JP5309288B2 - Computer program for machining error prediction, machining error prediction device, and device for correcting tool path based on the prediction result - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a device for predicting a machining error caused by the shape error or bending of a tool at a high speed. <P>SOLUTION: A tool M (trajectory of a cutting blade) at a position where a machining error is predicted and tool sweeping objects T1 and T2 in a predetermined range including the objects adjacent to the position and a shape W of an object to be cut are plotted in a visual field range V from the lower part by using the three-dimensional graphics function of a computer device 1. A pixel region where the tool M is displayed at the front in the plotted region is detected as a region of a machining surface remaining after the end of machining, and a machining error is predicted from the shape error or deflection quantity of the tool at a representative point (creation point C) of the machining surface. Thus, it is possible to sharply reduce man-hours and time required for manual correction after forming by correcting the tool path based on the predicted machining error. <P>COPYRIGHT: (C)2011,JPO&amp;INPIT

Description

  The present invention relates to prediction of errors in cutting, and particularly relates to a technique for predicting machining errors due to tool shape errors and bending using a three-dimensional graphics function of a computer apparatus.

  In general, molds used in the manufacture of press products, etc., work on the work by moving the rotary tool based on the tool movement path data (NC data) output from CAM (Computer Aided Manufacturing). It is made by doing. The CAM accurately fits the CAD (Computer Aided Design) model shape, which is the target shape, based on the tool information such as the ball end mill and the tool information such as its diameter and the processing conditions such as the feed pitch. So that the path is output (in the case of finishing).

  However, during actual machining, the tool bends due to the cutting force, and the tool itself has a dimensional error and an error in attaching to the rotating shaft of the machine tool. As shown in FIG. 5, it is inevitable that an error (processing error) occurs in the CAD model shape. For this reason, it is usually necessary to manually correct a portion with a large machining error in the die after cutting, and a great amount of man-hours and time are spent repeating this correction and measurement.

  There is an approach to solve the problem of tool deflection due to the cutting force by correcting NC data. The most basic one is to calculate the cutting volume per unit time, etc., and to change the feed rate based on the cutting volume to suppress the fluctuation of the cutting volume, but to reduce the cutting volume Therefore, it is necessary to reduce the feed rate of the tool, and the efficiency is lowered. Further, even if the feed rate is reduced, the tool will be bent, so that the machining error is not essentially solved. In addition, a ball end mill is generally used in die machining. However, in ball end milling, the direction and magnitude of the cutting force changes during one rotation of the tool, so it is difficult to accurately estimate the amount of tool deflection from the removed volume. there were.

  Therefore, a technique has also been proposed in which tool deflection is calculated in advance, and the coordinate value of the tool path is changed so as to cancel machining errors caused by the deflection (see Patent Documents 1, 2, and 3). That is, in Patent Document 1, the cutting force and the tool deflection are calculated using the structure calculation software, but it takes a lot of time to calculate the cutting force and the tool deflection using the structure calculation software. It is difficult to carry out the processing within a practical time for a cutting process for creating a product having a large and complicated shape such as.

  On the other hand, in Patent Document 2, the bending is calculated based on the SN (Surface Nominal) at the cutting point, the DSM (Dynamic Stiffness Matrix) of the machine tool, the cutting force, and the like. The calculation method of the machining error due to the above is not described. Further, in Patent Document 3, the correction amount is calculated from the normal vector of the machining surface and the cutting condition by an arithmetic expression based on an empirical rule or a database. An enormous amount of experimentation and evaluation is required for each posture relative to the rotational axis of the tool.

In consideration of these points, the inventors of the present application have proposed a method for predicting a cutting force by determining a cutting state of a tool at high speed in machining a complex shape having a free-form surface (Non-patent Document 1). 2). This is to predict the cutting 3 component force during one rotation of the tool at an arbitrary tool position at high speed by determining the cutting state of the tool using the three-dimensional graphics function of the computer device. The amount of deflection during one rotation of the tool can be calculated from the cutting force calculated by this method and the rigidity of the tool.
JP 2002-126634 A JP 2004-174620 A JP 2005-144620 A 2007 JSPE Autumn Conference Academic Lecture Proceedings pp67 Proceedings of 2008 Precision Engineering Society Ehime Regional Conference Lecture pp45

  However, since a tool such as a ball end mill itself exhibits a vibrational behavior, even when receiving a pulsed cutting force as shown in FIG. ), The amount of deflection greatly varies depending on the rotation angle. For this reason, it is not sufficient to calculate the cutting force during one rotation of the tool or during one blade feed, and to obtain the average amount of bending by this, and to predict machining errors due to this bending with high accuracy. In order to do this, it is extremely important to calculate the machining error from the amount of deflection at one point during one rotation of the tool.

  Specifically, first, FIG. 3 shows a schematic view of the side surface processing using a ball end mill as viewed from above along the rotational axis. While the tool rotates once, its cutting edge bites away from the work piece to perform cutting, and a part of the cut area (cutting area) is cut off by the next tool rotation. Therefore, not all of the cutting area remains as a machining surface.

  FIG. 4 shows a side view of processing a curved surface with a ball end mill. First, on the left side of the figure, the tool is sent from the front side of the paper to the back side (Y axis + direction). It is assumed that linear movement in a direction perpendicular to the paper surface is repeated, such as pick-feeding to the right side (X axis + direction) and feeding again in the depth direction of the paper surface. In this case, a part of the area cut in one path is cut off in the next path, and the area finally remaining as a processing surface is only a part.

  That is, of the tool deflection that changes during one rotation as described above, the machining error affects only when cutting the region that remains as the machining surface, that is, when creating the machining surface. In order to predict the machining error, it is necessary to know which position (rotational angle position) of the cutting blade is in contact with the work piece at that time. However, in machining a mold having a complicated shape such as a free-form surface, it is a tremendous amount of time to obtain the tool rotation angle when creating a machining surface at an arbitrary tool position using the tool path data by ordinary geometric calculation. It is difficult to say that it is realistic.

  Further, as a cause of the machining error, there is a shape error when the tool is rotated with the tool attached to the machine tool, in addition to the above-described tool deflection. In other words, the CAM uses a hemisphere of the radius of the ball end mill to be used to represent the trajectory of the cutting edge and create a tool path. However, an actual tool itself has a dimensional error and the rotation axis of the machine tool. This is because there is misalignment when mounting, and further, the trajectory of the cutting blade rotating for unbalance during actual machining deviates from the ideal shape.

  Regarding this point, measure the shape error by measuring the outer diameter of the cutting edge, that is, the tool diameter, in several cross sections (tool axis cross sections) perpendicular to the rotation axis in the state where the tool is rotated. However, in order to calculate the machining error due to the shape error using the data measured in this way, it is necessary to know which tool axis section the cutting edge is in contact with the work piece at the time of creation. It is not realistic to do this by ordinary geometric calculation.

  Because of these factors, there is no actual product that can predict machining errors due to tool shape errors or deflections within a practical time. It is an object of the invention.

  In order to achieve the above object, the present invention uses the three-dimensional graphics function of a computer device to cut the portion remaining as a machining surface after machining, for example, the cutting end position of a ball end mill. The rotation angle position around the tool rotation axis of the blade, the position in the direction of the rotation axis, and the like are calculated.

Specifically, the invention of claim 1 is directed to an apparatus for predicting a machining error of a workpiece by a rotating tool, the shape of the workpiece before machining by the tool, the cutting edge shape of the tool, and the Data input means for receiving input of predetermined data including information on the movement path by feeding, and the trajectory of the cutting blade at a position where a machining error is predicted on the movement path of the tool based on the predetermined data, the position A tool sweep body of a predetermined range including immediately before and immediately after, and the work piece are drawn in a visual field range in which the entire trajectory of the cutting blade can be desired from the work object side, and the tool sweep and drawing means for drawing in all of the field of view for the body, thus the pixel region passing track is displayed in front of the cutting edge in the rendered image, remaining after processing end processing From the machined surface detection means for detecting the machining area and the pixel area thus detected, the creation position, which is the position of the cutting edge for cutting the workpiece when creating the machining surface, is specified, and the tool position at this creation position is determined. Machining error calculating means for calculating an estimated value of the machining error based on at least one of the shape error and the deflection amount.

  When predetermined data such as the workpiece and tool specifications and the route thereof are input to the machining error prediction apparatus having the above-described configuration, the trajectory of the cutting edge at the predicted position of the machining error and a predetermined range including immediately before and after the position. The tool sweeper and the workpiece are each drawn by a drawing means in a predetermined visual field range where the tool is desired from the side of the workpiece. Strictly speaking, immediately before and immediately after the predicted position of the machining error means that the cutting blade is fed by one blade before it is fed by one tool, but it is not limited to this. A range within 2 to 3 blade feeding is assumed, and in this range, substantially the same result is obtained.

  Further, as the drawing means, for example, graphics hardware for drawing a three-dimensional polyhedron shape on a display can be used. In this way, among the necessary geometric operations, CPU (Central Processing) can be used. Unit) can perform processing such as coordinate transformation, projection, hidden surface removal, and nearest point calculation, which are difficult to perform in real time, at a high speed, thereby realizing a significant improvement in drawing speed.

  Then, it is possible to estimate the range of the processed surface remaining after the end of processing from the pixel region in which the trajectory of the cutting edge is displayed in the foreground in the image drawn as described above. Therefore, if this pixel area is detected by the machining surface detection means, the cutting edge position (machining surface creation position) of the tool when cutting this machining surface is specified. From the positional relationship, the rotation angle position around the tool rotation axis and the position in the direction of the rotation axis of the creation position can be calculated, and based on this, machining caused by the tool shape error or deflection at the creation position The error can be estimated.

  More specifically, for example, in the state where the tool is rotated as described above, the tool diameter, that is, the outer diameter of the trajectory of the cutting blade, is measured in several cross sections (tool axis cross sections) orthogonal to the rotation axis. A table in which the deviation from the ideal shape (shape error) is stored in advance in association with the position in the direction of the rotation axis of the tool by measurement or the like is created, and the machining error calculation unit is configured to store the tool at the specified creation position. The shape error may be calculated using the table (claim 2).

  Further, the machining error calculation means calculates the relationship between the rotation angle position around the rotation axis of the tool and the amount of deflection from the magnitude of the cutting force at the predicted machining error position and the rigidity of the tool (preferably dynamic rigidity). You may obtain | require and you may comprise so that the amount of tool deflection in the specified creation position may be calculated.

  When the passing locus of the cutting blade is formed in an axially symmetric shape centering on the rotation axis of the tool as described above, it is preferable that each pixel on the passing locus has a rotation angle position around the rotation axis of the tool and Predetermined pixel information is provided so that the value changes according to at least one of the positions in the rotational axis direction. In this way, the creation position can be easily identified from the pixel information of the detected machining surface area, and the rotation angle position and the position in the rotation axis direction of the tool can be identified directly from the pixel information of the creation position. This is advantageous for speeding up the processing.

  The axially symmetric shape representing the trajectory of the cutting edge can be represented as a semispherical surface approximately by a ball end mill, for example. However, the shape is not limited to this, and various shapes may be used depending on the shape of the cutting edge. is assumed. Specific examples of the pixel information include RGBA color information, depth, stencil index, and the like. However, the pixel information is not limited to one of these, and two or more of them may be used to determine the rotation angle position and rotation axis of the tool. What is necessary is just to assign to the identification information etc. which are a direction position or a tool.

  Preferably, the drawing means first draws the trajectory of the cutting edge at the predicted position of the machining error, the tool sweeping body up to immediately before the position, and the workpiece, and then In this configuration, the tool sweep body immediately after the predicted position is drawn (Claim 5), and the cutting edge in the image drawn before the tool sweep body after the predicted position is drawn is added. If the pixel area in which the trajectory is displayed in front is detected as the cutting area by the cutting blade (cutting area detecting means), the cutting force acting on the tool when cutting this area can be calculated (cutting) Force calculation means).

  Further, in this case, it is preferable to correct the viewpoint or visual field range in the drawing means based on the information on the cutting area (visual field correction means: claim 6). In other words, if the cutting area is viewed obliquely with respect to the normal line, the pixel area becomes smaller, so that the area remaining as the processing surface becomes smaller, and if it is less than one pixel, it is not drawn. There can be.

  Therefore, for example, when the cutting area is considered to be viewed obliquely, as in the case of the vicinity of the outer periphery of the trajectory of the cutting edge regarded as a hemisphere, the viewpoint should be approached so as to approach the normal direction of the cutting area. The viewpoint in the drawing means is either expanded by moving the pixel area, or narrows the field of view toward the cutting area even if the viewpoint remains as it is, and reduces the area per pixel to increase the resolution. Or correct the field of view. By doing so, it is possible to prevent the machined surface from being overlooked and to improve the accuracy of detecting the machined surface creation position.

  In addition, the machining error prediction apparatus according to the present invention divides the movement path of the tool, the tool sweep body in any one of the division paths, the tool sweep body up to this division path, and the target object. First sub-drawing means for drawing a workpiece from the side of the work piece with a tool desired, and a pixel region in which the tool sweep body of the divided path is displayed in the foreground in the drawn image. If there is a predicted position of the processing error in the pixel area, a predicted position setting unit that does not set the predicted position of the processing error in the divided path if there is no pixel area displayed in the foreground, (Claim 7).

  In this way, if the tool path set as NC data includes a divided path (path) that does not actually cut, machining is performed only for the path that actually cuts. An error can be predicted. In addition, when a cutting area and a non-cutting area are mixed in one pass, if a machining error is predicted in a non-cutting area, the calculation error becomes zero. If you do like that, there is no worry.

  Furthermore, since the above-described region to be cut can be specified as a range where the above-described cutting region and the surface of the machining surface can exist, in the subsequent processing, only the tool sweeper and the workpiece shape existing within the range are included. You may make it draw, and you may make it read pixel information only within the area | region. In this way, the processing speed can be further increased.

  And a second sub-drawing means for drawing the tool sweeper in each of the paths obtained by dividing the path of the tool and the work piece while drawing the tool from the work piece side. Estimate the angle formed by the tool rotation axis and the work surface for each path where the tool sweeper is displayed in the foreground in the image drawn by the above, and the machining error related to the divided path with this angle equal to or less than the specified value This prediction may be prohibited (claim 8).

  This is because when the angle between the rotation axis of the tool and the normal line of the work surface is close to 0 °, even if the tool is bent by the cutting force, there is almost no effect on the machining error. Therefore, the processing speed is increased by not predicting a machining error for a path whose angle is equal to or less than a predetermined value (for example, 30 to 60 °).

  The ninth aspect of the present invention provides a tool movement path so as to reduce the machining error based on the value of the machining error predicted by the machining error prediction device according to any one of the first to eighth aspects. By providing a path correcting means for correcting, in this way, by correcting the moving path of the tool in advance, it is possible to reduce processing errors and significantly reduce the man-hours and time required for correction after forming.

The present invention is also directed to a program for predicting a machining error of a workpiece by a rotating tool using a three-dimensional graphics function of a computer device, and the workpiece before machining by the tool. A data input step for receiving input of predetermined data, including information on the shape of the tool, the shape of the cutting edge of the tool, and the movement path of the tool, and machining errors on the movement path of the tool based on the predetermined data The trajectory of the cutting blade at a position where the cutting edge is predicted, the tool sweeping body within a predetermined range including immediately before and after the position, and the workpiece, respectively , the entire trajectory of the cutting blade from the workpiece side. draw the field of view that can desire, cutting edge in the drawing step and thus the drawn image to be drawn in all of the field of view for said tool sweeps body A machining surface detection step for detecting a pixel area in which the passing trajectory is displayed in front as a machining surface area remaining after machining is completed, and the workpiece is cut from the detected pixel area when the machining surface is created. A machining error calculating step of identifying a generating position that is a cutting edge position of the tool and calculating an estimated value of the processing error based on at least one of a shape error and a deflection amount of the tool at the generating position. (Claim 10).

  If such a computer program is executed by a general-purpose computer device, this computer device becomes the machining error prediction device according to the invention of claim 1 and the above-described operation is obtained.

  As described above, according to the machining error predicting apparatus and the like according to the present invention, the cutting edge of the tool for cutting the workpiece remains after the machining is finished by using the three-dimensional graphics function of the computer. Since the position (creation position) when creating a surface can be specified and the shape error and the amount of deflection of the tool at this creation position can be calculated at high speed, errors can occur even when machining dies with free-form surfaces. Can be accurately predicted, which is very advantageous for studying processing conditions. Further, by correcting the tool path based on the predicted machining error, it is possible to reduce the machining error and greatly reduce the man-hours and time required for correction after forming.

  Hereinafter, embodiments of the present invention will be described with reference to the drawings. In addition, the following description is only illustrations essentially and does not intend restrict | limiting this invention, its application thing, or its use.

(System overview)
The machining error prediction apparatus according to this embodiment can be configured by a general personal computer in addition to a general-purpose workstation. By installing and executing a required program in this computer apparatus, for example, errors in cutting of a mold can be predicted at high speed. Although not shown, the computer device is provided with a storage device such as a hard disk drive, and is connected to an image display device such as a display and an input device such as a keyboard and a mouse.

  Further, the computer apparatus of this embodiment includes graphics hardware for performing three-dimensional drawing at high speed. This is called, for example, a GPU (Graphics Processing Unit) and is used for drawing a three-dimensional polyhedron shape at high speed. Among the geometric operations necessary for this, a CPU (Central Processing Unit) performs real-time processing. It has the ability to compute difficult processes such as coordinate transformation, projection, hidden surface removal, and nearest point calculation in parallel.

  The graphics hardware may be configured as, for example, a board, a card, or another device, and may be mounted on a computer device or may be built in a chip set. In some cases, it may be configured separately from the computer apparatus. As an attempt to apply the calculation function of the graphics hardware to the problem of cutting, for example, for drawing a machining result at high speed (see Japanese Patent Laid-Open No. 2000-235407), for creating a tool path at high speed Are already known (see Japanese Patent Application Laid-Open No. 2001-242919), and those for determining interference of machine tools (see Japanese Patent Application Laid-Open No. 2006-244067).

  As an example, a case where a ball end mill is attached to an NC triaxial machine and a die is cut will be described below. In general, in a three-axis machine, a work piece fixed on a table or the like is cut by a tool such as an end mill from the upper part to the side part, and the tool has its rotation axis, that is, the tool axis in the vertical direction ( It is attached to the drive unit (in the Z-axis direction), and the cutting edge at the lower end thereof is positioned by numerical control (NC) with respect to three orthogonal axes including the horizontal plane (XY plane).

  Then, when the tool is moved based on the data (NC data) of the moving path, as described above with reference to FIG. 3 and FIG. A part of the cut area is cut off by the next one-blade feed, and the cutting area in one path of the tool sent in this way is also cut off by cutting in the adjacent subsequent path, and the machining surface is finished at the end of machining. The remaining area will be considerably smaller. As shown in FIG. 4, the final machining surface area is the lowermost part (Z value is the minimum value) of the overlapping tool paths.

  Therefore, as shown in FIG. 5, the tool M at the position where the machining error is predicted on the tool path P (movement path) set as NC data, that is, the trajectory of the cutting edge is defined as the rotation axis of the tool. The tool sweep body T1 is represented by a hemispherical surface having a center on the top, and is fed by one blade at that position (feed speed / (tool rotation speed × number of blades)), and the tool sweep body after feeding one blade. T2 and the workpiece shape W are each represented as a virtual polyhedron and drawn in a predetermined visual field range from below (the negative direction of the Z axis).

  In addition, since the tool is sent even while the cutting blade makes one revolution, strictly speaking, it can be said that the trajectory of the cutting blade is not a hemispherical surface. However, since the feed speed is generally very small with respect to the tool diameter, the passing trajectory of the cutting edge may be approximately regarded as a hemispherical surface.

  Further, the tool sweep bodies T1, T2 represent a sweep shape when the tool M is moved along the tool path P while rotating around the rotation axis, as shown in FIG. It is configured by combining a shape and a hemispherical shape. At this time, a conventionally known method (for example, see Japanese Patent Application Laid-Open No. 2009-020671) for reducing an error due to polyhedron is effective. Further, the workpiece shape W can be expressed as a polyhedron approximated by a large number of triangles, for example, as STL data, except for the lower surface thereof.

  Thus, when drawing, the tool M is provided with color information (pixel information, which will be described later in detail) which is different from the tool sweep bodies T1, T2 and the workpiece shape W. Further, the viewpoint is set in the negative direction of the Z axis with respect to all the tool sweep bodies T1, T2 and the workpiece shape W as shown in the figure, and the direction of the line of sight is the Z axis + direction. The center of the field of view coincides with the rotation axis of the tool M at the predicted position of the machining error, and the field of view is the diameter of the tool M both vertically and horizontally (see FIG. 12). The field-of-view range V thus set has a rectangular parallelepiped shape that is long in the Z-axis direction.

  Since only the lowermost image is displayed in the image drawn in the visual field range by the hidden surface removal function of the graphic hardware, the processing is performed from the pixel area where the color of the tool M is displayed in this image. It is possible to estimate the range of the processed surface remaining after the end. If the final machining surface can be detected from the cutting area in this way, the creation point (creation position) of this machining surface is specified as described below, and the tool rotation angle (rotation angle position) is determined from the positional relationship with the tool axis. Then, the tool contact angle (position in the direction of the rotation axis) is obtained, and thereby the machining error caused by the deflection or shape error of the tool M can be calculated.

-Process flow-
FIG. 7 shows an overall flow of processing for processing error prediction in this embodiment. First, in the data input step, tool specifications such as a tool type, a tool diameter, and a cutting edge shape are input to the computer apparatus 1 as necessary parameters. The cutting edge shape of the tool is identified by using, for example, a photograph viewed from the lower side of the tool as shown in FIG. The cutting force coefficient (specific cutting resistance) used for the calculation of the cutting force is obtained by performing a cutting experiment using the work to be used and a tool, and measured values of the cutting force, for example, as described in Non-Patent Documents 1 and 2. What is necessary is just to identify and optimize so that the difference | error with the cutting force prediction value estimated by the method may become small.

  As the rigidity data for calculating the tool deflection, either static stiffness or dynamic stiffness is used, but it is possible to predict the accurate tool deflection by using the dynamic stiffness. Dynamic rigidity is determined by, for example, exciting the tool tip using an impulse hammer, measuring the vibration of the tool tip using an acceleration sensor, laser Doppler vibrometer, etc. Obtain the transfer function (reciprocal of dynamic stiffness).

  As for the tool shape error, the tool diameter, that is, the diameter of the trajectory of the cutting blade is measured at several cross sections (tool axis cross sections) perpendicular to the rotation axis while actually rotating the tool, A table in which the deviation from the ideal shape is stored in advance in association with the position of the tool in the rotational axis direction (specifically, the tool contact angle q described later) is created. Alternatively, it is possible to identify the shape error by processing a resin workpiece that can be ignored by the cutting force and measuring the processing error at this time.

  Also, the workpiece (material) shape data (CAD data, block shape, etc.) and the tool path (NC data) output from the CAM are input. The tool path is composed of a number of paths (rows), and there are a linear movement path (NC data G01) and an arc movement path (NC data G02, G03). In the following, a linear movement path is assumed. To explain. An arc movement path can be handled as a set of minute linear paths. The path corresponds to the divided path described in the claims.

  In this example, a finishing process that determines the machining quality is assumed as a process for predicting a machining error. In general, since the intermediate finishing process is performed before the finishing process, it is considered that the processing state does not change much in one pass in the finishing process. Therefore, a method for predicting a machining error at one place in one pass will be described. When the change in angle between successive paths is small or when the moving distance is short, several paths may be combined into one, and conversely, when the moving distance is long, the paths may be further divided.

  Then, after predetermined data is input, it is determined whether there is a place where cutting is actually performed for each path (each line) of the tool path. This is because when a non-cutting and a cutting are mixed even in one pass, a machining error is predicted at a location during cutting. Specifically, as shown in FIG. 9 (a), the tool sweep body T0, the tool sweep body T1 up to that point, and the workpiece shape W are drawn from below with different colors. As shown by hatching in FIG. 5B, if there is a pixel area of the color of the tool sweeper T0 in the path of interest, this is determined as an area to be cut.

  If there is no area to be cut, the process returns without performing the subsequent processing for this pass, and the process moves to the determination for the next pass. On the other hand, if there is an area to be cut, a predetermined one point ( For example, the center of gravity or the vicinity of the center) is set as the predicted position of the machining error. In this way, even if a portion to be cut and a portion not to be cut are mixed in one pass, there is no worry that a machining error is erroneously predicted at the portion that is not cut.

  Further, as shown in FIG. 9B as an example, when the drawing is drawn with the horizontal width of the drawing field as the tool diameter, the vertical direction as the length of the path, and the tool feeding direction as the upward direction, Since the cutting area and the surface of the machining surface exist within the displayed width range of the area to be cut, the tool sweeper and workpiece shape that exist within that range in the subsequent processing Only color information may be drawn, and color information may be read only within the range. In this way, the processing speed can be increased.

  Then, the cutting area by the one-blade feed of the tool is detected at the machining error prediction position set as described above, and the cutting force acting on the tool when cutting this area is calculated. The equation of motion is solved from this cutting force and the dynamic stiffness of the tool, the relationship between the tool rotation angle and the tool deflection as shown in Fig. 2 (b) is obtained, and the tool deflection at the specified machining surface creation point as described below. Processing error is calculated from The machining error is also calculated from the tool shape error at the machining surface creation point.

  The machining error is calculated from the first to the last path of the tool path, and the predicted result of the machining error of the entire path is displayed on the display, and the tool path corrected in the opposite direction to this error is output. To do. By correcting the tool path in this way, it is possible to output a tool path that compensates for machining errors, thereby improving the accuracy of the cutting process.

  The above-described processing is realized by executing a required program (machining error prediction program) in the computer apparatus 1. In this sense, the computer apparatus 1 performs the data input described in the claims. Means, drawing means, cutting area detecting means, cutting force calculating means, machining surface detecting means, machining error calculating means, first sub-drawing means, predicted position setting means, and path correcting means.

(Processing error estimation)
Next, the processing error estimation method, which is a feature of the present invention, will be described more specifically with reference to the flowchart shown in FIG.

-Cutting force and tool deflection-
For the calculation of the cutting force, the methods described in Non-Patent Documents 1 and 2 are used. That is, first, in step S1 after the start in the illustrated flow, using the data input as described above, as shown in FIG. The tool sweep body T1 and the workpiece shape W before the blade feed are drawn in the visual field range V where the tool M is desired from below, with the feed direction of the tool M facing upward.

  The image thus drawn is as shown in FIG. In this example, the pixel area in which the color of the tool M is displayed as shown in black near the center is an area (cutting area) that is cut by one-blade feed of the tool M at this predicted position. The cutting area is detected from the color information of the tool M. In addition, the area shown by hatching on the right side of the image in FIG. (B) represents the workpiece shape W before finishing, and the left area represents the area cut by the tool sweeper T1, The cutting area is located in the vicinity of the boundary.

  In step S3, the cutting force acting on the tool M is calculated from the information on the cutting area detected as described above. That is, as described in Non-Patent Documents 1 and 2, the cutting edge of the tool is modeled as a set of minute cutting edges (see FIG. 8), and it is determined whether each minute cutting edge is cutting at each minute rotation angle. When it is determined that cutting is in progress, the minute cutting force is calculated from the cutting force coefficient, the cutting thickness, the cutting edge length, and the like, and this is integrated for all the minute cutting edges. By repeating this calculation, three component forces of cutting (Fx, Fy, Fz) during one rotation can be obtained.

  Subsequently, in step S4, the cutting force (Fx (t), Fy (t)) during one rotation of the tool and the compliance transfer function (Gx (ω), Gy (ω) which is the reciprocal of the dynamic stiffness of the tool. ) To calculate the tool deflection (Dx (t), Dy (t)) during one rotation as a response. In order to solve the equation of motion, for example, a numerical analysis method such as Runge-Kutta method can be used. In general, the response is the sum of free vibration and forced vibration. However, if it is assumed that the cutting force acts on the tool in the same state for several revolutions, only the forced vibration needs to be considered.

  At this time, from the product of the FFT processing result F (ω) and the compliance transfer function G (ω), the result of FFT processing of the tool deflection D (ω) = G (ω) F (ω) After that, by subjecting this to inverse FFT processing, the tool deflection during one rotation of the tool, that is, the relationship between the rotation angle position around the tool axis (tool rotation angle) and the deflection amount (see FIG. 2B) Can be calculated at high speed.

-Machining surface creation point and machining error-
Thus, while calculating the cutting force, in step S5, the tool sweep body T2 after one tool is fed from the predicted position of the machining error is added to the drawing result of FIG. 11 (see FIG. 5). ). A part of the workpiece shape W is cut off by the tool sweep body T2 thus added, and a part of the cutting area shown in FIG. 11 (b) is also deleted. The resulting image is as shown in FIG. 12, and the pixel area in which the tool color is displayed is very small.

  This pixel area represents the processed surface remaining after the processing is completed. In step S6, the processed surface is detected from the color information. Since it is not necessary to consider the area outside the workpiece shape, the color information of all the pixels is read only within the XY area of the workpiece shape, and the color information of the tool is identified therefrom. As the color information, generally four pieces of information of R, G, B, and A can be used, and information for recognizing the tool may be given to one of them (for example, A).

  Subsequently, in step S7, it is determined whether or not the field of view is to be corrected as will be described later. If necessary (YES), the field of view is corrected in step S8, and then the tool M and its sweep are performed as in steps S1 and S5. Drawing the bodies T1, T2, the workpiece shape W, etc. (redrawing) returns to the step S6. If not necessary (NO), the process proceeds to the step S9, and this is cut from the information of the detected machining surface. The machining surface creation point that is the position of the cutting edge at the time is specified, and the tool rotation angle θ, the tool contact angle q, and the machining surface normal vector n at this creation point are calculated.

  That is, in this example, as described above, the finishing of the mold is assumed, and the tool path is created with a fine pitch so that the finishing surface has a predetermined surface roughness or less, so that it remains as a processing surface. As described above, the pixel area is very small. Therefore, it can be considered that a machining error is determined from a tool deflection or a tool shape error at the moment of machining this point with a certain point in the pixel region as a representative point.

  Therefore, for example, the average value of the coordinates (X, Y coordinates) in the image area of the machining surface detected as described above is used as a creation point, and the passing locus of the cutting edge at the tool tip is represented as shown in FIG. On the hemisphere, the tool rotation angle θ from the cutting edge position with a tool rotation angle of 0 ° to the creation point C is obtained, and the angle between the line segment from the creation point C toward the center of the hemisphere and the rotation axis of the tool A tool contact angle q is obtained. A vector from the creation point C toward the center of the hemisphere is a normal vector n of the machining surface.

  The tool rotation angle θ and the tool contact angle q can also be calculated from the coordinates of the creation point C, that is, the positional relationship with the tool axis. In this embodiment, the color information of the tool pixel, for example, R and G, respectively. Information on the tool rotation angle θ and the tool contact angle q is given, and the tool rotation angle θ and the tool contact angle q can be obtained immediately and accurately simply by reading the pixel color information with a command of the computer device 1. .

  That is, although it appears as a light and shaded gradation in FIG. 13 (a), as shown on the lower side of this figure, each pixel on the hemisphere representing the tool has its color information R value determined by the tool rotation angle. It changes gradually according to the change of θ. Further, as shown in the upper side of the drawing, the value of the color information G gradually changes from the center of the hemisphere toward the lower end along the tool axis, that is, in accordance with the change of the tool contact angle q. It has become.

  Therefore, for example, the color information R and G of each pixel in the image area of the machining surface detected as described above is read and the average is obtained to immediately determine the rotation angle θ and the contact angle q of the tool at the machining surface creation point C. Can be sought. Further, the average value of the coordinates (X, Y coordinates) in the image area of the processing surface is set as the creation point C, and the rotation angle θ and the contact angle q of the tool can be obtained from the color information R, G of this point C.

  If the tool rotation angle θ at the machining surface creation point C is obtained in this way, the tool at the creation point is determined from the relationship between the tool rotation angle θ obtained in step S4 and the amount of deflection (see FIG. 2B). A bending error is obtained, and a machining error in the machining surface normal direction can be calculated from the tool deflection amount and the machining surface normal vector (step S10). Specifically, calculate the inner product of the tool deflection vector at the machining surface creation point (when considering only the translation direction, (Dx (X direction deflection), Dy (Y direction deflection), 0)) and the machining surface normal unit vector do it.

  In parallel, in step S11, a machining error resulting from the tool shape error at the tool contact angle q is calculated. This may be calculated using a table in which the shape error of the tool corresponding to the tool contact angle q at the machining surface creation point obtained as described above is measured and stored in advance.

  In this example, as described above, the tool deflection and the tool shape error are calculated from the information of one creation point on the machining surface. However, the tool rotation angle θ and the tool contact when the machining surface is actually cut are calculated. Since the angle q has a certain range, the machining error may be calculated for each angle, or the machining error may be calculated from the maximum value or the average value within this range. That is, the machining surface creation position specified for calculating the tool deflection or the like is not limited to one point, and may be a range having a certain extent.

-Correction of visual field-
Next, the visual field correction in step S8 will be specifically described. As an example, this is a method for dealing with a case where the color of the tool M is not displayed on the image when the tool sweep body T2 is added in the step S5. The color of the tool M is not displayed in the case where there is finally no remaining portion as the machining surface and the case where the portion is less than one pixel at the time of drawing. That is, when viewed from the bottom as in this example, when machining the side wall of the workpiece, the machining surface is viewed obliquely with respect to the normal line, so the portion remaining as the machining surface is Even if it exists, the pixel area becomes very small, and it may be less than one pixel.

  As a countermeasure method in such a case, the detection result (step S2) of the cutting area prior to the prediction of the machining error is used. That is, as described above, the portion that finally remains as the machining surface is included in the cutting region, and this cutting region is wider than the region of the machining surface, so it is not necessary to consider that this is not displayed. Therefore, as schematically shown in FIG. 14, when the machining surface is not displayed even though there is a cutting area, the viewpoint is moved so as to approach the normal direction of the cutting area, and viewed from the side as shown in the figure. By drawing an image, the pixel area of the processed surface is enlarged.

  Alternatively, the processing surface can be displayed by narrowing the field of view toward the cutting area and reducing the area per pixel to increase the resolution, in other words, by enlarging and drawing the cutting area even if the viewpoint remains unchanged. To do. By correcting the viewpoint or field-of-view range by any of these methods, it is possible to prevent the processing surface from being overlooked.

  Note that if the tool contact angle q is increased, the number of pixels is reduced even if the processed surface is displayed, so that the calculation error may be increased. Therefore, when there is a cutting region near the outer periphery of the tool (passage trajectory of the cutting edge) in the image from the negative Z-axis direction and the tool contact angle q is considered to be large (for example, exceeding 60 °), the machining surface Regardless of whether it is visible or not, it is possible to perform calculation with higher accuracy by changing the viewpoint or displaying the enlarged image in the same manner as described above.

  Steps S1 and S5 in the flow of FIG. 10 described above include a tool M at the machining error predicted position on the tool path (passage trajectory of the cutting blade), and a predetermined value including before and after feeding one blade at that position. This corresponds to a drawing step of drawing the tool sweep bodies T1 and T2 in the range and the workpiece shape W in the visual field range V from below. In this example, in step S1, the tool M at the predicted position of the machining error, the tool sweep T1 up to that point, and the workpiece shape W are drawn, and in the subsequent step S5, the tool sweep T2 after the predicted position. Is added.

  Step S2 corresponds to a cutting region detection step of detecting a pixel region where the tool M is displayed on the image plane before the tool sweep body T2 after the predicted position is added as a cutting region. Corresponds to a cutting force calculation step of calculating a cutting force acting on the tool from the information on the cutting region.

  Step S6 corresponds to a machining surface detection step of detecting a pixel area in which the tool M is displayed on the image plane after the tool sweep body T2 has been drawn in step S5 as a machining surface remaining after machining. Steps S9 to S11 correspond to a machining error calculation step in which the creation point of the machining surface is specified and the machining error is estimated and calculated from the tool shape error and the deflection amount.

  Further, step S8 is a step of correcting the viewpoint or visual field range when redrawing the image based on the information of the cutting area detected in step S2, and the computer apparatus 1 executes the step by executing this step. The visual field correcting means described in the claims is also configured.

(Tool path correction)
Finally, the tool path correction will be briefly described. In NC data, the tool path is usually specified by a straight line or arc connecting many constituent points in the X, Y, Z orthogonal coordinate system, and the constituent points are set in the opposite direction to the predicted machining error as described above. If moved, the tool path can be appropriately corrected. That is, in this embodiment, the machining error is calculated for each of a large number of paths constituting the tool path, and as shown in FIG. 15, the component point to be corrected is the intersection of the previous and subsequent paths. The constituent points may be corrected using the prediction results of the machining errors in one pass.

  Specifically, first, in each of the two front and rear passes, machining error vectors E (i) and E (j) having machining error magnitudes (positive and negative) in the normal direction of the machining surface are considered. The correction reference vectors E ′ (i) and E ′ (j) having the same magnitude as the error vectors E (i) and E (j) and opposite directions are calculated. Then, the direction of the correction vector E ′ (i + j) is set in the direction of the sum. The magnitude of the correction vector E ′ (i + j) is determined so that the difference between the movement of the two paths before and after the correction and the respective correction reference vectors E ′ (i) and E ′ (j) is as small as possible. do it. In the case of an arc movement path, it can be corrected by performing these calculations as a set of linear movement tool paths and obtaining a correction arc using the least square method or the like.

(Function and effect)
As described above, according to the machining error predicting apparatus according to this embodiment, the workpiece shape W and the tool M for cutting the work shape W using the three-dimensional graphics function of the computer apparatus 1, and the tool M sweep bodies T1 and T2 are respectively drawn in the visual field range V from below, and the pixel area in which the tool M is displayed in front of the drawn image is used as the area of the machining surface remaining after machining. Can be detected.

  Then, the creation point C, which is the cutting edge position of the tool at the time of creation, is specified in the region of the machining surface, and the estimated value of the machining error is accurately calculated based on the shape error and deflection amount of the tool at the creation point C. can do. Therefore, it is possible to accurately predict an error even when processing a die having a free-form surface, which is very advantageous for studying processing conditions.

  Further, if the tool path is corrected based on the predicted processing error, the processing error can be reduced and the man-hour and time required for correction after forming can be greatly reduced.

  In particular, in this embodiment, the tool M (cutting blade trajectory) to be drawn as described above is a hemispherical surface, and each pixel on the hemispherical surface gradually has a value depending on the tool rotation angle θ and the tool contact angle q. Since the color information R and G is changed so as to change, the tool rotation angle θ and the tool contact angle q can be directly calculated by reading the color information of the creation point C specified as described above. it can.

  Further, as described above, the tool M at the position where the machining error is predicted and the tool sweep body T1 up to the predicted position are drawn first, and from the image before the tool sweep body T2 after the predicted position is drawn, In the same manner as described above, the cutting region at the predicted position is detected, and the cutting force is calculated based on the detected region. The cutting force can be calculated at high speed.

  Moreover, since the machining surface is detected as described above from the image in which the tool sweep body T2 is added after the cutting area is obtained in this way, the drawing result for calculating the cutting force is used as the machining error. It is possible to shorten the processing time by using this for prediction.

-Example-
Below, the example which actually estimated the error of cutting using the prediction device of this embodiment is explained. What was carried out was contour line machining of a cylindrical surface as shown in FIG. 16, and the tool, work material, and cutting conditions are shown on the right side of the figure. In the machining of a mathematically defined shape such as a cylindrical surface, the tool rotation angle and the tool contact angle can be theoretically obtained, so that the calculation result can be verified. In the processing of the cylindrical surface, the intersection of the straight line connecting the center coordinate value of the ball end mill and the central axis of the cylindrical surface and the cylindrical surface shape becomes the processing surface creation point.

  FIG. 17 shows the difference between the calculated values of the tool rotation angle θ and the tool contact angle q from the theoretical values by drawing from the negative Z-axis direction. When the tool contact angle is 87.6 ° or more, the tool is not displayed and cannot be calculated. Also, the error increases from around 60 °. On the other hand, FIG. 18 shows a result calculated by changing the viewpoint when the tool contact angle is estimated to be 60 ° or more in the drawing from the negative Z-axis direction. It was confirmed that all values could be obtained by changing the viewpoint, and that the calculation error was reduced.

  FIG. 19 shows the result of predicting the machining error using the prediction method of this embodiment and the result of measuring the shape after machining with a three-dimensional measuring machine. The predicted value and the measured value are in good agreement. FIG. 20 shows measurement results of machining errors when the tool path is corrected based on the error prediction result and when the tool path is not corrected. It was confirmed that the machining error can be reduced by correcting the tool path based on the machining error prediction.

  Next, the prediction method of this embodiment was applied to the machining error prediction of the finishing process of the real mold model including the intermediate finishing process 99768 pass and the finishing process 286463 pass. First, a cutting area is determined for each pass of the finishing process, and an error prediction position is set in the cutting area. Then, the cutting force was calculated at that position, and thereafter, the portion remaining as the machining surface was calculated, and the machining error due to the tool deflection was predicted. Note that the drawing resolution was 512 × 512. The specifications of the personal computer used are: Core2Duo 3 GHz made by Intel, memory 2 GB, and GPU GeForce8800GTS made by nVIDIA. FIG. 21 shows calculation time and processing error prediction results. The machining error could be predicted in 0.070 seconds per pass for a large-scale mold model with a free-form surface.

  Further, in order to compare the prediction method of this embodiment with the conventional method (software processing), a machining error due to a tool shape error was predicted. The reason why the machining error caused by the tool shape error is used is to compare the calculation time of the portion remaining as the machining surface, excluding the calculation of cutting force and tool deflection. The predicted position is fixed at the midpoint position of the straight tool path, the tool contact angle at the machining surface creation point is calculated, and the machining error is calculated based on the tool contact angle. Only the calculation of the tool contact angle at the machining surface creation point was either software processing or the graphics hardware processing according to the embodiment.

  There are two types of error prediction targets, the finishing process (FIG. 22 (a)) of the above-described cylindrical surface machining tool diameter of 20 mm, and the process of the tool diameter of 30mm (NC data 5 in FIG. 21) of the above-mentioned actual mold model. ((B) in the figure). In the case of software processing, lattice points obtained by dividing the tool diameter by 512 are set in a circular area obtained by projecting a tool (sphere) at a position where a machining error is predicted on the XY plane. Then, the intersection of the straight line from each lattice point in the positive direction of the Z axis and the tool (sphere) is calculated, and the distance between the intersection and the lattice point is calculated. Similarly, the distance to all tool sweep bodies (cylinder, sphere) is calculated, and when the distance to all tool sweep bodies is larger than the distance to the tool, that is, when the tool is at the lowest position, It is determined that the position remains as the machining surface. If even one tool sweeper with a smaller distance than the tool is found, it is determined that the position does not remain as the machining surface.

  In order to shorten the calculation time as much as possible, the possibility of interference between the half-line and the tool sweeper was first determined, and the distance was calculated only for the potential tool sweeper. The specifications of the PC used are the same as in the previous section. The calculation was performed for all 809 passes for the cylindrical surface machining, and for the actual mold model, 128 points for every 1000 passes out of the tool route 125824 passes. The calculation time is as shown in FIG. 6C, and the speed is increased by 8.9 times in the cylindrical machining and 110.8 times in the machining of the real mold model. Therefore, it can be said that the machining error prediction method of the present invention is more effective in actual die machining having a large-scale tool path.

(Other embodiments)
The specific method for predicting the machining error according to the present invention is not limited to that of the above-described embodiment. For example, the prediction of the machining error is not limited to the finishing process of the mold, but also to the roughing process and the intermediate finishing process. Applicable. In that case, the machining state in one linear movement path may be determined, and the number of predicted locations may be increased depending on the machining state. Of course, the present invention can be applied to various cutting processes other than the mold.

  Further, the information to be given to the pixel of the tool is not limited to the tool rotation angle θ and the tool contact angle q, but may be other types of coordinate information. As an example, the color information R may have a tool contact angle q, while G may be an X coordinate and B may be a Y coordinate. Further, other pixel information can be used instead of color information.

  In the above-described embodiment, as described above with reference to FIG. 5, in addition to the tool M at the position where the machining error is predicted and the workpiece shape W, the tool up to the one-blade feeder at the predicted position. The sweeping body T1 and the tool sweeping body T2 after feeding one blade are drawn, and this is basic, but the tool sweeping bodies T1 and T2 are, for example, those up to 2 to 3 blades before feeding. It is good also as after-3 blade feeding.

  Furthermore, in the above-described embodiment, the description has been made mainly assuming a ball end mill as a tool used for processing. In this case, the trajectory of the cutting edge is approximately represented by a hemispherical surface. Rotating tools other than end mills (eg radius end mills, square end mills, etc.) are also envisaged. In this case, the trajectory of the cutting blade is approximated to various shapes depending on the shape, and all of them have an axially symmetric shape centered on the rotation axis of the tool. Of course, the tool sweeper is also configured according to the shape of the tool.

  Furthermore, in the above-described embodiment, the machining error is predicted for each of a large number of paths in the tool path. In general, since a large amount of tool path data is used in machining a die, all of them are used. However, when machining errors are predicted, it may take too much time. Therefore, it is desirable to exclude in advance parts where machining errors are small and the tool path does not need to be corrected.

  That is, the machining error due to tool deflection generally increases as the tool contact angle increases, and when the tool contact angle is close to 0 ° (ie, the angle between the tool axis and the normal of the work surface is close to 0 °). Sometimes), even if the tool is bent by the cutting force, it is considered that the machining error is hardly affected.

  Therefore, as schematically shown in FIG. 23, the color information of all the tool sweep bodies is changed according to the angle from the tool axis and the line number of the tool path prior to the prediction of the machining error, together with the workpiece shape from below. And drawing with the hidden surface removed from the viewpoint (corresponding to the second sub-drawing means). The image thus drawn has a shape after the movement of all the tools is completed, and in this state, by reading the color information of the pixels in the XY area of the workpiece shape, it finally becomes a machining surface. The line number of the tool sweeper and the angle from the tool axis at that time can be detected.

  Therefore, the maximum value of the tool contact angle, which is the maximum value from the tool axis, is calculated for each tool sweep number, and the maximum value is a predetermined value (for example, 30 to 60 °) or less. Is excluded from the processing error predicted portion (that is, the processing error prediction for this path is prohibited), thereby speeding up the processing.

  Note that the range of the visual field at the time of drawing may be determined according to the length expected to be left as a machining surface from the tool diameter and machining conditions. Moreover, you may calculate by enlarging about the part estimated that a tool contact angle is large.

Explanatory drawing of the tool path | route output from CAM and a processing error. Explanatory drawing of the cutting force during 1 rotation of a tool, and tool bending. Explanatory drawing which looked at the location which remains as a process surface from the top. Explanatory drawing which looked at the location which remains as a processing surface from the side. Explanatory drawing of the drawing method for detecting the area | region which remains as a process surface. Explanatory drawing of a tool sweep body. The schematic diagram which shows the whole flow of the prediction of the processing error which concerns on this invention. Explanatory drawing of modeling of a cutting-blade shape. Explanatory drawing of the method of determining whether there exists an area | region cut for every pass. The flowchart figure of the specific procedure of prediction of a process error. Explanatory drawing of the drawing method for detecting a cutting area | region. Explanatory drawing of the final process surface displayed on the drawn image. Explanatory drawing of the tool contact angle and tool rotation angle in a process surface creation point. Explanatory drawing of the change method of a viewpoint. Explanatory drawing of the correction method of a tool path | route. Explanatory drawing of the experimental method of an Example. The comparison figure of the calculated tool contact angle and the theoretical value of a tool rotation angle. FIG. 17 is a diagram corresponding to FIG. 17 when the viewpoint is changed. The comparison figure of the predicted value of a processing error, and a measured value. The comparison figure of the measurement error measurement value by the presence or absence of correction of a tool path. Explanatory drawing of the prediction time of a process error in a real die model, and a prediction result. The table | surface of the result of having compared the prediction method of the Example with software processing. Explanatory drawing of the selection method of the tool path | route which estimates a process error.

M Tool image (trajectory of cutting edge)
P Tool path T0 to T2 Tool sweep body W Workpiece shape 1 Computer device (data input means, drawing means, cutting area detecting means, cutting force calculating means, machining surface detecting means, machining error calculating means, first and second Sub-drawing means, predicted position setting means, path correction means, visual field correction means)
S1, S5 Drawing step S2 Cutting area detection step S3 Cutting force calculation step S6 Machining surface detection step S4, 9 to S11 Machining error calculation step S8 Field of view correction step

Claims (10)

An apparatus for predicting a machining error of a workpiece due to a rotating tool,
Data input means for receiving input of predetermined data, including information on the shape of the work piece before machining by the tool, the shape of the cutting edge of the tool, and the movement path by the feed of the tool;
Based on the predetermined data, the trajectory of the cutting edge at a position where a machining error is predicted on the movement path of the tool, a tool sweep body in a predetermined range including immediately before and after the position, and the work piece, and a drawing means, respectively to draw the visual field range that can desire overall passing track of the cutting edge from the side of the workpiece material, for the tool sweeping body to draw in all the visual field range,
A processing surface detection means for detecting a pixel area in which the trajectory of the cutting edge is displayed in the foreground in the image drawn by the drawing means as a processing surface area remaining after the processing;
A creation position that is a cutting edge position of a tool that cuts a workpiece when creating a machining surface is specified from the pixel area detected by the machining surface detection means, and at least a shape error and a deflection amount of the tool at the creation position are specified. A machining error prediction device comprising machining error calculation means for calculating an estimated value of a machining error based on one. In the processing error prediction device according to claim 1,
The shape error of the tool is obtained in advance as a deviation of the blade edge position with respect to the rotation axis, and a table stored in association with the position in the rotation axis direction is created,
The machining error predicting device, wherein the machining error calculating means calculates a tool shape error at the specified generating position using the table. In the processing error prediction device according to claim 1 or 2,
The machining error calculation means obtains the relationship between the rotation angle position around the rotation axis of the tool and the amount of deflection from the magnitude of the cutting force at the machining error predicted position and the rigidity of the tool, and at the specified creation position. A machining error predicting apparatus characterized by calculating a tool deflection amount. In any one processing error prediction apparatus of Claims 1-3,
The trajectory of the cutting edge is axisymmetric about the rotation axis of the tool, and each pixel on the trajectory has a rotation angle position around the tool rotation axis and a position in the rotation axis direction. In order to change the value according to at least one of the predetermined pixel information,
The machining error calculation means specifies a creation position from pixel information of the machining surface area detected by the machining surface detection means, and calculates at least one of a rotation angle position and a rotation axis direction position of the creation position. Characteristic processing error prediction device. In any one processing error prediction apparatus of Claims 1-4,
The drawing means first draws the trajectory of the cutting edge at the predicted position of the machining error, the tool sweeping body up to immediately before the position, and the work piece, and then immediately after the predicted position. Configured to draw a tool sweep,
Before the tool sweeper after the predicted position is drawn, a pixel area where the trajectory of the cutting edge is displayed in the image drawn by the drawing means is detected as a cutting area at the predicted position. Cutting area detecting means;
Cutting force calculating means for calculating a cutting force acting on the tool from the information of the cutting area;
A processing error prediction apparatus further comprising: In the processing error prediction device according to claim 5,
A machining error prediction apparatus, further comprising: a visual field correction unit that corrects a viewpoint or a visual field range in the drawing unit based on information on a cutting region detected by the cutting region detection unit. In any one processing error prediction apparatus of Claims 1-6,
The movement path of the tool is divided, and the tool sweep body in any one of the division paths, the tool sweep body up to the division path, and the workpiece are respectively separated from the workpiece side. A first sub-drawing means for drawing a tool,
In the image drawn by the first sub-drawing means, if there is a pixel area where the tool sweep body of the divided path is displayed in front, the predicted position of the processing error is set in this pixel area, while A predicted position setting unit that does not set a predicted position of a processing error in the divided path if there is no displayed pixel region;
A processing error prediction apparatus further comprising: In any 1 processing error prediction device of Claims 1-7,
A second sub-drawing unit that divides the movement path of the tool, draws the tool sweeper in each of the divided paths, and the work piece while looking for the tool from the work piece side;
With respect to the divided path in which the tool sweeper is displayed on the front side in the image drawn by the second sub-drawing means, the angle formed by the rotation axis of each tool and the work surface is estimated, and this angle is a predetermined value. Prediction prohibiting means for prohibiting prediction of machining errors related to the following divided paths;
A processing error prediction apparatus further comprising:   A path correcting means for correcting the moving path of the tool so as to reduce the machining error based on the value of the machining error predicted by the machining error prediction apparatus according to claim 1. A tool path correcting device. A program for predicting a machining error of a work piece by a rotating tool using a three-dimensional graphics function of a computer device,
A data input step for receiving input of predetermined data, including information on a shape of a workpiece before machining by the tool, a cutting edge shape of the tool, and a movement path by feeding the tool;
Based on the predetermined data, the trajectory of the cutting edge at a position where a machining error is predicted on the movement path of the tool, a tool sweep body in a predetermined range including immediately before and after the position, and the work piece, A drawing step of drawing in the entire visual field range where the entire trajectory of the cutting blade can be desired from the side of the workpiece, and drawing the tool sweeper in the entire visual field range ;
A processing surface detection step of detecting a pixel region in which the trajectory of the cutting blade is displayed in the foreground in the image drawn in the drawing step as a region of the processing surface remaining after the processing;
From the pixel area detected in the machining surface detection step, a creation position that is a cutting edge position of a tool that cuts a workpiece when creating the machining surface is specified, and the shape error and the amount of deflection of the tool at the creation position are determined. A computer program for predicting a machining error, comprising: a machining error calculating step for calculating an estimated value of a machining error based on at least one.