JP2012162226A - Relational expression estimation support method and relational expression estimation support apparatus - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a relational expression of a tire for changing the value of attribute information with high accuracy.SOLUTION: A relational expression support system 10 for estimating a relational expression on a tire, from a measured value obtained by experiments on a plurality of tires, includes: a multivariate analysis section 102 for estimating the relational expression by calculating a coefficient of an explanatory variable, if the explanatory variable is selected when a correlation graph is selected; a prediction value calculation section 103 for calculating a prediction value by substituting the measured value for the estimated relational expression; and a display processing section 105 in which, when an objective variable to be calculated is input via an input section 301, a correlation graph indicating correlation between the explanatory variable in each measured value and the objective variable is displayed on a display section 302, and a screen for comparing the measured value of the objective variable with the prediction value is displayed on the display section 302 of a terminal 3.

Description

  The present embodiment relates to a relational expression estimation support method and a relational expression estimation support apparatus that estimate a relational expression using tire attribute information as an explanatory variable.

  As a modeling tool for calculating and simulating tire performance and characteristics, a physical model that is a numerical model and data measured by experiments are used, using relational expressions that theoretically formulate tire physical phenomena. An empirical model for modeling using a relational expression based on a specific approximate expression has been used.

  Non-Patent Document 1 shows the relationship between cornering power (CP), which is one of the indexes indicating tire performance, and the load, and the relationship between cornering power and tire internal pressure.

Edited by Bridgestone Corporation, “Fundamentals and Practice of Automobile Tires”, 1st edition, April 10, 2008, Tokyo Denki University Press, p. 128, p129

However, physical models are suitable for understanding phenomena related to tire performance and qualitatively estimating and discussing performance, but are not accurate enough to quantitatively predict actual tire characteristics. There is a problem that the grounds for supporting the estimated value are also poor.
On the other hand, the empirical model is modeled on the basis of data obtained from an experiment conducted using a specific tire, and is a model having sufficient accuracy as a model for the tire on which the experiment was conducted. Can be obtained. However, it is difficult to accurately predict the characteristics when the attribute information values such as the tire dimensions and air pressure are changed and the attribute information values are different from the experiment. It is necessary to newly acquire experimental data and estimate a tire model by performing an experiment according to the change.

  Therefore, an object of the present invention is to estimate a relational expression that is highly accurate and can cope with a change in the value of attribute information when estimating the relational expression using tire attribute information as an explanatory variable.

  The invention according to claim 1 of the present invention that solves the above-described problem is a relational expression estimation support method by a relational expression estimation support device that estimates a relational expression related to the tire using tire attribute information as an explanatory variable, The relational expression support apparatus stores, in a storage unit, an actual measurement value of the attribute information obtained as a result of an experiment on a plurality of the tires, and an actual measurement value of performance information that is a result of a performance experiment on the attribute information. The objective variable is selected from the performance information via the input unit, and the actual value of the attribute information and the performance information selected as the objective variable corresponding to the actual value of the attribute information. A correlation graph showing the correlation with the actual measurement value is displayed on the display unit, and the explanatory graph used in the relational expression is selected by selecting the correlation graph via the input unit. The original relational expression using the explanatory variable is input via the input unit, and the relational expression is estimated by calculating the coefficient of the explanatory variable by multivariate analysis, and the estimation A display screen for comparing the predicted value of the performance information calculated by substituting the measured value into the relational expression and the measured value of the performance information corresponding to the predicted value is displayed on the display unit. To do.

  According to the first aspect of the invention, since the relational expression estimated by multivariate analysis is used based on the data of a plurality of tires, the performance value of the vehicle, the specific value, etc. are calculated based on the biased experimental data. Can reduce the risk.

  The invention according to claim 2 is the relational expression estimation support method according to claim 1, wherein the multivariate analysis is single regression analysis or multiple regression analysis.

  According to the second aspect of the invention, general single regression or multiple regression analysis can be used.

  The invention according to claim 3 is a relational expression estimation support apparatus that estimates the relational expression related to the tire using the attribute information of the tire as an explanatory variable, and includes the attribute information obtained as a result of an experiment related to the plurality of tires. The measured value and the measured value of the attribute information through the storage unit storing the measured value and the measured value of the performance information that is the result of the performance experiment in the attribute information, and the input unit The explanatory variable used in the relational expression is selected by selecting a correlation graph indicating the correlation between the measured value of the performance information selected as the objective variable corresponding to the measured value of the information, and When the original relational expression using the explanatory variable is input via the input unit, the relational expression is estimated by calculating the coefficient of the explanatory variable by multivariate analysis. A multivariate analysis unit, a predicted value calculation unit that calculates a predicted value of the performance information by substituting the actual measurement value into the estimated relational expression, and the correlation graph is displayed on the display unit. And a display processing unit that displays a comparison screen between the predicted value and the measured value of the performance information corresponding to the predicted value on the display unit.

  According to the invention of claim 3, since the relational expression estimated by multivariate analysis is used based on the data of a large number of tires, the vehicle performance value, specific value, etc. are calculated based on biased experimental data. Can reduce the risk.

  According to the present invention, when estimating a relational expression using tire attribute information as an explanatory variable, it is possible to estimate a relational expression that is highly accurate and can respond to a change in the value of the attribute information.

It is a figure which shows the structural example of the relational expression estimation assistance system which concerns on this embodiment. It is a flowchart which shows the procedure of the measurement data storage and correlation graph calculation process which concern on this embodiment. It is a figure which shows the example of the measurement data regarding cornering power. It is a figure which shows the example of the measurement data regarding vertical spring rigidity. It is a flowchart which shows the procedure of the relational expression estimation process which concerns on this embodiment. It is a figure which shows the example of the relational expression input / display screen which concerns on this embodiment. It is a figure which shows the example of the correlation graph selection screen which concerns on this embodiment (1 variable). It is a figure which shows the example of the correlation graph selection screen which concerns on this embodiment (2 variables). It is a figure which shows the example of the verification result screen regarding cornering power. It is a figure which shows the example of the verification result screen regarding vertical spring rigidity. It is a figure which shows an example of the development examination screen which concerns on this embodiment.

  Next, modes for carrying out the present invention (hereinafter referred to as “embodiments”) will be described in detail with reference to the drawings as appropriate. In the present embodiment, the tire model is a virtual tire having performance values such as cornering power and longitudinal spring rigidity obtained as a result of substituting a certain value for the explanatory variable of the estimated relational expression. is there.

[System configuration]
FIG. 1 is a diagram illustrating a configuration example of a relational expression estimation support system according to the present embodiment.
A relational expression estimation support system (relational expression estimation support apparatus) 10 performs a calculation server 1 for estimating a relational expression in a tire, a database (storage unit) 2 for storing data, and displaying an estimated relational expression. Terminals 3 are connected to each other via a network or the like.
The calculation server 1 includes a processing unit 100, and the processing unit 100 includes a correlation calculation unit 101, a multivariate analysis unit 102, a predicted value calculation unit 103, a verification unit 104, and a display processing unit 105.
The correlation calculation unit 101 generates a correlation graph, which will be described later, based on the actual measurement values obtained as a result of experiments on tires stored in the database 2. The multivariate analysis unit 102 calculates the coefficient of the relational expression related to the tire input from the terminal 3 by multivariate analysis such as multiple regression analysis. The predicted value calculation unit 103 calculates a predicted value using the calculated relational expression. The verification unit 104 displays a screen for verifying the accuracy of the relational expression on the display unit 302 of the terminal 3 via the display processing unit 105. The display processing unit 105 displays various screens on the display unit 302 of the terminal 3. Here, multivariate analysis includes single regression analysis.

  The processing unit 100 and each of the units 101 to 105 have a program stored in a ROM (Read Only Memory) or HDD (Hard Disk Drive) (not shown) developed in a RAM (Random Access Memory), and a CPU (Central Processing Unit). ) To be embodied.

The database 2 includes actual measurement data 201 that is data of actual measurement values, correlation graph data 202 such as a correlation graph, estimated data 203 that is data of predicted values calculated using a relational expression estimated by the multivariate analysis unit 102, and The verification result data 204 used for verification of the relational expression such as an error between the predicted value and the actual measurement value is stored.
The terminal 3 includes an input unit 301 for inputting information and a display unit 302 for displaying information.
In the present embodiment, the calculation server 1, the database 2, and the terminal 3 are different devices, but may be a single device.

[Measurement data storage and correlation graph calculation processing]
Next, actual measurement data storage / correlation graph calculation processing according to the present embodiment will be described along FIG. 2 with reference to FIGS. 1, 3 and 4 as appropriate.
FIG. 2 is a flowchart showing a procedure of actual measurement data storage / correlation graph calculation processing according to the present embodiment.
First, the user stores each measured value acquired as a result of the experiment on the tire in the database 2 as measured data 201 (S101).
Then, the correlation calculation unit 101 of the calculation server 1 plots the measured values with the values of the measured data 201 of the database 2 as the horizontal axis, the objective variables set in advance as the vertical axis, and from the plotted measured values. A correlation graph such as a regression line is calculated, and the calculated correlation graph is stored in the database 2 as correlation graph data 202 (S102). Thus, by calculating the correlation graph or the like in advance, the correlation graph can be displayed without waiting for the user when displaying the correlation graph on the display unit 302 of the terminal 3 as described later. Note that the correlation calculation unit 101 may perform the process of step S102 after receiving a display instruction from the user.
At this time, the user determines in advance how much the correlation graph is to be created, such as obtaining up to the first order correlation, obtaining up to the second order correlation, using one variable, or using two variables. Alternatively, it may be set at any time as desired by the user.

(Actual measurement data)
3 and 4 are diagrams showing examples of actually measured data according to the present embodiment.
FIG. 3 shows actual measurement data related to cornering power (CP), and FIG. 4 shows actual measurement data related to longitudinal spring stiffness (Kz). Here, the actual measurement value stored in the actual measurement data 201 is a statistically significant number of measurement results obtained from experiments on a plurality of types of tires having characteristics suitable for creation of a target tire model. It is desirable to be.
As shown in FIG. 3, as actual measurement data 201, values of attribute information such as tire width, flatness ratio, rim diameter, air pressure, vertical load, and cornering power are stored as performance information values for the respective values. In addition, the vertical load ² calculated by the correlation calculation unit 101, the value of tire width × vertical load, and the like are also stored.
As shown in FIG. 4, the tire width and the flatness ratio are not stored independently, but may be stored as a value obtained by multiplying the vertical load or the vertical load ² by all. In the present embodiment, it is assumed that the actual measurement data 201 relating to the longitudinal spring stiffness has the same configuration as that in FIG. In FIG. 4, the tire deformation amount is stored because the longitudinal spring stiffness is a value obtained by taking the reciprocal number after differentiating the tire deformation amount.
In the present embodiment, experimental values are stored as the actual measurement data 201, but catalog data may also be used.

[Relationship estimation process]
Next, the relational expression estimation processing according to the present embodiment will be described along FIG. 5 with reference to FIGS. 1 and 6 to 10 as appropriate.
FIG. 5 is a flowchart showing a procedure of relational expression estimation processing by multivariate analysis according to the present embodiment.
First, the user activates the processing unit 100 of the calculation server 1 via the terminal 3.
Next, the user selects an objective variable to be estimated for the relational expression such as cornering power via the terminal 3 (S201), and transmits a request for displaying a correlation graph with the value of the actual measurement data 201 to the calculation server 1. The arithmetic server 1 displays a correlation graph between the requested objective variable and the actual measurement data 201 on the display unit 302 of the terminal 3, and the user refers to the displayed correlation graph and then inputs the correlation graph via the input unit 301. An explanatory variable is selected (S202).

(Relational expression input / display screen)
FIG. 6 is a diagram illustrating an example of a relational expression input / display screen displayed on the display unit of the terminal in the step S201.
The relational expression input / display screen 600 has an objective variable selection area 601, a relational expression input area 602, and a coefficient display area 603.
The objective variable selection area 601 is provided with an objective variable selection window 611. When the “▼” button beside the objective variable selection window 611 is clicked, a list of objective variables stored in advance in the calculation server 1 is displayed. A pull-down menu is displayed, and the user selects a target variable from the pull-down menu. Alternatively, the user may designate an objective variable that is not stored by directly inputting the objective variable into the objective variable selection window 611. When an “OK” button in the objective variable selection area 601 is selected and input, a correlation graph display request is sent from the terminal 3 to the calculation server 1.
The relational expression input area 602 and the coefficient display area 603 will be described later.

(Correlation graph selection screen)
FIG. 7 is a diagram illustrating an example of a correlation graph selection screen displayed on the display unit of the terminal at the stage of step S202. Each correlation graph displayed on the correlation graph selection screen 700 is obtained by displaying the correlation graph data 202 of the database 2 on the display unit 302 of the terminal 3.
As shown in FIG. 7, the correlation graph selection screen 700 displays a correlation graph with the corresponding objective variable as the vertical axis (cornering power (CP) in FIG. 7) and each explanatory variable as the horizontal axis. Note that the correlation graph is calculated in step S102 of FIG. In each correlation graph, measured values obtained as a result of experiments on a large number of tires are plotted as points 711, and a correlation graph for the measured values is shown as a graph 712.
A check box is provided at the upper left of each correlation graph. The user can select each explanatory variable that looks likely to be correlated with the objective variable to be estimated for the relational expression by looking at each correlation graph.

  In the example of FIG. 7, it is estimated that there is a positive correlation between the tire width (W) and the cornering power (CP), and a negative correlation between the flatness ratio (H) and the cornering power (CP). It is estimated that there is a correlation, and it is predicted that there is a second-order correlation between the vertical load (Fz) and the cornering power (CP). However, it is difficult to recognize a correlation between air pressure (P) and cornering power (CP). Therefore, the user checks the check box in the graph of tire width (W) -cornering power (CP), flatness ratio (H) -cornering power (CP), vertical load (Fz) -cornering power (CP). Thus, the tire width (W), flatness ratio (H), and vertical load (Fz) are selected as explanatory variables.

For example, an item “select all” is provided in the pull-down menu that appears when editing of the menu bar is clicked, and when the user selects this item, the check boxes of all correlation graphs are checked. May be.
Further, for example, an item for “check a highly correlated graph” is provided in a pull-down menu that appears when editing of the menu bar is clicked, and when the user selects this item, the display processing unit 105 performs a correlation coefficient. A check box for an explanatory variable whose value is greater than or equal to a predetermined value may be checked.

In addition, in the example of FIG. 7, the correlation between the air pressure (P) and the cornering power (CP) is not recognized, so it is not checked, but there is actually a correlation between the air pressure and the cornering power. If the user determines that there is a possibility that no correlation is apparent on the correlation graph because the correlation between the air pressure and the cornering power has been offset by the influence of other explanatory variables, the air pressure (P) And a correlation graph between cornering power (CP) can be checked.
As described above, when the user selects the correlation graph of the explanatory variables, for example, it is possible to select an explanatory variable that does not seem to be correlated at first glance, and it is possible to estimate the relational expression with high accuracy.

In the example of FIG. 7, a correlation graph with one explanatory variable is shown. However, a correlation graph with two explanatory variables can be displayed as shown in FIG.
Further, the selected objective variable and an arbitrary explanatory variable may be fixed, and a correlation graph between the explanatory variable different from the fixed explanatory variable and the objective variable may be displayed.

Next, when the user selects and inputs a coefficient calculation button (not shown) on the correlation graph selection screens 700 and 800, the display processing unit 105 activates the relational expression input / display screen 600 shown in FIG. Prompt. Here, the user inputs a relational expression in the relational expression input area 602 of the relational expression input / display screen 600 (S203 in FIG. 5).
At this time, in the relational expression input window 621 of the relational expression input area 602, the primary combination formula of the explanatory variables selected in step S202 is displayed as a default, and the user via the input unit 301 of the terminal 3 The relational expression may be input by rewriting this relational expression. At this time, the user can edit the relational expression while viewing the correlation graph selection screen 700 of FIG. For example, since it is estimated from FIG. 7 that the tire width (W) and the cornering power (CP) have a positive correlation, the tire width in the relational expression is assumed to be first order, and the flatness ratio (H) and the cornering power (CP) Is assumed to have a negative correlation, it is assumed that there is a negative correlation before the explanatory variable of the flatness, or that there is a second-order correlation between the vertical load (Fz) and the cornering power (CP). Therefore, it is possible to perform editing such as adding a term in which the explanatory variable of the vertical load is squared.
Alternatively, when the user clicks the “▼” button next to the relational expression input window 621, a pull-down menu in which main relational expressions that are combined in advance by the processing unit 100 of the calculation server 1 are displayed. A relational expression in the pull-down menu may be selected.

The objective variables and explanatory variables in the relational expression input area 602 in FIG. 6 are as follows.
W = tire width (mm), H = tire flatness (%), D = rim diameter, P = air pressure (kpa), Fz = vertical load (kN)
Note that c1 to c10 are coefficients.

Next, when the user selects and inputs the OK button in the relational expression input area 602, the multivariate analysis unit 102 of the calculation server 1 calculates the coefficient of each explanatory variable in the relational expression using multiple regression analysis and the contribution rate. Is calculated (S204 in FIG. 5) to estimate the relational expression. At this time, the multivariate analysis unit 102 determines the degree of freedom (double) adjusted contribution rate, presence / absence of abnormal values, variance of residuals, regression coefficient (B coefficient), determination coefficient (R-square value), and multicollinearity. Presence / absence may be calculated. Each calculated coefficient and contribution rate are stored in the estimation data 203.
Then, the multivariate analysis unit 102 of the arithmetic server 1 displays the coefficient of the explanatory variable in the calculated relational expression (that is, the coefficient of the estimated data 203) in the coefficient display area 603 of the relational expression input / display screen 600 (FIG. 6). (S205).
In the coefficient display area 603 shown in FIG. 6, the calculated value of each coefficient is displayed. In the example of FIG. 6, a check box is displayed beside each coefficient. For example, when the user has determined that the value of the coefficient is too small and the explanatory variable of the coefficient has a small contribution to the objective variable, the corresponding explanatory variable is removed by unchecking the check box next to the coefficient. Can be removed from the relational expression.
In the coefficient display area 603 in FIG. 6, only the coefficient value is shown, but the contribution rate may be displayed together.

Subsequently, when the user selects and inputs a “verification” button on the menu bar of the relational expression input / display screen 600, the verification unit 104 of the calculation server 1 substitutes the value of each measured data 201 into the relational expression, A predicted value of the variable is calculated, and further a verification process for calculating an error distribution (error distribution) between the predicted value and the actually measured value is performed (S206). At this time, the user can also instruct to calculate a predicted value by changing a specific explanatory variable in several stages. The result of the verification process is stored in the database 2 as verification result data 204.
Then, the display processing unit 105 of the calculation server 1 uses the verification result screen 900 shown in FIG. 9 as a verification result screen 900 shown in FIG. The information is displayed on the display unit 302 (S207).

The user looks at the verification result screen 900 to determine whether or not the prediction accuracy is appropriate (S208). When the user determines that the prediction accuracy is appropriate (S208 → Yes), for example, When the “end” item (not shown) of the pull-down menu that appears when the “file” menu is clicked is selected and input via the input unit 301, the arithmetic server 1 ends the processing.
If it is determined in step S208 that the user is not valid (S208 → No), the user, for example, a “recalculate” item (not shown) in a pull-down menu that appears when the “file” menu on the verification result screen is clicked. Select and input. Then, the calculation server 1 returns the process to step S203, and the user re-inputs the calculation formula.

(Verification result screen)
9 and 10 are diagrams illustrating an example of the verification result screen according to the present embodiment. The verification result screens 900 and 900 a shown in FIGS. 9 and 10 are obtained by displaying the verification result data 204 of the database 2 on the display unit 302 of the terminal 3.
FIG. 9 is a diagram illustrating an example of a verification result screen regarding cornering power.
The predicted value shown in FIG. 9 is a function input in the relational expression input area 602 in FIG. 6 and is calculated using a relational expression in which the coefficient value is the value displayed in the coefficient display area 603. is there.
The verification result screen 900 has a verification result display area 901 and an error distribution display area 902.
In the verification result display area 901, a prediction result graph in which the predicted value and the actually measured value are displayed together is displayed. In the example of FIG. 9, nine prediction result graphs are displayed for each tire size. In the example of FIG. 9, the actual measurement values are displayed as plot points, and the predicted values based on the relational expressions are shown as curve graphs. Incidentally, numerical values such as “195 / 65R5” in each prediction result graph indicate the tire size, and each prediction result graph has a ground contact load (Fz (kN)) on the horizontal axis and a cornering power (CP (kN) on the vertical axis. / Deg)).
In the prediction result graph of FIG. 9, a plurality of lines are displayed for one tire size, and each line shows a graph of predicted values when the tire internal pressure is changed.

The error distribution display area 902 displays the error distribution by displaying the error between the predicted value and the actual measurement value under the same conditions as a histogram.
In the example of FIG. 9, the distribution generally follows a normal distribution, and the calculated relational expression can be said to be generally good.

FIG. 10 is a diagram illustrating an example of a verification result screen regarding vertical spring stiffness.
In the verification result screen 900a of FIG. 10, the error distribution display area is omitted.
In the verification result graph in the verification result display area 901a, the vertical axis represents the vertical load (Fz (kN)), and the horizontal axis represents the tire deformation (Dz (mm)).
Since the vertical spring stiffness is a value obtained by differentiating the amount of deformation of the tire with the vertical load and further taking the reciprocal, the user can observe the vertical load and the vertical spring by observing the inclination (tangential inclination) of the graph of FIG. You can see the rigidity relationship. Further, the verification unit 104 differentiates the tire deformation amount expression by the vertical load, calculates the value of the longitudinal spring stiffness from the inverse formula, and displays this on the verification result screen 900a, thereby displaying the longitudinal spring stiffness value. The value may be directly observed.
The prediction result graph in FIG. 10 is displayed for each tire size in the same manner as in FIG. 9, and the graph of the prediction value when the tire internal pressure value is changed is indicated by three lines.

In addition, the relational expression used for prediction of the longitudinal spring rigidity (Kz) in FIG. 10 is the following relational expression.
Kz = 1 / {(c1 · W + c2 · H + c3 · D + c4 · P + c5) + 2 · (c6 · W + c7 · H + c8 · D + c9 · P + c10) · Fz}
However,
c1 = −0.01866
c2 = 0.072462
c3 = −0.12321
c4 = −0.01749
c5 = 11.331117
c6 = 0.01818
c7 = -5.1E-05
c8 = 0.06885
c9 = 0.00061834
c10 = −0.77294
Kz is the longitudinal spring stiffness (kN / mm). What each explanatory variable indicates is the same as the cornering power.

  Note that the verification result screens 900 and 900a illustrated in FIGS. 9 and 10 are examples, and other information can be displayed.

[Development study]
The user conducts tire development studies using the relational expression determined from the above processing.
FIG. 11 is a diagram illustrating an example of a development review screen when the user conducts development study using the estimated relational expression after the process illustrated in FIG. 5 is completed. This development review screen 1100 assumes that the display processing unit 105 of the arithmetic server 1 displays on the display unit 302 of the terminal 3, but a device different from the arithmetic server 1 displays on the display unit 302 of the terminal 3. You may make it do.
In the development study screen 1100 of FIG. 11, the horizontal axis is cornering power (CP (kN / deg)), and the vertical axis is vertical spring stiffness (Kz (kN / mm)).
Here, the cornering power is one of the main indicators of the vehicle performance, and the longitudinal spring stiffness is one of the indicators that affect the riding comfort of the vehicle. These two are related, and it has been difficult to set each of them independently by the conventional setting methods. In general, the lower right corner of the screen of FIG. 11 tends to be preferable, but specifically, it is required to keep both within an appropriate range.
A point 1101 in the development study screen 1100 is a plot of the result of substituting the same explanatory variable values into the respective relational expressions of cornering power and longitudinal spring stiffness.
Lines 1102 to 1106 are obtained by changing the values of the explanatory variables using the point 1110 at the center of the screen as a reference point.
Of these, the line 1102 is obtained by changing the tire width, the line 1103 is obtained by changing the flatness, the line 1104 is obtained by changing the rim diameter, and the line 1105 is changed by changing the tire internal pressure. The line 1106 is obtained by varying the vertical load.

  In this way, by using the relational expression estimated by the method according to the present embodiment, it is possible to accurately examine how much the explanatory variable is changed and how much influence is exerted on the tire performance. it can. In the present embodiment, when estimating the relational expression, the relational expression is estimated based on the actual measurement values obtained from the empirical expressions for a large number of tires.

  In addition, although the example which calculates | requires the cornering power and vertical spring rigidity of a tire is shown in this embodiment, you may apply to the relational expression which calculates | requires values other than these, and may apply to other than a tire.

In this embodiment, multiple regression analysis using quantitative variables is used as multivariate analysis for obtaining the coefficients of explanatory variables. For example, if multiple regression analysis using dummy variables is used, radial analysis is performed. Qualitative variables such as “1” for tires, “2” for bias tires, “3” for solid tires, “1” for normal tires, “2” for studless tires, etc. may be used as explanatory variables. it can. Or, dummy variable 1 is radial tire “1”, bias tire “0”, solid tire “0”, dummy variable 2 is radial tire “0”, bias tire “1”, solid tire “0” A dummy variable may be set like this.
Furthermore, although multiple regression analysis is used in this embodiment, single regression analysis may be used as long as there is one explanatory variable.
In this embodiment, linear multiple regression analysis is used as multivariate analysis for obtaining the coefficients of explanatory variables. However, nonlinear multivariate analysis may be used.

[Summary]
According to this embodiment, by performing multivariate analysis based on experimental data in a plurality of tires, the relationship between the tire performance value and the characteristic value using attribute information such as tire dimensions and air pressure as explanatory variables The equation can be estimated.
By doing so, it is possible to generate a tire model in the case where the value of the attribute information, which has been fixed in the empirical formula model so far, is changed.
In addition, since a statistically determined model is generated from a plurality of tire data, an average performance value / characteristic value for each attribute information value can be determined, and accuracy can be improved.

As described above, according to the present embodiment, the value of the assumed tire attribute information is set, and when the value is changed, the performance and characteristics of the vehicle are quantitatively calculated. It can be simulated.
In addition, it is possible to generate a tire model for which no experiment has been performed and data has not been acquired.
The accuracy is sufficient and the predicted value can be supported.
It is also possible to estimate a tire model that does not exist.

In addition, according to the present embodiment, since a relational expression based on a large number of tire data is used, it is possible to reduce the risk of calculating a vehicle performance value or a specific value based on biased experimental data. it can.
Furthermore, since the value of performance and the like generally changes with the vehicle type, the performance value of the tire is often changed even at the same size at each stage of development. At this time, if a tire having an uneven performance value is set at the initial stage of development, subsequent changes and adjustments may be difficult. According to the relational expression obtained in the present embodiment, an average model can be obtained with respect to variations in tire performance values, the width of change, and the like. For this reason, by setting an average model obtained based on the relational expression obtained in the present embodiment at the initial stage of development, subsequent changes and adjustments are relatively easy.

1 computation server 2 database (storage unit)
3 terminal 10 relational expression estimation support system (relational expression estimation support device)
DESCRIPTION OF SYMBOLS 100 Processing part 101 Correlation calculation part 102 Multivariate analysis part 103 Predicted value calculation part 104 Verification part 105 Display processing part 201 Actual measurement data 202 Correlation graph data 203 Estimated data 204 Verification result data 301 Input part 302 Display part

Claims (3)

A relational expression estimation support method by a relational expression estimation support device that estimates the relational expression related to the tire using the attribute information of the tire as an explanatory variable,
The relational expression support device is:
The measured value of the attribute information obtained as a result of the experiment on the plurality of tires and the measured value of performance information that is the result of the performance experiment in the attribute information are stored in the storage unit,
An objective variable is selected from the performance information via the input unit,
A correlation graph indicating the correlation between the measured value of each of the attribute information and the measured value of the performance information selected as the objective variable corresponding to the measured value of the respective attribute information is displayed on the display unit,
By selecting the correlation graph via the input unit, the explanatory variable used in the relational expression is selected,
An original relational expression using the explanatory variable is input via the input unit,
By calculating the coefficient of the explanatory variable by multivariate analysis, the relational expression is estimated,
Displaying on the display unit a comparison screen between the predicted value of the performance information calculated by substituting the measured value into the estimated relational expression and the measured value of the performance information corresponding to the predicted value. Characteristic relational expression estimation support method. The relational expression estimation support method according to claim 1, wherein the multivariate analysis is single regression analysis or multiple regression analysis. Using the tire attribute information as an explanatory variable, a relational expression estimation support device that estimates a relational expression related to the tire,
A storage unit that stores an actual measurement value of the attribute information obtained as a result of an experiment on a plurality of the tires, and an actual measurement value of performance information that is a result of a performance experiment in the attribute information;
A correlation graph showing the correlation between the measured value of each of the attribute information and the measured value of the performance information selected as the objective variable corresponding to the measured value of each of the attribute information via the input unit. When selected, the explanatory variable used in the relational expression is selected, and when the original relational expression using the explanatory variable is input via the input unit, multivariate analysis is performed. A multivariate analysis unit that estimates the relational expression by calculating a coefficient of the explanatory variable;
A predicted value calculation unit that calculates a predicted value of the performance information by substituting the measured value into the estimated relational expression;
A display processing unit that displays the correlation graph on a display unit, and displays a comparison screen between the predicted value of the performance information and the actual measurement value of the performance information corresponding to the predicted value;
A relational expression estimation support apparatus characterized by comprising: