Professional Documents
Culture Documents
Design For Casting
Design For Casting
1 Casting Design Today Casting continues to be the most preferred process to create intricate shapes in metal, but is also one of the most difficult to precisely model and control for achieving consistent quality. The range of variation in terms of geometric, material and process characteristics, and their uncharted effects on manufacturability makes every casting project a new challenge. It is no surprise that the foundry industry continues to suffer from poor utilization of material, energy and human resources: the average scrap rate is as high as 7% [1] and the average lead time to produce the first casting for approval is 10-14 weeks [2]. Today, casting engineers have access to a range of software, starting from database management and design calculations to process simulation and computer-integrated manufacturing, in increasing order of difficulty of installation and use. Of these, computer simulation of casting process has emerged as a powerful tool for achieving quality assurance without time consuming trials. Software packages for simulating the solidification of molten metal in the mold enable predicting the location of shrinkage defects and optimizing the design of feeders to improve the yield; more advanced packages perform coupled simulation of mold filling and casting solidification [3]. It has been reported that simulation studies can reduce casting defects, manufacturing costs and lead time by as much as 25% [4]. Already, an estimated 1000 foundries (among 33,500 worldwide) are using simulation software to improve their performance and the number of simulation users is steadily increasing. However, casting engineers only produce what product engineers design. Designers mainly focus on product function - creating an optimal shape to withstand operating stresses - aided by excellent Finite Element Analysis software available for this purpose. They are largely unaware of the casting processes, their capabilities and limitations [5]. This results in either over design (for
Dr. B. Ravi is an Associate Professor of Mechanical Engineering, In-Charge of the Casting Simulation Laboratory and Co-Founder of Rapid Prototyping Cell. He can be reached at: Department of Mechanical Engineering, Indian Institute of Technology, Bombay 400076. Email: [email protected]
example, unnecessary thick ribs leading to heavier castings) or under design (for example, inadequate fillet radius leading to casting defects). Many product features (for example, undercuts) require complex tooling, and others (for example, isolated mass concentration) require additional steps during casting (say, placement of chills). All these mean higher cost, lower quality assurance and longer lead time. Often, a product design causes severe problems at the casting stage, and the foundry may request the customer to either modify the design or pay a higher price to offset the costs of tooling modification, increased scrap and additional operations (such as heat treatment and machining). Significant design modifications at this stage could mean additional burden of redoing the tooling and planning the process again, besides losing the time and resources already spent on these activities. Compared to the above, modifying the design of a cast product while it is still on the drawing board is as easy as erasing and redrawing a line. If the ratio of benefit to cost is considered, then it is immediately apparent that dramatic saving of resources can be achieved by predicting potential problems at the design stage itself, and preventing them through suitable changes to product features. This approach is termed Design for Manufacture. 2 Design for Manufacture of Cast Components Design for Manufacture (DFM) has caught the attention of both researchers and practicing engineers over the past one decade to bring about dramatic improvements in the ease of manufacture. The most popular method of DFM is reusing prior experience coded as design guidelines. Automobile companies were among the first to apply the principles of DFM to simplify product designs, particularly for machined parts, sheet metal parts and assemblies. One of the earliest examples was the Viper car project at Chrysler which was developed from scratch within 3 years and within budget. Gradually this approach has spread to the rest of engineering companies, many of them reporting savings of 20% of costs and 50% of time, besides improved quotations from vendors [6]. However, the guidelines may not be available for a radically new design (especially when new materials and processes are involved) and sometimes conflict with each other. Therefore many companies rely on design review committees (concurrent engineering team) comprising of experts in design, prototyping, tooling, manufacturing, marketing, maintenance and other areas evaluate the design and suggest modifications before it is finalized and released for further action. This is however, a reactive approach in which designers merely try to confirm the manufacturability of a particular design or improve it slightly. Product engineers need to be proactive and take advantage of the range of cast metals and processes to create a product which best satisfies the final requirements. A team of researchers from Concurrent Technologies Corp prescribed a combination of simulation-based DFM and Concurrent Engineering [7]. They visualized how castings will be designed and manufactured through electronic exchange of relevant information between designers and foundry engineers who may be using different CAD software specific to their tasks. However, simulation-based DFM of cast products calls for an in-depth knowledge of the process, as well as significant time and effort for design improvements. For this reason, very few engineering companies actually practice Design for Casting, including those who have captive foundries. Virtual prototyping of not only the casting process, but also the entire business has also been proposed [8], however any practical implementation of any such systems has not been reported so far. Two separate initiatives in 1996 helped understand the needs of casting engineers. The first was the Foundry Benchmarking Survey conducted by the American Metalcasting Consortium and sponsored by the Department of Defense [2]. The survey team was led by Prof. R.C. Creese at the West Virginia University. This study showed that the average lead time for the first article of approval is about 10 weeks for aluminum and steel foundries, 11 weeks for ductile iron and 14 weeks for gray iron foundries. Tooling development emerged as the most important factor, taking up as much as 80% of the total lead time. This showed that significant savings in overall lead time could be achieved by compressing the tool development time alone. The second initiative was the setting up of a web site called Casting 2000 at the Concurrent Engineering Research Center in Morgantown, West Virginia during 1996-97. It described a conceptual framework for assisting and linking casting life-cycle engineers (product designers, tool makers, foundry engineers and supply managers) for better and faster decision making. The
web site received encouraging feedback from all parts of the world, and the suggestions helped in designing the information backbone of a future software system. A prototype of the system, named AutoCAST, was developed by Advanced Reasoning Technologies Pvt. Ltd., with technical inputs from the author, Professor Creese and other experts, and shown to casting engineers [9]. Their suggestions were incorporated in the software to improve it further in terms of features, programs, database structure and user interface. 3 Intelligent Assistant for Casting Designers AutoCAST simulates the way casting engineers perform various tasks in casting design. The present work integrates the latest research work in parting line and cores [10], feeding [11], gating [12] and information management [13]. The entire casting design is analyzed using a set of castability criteria [14]. Depending on the results of the analysis, suitable design changes to improve castability are suggested through guidelines [15]. The software runs on Pentium computers with Windows-NT or Windows-95/98 systems. It comprises nine integrated programs for various tasks, linked by a casting project database manager. These enable casting design, castability assessment and concurrent engineering for the projects. Casting Design The design of the casting, which involves several decisions such as the orientation of the part in the mold, location of the parting line, design of cores, location and design of risers and gating channels and other casting elements such as feed-aids and filters, requires prior experience. In AutoCAST, this experience has been coded as a knowledge base of design heuristics. Further, to apply these heuristics to a given part, we have developed an advanced geometric reasoning engine, which can automatically understand the shape of a cast part and suggest where to locate the parting line, riser or ingate. This helps the casting designer in speeding up his tasks. The starting point for casting design is the 3D solid model of the part to be produced. The solid model can be created in CAD packages such as AutoCAD, I-DEAS, Pro-Engineer, SolidWorks or Unigraphics and exported in a standard format. Starting from this solid model, AutoCAST generates the complete casting design through several steps, which include casting orientation, mold box, parting line, cores, risers and gating. At the end of each step, the user has the option to provide his own suggestion for the corresponding design parameters. Given a part model, one iteration of casting design usually takes 20-30 minutes on a Pentium computer. Castability Assessment To assess the influence of casting design on castability which is reflected in terms of quality assurance, costs (tooling and manufacturing) and lead time we need to simulate the casting process. The latest approaches for analyzing the casting process, which combine the physics of solidification and filling with practical heuristics, have been implemented. This gives reliable results without the need for powerful computers. Casting process analysis is supplemented with a criteria-based check of castability. This involves analyzing the shape of the part and the casting, as well as the results of solidification and mold filling analysis with a set of criteria. There are about twenty criteria, which include: part wall thickness, parting flatness, core length to diameter ratio, feeding efficiency, fettling of risers, difference between ideal and actual filling time, and others. Each criterion returns 100 if the casting design is ideal for the specific aspect, 0 if the casting cannot be produced, or an intermediate value. This provides a quantitative measure of the health of the casting design for a set of aspects. Further, weights can be attached to the criteria and a cumulative assessment of the casting design can also be obtained. Based on the results of criteria-based castability analysis, the software suggests illustrated guidelines for improvement. Each guideline comprises a pair of pictures, showing the problem and a possible solution, along with a written explanation.
Concurrent Engineering Concurrent engineering, which has emerged as a team-driven approach to dramatically reduce lead time, costs and defects, requires a systematic approach for better communication, cooperation and collaboration between casting buyers and suppliers. This can be achieved by a casting project data management system included in AutoCAST. Over 2000 different pieces of information about a casting project, starting from administrative information such as customer address and order details, to information about the product, tooling (parting, cores, risers, gating, etc.), process (activities and tasks), materials and quality are systematically stored in a database. Some of the data is input by the user or copied from libraries (metal properties, mold box sizes, etc.) and other data is generated by the software during execution. Any part of the project database can be exported into a separate file and attached to emails. This facilitates better and faster communication between casting buyers and suppliers. They can engage in a constructive dialogue to plan tasks in advance, quickly inform about any changes, identify cost drivers and adjust the product design and process parameters for an overall cost reduction. 4 Casting Design - An Example An example session of casting design and analysis using AutoCAST is described here. The casting is an aluminum alloy unloader valve. The valve was first modeled using Pro-Engineer software (from Parametric Technologies Corp, USA) and exported as a standard exchange file in STL format. A new casting project was started in AutoCAST with preset default values. The part STL model was imported and the alloy properties were copied from the library. An appropriate mold box was selected and modeled. The part is displayed as a shaded image inside a transparent mold in the main window and the computed properties are shown in the database window. The program suggested gravity die casting as the most appropriate process for the given casting alloy, geometric characteristics, quality specifications and lot size. Based on the capabilities of the process, the part features were analyzed. The criteria for complexity and wall thickness returned poor assessments and the guideline suggested increasing the minimum wall thickness. The casting orientation and parting line suggested by the program were accepted. Cored features were automatically extracted and modeled. Minimum core diameter was found to be adequate during the analysis of cores. Solidification analysis revealed two isolated hot spots inside the casting. After feeder design and modeling, it was found that one of the hot spots is difficult to remove even with a large sized feeder. The actual casting experienced severe shrinkage defects at the same location. Since the hot spot is caused at the junction of thick sections and cores, coupled with the difficulty in reaching the defect region to apply either feeders or feedaids, modifications to part design were found necessary. The gating design was done as suggested by the program; the filling simulation showed the sequence of filling the mold and a narrow difference between the optimal (designed) and actual filling time. 5 Conclusion Design for Manufacture (DFM) has been successfully applied to machined, sheet metal and plastics parts to significantly reduce manufacturing costs and lead time. So far, it did not find wide acceptance in the casting sector owing to a lack of suitable software tool for this purpose. The AutoCAST software described in this paper aims to fill this gap. Early benchmarking results and beta tests have shown that it can be easily used by design engineers, who have limited knowledge of the casting process, to improve the design for ease of casting. This will enable product engineers to design more cost-effective parts through a better appreciation of the problems faced by casting engineers. It will also improve the level of communication between product, tooling and foundry engineers, leading to better and faster decision-making.
References 1. D. Kaniki, The Current State of Casting Yield, Transactions of the AFS, Vol.106, 1998. 2. R.C. Creese, Benchmarking and Lead Time Reduction, AMC Benchmarking Project Report, West Virginia University, Oct 1996, pp.1-137.
3. AFS, Sources of Casting Modeling Software, Modern Casting, Vol.87, No.10, 1997, pp.2729. 4. D.H. Jensen, Survey Provides Profile of Casting Design Software Use, Modern Casting, Vol.86, No.9, 1996, pp.29-33. 5. M.A. Gwyn, Cost-Effective Casting Design, Modern Casting, Vol.88, No.5, 1998, pp.32-36. 6. P. Dvorak, Manufacturing Puts a New Spin on Design, Machine Design, Vol.66, No.16, 1994, pp.67-74. 7. C.D. Orogo, H.D. Callihan, G.K. Sigworth and H.A. Kuhn, A Vision of Computer-Aided Casting in the Year 2000, Modern Casting, Vol.83, No.10, 1993, pp.20-24. 8. G.F. Ruff, Metalcasting - Virtual Reality and Strategies for Growth, Modern Casting, Vol.88, No.7, 1998, pp.62-69. 9. B. Ravi and R.C. Creese, Collaborative Design and Manufacture of Cast Products, Proceedings, 45th Indian Foundry Congress, Mumbai, January 1997. 10. M. Nainy-Nejad, R.C. Creese and B. Ravi, Research Issues in Computer-Aided Parting Design, Transactions of the AFS, 105, 1997. 11. Ravi and M.N. Srinivasan, Casting Solidification Analysis by Modulus Vector Method, International Cast Metals Journal, Vol.9, No.1, 1996, pp.1-7. 12. B. Ravi, Intelligent Design of Gating Channels for Casting, Materials Science and Technology, Vol.13, No.9, 1997, pp.785-790. 13. B. Ravi and M.M. Akarte, Casting Information Management, Transactions of the AFS, Vol.104, 1996, pp.217-223. 14. B. Ravi and M.N. Srinivasan, Features-Based Castability Evaluation, International Journal of Production Research, Vol.33, No.12, 1995, pp.3367-3380. 15. B. Ravi, Knowledge-Based Casting Design, Indian Foundry Journal, Vol.42, No.10, 1996, pp.7-17.
Fig.1 Casting design and analysis for an aluminum valve using AutoCAST.