are required, RP method isn t economic, especially if it is required surface roughness. For simple and small part, using of the multiple cavity die is economical. This die can save 50% in production cost due to the reduction of production cycle time and elimination of connection runner to model. It is also reduce probable failure of small part and brittle material due to the reduction of displacement. The machining of cavity can t sure all models are the same, but the entire models are the same by using RT [5]. Furthermore, RT is allowing casting foundry to apply contraction parameter in production [6]. Difference between traditional investment casting with casting by RP and indirect RT shows in chart 1, 2 and 3. In this paper manufacturability of turbine blade die from composite material utilizing rapid tooling are investigated. For studying of this method with other RP and RT, a turbine blade die is made by silicon rubber RT method and MJM RP method. A second stage turbine blade of Ruston TA1750 is selected to make die for injection molding of wax.
2- Geometric Modeling 2-1-Reverse Engineering A computer numerical control CMM (Coordinate Measuring Machine) is used to measure the blade. This CMM is bridge type and made by Mitutoyo Company model FN1106 with software Geopak. A V-block fixture with precision pins is used for clamping blade on the CMM table. Blade is positioned on CMM machine as shown in Fig.1. By defined coordinate system, six sections of the blade are measured as shown in Fig. 2. Each section has measured by 167 points. Platform of the blade is measured in seven sections. A profile projector is also used to measure the profile of the blade root. Fig. 1: Measuring the turbine blade by CMM Fig. 2: Profiles of measured sections 2-2- Modeling A turbine blade has three parts called as airfoil, platform and root. We need to make a casting model of the blade because the platform and root of the blade required machining process. Therefore, for making die, one need casting model with machining allowance as shown in Fig. 3. For modeling of airfoil, first 2D profiles were drawn by the output of CMM data on each section, and then airfoil surface was made by lofting these profiles. Platform and root of the blade are made by extrude. Airfoil and platform surface of the blade was connected by a fillet surface. In order to precede the procedure, surface model of the blade was converted to solid model by a commercial software as shown in Fig. 4. The maximum error of the geometrical model is about 0.025 mm as shown in Fig. 5.
Fig. 3: Casting model of turbine blade Fig. 4: Surface model of the blade was converted to solid model by a commercial software Fig. 5: Maximum geometrical error of the model 3-Making Model 3-1-Making SLA Model Among RP methods, SLA has better feature definition and can create better surface roughness [7]. Turbine blade model is made by SLA method by the following procedure; 1. Convert CAD model in the STL format and send to 3DLightyear software. 2. Positioning workpiece on the SLA5000 platform 3. Positioning support 4. Slicing workpiece and support by thickness 0.15mm 5. Preparing SLA5000: content define recoat in any slice and Z-wait 6. Making workpiece 7. Remove SLA model from machine chamber and remove support
The maximum error of SLA model is about 0.017 mm as shown in Fig. 6. These deviation is measured by a CMM supplied PC-DMIS software to find the accuracy of SLA model. Fig. 6: Measuring result of a slice of SLA model 2-3-Making Model with MJM The MJM (Multijet-Modeling) machine has a headstall that it has 354 nozzles that move in 2 axes X and Y. Thickness of each layer supplied by Z axis movement. From these nozzles exit thermo polymer material and pure in the necessary positions[1]. Among different RP methods, MJM has most compatibility with investment casting process and it can be used to produce wax models based on standard of investment casting process. The investment casting die made by MJM model has excellent surface roughness [8]. Therefore CAD model of the blade import in ThermoJet software and four wax models produce with layer thickness 0.15 mm, as shown in Fig 7. Fig. 7: MJM model Fig. 8: Composite die with aluminum backing and shoe
4-Making Die 4-1-Making Composite Die EP310, composite material made by MCP-HEK Company, is used for making the blade composite die. After determination of the die parting line and mixture of resin and hardener, half of the die is made. After hardening half of the die, first separator agent is added on the die surface and second half of the die is made. Two halves of the die are polished after post curing. Finally runners are added by machining process as shown in Fig.8. Fig. 9: Composite die for injection wax Fig. 10: Samples of injected wax in Composite Die 4-1-1-Test Composite Die Finally after preparing of composite die, 75 wax models are made by composite die in Mavadkaran Company by injection molding at 44 bars as shown in Fig. 9, 10. 4-2-Making Silicon Rubber VTV and Making Wax For making die, liquid silicon rubber VTV750 has to be mixed with hardener and degassed. In this case a runner added in SLA model and VTV 750 pure on the SLA model, after degassing and hardening die, two waxes was pure in the die as shown in Fig. 11. Fig. 11: Silicon rubber die with wax model
5-Making Casting Blade 60 wax models produced with composite die, 2 wax models produced with silicon rubber die and 2 wax models made by MJM method are assembled in two 32- clusters and caste to investigate investment casting parameters and to insure about the accuracy of alloy and wax contraction as shown in Fig. 12. Fig. 12: Two 32-clusters and caste for investment casting Fig. 13: Measured wax model by optical system 6-Dimensional accuracy of Injected Wax model and Casting Blade An optical measuring system, termed as ATOS machine made by GOM Company, is used to control the dimensional accuracy of the injected wax model, results satisfy the required tolerance as shown in Fig. 13. The chord and maximum thickness of blades are measured by universal measuring tools. It is observed that the accuracy of blade profile is in tolerance range ±0.1 mm as shown in Fig. 14. The surface roughness of the samples are also measured and it is shown that the injected wax models in composite die have better surface roughness compare with the other methods, as shown in Table 1. Table 1: Surface roughness of different models in investment casting Type of model Blade concave side(µm Ra) Blade convex side (µm Ra) SLA model 5.08 4.24 MJM model 4.26 4.06 Casting wax in VTV die 5.20 5.02 Injected wax in composite die 1.94 1.89 Casting piece from composite die wax 2.96 2.48
7-Comparing Cost and Time Production time and cost of different methods for making models for investment casting die are displayed in Fig. 14. It is shown that MJM method had minimum production time and cost respect to the others. (a) (b) Fig. 14: compare different methods for making model for investment casting, (a) production time, (b) production cost. Conclusion Among RP and three RT methods considered in this research, MJM method has most economic justification and shorter lead time in low volume production. In this case, the cost of a SLA model of the blade is about four times of the MJM model one. But in high volume production the MJM method is no more economical and RT methods are more considerable. Silicon rubber die is economical compare with composite die. Silicon rubber die has more flexibility in case that the workpiece with a negative slop can exit from die, especially blades have shrouds or a platform has radius or vanes have inner shroud and outer shroud. For building a die by composite material or traditional method,
one need to place an especial mechanism in die that wax model can exit from die without problem. But in silicon rubbers no more need this mechanism. But major problem of silicon rubber die is low heat transfer which causes this method non productive, therefore this method is appropriate for limited production. Finally by comparison of three die making methods which investigated in this research, we found that composite die is appropriate method for limited production. And composite die in comparison with aluminum die has the following advantage. 1. Solve problem of miss alignment of two halves of die that cause injected wax need to be modified. 2. Solve problem of tiny section machining that it need EDM machining and in most case, this process causes inaccuracy of the die. 3. Reduction in cost and time while for this blade observe in comparison with traditional method we save 50% in time and 60% in cost. 4. Develop casting technology in short time: because it is very difficult to determine the coefficient contraction. In case the die is made by traditional method, the die modification process is very time consuming. 5. Develop machining technology in short time: because for production of any blade, it required some test cut in order to solve machining problems. Reference 1. Kunwoo LEE,1998 Principle of CAD /CAM / CAE, Addison Wesley. 2. D.King, T. Tansey, 2002, alternative material for rapid tooling, Journal of materials Processing technology, Vol. 121,pp. 313-317. 3. A.Rosochowski, A.Matuszak, 2000, Rapid tooling : The state of art, Journal of materials processing technolog, Vol. 106,pp. 191 198. 4. Thierry Dormal, December 2002, Rapid tools for wax injection, Sare consortium, issue 4. 5. RAW.San, L.wood, David lee von Ludwing, 1952, Invest ment casting for Engineers, Rein hold publishing corporation 6. Jacobs, Paul f, 1996, Stereo Lithography and other RP & M technologies, Society of manfacturing Engineering in cooperation with rapid prototyping Association of SME 7. T.Riek, p.chris todoulou, S loose, 1996, Comparing rapid prototy ping pattern for investment casting on Australian, 9 th world conference on investment casting. 8. S.J. Zhang, VH Raja, KJ fernandes,c Ryall, D wim penny, 2003, Rapid prototyping models and their quality evaluation using reverse engineering, Journal of Mechanical Engineering science, Vol. 217, pp 81-96.