JP2009080076A - Method and device for evaluating shape of molding - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a technique accurately and rapidly performing shape evaluation of a molding while reducing a burden on an arithmetic processor. <P>SOLUTION: Shape data consisting of many data point sequences is obtained by measuring the surface shape of the molding, and a basis function for making connection between adjacent data points for each data segment is defined. Next, after thinning the data points, the basis function for making connection between adjacent data points for expanded each data segment is redefined, and the error between the redefined basis function and each thinned data point is evaluated. The thinned data points are restored in the data segment where the error exceeds an allowance, but the basis function and data amount are compressed by further repeating the thinning of data points in a data segment where the error is the allowance or lower, and the shape of the molding is evaluated. <P>COPYRIGHT: (C)2009,JPO&INPIT

Description

  The present invention relates to a molded product shape evaluation method and apparatus suitable for automatically evaluating the surface shape of a molded product such as a press molded product or an injection molded product.

  Dimensional accuracy analysis of a press-molded product or the like is an operation necessary for optimizing a molding die shape and molding conditions. For this reason, conventionally, the shape of a molded product is measured by a shape measuring machine, the shape data of a specific section is extracted and taken out to a computer terminal or the like, and the shape is evaluated by manual data manipulation. Such work requires, for example, about 30 minutes, but there is no problem when there are a small number of products to be evaluated. However, for example, when it is desired to evaluate the shape of the total number of molded articles that are continuously molded, it is impossible to perform such a manual operation. Thus, attempts have been made to automate the evaluation of the shape of molded products, but automation has been difficult for the following reasons.

  First, the shape data obtained by the shape measuring machine is a data point sequence of equal intervals or random intervals, and does not have numerical information of shapes necessary for automation. That is, the shape data is merely a set of coordinates representing the surface shape and does not itself have numerical information on the shape (for example, the gradient of the forming surface, the length of the straight portion, the amount of springback, etc.).

  Secondly, since the shape data obtained by the shape measuring machine is generally a data point sequence of several hundred to several thousand points, the capacity of the storage device may be squeezed and a long time is required for the calculation processing. Therefore, it was impossible to evaluate the shape of many molded products at high speed. In order to avoid this problem, it is conceivable to reduce the number of data points of the shape data. However, if the number of data points is simply reduced, necessary shape data is lost and the detailed shape cannot be expressed. That is, when performing shape evaluation, it is required to shorten the data interval of the necessary portion and increase the data interval in the straight portion, but if the number of data points is simply reduced, the necessary shape data will be lost.

In Patent Document 1, the shape of a press-molded product or the like is measured by a shape measuring machine to obtain an approximate curve, an error between the measured value and the approximate curve is calculated, and when the error exceeds an allowable range, A shape evaluation method for determination is described. However, since this method handles the shape data of a large number of points as they are without reduction, there is still a problem that it takes a lot of time for the arithmetic processing in addition to squeezing the capacity of the storage device.
Japanese Patent Laid-Open No. 6-288863

  Therefore, the object of the present invention is to solve the above-mentioned problems of the prior art, and to accurately and quickly automatically evaluate the shape of a molded product such as a press-molded product, and to the portion of the data that does not affect the accuracy. It is an object to provide a molded product shape evaluation method and apparatus capable of automatically reducing the number of points and reducing the burden on an arithmetic device.

The method for evaluating the shape of the molded product of the present invention made to solve the above problems is as follows.
A first step of measuring the surface shape of the molded product with a shape measuring machine to obtain shape data comprising a number of data point sequences;
A second step of defining a basis function connecting adjacent data points for each data interval;
A third step of redefining a basis function connecting adjacent data points for each enlarged data section after thinning data points;
A fourth step of evaluating an error between the redefined basis function and each thinned data point, and restoring the thinned data point in a data section where the error exceeds an allowable value;
A fifth step of further thinning out the data points in the data section where the error is less than or equal to the allowable value;
A sixth step of redefining a basis function connecting the adjacent data points thereafter;
A seventh step of compressing the basis function and the data amount by repeating the fourth to sixth steps a plurality of times;
An eighth step of outputting the compressed shape data from the data output unit;
It is characterized by comprising.

  In the present invention, data points can be thinned out, for example, every other point. Moreover, it is preferable to use any one of a primary spline function, a secondary spline function, and a cubic spline function as a basis function. As an error evaluation method, the least square error or the maximum error can be used.

  In addition, the molded product shape evaluation apparatus of the present invention includes a shape measuring machine for measuring the surface shape of a molded product, an arithmetic processing device for shape data composed of a large number of data point sequences output from the shape measuring machine, and an external device. This data processing unit includes a basis function defining unit that defines a basis function that connects adjacent data points for each data section, a data reduction unit that performs data point thinning, a basis function, And an error evaluation unit that evaluates the error between each thinned data point, and in the data section where the error is less than or equal to the allowable value, it further repeats the thinning of the data points to compress the basis function and the data amount It is characterized by being. A laser displacement meter can be used as the shape measuring machine.

  According to the present invention, the error is evaluated while thinning the shape data consisting of a large number of data point sequences measured by the shape measuring machine, and the data thinning is stopped in the data section where the error exceeds the allowable value. In the data section where the error is less than the allowable value, the basis function and the data amount are compressed by further repeating the data thinning. As a result, a lot of shape data remains in the part with a small shape change, and the number of shape data is greatly reduced in the part close to the plane, and the total number of data points is reduced to the first tenths. Can do. Therefore, shape evaluation can be performed automatically accurately and quickly, and the burden on the arithmetic unit can be greatly reduced.

Embodiments of the present invention will be described below.
(First step)
In this embodiment, the shape evaluation of the press-molded molded product 1 having the cross-sectional shape shown in FIG. 1 is performed using a shape measuring machine such as a laser displacement meter. The molded product 1 is placed on the base 2 of the shape measuring machine, and the shape data is measured at predetermined intervals while the sensor head 3 travels along a rail (not shown) provided on the upper surface thereof.

  This molded product 1 is formed by connecting the straight portion 4 (plane portion) of the base portion and the tip portion 6 with a gently curved surface 5 and having a tip R at the most distal end portion. The object is the amount of warping of the tip portion 6 with respect to the flat portion 4 serving as the reference line. In order to perform automatic calculation, the flat portion 4 serving as the reference line and the starting position of the tip R are determined and then It is necessary to calculate the distance between Since the amount of warpage varies depending on the amount of springback of the molded product, it becomes an important evaluation index in performing molding management.

  The measurement by the shape measuring machine is normally performed at a constant pitch. In this embodiment, as shown in FIG. 2, shape data consisting of a sequence of 800 data points is output from the shape measuring machine to the arithmetic processing unit 7. The shape data may be output to the arithmetic processing unit 7 after being subjected to data processing such as averaging in advance in the shape measuring machine. This data point sequence is shown without breaking the shape of the tip R, but on the other hand, there are more data point sequences than necessary on the flat part 4, and this is wasteful.

  For reference, an example in which the number of data points is simply reduced to 1/10 is shown in FIG. In FIG. 3, the number of data points of the flat part 4 is reduced successfully, but the tip R is deleted evenly to the data points necessary to reveal its shape, so that the shape has collapsed and the shape is correctly evaluated. Can not do it. Therefore, in the present invention, in order to reduce the number of shape data as much as possible in a portion close to a plane as much as possible while leaving a lot of shape data in a portion with a fine shape change, and to greatly reduce the number of data points as a whole, The second to seventh steps shown below are executed in the arithmetic processing unit 7.

  Note that the arithmetic processing unit 7 is a computer device, and as shown in FIG. 4, a basis function constructing unit 11, a data reducing unit 12 for thinning out data points, and a basis function between each data point thinned out. And an error evaluation unit 13 for evaluating the above error, and a data output unit 14 to the outside. The function of each part will be described sequentially.

(Second step)
After shape data consisting of a large number of data point sequences is input from the shape measuring machine to the arithmetic processing unit 7 in the first step, first, the basis function construction unit 11 defines basis functions that connect adjacent data points for each data section. To do. The situation is as shown in FIG. 5, and first, a basis function is defined for the interval between all data points input to the arithmetic processing unit 7. Here, the basis function is a set of functions that are defined for each section divided from the original section data and satisfy the boundary condition.

  A spline function can be used as a specific basis function. As summarized in Table 1, it is preferable to select one of a primary spline function, a secondary spline function, and a cubic spline function. The primary spline function is simply a polygonal line connecting adjacent data points with a straight line, and is the simplest to implement. In this embodiment, this linear spline function is adopted as a basis function. However, depending on the size evaluation method, it may be desired to obtain a derivative at the boundary. In this case, it is necessary to select a quadratic spline function or a cubic spline function. A quadratic spline function connects adjacent data points with a parabola, but is not used in practice. On the other hand, since the cubic spline function is essential for the curvature evaluation, the cubic spline function should be adopted when the curvature evaluation is desired. Table 1 shows the expressions and boundary conditions of these spline functions. Note that a method for defining a spline function between adjacent data points is known and can be easily obtained from a formula of the spline function.

(Third step)
In the second step, since all data points are used and the basis functions are defined for each data section, if there are 800 data points, 799 basis functions are defined. However, since the number of data points is too large and the burden on the arithmetic processing unit 7 is heavy, the data reduction unit 12 thins out the data points in the third step.

  The situation is as shown in FIG. 6, and in the example of FIG. 6, data points are thinned out every other point. In other words, even-numbered data points are thinned out. As a result, the data sections Φ1 and Φ2 are merged (integrated) and expanded to the data section Φ1 + 2, and the data sections Φ3 and Φ4 are merged and expanded to the data section Φ3 + 4. Has been. Note that the thinning method is not necessarily limited to every other point, and can be performed every other point.

  Since the odd-numbered data points remain when the data points are thinned out as described above, the basis function forming unit 11 redefines the basis functions connecting the adjacent data points for each enlarged data section. In other words, the basis functions of adjacent data sections are merged. This is shown in the upper right diagram of FIG. As a result, the number of data points and the number of basis functions are halved, but on the other hand, an error occurs between the redefined basis function and the thinned data points.

(4th step)
Next, the error evaluation unit 13 evaluates an error between the redefined basis function and each thinned data point. As an error evaluation method, a least square error or a maximum error in a finite interval can be used. These are defined in the section when there are N measurement data sets [x1, y1,... XN, yN] on the section Ii = [ai, bi] shown in FIG. This represents an error from the basis function. Table 2 shows the mathematical definition and meaning of least square error or maximum error. In this embodiment, the error is evaluated for all sections using the least square error.

  Note that the measurement standard deviation σ of the shape measuring machine can be used as a guideline for setting the allowable value. If σ = 0.03 to 0.10 mm, the data deterioration according to the present invention can be ignored. When the error is evaluated in this way, it is possible to distinguish between a data interval in which the error is less than or equal to an allowable value and a data interval in which the error exceeds the allowable value. Therefore, since the data point thinning is not appropriate for the data section where the error exceeds the allowable value, the thinned data point is restored. This state is shown in the lower left diagram of FIG.

(5th step)
After the first reduction in the number of data points and basis functions is completed in this way, in the fifth step, the data reduction unit 12 further thins out data points for a section where the error is less than or equal to the allowable value.

(6th step)
In a sixth step, a basis function that connects adjacent data points is redefined in the section where the thinning is performed.

(7th step)
Thereafter, returning to the fourth step, the error is evaluated, and for the data section where the error exceeds the allowable value, the thinned data point is restored, and for the section where the error is less than the allowable value, the data point is further reduced. The fifth and sixth steps of thinning out and redefining the basis function are repeated a plurality of times. The above flow is shown in FIG.

(8th step)
In the eighth step, the compressed shape data is output from the data output unit 14 to the outside.

  As a result of executing the above steps, the data points are greatly reduced in the portion where the surface shape of the molded product is flat, but the data points are not reduced in the portion where the surface shape is changed. The final state is reached as shown below. At the top left of FIG. 6, the number of data points was 11 and the basis function was 10, but in the final state, the number of data points was 4 and the basis function was 3. Moreover, it can be seen that the overall shape features remain. By using shape data with a greatly reduced number of data in this way, the load on the arithmetic unit is reduced, and shape evaluation can be performed automatically. Specific results are shown in the examples.

  A molded product (steel type: GA590) having the shape shown in FIG. 8 was molded by a press molding machine having a molding load of 140 tons and a wrinkle pressing force of 15 tons. The purpose of shape evaluation of this molded product is to grasp the amount of warping of the tip portion 6 when the straight portion 4 of the base is used as a reference plane. For shape measurement, a laser displacement meter manufactured by Keyence Corporation was used to measure the cross-sectional shape of the central portion. Here, the number of data points to be output is 800 points.

  A linear spline function is used as the basis function, the allowable error is 0.1 mm, which is the measurement variation of the laser displacement meter, the least square error is adopted as the error evaluation method, and the number of data points is reduced according to the steps shown in the embodiment. Was done. The final result is shown in FIG.

  In FIG. 10, although the number of data points is greatly reduced from the first 800 points to 40 points, the portion where the shape change suddenly is fine, the flat portion is rough, and the overall efficient approximation is achieved. Has been made. Further, since the shape change is expressed by the interval of the section, the “reference line” necessary for automatic calculation of the amount of warping of the tip end portion 6 when the straight line portion 4 of the base is used as the reference plane and the “evaluation point” of the tip end "Can be automatically detected. That is, the maximum line segment in the section [200, 400] in FIG. 9 is set as the “reference line”, and the end point of the minimum line segment in the section [0, 100] is set as the “evaluation point”. Can be automatically calculated as the “warping amount”.

  As a result, when the conventional manual shape evaluation is performed, the number of data points is 800 and the required time is 30 minutes. However, according to the present invention, automatic calculation is possible, the number of data points is 40, and the required time is It was 1 minute. In this way, it was confirmed that the data capacity and required time can be reduced to several tenths.

It is explanatory drawing of a shape measuring machine. It is a graph which shows the data point sequence output from the shape measuring machine to the arithmetic processing unit. It is a graph which shows the data point sequence which reduced the number of data points to 1/10 simply. It is a block diagram which shows an apparatus structure. It is explanatory drawing of the basis function definition in a 2nd step. It is explanatory drawing of the reduction process of the data score in this invention. It is a flowchart of the reduction process of the data score in this invention. It is a perspective view of the molded article used as the evaluation object in an Example. It is a graph which shows the data point sequence of the final result in an Example.

Explanation of symbols

DESCRIPTION OF SYMBOLS 1 Molded product 2 Base of shape measuring machine 3 Sensor head 4 Straight line part 5 Curved surface 6 Tip part 7 Arithmetic processor 11 Basis function structure part 12 Data reduction part 13 Error evaluation part 14 Data output part

Claims (6)

A first step of measuring the surface shape of the molded product with a shape measuring machine to obtain shape data comprising a number of data point sequences;
A second step of defining a basis function connecting adjacent data points for each data interval;
A third step of redefining a basis function connecting adjacent data points for each enlarged data section after thinning data points;
A fourth step of evaluating an error between the redefined basis function and each thinned data point, and restoring the thinned data point in a data section where the error exceeds an allowable value;
A fifth step of further thinning out the data points in the data section where the error is less than or equal to the allowable value;
A sixth step of redefining a basis function connecting the adjacent data points thereafter;
A seventh step of compressing the basis function and the data amount by repeating the fourth to sixth steps a plurality of times;
An eighth step of outputting the compressed shape data from the data output unit;
A method for evaluating the shape of a molded product, comprising:   2. The method for evaluating a shape of a molded product according to claim 1, wherein thinning of data points is performed every other point.   The shape evaluation method for a molded product according to claim 1, wherein any one of a primary spline function, a quadratic spline function, and a cubic spline function is used as a basis function.   2. The method for evaluating a shape of a molded product according to claim 1, wherein a least square error or a maximum error is used as an error evaluation method.   A shape measuring machine for measuring the surface shape of a molded product, a shape data arithmetic processing device comprising a large number of data point sequences output from the shape measuring device, and an external data output unit. Is the basis function component that defines the basis function that connects adjacent data points for each data interval, the data reduction unit that thins out the data points, and the error between the basis function and each data point that is thinned out An error evaluation unit that evaluates the data, and in the data section where the error is less than or equal to the allowable value, the data point is further thinned, and the function of compressing the basis function and the data amount is provided. Shape evaluation device.   6. The molded product shape evaluation apparatus according to claim 5, wherein the shape measuring machine is a laser displacement meter.

Images