Patent Application
20160034614
The present invention relates generally to the predicted mechanical properties of cast components and, more particularly to systems, methods, and articles of manufacture to provide an integrated computational way to generate thermodynamic, thermal-physical, and mechanical material properties for cast aluminum alloy components based on the property requirements for such components.
Many critical structural applications utilize cast components or products. This is especially true for automotive and related transportation systems, where engines, transmissions, suspension systems, load-bearing primary structures, seating components, interior support structures or the like have all benefited from the low-cost manufacturing associated with casting. Casting processes are often the most cost effective method to produce geometrically complex components and offer net shape or near net-shape capability in comparison with other manufacturing processes. Such casting processes are especially beneficial when used in conjunction with lightweight structural materials, such as aluminum-based alloys, where high strength to weight ratios, good corrosion resistance, and relatively low raw material cost are useful design parameters.
Relatively recent advancements in computer-based tools have enabled improvements in component design for components made through casting. Computer aided engineering (CAE)—which may also include computer-aided analysis (CAA), computer aided design (CAD), computer aided manufacturing (CAM), computer-aided planning (CAP), computer-integrated manufacturing (CIM), material requirements planning (MRP) or the like—can be utilized to not only predict how to design and manufacture a complex cast component, but also predict how the component will perform in its intended operating environment.
Efforts have been made to integrate some of these traditionally discrete, independent disciplines as a way to reduce long casting development cycles, as well as improve casting quality, reliability and other indicia of component integrity. One such effort is known as Integrated Computational Materials Engineering (ICME), which focuses on employing computer-based tools to improve the development of cast components by linking processes and structures to their corresponding properties to computationally simulate component performance prior to undertaking any actual fabrication-related activities. Despite the advantages associated with ICME and related approaches, initial simplifying assumptions must still be made with regard to casting design, process modeling and optimization, as well as prediction of defects, microstructure and product performance. Particularly problematic is that certain properties (for example, the material properties) are conventionally assumed to be substantially uniform through the object being simulated. Unfortunately, many such objects do not exhibit such uniformity in their material properties, especially those where highly complex shapes or significant differences in component thickness are present. For example, automotive engine blocks have numerous thick and thin regions that hamper the ability to assess material properties and accurately conduct related durability and life prediction analyses. Neglecting the effect of material property variations arising out of particular casting configurations manifests itself in inaccuracies in casting process simulations, including the determination of long-term component durability predictions.
As such, systems, methods and articles of manufacture to accurately account for material properties of casting process simulation are lacking. Likewise, CAE and related analysis methods used to conduct durability analyses for cast aluminum components could be improved based on a better prediction of these underlying material properties.
The present invention enables more accurate prediction of material properties that can be used in casting process simulation studies. The present invention allows a modeler to combine properties from various databases—including, but not limited to, a material property database, a thermodynamic database, and a defects and microstructure database—with various integrated modules to predict the properties of a selected aluminum-based material that will be used in a casting operation to manufacture a particular component.
According to an aspect of the present invention, a device for predicting properties of a material used in a cast aluminum component is disclosed. The device includes computational elements made up of a data input, a data output, one or more processing units and one or more data-containing and instruction-containing memories that are cooperative with one another through a data communication path. Various functional (i.e., computation) modules are configured to be programmably cooperative with one or more of these computational elements such that upon receipt of data pertaining to one or more of the component, casting process and material being modeled, the device subjects the data to the functional modules in order that generated output data provides performance indicia of the material selected for the particular component and process. The modules include at least, but not limited to, (1) a thermodynamic phase calculation module, (2) a thermal-physical property module, (3) a mechanical property prediction module and (4) a materials selection/alloy design module.
According to another aspect of the present invention, an article of manufacture is disclosed. The article includes a computer usable medium with computer readable program code embodied therein for a plurality of modules programmably cooperative with one another to generate various material (including thermodynamic, thermal-physical and mecahnical) properties of an aluminum-based alloy for use in one or more of casting design, casting process simulation and CAE nodal property mapping and durability analyses for a particular cast component being modeled. The modules are similar to those discussed above in conjunction with the previous aspect.
The following detailed description of specific embodiments can be best understood when read in conjunction with the following drawings, where like structure is indicated with like reference numerals and in which:
The embodiments set forth in the drawings are illustrative in nature and are not intended to be limiting of the embodiments defined by the claims. Moreover, individual aspects of the drawings and the embodiments will be more fully apparent and understood in view of the detailed description that follows.
Referring first to
Referring with particularity to
It will also be appreciated by those skilled in the art that there are other ways to receive data and related information besides the manual input approach depicted in input 120 (especially in situations where large amounts of data are being input), and that any conventional means for providing such data in order to allow processing unit 110 to operate on it is within the scope of the present invention. As such, input 120 may also be in the form of high-throughput data line (including the internet connection mentioned above) in order to accept large amounts of code, input data or other information into memory 140. The information output 130 is configured to convey information relating to the desired casting approach to a user (when, for example, the information output 130 is in the form of a screen as shown) or to another program or model. It will likewise be appreciated by those skilled in the art that the features associated with the input 120 and output 130 may be combined into a single functional unit such as a graphical user interface (GUI), such as that shown and described in conjunction with an expert system in U.S. Pat. No. 7,761,263 entitled CASTING DESIGN OPTIMIZATION SYSTEM (CDOS) FOR SHAPE CASTINGS that is owned by the Assignee of the present invention and incorporated herein by reference.
In one form, input into the computer 100 may be through numerous databases, including one for alloy compositions and designation database 600, a thermodynamic database 700 and a materials property database 800. These databases and their cooperation with the various modules will be discussed in greater detail below. Two additional modules—defect & microstructure module 900 and casting process simulation module 1000—are configured to operate independently from the computational modules 200, 300, 400 and 500 of the present material property predictor system. Their purpose is to provide detailed information on defects and microstructure (such as dendrite arm spacing (DAS)) to the mechanical property module 500 that is discussed in more detail below. Details of the casting process simulation module 1000 and the defect & microstructure module 900 have been disclosed in two prior patents owned by the Assignee of the present invention and incorporated herein by reference: U.S. Pat. No. 8,355,894 entitled METHOD FOR SIMULATING CASTING DEFECTS AND MICROSTRUCTURES OF CASTINGS and U.S. PAT NO. 8,655,476 entitled SYSTEMS AND METHODS FOR COMPUTATIONALLY DEVELOPING MANUFACTURABLE AND DURABLE CAST COMPONENTS. Within the present context, the integration among the various modules 200 through 500 takes place in conjunction with input received by one or more of the aforementioned databases 600 through 800, as well as the external modules 900 and 1000. An example of such interaction is shown by the connecting arrows between the modules, where the thermal-physical property module 300 (discussed in more detail below) can receive data from the computer input 120 for data that corresponds to the chosen material from database 600, as well as exchange data with the thermodynamic calculation module 200.
The first of the functional modules is the thermodynamic calculation module 200. In one form, the thermodynamic phase fractions and phase diagrams of module 200 are calculated using the known calculation of phase diagram (CALPHAD) method, where inputs from the alloy compositions and designation database 600 and thermodynamic database 700 also include solidification (i.e., cooling rate) conditions. Significantly, unlike conventional thermodynamic approaches that only deal with equilibrium and partial non-equilibrium conditions, module 200 incorporates a third solidification condition (i.e., non-equilibrium) capable of performing solid back diffusion calculations as a way to predict actual phase fractions and phase diagrams in real casting conditions. In this way, equilibrium (lever rule) solidification assumptions—which hold that the solid-liquid interfaces move infinitely slow such that the compositions of the solid and liquid phases are uniform and always have the equilibrium compositions such that the diffusion coefficients are infinitely large in all phases so that the compositions of the solid and liquid phases at any temperature correspond to those given by the phase diagram—can by the present invention now be adjusted to account for non-equilibrium conditions in the actual casting. Likewise, the Scheil model normally refers to solidification of an alloy under partial non-equilibrium conditions in such a way that no diffusion occurs in the solid phase while exhibiting complete diffusion in the liquid phase. The assumptions made in the Scheil model are (in addition to no diffusion in the solid and complete diffusion in the liquid (uniform liquid composition)), local equilibrium at the solid/liquid interface, planar interface with negligible undercooling and no density difference between liquid and solid. The present inventors have determined that the actual solidification process is neither equilibrium nor partial non-equilibrium, noting with particularity that there is diffusion in the solidified metal, and moreover that the density is also different between the liquid and solid in the solidifying interface. The present solid back diffusion that is taken into consideration in module 200 corrects the simplifications made in the lever rule and Scheil models discussed above.
The thermodynamic database 700 of
Referring next to
Referring with particularity to
C
Lj*(L−xs)+∫0x
shows that features not accounted for (or improperly accounted for) in the underpredicting Scheil model S and the overpredicting lever rule model LR can be considered. In the equation, CLj* is the element j concentration in liquid at the solid/liquid interface, CS j is the element j concentration profile in solid, C0 j is the element j concentration in bulk material, L is the total length of the volume element which is half of the DAS, xs is the length of the volume element solidified and dx is the solid/liquid interface advanced during each time step. More accurate casting simulation is made possible because assumptions associated with each approach are combined to preserve the best attributes of each, while removing or reducing the negative externalities associated with such assumptions. For example, in the lever rule approach, it is assumed that there exists infinite diffusion in both liquid and solid, although in reality such infinite diffusion is never possible Likewise, in the Scheil approach, it is assumed that there is no diffusion in solid (which is not entirely accurate, either). The present inventors' back diffusion assumption takes into consideration a limited (finite) diffusion in the solid.
The comparison of the solute content evolution in the aluminum matrix during solidification shown—a expected—reveals that the lever rule model LR predicts high and uniform solute content in solid even from the start of solidification. At the end of solidification, the solute is uniform across the whole casting and there is no segregation. As stated above, this is never the case in practice. For the Scheil model S, the predicted solute content is lower in the first solidifying aluminum matrix and more in the final part; this too has been proven to be wrong in practice. The predicted solute content in the solidifying matrix by the back diffusion model BD is somewhere between lever rule and Scheil models LR, S; the present inventors have found that the predicted solute content profile using this approach is very close to reality.
The second of the functional modules is the thermal-physical property module 300. Referring next to
The following table highlights some of the thermal-physical properties that are generated as part of the module 300.
Significantly, in a KNN classification, the output is a class membership. An object is classified by a majority vote of its neighbors, with the object being assigned to the class most common among its k nearest neighbors (where k is a positive (and typically small) integer). In situations where k=1, then the object is simply assigned to the class of that single nearest neighbor. In a KNN regression, the output is the property value for the object. This k value is the average of the values of its k nearest neighbors. Likewise, the “Best ARE” column is the averaged relative error, while the column “Best Method” means for each thermal physical property there is one best method (either Weighted KNN or Basic KNN). In addition, with regard to “Weighted KNN” method, both for classification and regression, it can be useful to weight the contributions of the neighbors, so that the nearer neighbors contribute more to the average than the more distant ones. For example, a common weighting scheme consists in giving each neighbor a weight of 1/d, where d is the distance to the neighbor.
The third of the functional modules is the materials selection or alloy design module 400. This module offers the capability to select the alloy and related casting process based on the targeted mechanical and thermal physical properties at both room and elevated temperatures, as well as between one of optimized aluminum alloy compositions and target/required physical and mechanical properties. The mechanical properties include at least tensile and fatigue properties. The thermal properties include at least density, thermal conductivity, specific heat, coefficient of thermal expansion, Young's modulus or the like. The selection of the alloy to meet the targeted properties is accomplished by using intelligent searching engine. In the present context, an intelligent searching engine uses expert system technology to provide needed information from the knowledge database. One example of such a system is an inference engine which is a tool from the field of artificial intelligence, where the knowledge base stored facts about the subject and the inference engine applied logical rules to the knowledge base and deduced new knowledge. The iterative nature of the process allows additional rules within the inference engine to be triggered. Moreover, inference engines may work primarily in one of two modes: forward chaining and backward chaining, where the former starts with the known facts and asserts new facts and the latter with goals from which it works backward to determine what facts must be asserted so that the goals can be achieved. An example of the use of such forward chaining to perform casting design may be found in aforementioned U.S. Pat. No. 7,761,263. In one preferred form, the present inventors have determined that alloy selection and design in the present invention may also take advantage of the forward chaining method.
Referring next to
The fourth of the functional modules is the mechanical property module 500. The global uniform mechanical properties are predicted based on the materials property database 800 from various sources such as known material property handbooks; such information may be provided by the alloy compositions and designation database 600 discussed above. In contrast, the local mechanical properties may be calculated by taking into consideration multi-scale defects and microstructures on a node-by-node basis; information may come from the defects & microstructure module 900. The nodal-based multi-scale defect (for example, porosity) and microstructure (for example, DAS) information is needed to establish the localized material property prediction. Module 500 can either search for material properties from the materials property database 800 for a given alloy (composition) provided by input from the alloy compositions and designation database 600, or perform nodal property calculations for each node based on information taken from the defect & microstructure module 900 and alloy compositions and designation database 600. It should be noted that the searched material properties will be generic and uniform property data.
In addition to the input from the defects & microstructure module 900, module 500 receives input from the casting process simulation module 1000 (also called casting modeling, casting simulation or the like) such that the detailed mold filling and solidification processes are simulated. The velocity, thermal and pressure information calculated during casting process is used for prediction of defects and microstructure. The casting process simulation module 1000 may be in the form of numerous commercially-available software packages, including MAGMA, ProCAST, EKK, WRAFTS, Anycasting or the like. Such software typically has several modules that can simulate casting mold filling, solidification, core molding (blowing) and related functions, which combine to determine the distribution of defects and microstructures in a casting. The casting simulation is also configured to deliver nodal numbers as well as their corresponding nodal coordinates (for example, x, y and z coordinates from a Cartesian coordinate system) to one or more of the modules 200 through 500.
Referring with particularity to
where σa represents the applied stress or fatigue strength at a given life cycle, σl represents the infinite life fatigue strength, C0 and C1 are material-dependent empirical constants, aECD is an equivalent circle diameter of a defect or pore formed in the casting, Nf is fatigue life, UR(aECD) is a crack closure correction and Keff th is an effective threshold stress intensity factor of a material used in the casting. It will be appreciated by those skilled in the art that exemplary coefficients and constants (not shown) may be used in conjunction with the fatigue life model. The specimens tested (shown as the geometric shapes corresponding to squares, diamonds and circles) include those respectively with and without skin, as well as an engine block bulkhead region; comparable modeled material property predictions are shown with solid line and two different dashed lines.
In one form, the nodal mapping and calibrating function (sometimes referred to herein as MATerial GENeration, or MATGEN) includes reading the node number and corresponding nodal coordinates (such as the aforementioned {x, y, z} coordinates in a Cartesian system) of the cast aluminum component of interest; details of this system may be found in U.S. Pat. No. 8,666,706 that is incorporated herein by reference and owned by the Assignee of the present invention. Such a material property generation program can read in (or otherwise accept, such as in text format) nodal level values from a casting process simulation software (such as the one or more of the ones mentioned above) that may include routines to consider the casting defects & microstructure module 900. Thus, upon generation of the localized (i.e., node-by-node) material properties that include the effects of porosity and DAS, module 500 can output the information for subsequent designer or modeler use. In one preferred form, the nodal mapping and calibrating function of MATGEN may be used in conjunction with the present invention, in particular being a part of module 500 as well as the substantial entirety of modules 900 and 1000. In a more preferred form, the nodal-based property calculations are actually performed by MATGEN.
Referring again to
In summary, specific attributes of the present invention include multiple abilities, including the ability to (1) integrate all of the prediction capabilities into a single computational platform, (2) take solid back diffusion into consideration when conducting phase calculations, (3) employ a k-nearest neighbor model for the module used to make thermal-physical property calculations, and (4) generate local mechanical property (including multi-axial fatigue, etc) data in order to (5) optimize the selection of a material for a particular component.
It is noted that recitations herein of a component of an embodiment being “configured” in a particular way or to embody a particular property, or function in a particular manner, are structural recitations as opposed to recitations of intended use. More specifically, the references herein to the manner in which a component is “configured” denotes an existing physical condition of the component and, as such, is to be taken as a definite recitation of the structural factors of the component. Likewise, for the purposes of describing and defining embodiments herein it is noted that the terms “substantially”, “significantly” and “approximately” are utilized herein to represent the inherent degree of uncertainty that may be attributed to any quantitative comparison, value, measurement, or other representation, and as such may represent the degree by which a quantitative representation may vary from a stated reference without resulting in a change in the basic function of the subject matter at issue.
Having described embodiments of the present invention in detail, and by reference to specific embodiments thereof, it will be apparent that modifications and variations are possible without departing from the scope of the embodiments defined in the appended claims. More specifically, although some aspects of embodiments of the present invention are identified herein as preferred or particularly advantageous, it is contemplated that the embodiments of the present invention are not necessarily limited to these preferred aspects.
