A new approach for predictive modelling of sheet formability and one-step CAD/CAM of press tools for multiple-stage deep drawing

Student thesis: Doctoral Thesis

View graph of relations


  • Wu Man SING


Awarding Institution
Award date13 Oct 1997


Predictive Modelling of Sheet Formability In sheet metal forming, it is essential that the formed components are free from undesired thinning and fracture as a result of necking or shear. Before the design and manufacture of a die, it is customary to perform a formability study for the specified work material because the subsequent and associated tooling and production setup costs are very high. Following the successful introduction of the concept of forming limit diagram by Keeler and Goodwin in the sixties, several models have been proposed to predict the formability of sheet metals. However, these models generally involve the knowledge and determination of some conceptual parameters or the underlying instability mechanism necessitates complex computation. The aim of this study is to model the formability using a minimum number of parameters that are determinable readily through standard laboratory tests. It is specifically targeted to accommodate the influence of the mechanical properties of the sheet metal, strain history and process parameters, or a combination of these on the formability. In addition, the proposed formability model should be easily understood and be applicable to shop floor situations as well as adaptable in a computer integrated manufacturing system. The formability models generally assume that the development of surface undulations, necking, voids or slip bands before fracture is due to attainment of a critical stress, equivalent stress, critical strain, equivalent strain, work done, or deformation energy. An yield criterion with the associated flow rule and a strain hardening equation are the basic elements to formulate an ideal formability model, and such a model is hardly available till now. The selection of a formability model depends on the type of material, the properties of the material and the imposed processing conditions. However, the formability models that have been reported so far are strongly influenced by the constitutive equations being employed. By converting the forming limit strains into limit stress states, Embury, LeRoy, Anieux, and Gronostaski have demonstrated that the resultant forming limit stress curve (FLSC) or stress map can provide a better scope to study the material failure. Especially, the FLSC is independent of the influence of strain-path and ma> be closely related to the expanded yield locus. Not much work has been reported that provides any direct relationship between the FLSC and the expanded yield locus. The construction of an yield locus, regardless of its shape (either a piecewise linear yield locus or a smooth yield locus), is based on critical stress: critical strain, or defonnation energy. As the yield criterion marks the commencement of plastic flow and failure connotes its termination, it is worthwhile to speculate on the possible existence of a direct relationship between these two boundary states. There is a possibility that the FLSC can be predicted using an yield criterion and its associated flow relationship \\it11 minimum modification of the yield locus. Based on such an FLSC. the forming limit curve (FLC) for pre-strain or strain-path-dependent cases can also be predicted by adopting a relevant strain hardening equation. In this study, an FLSC is developed by associating an uniaxial localized instability criterion with Hill's expanded yield locus and linearising the expanded yield locus between the uniaxial tension and equi-biaxial tension region. It will be demonstrated that the expanded Hill's yield locus can be distorted from an elliptical locus to a rectilinear locus by bringing minor or insignificant changes in the values of straining hardening exponent (n), anisotropy (R) and/or strength coefficient (K), which are easily encountered in real forming situations. Also, a similar distorted locus may be achieved due to variations in the processing conditions of temperature and strain rate as well as changes occurring in strain rate sensitivity of the material during deformation. Four special cases have been suggested by Hill, and Hosford's yield criterion can be treated as a fifth case of the Hill's yield criterion. It will be shown that these cases have significant effects on the predicted FLC. The effect of the mechanical properties of the sheet material as well as the processing conditions on the FLC will be theorized through the proposed FLC prediction method. It is known that different strain hardening models are used to simulate the experimental results obtained on different materials. It should be expected that the strain hardening model exerts considerable influence on the prediction of FLC. The possible role of strain hardening models in the prediction of FLC will be hypothesized using the proposed model. In addition, it will be shown that pre-strain and strain-path-dependent FLC can be predicted with the use of a suitable strain hardening model. It is believed that the strain-rate hardening has a critical effect on the development of sheet metal fracture. Such effect can be demonstrated using the present model with regard to fracture strains, and the results of the model are also compared with some published experimental data. One-Step CAD/CAM of Press Tools for Multiple-Stage Deep Drawing Design and manufacture of dies for sheet metal forming involve several stages, namely, formability of the work material, process layout planning, die design, bill of materials, tolerance allocation for various die components, die production planning, scheduling, part programming etc. Until recently, several of these functions are performed manually and separately. Such a system responds slowly and is human-dependent. Incomplete information may be exchanged between different activities and inflexibility, both on the organization and staff involved. is apparent. The need for dynamic responses, fast delivery times and smooth transactions between the design and manufacturing sections necessitates the establishment of a quick die design and manufacturing system. In the current work, a system that provides almost full automation in terms of the flow of information between various phases of die development is proposed, which also relates the product features to the down stream processes right up to part programming. The work force engaged in die design and manufacture of dies is highly skilled in design, machine Operations, part programming and fitting. Traditionally, these professional staff progress through an apprenticeship to a master of their own skills. In the conventional approach to die development, first the product design with precise specifications is received from the marketing or product design department. The die designer(s) will make the process layout / strip layout together with initial die design planning based on the specifications. After consultation with the customer and various internal departments, a specific die design will be formulated. Component drawings of the die along with detailed tolerancing specifications will be issued to various related personnel and departments. Then, the inventory department will order appropriate raw or semi-finished stock of tool materials. Process planners in the machine shop will arrange the sequence of the machining operations and develop the necessary parts programs for CNC machines. Finally, the finished die components will be assembled and taken for trial run. The organization of these workshops is generally loose and decentralized. The development of a die is usually one-of-a-kind and takes a long time. Currently, this industry is confronted with a number of difficulties, namely, inability to attract or interchange the work force, ever reducing delivery time and cost, increased turn-over of manpower, upgrading to emerging technology and new concepts in manufacturing system. While adopting new technology and concepts, standard die designs and standard components are some possible solutions to reduce the processing time, they should be assessed for making proper investment on machinery. There is a belief that the whole die development process involving process layout planning, die structure design, die assembly design, tolerance allocation for individual die components and part programming for these die components can be completely computerized and integrated into a quick die design and manufacturing system. Axisymmetrical deep drawing process is chosen in this research to develop a platform for such an automated prototype system. Once the product design is provided to the system, appropriate process layout, die design, bill of materials and parts programs for non-standard die components should be generated immediately so that the die design and manufacturing methods can be studied concurrently by the customers as well as various internal departments. With a fully computerised system, the whole organization can achieve a complete horizontal and vertical integration and the difficulties mentioned earlier can be lessened. This can be achieved by blending the application of information technology with existing manufacturing methods. A great deal of practical experience, which is accumulated on day-to-day trouble shooting, can be converted into computer programs using different methodologies like knowledge-based, neural-network systems, fuzzy logic, decision table method etc. A decision table, which is similar to tables used frequently in text books or handbooks, is easy for a human to read and understand. Each column represents one rule and a new rule may be added by inserting a column. An old rule can be deleted or modified when the programmer or user finds it necessary. A processor that converts a decision table into a program can be made along with the provision of interactive screen editing or file editing. When the knowledge on die design and manufacturing can be converted as decision tables, it serves the purpose of continuous improvement. Hence, the work force can design and maintain their own rules, procedures and technical data. In the proposed system, first the available knowledge of process layout planning will be formulated as a number of decision tables, and the product features will be mapped to these decision tables. A set of highest feasible forming process layout parameters for the purpose of die design will emerge from the process layout sub-system. The product features as well as the process layout will be transferred to downstream processes. Various die structures and die design methods will be standardised using group technology and computerised using parametric method. The product features and process layout parameters can be directly linked and mapped to these die structures and die designs. After a die design is determined, the relevant information will be passed to the die-material selection sub-system. The material selection will depend on the functions of the die component, the material properties, manufacturing cost, production quantity and precision requirement on the product. The next step is to allocate tolerances for all the die components based on the functions of the die components and specifications of the product. The proposed tolerance allocation sub-system is an important bridge between the die design and the process planning stages for individual die components. It resides invisibly inside the computer integrated die development environment to achieve a true paperless CAD/CAM system. Agreed international standards on tolerances and fits are to be closely followed. An algorithm will be proposed to analyse the dimensional relationships of the die components with the accuracy requirement of the products, to ensure required robustness in the die. The relevant rules and parameters will be contained in a series of decision tables/decision matrices. A simple precision die is built up of a punch, a die and a number of die plates. A series of holes are to be machined in the plates. These holes can be round, irregular and threaded holes. Analysis showed that there are about 20 to 30 basic hole shapes that are mostly used in the die plates. When some standard codes are assigned to these particular hole shapes, the corresponding part programs to produce these holes can be saved in a computer database. When a code is selected by the user, the part program can be restored and generated automatically. Lastly, it is proposed that the selection of these codes can be done automatically based on the product features. the die assembly design and the tolerance allocation. According to the selection codes, CNC part programs for the die plates can be generated automatically. When the mapping between product features and the features from the various design processes is done using rules or decision tables, it may be possible to completely avoid expensive feature recognition and extraction methods (involving graphic systems) that are used for universal situations. The proposed methodology leads to the possibility of generating CNC part programs as soon as the sheet-metal product is recognized by a quick die design and manufacturing system.

    Research areas

  • Sheet-metal work, Formability, Metals