Design support device and design support method

Abstract

To enable a user of a design support device to easily confirm the risk of a design when design change or the like is performed, and shorten an examination period of the design change and avoid defects of a product by the design change.SOLUTION: A design support device has an arithmetic processing part for evaluating a design risk of a model to be separately designed using design data of an existing similar model, calculating a predicted value and an analysis value of a performance parameter for a changed design parameter of the model to be separately designed using the design data of the similar model, and evaluating the design risk on the basis of an error between the predicted value and the analysis value.SELECTED DRAWING: Figure 1B

Description

Technical Field

[0001] The present invention relates to a design support device and a design support method for supporting design by visualizing design risks in a design space.

Background Art

[0002] In the manufacturing industry, products are designed to meet product specifications set based on the results of market research or product specifications set based on customer requests. At this time, it is rare to design a completely new product, and in many cases, an existing product is made into a new product by changing the design. When there are multiple existing products, an existing product to be used as a base is selected after considering the design change policy.

[0003] In considering the design change policy, analysis is often used to evaluate the impact of design changes. Especially in recent years, with the spread of software such as stress analysis and fluid analysis, from the perspective of shortening the design period, not only the impact evaluation of design changes but also whether the product meets the specifications is often evaluated by analysis before prototyping.

[0004] However, when there are many parameters to be changed in the design, the number of analyses may become extremely large and it may take a lot of time for the analysis. Also, in the case of 3D analysis, the calculation time per analysis case is long, and it is difficult to perform many analyses within the limited time available for design.

[0005] Therefore, techniques have been developed to create approximation functions using polynomial approximation, response surface method, machine learning, etc. from a small number of analysis results and estimate the analysis results under unanalyzed conditions.

[0006] Conventionally, as an invention of this type, there is one described in Patent Document 1. Patent Document 1 describes a reliability analysis method for assisting in analyzing the correlation between each failure mode in a multi-objective design space and the causal and correlative relationships hidden between each design variable.

Prior Art Documents

Patent Documents

[0007] [Patent Document 1] Japanese Patent Application Laid-Open No. 2005-202515 [Summary of the Invention] [Problems to be Solved by the Invention]

[0008] In Patent Document 1, within the range of the response surface, design uncertainties such as dimensional variations are considered, and a highly reliable design plan is selected. However, it is a method of obtaining Pareto solutions in which the objective function satisfies the specifications from the response surface and selecting a more reliable design plan from among them, and it was not possible to evaluate the distribution of design plans with low reliability, that is, high design risks. [Means for Solving the Problems]

[0009] In order to solve the above problems, the design support device of the present invention evaluates the design risk of a model to be designed separately using the design data of existing similar models, calculates the predicted value and the analytical value of the performance parameter with respect to the design parameter after change of the model to be designed separately using the design data of the similar models, and has an arithmetic processing unit that evaluates the design risk based on the error between the two. [Effects of the Invention]

[0010] When making design changes, etc., the user of the design support device can easily confirm the design risk, shortening the period for considering the design change policy and avoiding product defects due to design changes.

[0011] Problems, configurations, and effects other than those described above will be clarified by the description of the embodiments for carrying out the following invention. [Brief Description of the Drawings]

[0012]

Figure 1A

Figure 1B

Figure 1C

Figure 1D

Figure 1E

Figure 2A

Figure 2B

Figure 3

Figure 4

Figure 5

Figure 6

Figure 7

Figure 8

Figure 9

Figure 10

Figure 11

Figure 12

Mode for Carrying Out the Invention

[0013] (Summary of the Invention) When a design is changed, it may fall outside the range of the existing product's design space. Here, the design space is the scope of design consideration that takes multiple design parameters as axes and represents combinations of design parameters. Design parameters are parameters that designers can change, such as the dimensions and materials of parts. Also, parameters related to physical phenomena determined by the design are called performance parameters. For example, stress, flow velocity, temperature, etc. are performance parameters.

[0014] The outside of the existing product's design space is a range that has never been designed before, and there is a possibility of different physical phenomena from the existing product, which may lead to product defects, so sufficient consideration is required.

[0015] However, if considered after the product specifications are clear, it takes time for the design, and designers with little experience may not notice the design risks, which may cause delays in delivery and rework from the manufacturing and inspection processes in the subsequent processes.

[0016] The present invention has been made in view of the above, and when the design range is outside the conventional design space, the value predicted from the approximation function is compared with the result analyzed from the design parameters outside, and the error is compared with the threshold value to evaluate the risk of the design space. As the threshold value, actual values such as the error at the time of occurrence of defects in past designs and the error at the time of design rework are used.

[0017] And by visualizing the design risk outside the existing product's design space for the designer, the designer can easily confirm the risk of the design space when making a design change and can formulate a design change policy at an early stage.

[0018] Note that in this embodiment, design changes include separate new designs that reuse part of the existing design data and new developments.

[0019] Hereinafter, embodiments of the present invention will be described in detail with reference to the drawings. The embodiments are examples for explaining the present invention, and for the sake of clarity of explanation, appropriate omissions and simplifications have been made. The present invention can also be implemented in various other forms. Unless otherwise specifically limited, each component may be singular or plural.

[0020] In the drawings, the positions, sizes, shapes, ranges, etc. of the respective components shown may not represent the actual positions, sizes, shapes, ranges, etc. in order to facilitate understanding of the invention. For this reason, the present invention is not necessarily limited to the positions, sizes, shapes, ranges, etc. disclosed in the drawings.

[0021] As examples of various information, it may be described in expressions such as "table" and "list", but the various information may also be represented by other data structures. For example, various information such as "XX table" and "XX list" may also be referred to as "XX information". When explaining identification information, expressions such as "identification information", "identifier", "name", "ID", "number", etc. are used, but these are mutually replaceable.

[0022] When there are a plurality of components having the same or similar functions, they may be described by attaching different subscripts to the same reference numeral. Also, when it is not necessary to distinguish these plurality of components, the subscripts may be omitted in the description.

[0023] In the embodiments, the processes performed by executing a program may be described. Here, a computer executes a program by a processor (e.g., CPU, GPU), and performs the processes defined by the program while using a storage resource (e.g., memory) and an interface device (e.g., communication port), etc. Therefore, the entity that performs the processes by executing the program may be the processor. Similarly, the entity that performs the processes by executing the program may be a controller, a device, a system, a computer, or a node having a processor. The entity that performs the processes by executing the program may be an arithmetic unit, and may include a dedicated circuit that performs specific processes. Here, the dedicated circuit is, for example, an FPGA (Field Programmable Gate Array), an ASIC (Application Specific Integrated Circuit), a CPLD (Complex Programmable Logic Device), etc.

[0024] The program may be installed in a computer from a program source. The program source may be, for example, a program distribution server or a storage medium readable by a computer. When the program source is a program distribution server, the program distribution server includes a processor and a storage resource that stores the program to be distributed, and the processor of the program distribution server may distribute the program to be distributed to other computers. Also, in the embodiments, two or more programs may be realized as one program, or one program may be realized as two or more programs.

Embodiment

[0025] FIG. 1A is a block diagram of a design support device 100 according to Embodiment 1 of the present invention, which is realized by a general computer and includes a central control unit (CPU) 101 that performs arithmetic processing, an input device 102 such as a keyboard and a mouse, an output device 103 such as a display device, a main storage device 110, and an auxiliary storage device 120.

[0026] The main memory device 110 includes an approximation function generation unit 111, a design change analysis unit 112, an analysis execution unit 113, an error analysis unit 114, a design space validity analysis unit 115, and a design space validity visualization unit 116, which are programs. Hereinafter, when the subject is described as "○○ unit", it is assumed that the central control device 101 reads each program from the main memory device 110 and realizes the functions of each program (details will be described later) on the main memory device. However, it may be realized by a dedicated circuit or the like.

[0027] The auxiliary storage device 120 includes an analysis result database (DB) 121 of similar models, a design change history DB 122 of similar models, a customer needs DB 123, and an error threshold DB 124. Note that some or all of the various DBs provided in the auxiliary storage device 120 may be provided in a storage device external to the design support device 100.

[0028] Next, the configuration of each unit in the main memory device 110 and the content of each DB in the auxiliary storage device 120 will be described with reference to FIG. 1B. The arrows in FIG. 1B indicate the processing flow and data flow in the present design support device 100.

[0029] The analysis result DB 121 of similar models stores the analysis results of similar models of the product to be designed (in this embodiment, the "product" includes "parts" as sales products used in the manufacture of specific products). For example, during design consideration, when changing the values of design parameters such as the dimensions and materials of the parts to be used and searching for optimal values, the results of parameter studies, or when assuming a change in materials and changing the material property values, the analysis results of stresses, flow rates, temperatures, etc. expected in the product are stored in sets with the design parameters. Here, similar models include not only existing models that are the same as the models to be newly or separately designed and their components but only differ in design parameters, but also existing models with the same main components.

[0030] For example, as a similar model, an example of the analysis result DB121 of similar models for a transmission device (a part thereof) having a pulley 1001 with a radius "R" and a belt 1002 with a dimension (thickness) "h" shown in FIG. 10 is shown in FIG. 1C. In FIG. 1C, the model name 121A is a name or ID for identifying a similar model, the dimensions and materials of the pulley and the belt are design parameters 121B, and the bending stress 121C is a performance parameter that changes depending on the change of the design parameter 121B.

[0031] The approximate function generation unit 111 creates a function that approximates the relationship between the design parameters and the performance parameters of the analysis result DB121 of similar models. The approximate function is, for example, a function created by polynomial approximation of the relationship between the design parameters and the performance parameters.

[0032] The design change history DB122 of similar models records the change history of the design parameters changed due to past customer requirements and design revisions of similar models. FIG. 1D is an example of the design change history DB122 assuming the transmission device shown in FIG. 10, in which the model name 122A, the change content 122B of the model, and the change date 122C are recorded. The reasons for design changes, such as design revisions due to defects and customer requirements, may also be added and recorded.

[0033] The customer needs DB123 records the customer needs collected through market research results and customer hearings. FIG. 1E is an example of the customer needs DB123 assuming the transmission device shown in FIG. 10.

[0034] The design change analysis unit 112 predicts the type of design parameters and the range (value) to be changed where design changes are likely to occur from the design change history DB122 and the customer needs DB123 of similar models.

[0035] Based on the prediction result of the design change analysis unit 112, the analysis execution unit 113 creates an analysis model with design changes and executes the analysis.

[0036] The error analysis unit 114 compares the result of analysis using the value of the changed design parameter with the result predicted from the approximation function using the value of the changed design parameter, and calculates the error therebetween.

[0037] The design space feasibility analysis unit 115 compares the error calculated by the error analysis unit 114 with the threshold value stored in the error threshold DB 124 described later, and estimates a design risk such that there is a high possibility that a defect occurs when the threshold value is exceeded.

[0038] The error threshold DB 124 stores threshold values defined in consideration of cases such as the error between the predicted value by the above-described approximation function and the analysis value by the analysis described later for the performance parameters in the designs of past similar models, the error when a product defect occurs, and the error when there is a design rollback. Further, a plurality of values may be defined for the threshold according to the degree of the defect, or the threshold may be defined for each performance parameter to be evaluated.

[0039] The design space feasibility visualization unit 116 provides a GUI that clearly shows the design risk estimated by the design space feasibility analysis unit 115 to designers and the like using the output device 103 or the like.

[0040] Next, the processing flow of the design support apparatus 100 according to Embodiment 1 of the present invention will be described. FIGS. 2A and 2B are flowcharts showing the processing procedure when the design support apparatus 100 visualizes the risk of the design space. Hereinafter, each step in FIGS. 2A and 2B will be described.

[0041] Step S201: Search the analysis result DB 121 of similar models, extract the analysis results (design parameters and performance parameters therefor) of models similar to the device to be verified this time (product to be designed), and generate an approximation function using them. The approximation function may be generated by polynomial approximation or the like for the relationship between the design parameter and the performance parameter as described above, or may be generated using a Kriging model or a regression model by machine learning.

[0042] Step S202: Refer to the design change history DB122 and customer needs DB123 of similar models to predict the design parameters that are likely to be changed. If the design parameter is a numerical value, predict the numerical range; if it is a material or the like, predict the material name or the like.

[0043] Figure 2B is a flowchart showing the process of Step S202 in more detail. First, extract the frequently requested requirements from the customer needs DB123 (Step S2021). Next, extract the design parameters with high change frequencies from the design change history DB122 (Step S2022). Then, estimate the design parameters and their ranges that are likely to be changed corresponding to the customer needs extracted in Step S2021 (Step S2023).

[0044] Note that the customer needs extracted in Step S2021 are not limited to one, and there may be multiple frequently requested ones. In that case, the processes of Step S2022 and Step S2023 are performed for each of the extracted multiple requests.

[0045] Also, instead of using the design parameters predicted in Step S202, the design change analysis unit 112 may use the values directly set by the user via the input device 102 or the like.

[0046] Return to Step S203 in Figure 2A. Here, create an analysis model based on the design parameters corresponding to the predicted design change content, and perform an analysis of the performance parameters for it. For example, when calculating the bending stress 121C in the example of the analysis result DB121 shown in Figure 1C, perform stress analysis. As the stress analysis, known methods can be adopted. For example, there are numerical stress analysis using the finite element method and theoretical stress analysis using material mechanics, and each can be used alone or in combination with multiple methods.

[0047] Step S204: Apply the design change content (predicted change value of design parameters) predicted in step S202 to the approximate function generated in step S201 to predict the performance parameters. Then, calculate the error by comparing with the value of the performance parameters obtained by analysis in step S203.

[0048] Step S205: Determine whether the error between the performance parameters by the approximate function and the performance parameters by analysis calculated in step S204 is equal to or less than a predetermined threshold stored in the error threshold DB124.

[0049] Step S206: If the error is smaller than the threshold (\"YES\" in step S205), regarding the design change content predicted this time, since the value of the performance parameters can be accurately predicted by the approximate function based on past performance, it is determined that the events that may occur due to the design change are also within the assumption, and it is visualized, for example, on the display of the output device 103, as a place with a low risk in the design space (a combination or region of design parameters with a low risk).

[0050] Step S207: If the error is larger than the threshold (\"NO\" in step S205), from the approximate function based on past performance, since the predicted value of the performance parameters based on the design change deviates greatly from the analysis value, it is determined that an unexpected event may occur due to the design change, and it is visualized as a place with a high risk in the design space.

[0051] Through the above processing flow, the risk of the design change predicted based on the needs can be easily estimated. Also, these processes may be performed at the timing determined by the designer, but in order to formulate the design change policy in a short time, it is advisable to automatically execute them at the timing when the customer needs DB123 is updated or at a predetermined regular timing.

[0052] FIG. 3 is a diagram showing a method of calculating the error between the analysis result obtained by the error analysis unit 114 and the result predicted from the approximation function, and determining the risk of the design by the design space validity analysis unit 115, and explains the case where the error between the analysis result and the prediction result is small and below the threshold (corresponding to steps S204 to S205 to S206 described above).

[0053] In FIG. 3, the horizontal axis represents the design parameter, the vertical axis represents the performance parameter, and a plurality of analysis result points 301 are plotted. The analysis results are the analysis results stored in the analysis result DB 121 of similar models. The approximation curve 302 is a curve that approximates a plurality of analysis result points 301, and is, for example, a polynomial approximation, a Kriging model, a regression model by machine learning, or the like.

[0054] Assume that the value of the design parameter has changed from 304 to 305 due to a design change, the analysis value of the performance parameter analyzed at the design parameter value 305 is 306, and the value obtained by extending the approximation curve 302 and using it as the predicted value of the performance parameter corresponding to the design parameter value 305 is 307.

[0055] In this case, the error 308 between the analysis result 306 and the prediction result 307 is very small and is below the threshold stored in the threshold DB 124. It can be determined (estimated) that the possibility of product defects occurring due to the changed design parameter 305 is low and the risk of the design is low. Note that the risk of the design is, for example, the possibility of product defects occurring or the possibility that the product fails inspection due to manufacturing variations.

[0056] On the other hand, FIG. 4 explains the case where the error between the analysis result and the prediction result is large and exceeds the threshold (corresponding to steps S204 to S205 to S207 described above).

[0057] In this case, the error 402 between the result 401 analyzed at the value of the design parameter 305 and the prediction result 307 by the approximation curve is larger than the error 308 in FIG. 3 and exceeds the threshold stored in the threshold DB 124.

[0058] That is, when the design parameters are changed from the design parameters 304 to the design parameters 305, the values of the performance parameters change rapidly, and the tendency is different from the prediction curve by the approximation function 302, suggesting that some phenomenon or the like occurring inside the manufactured product may have changed. It can be determined that the fact that the phenomenon has changed may include risks that have not been fully considered in the previous design.

[0059] Next, an example of presenting the estimated design risk to the designer using the output device 103 or the like in the above-described steps S206 and S207 will be described. FIG. 5 is an example of a method for visualizing the design risk performed by the design space validity visualization unit 107, assuming a design risk evaluation for a combination of two (x1, x2) design parameters.

[0060] That is, the design parameters x1 and x2 are taken as the x-axis and the y-axis, respectively, and a plurality of design parameters 501 of similar models and the design parameters 502 after the design change are plotted. Here, the design parameters 502 after the design change indicate the design parameters estimated to have a high design risk by the above-described method.

[0061] On the other hand, a plurality of design parameters 501 of similar models indicate design parameters with a low design risk.

[0062] Therefore, the boundary between these plurality of design parameters 510 and the design parameters 502 with a high design risk is obtained by a two-class classification method using machine learning, and the design risk is visualized as a region within the design space.

[0063] For example, the probability that the design parameters 502 become a high risk is "1.0", the possibility of becoming a low risk is "0.0", the boundary 503 is a boundary where the probability of becoming a high risk is "0.7" and the probability of becoming a low risk is "0.3". The boundary 504 is a boundary where the probability of becoming a high risk is "0.5" and the probability of becoming a low risk is "0.5". The boundary 505 is a boundary where the probability of becoming a high risk is "0.3" and the probability of becoming a low risk is "0.7".

[0064] Also, the line 510 indicating the outer edge of the region on the side of the design parameter 510 of similar models can be said to be a boundary where there may be more or less design risks.

[0065] By visualizing the boundary in this way, it is possible to visualize the risk of the design range to be changed.

[0066] FIG. 6 is an explanatory diagram of another method for visualizing design risks. An approximate curve 601 connecting the analysis results 301 of similar models including the analysis point 401 when the value of the design parameter is 305 is calculated. The error from the extended approximate curve of the analysis results 301 of similar models becomes the distribution of the error from the design parameter 304 to 305. Here, three errors are shown as the error value p1 (602a), the error value p2 (602b), and the error value p3 (602c). Also, by displaying the value of the threshold th (610) in the vicinity as well, it is possible to visually present how much margin can be secured with respect to the threshold 610 when the design parameter is changed from 304 to 305.

[0067] FIG. 7 is an explanatory diagram of yet another method for visualizing design risks. Taking the design parameters x1 and x2 as axes in the same way as in FIG. 5, a plurality of design parameters 501 of similar models and the changed design parameter 502 are plotted, and further, a curve (701a, 701b, 701c) connecting the points of the combination of the design parameters (x1, x2) where the magnitude of the error between the predicted value and the analyzed value is the same when the design parameters (x1, x2) are changed is added.

[0068] This is, for example, a line connecting the points with the same magnitude of error corresponding to the errors 602a, 602b, and 602c in FIG. 6. From FIG. 7, it can be visually recognized that when the error is large, it is highly likely to exceed the threshold, and the design risk changes according to the error value, and it is possible to visualize the design risk according to the error.

[0069] As yet another visualization method for design risks, FIG. 8 is a schematic diagram of a GUI that varies thresholds for Δy1, Δy2, and Δy3, which are errors in performance parameters y1, y2, and y3, with sliders 801y1, 801y2, and 801y3 in the visualization diagram of design risks shown in FIG. 5.

[0070] The user of the design support device can clearly visualize how the design risk changes when changing the threshold of any performance parameter, and can perform a design that avoids a high-risk range during design changes. As the threshold, for example, there are a threshold 803 when a defect may occur in the market after the product is shipped, and a threshold 802 when a defect may occur during prototype verification in the factory and the design may have to be reworked. These are stored in the error threshold DB124.

[0071] Sliders 801y1, 801y2, and 801y3 all have values smaller than the threshold 803 when a defect occurs and the threshold 802 when reworking. The boundary line of the design risk in this case is 804. The area inside the hatched 804 is judged to have a low design risk, and the area outside is judged to have a high design risk.

[0072] FIG. 9 shows the boundary line 901 of the design risk when sliders 801y1 and 801y3 are set as the threshold 802 for design rework and slider 801y2 is set as the threshold 803 for defect occurrence. It means that the area inside the hatched 901 has a low design risk and the area outside has a high design risk. When the threshold is increased with the slider, it can be visually confirmed that the area with a low design risk expands.

[0073] In this way, even when it is difficult to determine an appropriate threshold, by having the user of the design support device determine how much to set the threshold and visualizing the corresponding design risk, it is possible to estimate the design risk while adjusting the set value of the threshold, so that the policy for design changes can be determined quickly.

Example

[0074] Hereinafter, in Example 2, the case will be described by taking the design of the transmission device as an example.

[0075] One of the mechanical devices is a transmission device that transmits the rotation of a motor using a belt and a pulley. FIG. 10 is a part of the transmission device also used in the description of the analysis result DB121 and the like of a similar model in Example 1, and is composed of a pulley 1001 with a radius R and a belt 1002 with a thickness h. When the pulley 1001 rotates, the belt 1002 repeatedly receives bending stress when passing through the pulley, so damage such as cracks in the belt occurs due to fatigue. Therefore, the design of the stress applied to the belt is important.

[0076] FIG. 11 shows the relationship between the pulley radius and the bending stress of the belt. An approximate function 1102 is created using a plurality of analysis results 1101 of similar models. When it is estimated that the pulley will be miniaturized based on the design change history and customer needs, it is predicted that the pulley radius will be miniaturized from 1103 to 1104, the belt bending stress is analyzed, and the analysis result 1106 and the prediction result 1107 obtained by extending the approximate curve 1102 are obtained. Then, these differences are compared with the threshold value to determine the design risk.

[0077] FIG. 12 is a visualization diagram of the design risk in the transmission device design. There are a plurality of analysis results 1201 of similar models and the analysis result 1202 when the design is changed. Similar to FIGS. 8 and 9, the design risk boundary line 1203 can be visualized. At this time, similar to FIGS. 8 and 9, it may be visualized while adjusting the threshold value.

[0078] As described above, when designing to miniaturize the pulley according to customer requirements, the design range with low risk can be easily confirmed, so the policy of design change can be formulated in a short period of time, and product defects due to design change can be avoided.

Explanation of symbols

[0079] 100: Design support device 101: Central control device 102: Input device 103: Output device 110: Main memory device 111: Approximation function generation unit 112: Design change analysis unit 113: Analysis execution unit 114: Error analysis unit 115: Design space validity analysis unit 116: Design space validity visualization unit 120: Auxiliary storage device 121: Analysis result DB of similar models 122: Design change history DB of similar models 123: Customer needs DB 124: Error threshold DB

Claims

1. A design support device for evaluating the design risk of a model to be separately designed using the design data of existing similar models, comprising an arithmetic processing unit that calculates a predicted value and an analytical value of a performance parameter with respect to a design parameter after change of the model to be separately designed using the design data of the similar models, and evaluates the design risk based on the error between the two. The design support device is characterized by this.

2. The design support device according to claim 1, an analysis result database that stores performance parameters obtained by analyzing the design parameters included in the design data of the similar models, an approximation function generation unit that generates an approximation function approximating the relationship between the design parameters and the performance parameters stored in the analysis result database, an analysis execution unit that calculates an analytical value of a performance parameter with respect to the design parameter after change, an error analysis unit that calculates a predicted value of a performance parameter with respect to the design parameter after change using the approximation function, and calculates an error from the analytical value of the performance parameter with respect to the design parameter after change, and having, wherein the arithmetic processing unit evaluates the design risk based on the error calculated by the error analysis unit. The design support device is characterized by this.

3. The design support device according to claim 2, a design change history database that stores the change history of the design parameters of the similar models, a customer needs database that stores customer needs for the similar models, a design change analysis unit that predicts the type and value of the design parameter after change using the information stored in the design change history database and the customer needs database, and having, wherein the analysis execution unit and the error analysis unit use the design parameter after change predicted by the design change analysis unit. The design support device is characterized by this.

4. The design support device according to claim 3, having an error threshold database in which a threshold value defined based on cases in past products is stored for the error between the predicted value and the analytical value of the performance parameter of the similar models, wherein the arithmetic processing unit has a design space feasibility analysis unit that compares the threshold value stored in the error threshold database with the error calculated by the error analysis unit to estimate the design risk. The design support device is characterized by this.

5. The design support device according to claim 4, A design support apparatus, comprising a visualization unit that displays, on an output device, the design risks estimated by the design space feasibility analysis unit. **Claim 6** The design support apparatus according to claim 5, wherein the visualization unit plots the design parameters of the similar models and the changed design parameters estimated to have a high design risk, generates a boundary line of the changed design parameters having the same probability of the height of the design risk between the two, and displays the boundary line on the output device. **Claim 7** The design support apparatus according to claim 5, wherein the visualization unit generates a boundary line of the changed design parameters where the error between the predicted value and the analyzed value of the performance parameter is the same as that of the design parameters of the similar models, and displays the boundary line on the output device. **Claim 8** The design support apparatus according to claim 6, wherein the visualization unit displays a slider capable of changing the threshold value, and generates the boundary line using the threshold value changed by the slider. **Claim 9** A design support method for evaluating the design risk of a model to be designed separately using the design data of existing similar models, wherein the predicted value and the analyzed value of the performance parameter for the changed design parameters of the model to be designed separately are calculated using the design data of the similar models, and the design risk is evaluated based on the error between the two.

Images