JP2006268422A - Analysis navigation system - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To solve the problem that it is difficult for a beginner to set a proper analysis condition meeting an individual problem even though it is necessary to set the analysis condition in order to elicit a significant answer for various analysis objects in a development field. <P>SOLUTION: In a configuration comprising an analysis work procedure storing means 8, a work guiding means 9, an information displaying means 10, an inputting means 11, a CAE analyzing means 12, an analysis example storing means and an example referring means 14, even the beginner can determine convergence of analysis at a glance by a convergence determination index 15 attached to an analysis result, uses the convergence determination index 15 to classify/examine stored analysis examples and examines advantages and disadvantages of a convergence state and the trend of the analysis condition to thereby be able to correct the analysis condition of an analysis problem with bad convergence. <P>COPYRIGHT: (C)2007,JPO&INPIT

Description

  The present invention makes it easy for anyone to use CAE analysis work such as computer analysis of equipment, for example, structural analysis for obtaining strength and vibration characteristics of a structure, and thermal fluid analysis for obtaining airflow distribution and temperature distribution in a room. It is related with the analysis navigation system for doing.

  When using CAE analysis technology in product design and R & D, not only designers and researchers themselves, but analysis experts use complex simulation software to perform calculations such as strength analysis and fluid analysis. . This is because in CAE analysis, complex physical phenomena of mechanics and fluid are expressed by differential equations, and based on that, it is a mechanism to obtain numerical answers (solutions) by advanced mathematical methods. . In other words, in order to make full use of CAE analysis, understanding of differential equations and understanding of mathematical promises required for numerical analysis are required, so not everyone can easily perform CAE analysis.

  However, in order to analyze a physical phenomenon, understanding the physical phenomenon itself is as important as or better than understanding a numerical analysis method. This is because the current CAE analysis is only an approximate calculation and does not faithfully reproduce the reality, so that the work of modeling the phenomenon becomes indispensable. This may be solved by significant progress in both hardware and software of the computer, but at present it cannot be realized in units of several years.

  Therefore, it is necessary to make it possible for engineers who are not experts in CAE analysis, that is, designers who are familiar with physical phenomena but are not familiar with mathematical numerical analysis methods, to easily use CAE analysis. It is. In general, not only CAE analysis but also improvement of operability in various analyzes is a technique necessary for improving work efficiency, and various techniques for improving operability have been proposed. (For example, refer to Patent Document 1). The technique disclosed in this document stores and saves an analysis operation procedure, and automatically performs analysis based on the stored contents. Thus, once the analysis operation procedure is established and saved, the trouble of repeating the same operation procedure every time the analysis is performed is eliminated, leading to an improvement in the operability of the analysis.

  However, in this technology, the determination of the first operation procedure has to be performed by an expert as before, and it was not easy for everyone to use from the beginning. Therefore, an analysis system as shown in FIG. 5 has also been proposed so that even a designer who does not have specialized knowledge of CAE analysis can easily perform CAE analysis from the beginning (for example, a patent). Reference 2). This is a CAE analysis system in which an interactive application program that defines a target product for CAE analysis is stored, and is stored in the field work procedure and design know-how file storage means 2 stored in the field work procedure file storage means 1. A design that has an input screen display means 3 and an input means 4 for displaying a design information input screen according to the design know-how of the stored target product, and that is input according to the design information input screen between the CAE procedure file storage means 6 An interface unit 5 is provided in which the procedure is replaced with a CAE procedure and transferred to the CAE procedure file storage unit 6.

The CAE procedure file replaced with the CAE procedure by the interface unit 5 and transferred to the CAE procedure file storage unit 6 starts the CAE analysis unit (CAE engine) 7 to execute the CAE analysis. That is, the operator can perform the CAE analysis only by inputting in accordance with the familiar field work procedure.
JP-A-10-303267 JP 2001-290846 A (page 3, right column, lines 16 to 44, FIG. 1)

  However, in this conventional technique, design information input in a routine manner and a routine CAE analysis are linked only to a specific design work for which a field work procedure has been established. There is a problem that beginners cannot use a wide variety of CAE analyzes required in design and research and development with many elements.

  For beginners, it is also effective to determine analysis conditions with reference to past analysis cases in order to set appropriate analysis conditions for various analysis targets in the actual development site and derive meaningful answers. Means. However, a database of analysis cases that can be easily used is not established, or even if analysis results are simply accumulated, there are problems such as convergence judgment unique to numerical analysis, and the analysis conditions for the analysis results are appropriate It is difficult for beginners of CAE analysis to determine whether or not, and it is difficult to utilize past analysis cases in order to select optimal analysis conditions.

  In order to achieve the above object, the analysis navigation system of the present invention describes an analysis work procedure storage means storing an interactive analysis work procedure using CAE analysis software, and an analysis work procedure based on the analysis work procedure. Work guidance means, information display means for displaying procedure explanations and interactive analysis work, input means for inputting design information and analysis information necessary for analysis work, and input analysis information CAE analysis means for executing analysis calculation based on the above, analysis case storage means for storing information related to the analysis performed based on the analysis work procedure, and extracting specific analysis information from a plurality of stored analysis cases, Display case reference means for displaying, and the analysis case storage means calculates at least one convergence determination index representing the convergence status of the analysis calculation and adds it to the analysis information In both cases, the case reference means cooperates with the work guidance means to determine and classify the convergence of the stored analysis cases using the convergence determination index, and compare analysis information of a plurality of analysis cases. Is displayed.

  With this configuration, by adding a convergence determination index indicating whether or not the analysis is successfully converged with respect to the analysis result, the validity of the result can be determined at a glance when a beginner views the data. Furthermore, by classifying the stored analysis cases using the convergence judgment index and comparing the analysis conditions, it is possible to grasp the condition of the convergence condition and the tendency of the analysis conditions. Appropriate analysis conditions can be set for new analyzes that are not available.

  As described above, according to the analysis navigation system of the present invention, since a convergence determination index is added when analysis cases are accumulated in the analysis case storage means, it is easy even for beginners of CAE analysis by looking at the convergence determination index. It is possible to determine the success or failure of the convergence of the analysis result, that is, it is possible to immediately know whether or not the calculation result is meaningful as an analysis.

  Furthermore, by classifying the stored analysis cases using the convergence judgment index and comparing the analysis conditions, it is possible to know the relationship between the convergence and the analysis conditions. In contrast, it is possible to select an appropriate analysis condition with good convergence.

  Also, in the case of fluid analysis, by using evaluation index values that combine mesh quality, mesh size, flow velocity, calculation time increment, etc., whether the poor convergence is due to mesh quality or mesh size In addition, it is possible to easily determine whether it is caused by calculation time increments.

  Embodiments of the present invention will be described below with reference to the drawings.

(Embodiment 1)
FIG. 1 is a conceptual configuration diagram of an analysis navigation system according to Embodiment 1 of the present invention.

  In FIG. 1, 8 is an analysis work procedure storing means for storing an interactive analysis work procedure using CAE analysis software, 9 is a work guide means for explaining the analysis work procedure based on the analysis work procedure, and 10 is a description of the procedure. Information display means for advancing the analysis work in a display or interactive format, 11 is an input means for inputting design information and analysis information necessary for the analysis work, and 12 executes analysis calculation based on the input analysis information. CAE analysis means, 13 is an analysis case storage means for storing information related to the analysis performed based on the analysis work procedure, and 14 is a case reference means for extracting and displaying specific analysis information from a plurality of stored analysis cases. It is. Furthermore, 15 is a convergence determination index for determining the degree of convergence determined from the analysis result when storing the analysis example.

  Next, the flow of the CAE analysis work will be described using the flowchart of FIG. 2 and the operation of the analysis navigation system will be described.

  The CAE analysis work starts with the creation of a shape model 16 when the object and purpose of analysis are determined. For example, when strength analysis or vibration analysis of a structure is performed, the shape of the structure is created using a general-purpose CAD or a modeler dedicated to analysis. In the case of fluid analysis, it is necessary to create a shape model of a space portion where a fluid flows, not a solid portion. This shape model creation is also instructed by the work guidance means 9 of the present analysis navigation system, but recently, specific work is often performed using general-purpose 3D-CAD, and the shape model is included in the analysis system. Creating is less.

  Next, mesh division 17 is performed. Since the CAE analysis discretizes a continuum such as a solid or a fluid and performs mathematical processing, the work of mesh division 17 that divides the analysis region into small regions is necessary. A hexahedron or tetrahedron is often used as a mesh that divides the volume into small regions, but there are several types other than these, and they are often used in combination. Recently, the mesh division 17 work itself can be processed fully or semi-automatically using a dedicated tool, and the work itself is possible even if it is not a dedicated analysis expert with special know-how. Yes. Therefore, if a mesh creation tool is activated in accordance with instructions from the work guide means 9 and a shape model file prepared in advance is read, mesh creation can be performed almost automatically. Of course, it is also possible to select the type of mesh or to partially specify the size of the mesh.

  When the mesh division 17 is completed, the boundary condition setting 18 is performed. This prescribes the physical constraint conditions in the analysis area, such as the fixing method of the structure, the point and size of the external force, and the inflow / outflow speed of the fluid. Ingenuity has been made. In this analysis navigation system, the work guidance means 9 presents these various boundary conditions to the worker, and the worker selects the necessary boundary conditions and sets conditions by inputting numerical values such as force and flow velocity.

  Subsequently, initial condition setting 19 is performed. The mathematical method used for CAE analysis is to express various physical phenomena as partial differential equations, and to discretize the equations and solve them numerically, and to specify the state at the start of the calculation, It is necessary to give initial conditions. In normal steady-state analysis, if the above boundary conditions are set appropriately, it does not matter much because it converges to a constant solution regardless of how the initial conditions are given. Various ideas may be made. Of course, in the case of non-stationary analysis, an appropriate setting according to the analysis purpose is necessary. In this analysis navigation system, the work guidance means 9 prompts the user to set initial conditions, and inputs numerical values according to the instructions. In the case of steady analysis, a method in which default values are set in advance and the operator is not aware of the initial value setting is also possible.

  The above boundary conditions and initial conditions are conditions that represent actual physical phenomena and must be determined by the operator according to the contents to be analyzed. The next calculation control condition setting 20 is a partial differential equation. This is a condition setting for the mathematical method for solving In other words, if these conditions are automatically determined without the operator being conscious, it is very convenient to use. The conditions set here vary depending on the CAE analysis software used, and cannot be said to be uniform. In case of non-stationary calculation, the calculation time step size is set. In this analysis navigation system, by setting default values for these calculation control conditions in advance, the burden on the operator can be reduced. However, these default values may not be valid in all cases, so there may be cases where the calculation diverges or terminates abnormally, or when the results end normally but the results are not expected, and corrections described later are necessary. Become.

  After performing the various condition settings as described above, the calculation execution 21 is started. The calculation of CAE analysis is roughly divided into normal termination and abnormal termination. The normal end is a state in which the calculation is executed to the end according to the programmed calculation processing procedure, and the abnormal end is a state in which an unexpected state occurs during the calculation and the calculation cannot be continued, and the process is ended. When the process ends abnormally, a message is usually displayed from the analysis software, and the occurrence of the abnormal end can be easily known.

  On the other hand, in the case of normal completion, the calculation itself is normally completed, but it is necessary to check whether the mathematical error is small enough to be ignored. Usually, in order to judge such convergence, physical quantities such as flow velocity and pressure at a predetermined position (monitor point) are often displayed in a graph during calculation. Looking at this graph, it can be determined that the value is sufficiently converged if the value is constant. In this analysis navigation system, the operator can be instructed through the work guidance means 9 for setting of such a convergence determination method.

  If it is confirmed by such convergence judgment that the calculation result is correct, various post-processing 23 are utilized to visualize and analyze the phenomenon, to elucidate the phenomenon in question and make use of the knowledge obtained. Make a new design. Usually, since the analysis result is also used for the subsequent examination, the analysis result storage 24 is performed as a general procedure. In this case, analysis results that are not sufficiently converged may be stored for reference. Here, in the present analysis navigation system, the convergence determination index 15 is calculated as a characteristic value representing the convergence described above, and is stored together with the analysis result. The convergence status varies depending on the phenomenon to be analyzed, and since it cannot be determined unconditionally, it is desirable to make a judgment by graphing the monitor point value as described above. In order to judge the sex, it is easier to understand the convergence value more generally. Thus, by defining the convergence determination index 15 and storing it together with the analysis result, any worker who sees the result can easily know the convergence. Various definitions can be considered for the convergence determination index 15. For example, the convergence determination index 15 can be determined as follows using a value called a residual calculated by many CAE analysis software and a convergence determination reference value.

Convergence determination index A = (residual) / (convergence determination reference value A)
If this value is 1 or less, it has converged, but if it is 1 or more, it can be said that the convergence is insufficient. That is, the degree of convergence can be easily determined from this value.

  Alternatively, it can be expressed as follows using the physical quantity value of the nth and (n + 1) th iteration and the predetermined convergence criterion value at each calculation point in the analysis region.

Convergence determination index B = Σ ([n + 1] th physical quantity−nth physical quantity)
/ Σ ([n + 1] th physical quantity) / (convergence judgment reference value B)
Here, the symbol Σ means taking the sum at each calculation point in the analysis region. Also in this case, since the degree of convergence can be quantitatively expressed by the size of the convergence determination index B, anyone can easily make a convergence determination. Of course, in addition to these, it is also possible to define the convergence determination index 15 according to the analysis target.

  On the other hand, if the calculation diverges or the result is not desirable, some condition settings are changed and recalculated. An operator who has little analysis experience is most concerned about where and how to correct in such a case. Here, the present analysis navigation system has a function of presenting an example of appropriate analysis conditions, particularly calculation control conditions, to the operator using the past analysis cases accumulated by the analysis case storage means 13 and the case reference means 14. Have.

  This procedure will be described with reference to FIG. First, the case reference means 14 is used to extract only cases where the convergence determination index 15 is small, that is, good convergence (procedure 25). The degree to which the convergence determination index 15 is selected depends on the individual analysis software, so it cannot be generally stated. However, when the definition such as the convergence determination index A is made, “0” is selected. A value of about “.001” is often used.

  Next, the calculation control conditions are presented to the operator via the work guidance means 9 and the information display means 10 (procedure 26). If there are multiple cases, the calculation control conditions can be displayed in the form of a list. The operator compares the various calculation control conditions shown with the calculation control conditions set by the worker, and examines optimization of the conditions by paying attention to the greatly different points (procedures 27 and 28).

  In this way, the calculation control condition is corrected based on the past analysis examples, and the calculation is executed again. If there is a problem in the result, the above-described procedure is repeated again to correct the calculation condition, thereby determining a calculation control condition with good convergence.

  On the other hand, by extracting only cases with poor convergence and finding common points, it is possible to modify the condition settings for the analysis that you are doing.

  As described above, in the analysis navigation system of the present invention, when the calculation is not successful, the past analysis cases are selected and extracted using the convergence determination index 15 stored together with the analysis result, and various analysis conditions are set. Can be displayed. With reference to these condition setting examples, the worker can correct the analysis conditions that did not work by himself.

  In the above description, an example in which past analysis cases are extracted using only the convergence determination index 15 has been described. However, a case of high similarity by adding an analysis object, a fluid type, a boundary condition type, and the like as keywords. It is also possible to extract only those and determine the calculation control conditions.

  The analysis result storage 24 is usually intentionally performed by the operator. However, in order to store a large number of analysis cases, the convergence determination index 15 is automatically calculated when the analysis is completed, and the analysis is performed. It is possible to use the case where the analysis information is stored by the case storage means 13. By storing not only analysis cases with good convergence but also bad analysis cases, it becomes possible to clarify the influence of calculation control conditions on convergence.

(Embodiment 2)
FIG. 4 is a conceptual configuration diagram of the analysis navigation system according to the second embodiment of the present invention. The difference in configuration between the second embodiment and the first embodiment described above is that the analysis example storage unit 13 calculates not only the convergence determination index 15 but also a predefined evaluation index value 31 and stores it together. Is a point. Furthermore, the case reference means 32 of the second embodiment has a function that can handle the convergence determination index 15 and the evaluation index value 31 in association with each other.

  The evaluation index value 31 is an index value obtained by combining elements that determine convergence according to the type of analysis, and is represented by a combination of one or more numerical values (in this case, a numerical value corresponding to the element). The numerical value of is closely related to the convergence of the analysis. For example, in the case of fluid analysis, a combination of the following parameters can be considered.

Evaluation index value A = (mesh quality, mesh size / flow velocity)
In general, the mesh quality represents the distortion of the mesh. When the spatial mesh is a hexahedron, the mesh quality is defined as a cube. When the spatial mesh is a tetrahedron, the mesh quality is defined as closer to a regular tetrahedron. There are various defining formulas, but as a simple example, the ratio between the maximum value and the minimum value of the ridgelines constituting each mesh is used. In other words, if the ratio between the maximum value and the minimum value of the ridgelines of hexahedrons and tetrahedrons is 1, they are cubes and regular tetrahedrons, and the mesh quality is very good. Usually, when the mesh quality is 0.2 or less, the mesh is considerably distorted.

  The second mesh size is often represented by the volume of each mesh, but can also be defined by a length such as the minimum dimension on the coordinate axis.

  As the third flow velocity, a flow velocity, an average flow velocity, a maximum flow velocity, or the like at each point in the analysis region can be used.

  All of these parameters have a great influence on the convergence. The worse the mesh quality, the worse the convergence and the easier the calculation to diverge. Or, even if the divergence does not occur, the convergence determination index 15 such as a residual does not become small. In addition, the mesh size and the flow velocity influence the convergence in relation to each other, and as the ratio between the flow velocity and the mesh size increases, the stability decreases and the light tends to diverge.

  In order to obtain more detailed information, the mesh quality, mesh size, and flow rate may be divided into minimum value, maximum value, average value, and the like as evaluation index values.

  Thus, by comparing and examining the values of the parameters included in the evaluation index value 31 for various analysis cases, it can be understood what should be improved in order to enhance the convergence. That is, when the convergence determination index 15 is large and the convergence state is not good, the evaluation index value 31 is examined. If the mesh quality is poor, it is found that there is a high possibility that there is a problem with the mesh. In this case, it is necessary to recut the mesh. Alternatively, it is understood that there is a possibility that the setting of the calculation control condition may be changed when it is determined that there is no problem in the mesh quality and there is a problem in the ratio between the flow velocity and the mesh size. In other words, measures such as reducing the relaxation coefficient of the repeated calculation are considered to be effective.

  In the case of unsteady analysis, the following parameters can also be used.

Evaluation index value B = (flow velocity * calculation time step / mesh size)
The ratio between the product of the flow velocity and the calculation time step and the mesh size is a value of the same dimension as the so-called Courant number, and is an index for evaluating the mathematical stability of the calculation. The commonly used Courant number is defined for each mesh in the analysis region, but can be defined as an evaluation index value for the entire analysis region by weighting with the mesh size. If the evaluation index value B is larger than 5 when the convergence is poor in non-stationary calculation, stable analysis can be performed by changing the calculation step time and mesh size so that the evaluation index value B is 1 or less. Is often possible.

  As described above, by using the evaluation index value 31 in addition to the convergence determination index 15, whether the poor convergence is caused by the mesh quality, the relaxation coefficient of the repeated calculation, or the calculation time increment. It becomes easy to specify what to do. Instead of simply comparing the calculation control conditions and adjusting different points, the calculation control conditions can be adjusted by narrowing the point of interest by the evaluation index value 31.

  The analysis navigation system of the present invention can easily determine the validity of the convergence of analysis results even for beginners of CAE analysis using the convergence determination index added to the accumulated analysis cases. Using this analysis navigation system, workers can perform stable analysis with good convergence by extracting examples with good convergence from past analysis cases and referring to the setting conditions. .

  Also, by classifying the stored analysis cases using the convergence judgment index and comparing the analysis conditions, it is easy to know the relationship between the convergence and the analysis conditions. Even for this, it is possible to set appropriate analysis conditions and meshes.

  These features can be applied to a wide range of fields in which CAE analysis such as structural analysis and fluid analysis is performed.

Conceptual configuration diagram of an analysis navigation system according to Embodiment 1 of the present invention Standard workflow diagram for CAE analysis Flowchart of analysis condition correction procedure in Embodiment 1 of the present invention The conceptual block diagram of the analysis navigation system in Embodiment 2 of this invention Conceptual configuration diagram of a conventional analysis navigation system

Explanation of symbols

8 Analysis work procedure storage means 9 Work guide means 10 Information display means 11 Input means 12 CAE analysis means 13 Analysis case storage means 14 Case reference means 15 Convergence determination index 31 Evaluation index value 32 Case reference means

Claims (4)

Analysis work procedure storage means for storing interactive analysis work procedures using CAE analysis software;
Work guide means for explaining the procedure of the analysis work based on the analysis work procedure;
An information display means for displaying a description of the procedure and an explanation for proceeding with the analysis work interactively;
Input means for inputting design information and analysis information necessary for analysis work;
CAE analysis means for executing analysis calculation based on the input analysis information;
Analysis case storage means for storing information related to the analysis performed based on the analysis work procedure by the CAE analysis means;
A case reference means for extracting and displaying specific analysis information from a plurality of analysis cases stored in the analysis case storage means,
The analysis case storage means calculates at least one convergence determination index representing the convergence status of the analysis calculation and adds it to the analysis information.
The case reference means, in cooperation with the work guidance means, uses the convergence determination index to determine and classify the convergence of the stored analysis cases, and compares and displays the analysis information of a plurality of analysis cases Analytical navigation system characterized by The analysis case storage means calculates at least one predetermined evaluation index value and stores the evaluation index value together, and the case reference means converges with a specific evaluation index value from a plurality of stored analysis cases. The analysis navigation system according to claim 1, wherein a determination index is extracted and displayed. The analysis navigation system according to claim 2, wherein the evaluation index value is used in combination with mesh quality, mesh size, and flow velocity in fluid analysis. The analysis navigation system according to claim 2 or 3, wherein the evaluation index value is used in combination with a mesh size, a flow velocity, and a calculation time step size in the case of unsteady calculation in fluid analysis.

Images