Tool defect detection method

Abstract

A tool defect determination method for determining a defect in a tool during machining of an NC machine tool 1, the tool defect determination method comprising: setting a threshold on the basis of the difference between an integrated value of a load generated in the tool in an immediately preceding predetermined section and an average value of the load; and determining presence / absence of abnormality in the tool on the basis of the difference between an integrated value of an N-th load and the average value of the load by comparison with the threshold. By improving calculation accuracy of the load generated in the tool during machining, determination accuracy of presence / absence of the tool defect is improved.

Description

Tool Defect Detection Method

[0001] The present invention relates to a tool defect detection method used for detecting defects of tools in a numerically controlled machine tool.

[0002] Conventionally, as a method for detecting defects of tools in an NC machine tool, there is one described in Patent Document 1. In Patent Document 1, as a method for determining an abnormality of a cutting tool, a waveform of a cutting load (load integrated value) is used as a load state value, and an average movement of the load state values in a plurality of machining cycles up to immediately before is determined and obtained as a determination threshold value, and the load state value and the determination threshold value are compared to determine an abnormality of the cutting tool.

[0003] Japanese Patent Application Laid-Open No. 2004-130407

[0004] However, in the above-described conventional tool defect detection method, when an abnormality such as breakage occurs in the tool near the end of the cutting section, the load integrated value (area of the cutting waveform) becomes almost the same as that in the normal state, so there is a risk of false detection that it is normal despite an abnormality in the cutting tool.

[0005] The present invention has been made in view of the above-described conventional situation, and an object thereof is to provide a tool defect detection method capable of improving the determination accuracy of the presence or absence of tool defects by enhancing the calculation accuracy of the load generated in the tool during machining. [[ID=十六]] [[ID=十七]]

[0006] The tool defect detection method according to the present invention is a method for determining a defect of a tool during machining of an NC machine tool. This tool defect detection method sets respective threshold values based on the integrated value of the load generated in the tool in a predetermined section immediately before and the difference from the average value of the load, and determines the presence or absence of an abnormality of the tool based on the difference between the integrated value of the load at the Nth time and the average value of the load by comparing with these threshold values.

[0007] Since the tool defect detection method according to the present invention adopts the above-described configuration, it is possible to enhance the calculation accuracy of the load generated in the tool during machining and improve the determination accuracy of tool defects. For example, even when an abnormality such as breakage occurs in the tool near the end of the machining section, the abnormality can be surely detected.

[0008] This is a block diagram showing NC machine tools and control devices applicable to tool defect detection methods. This is a flowchart showing the control process in the control device shown in Figure 1. This is a graph showing the relationship between cutting load and spindle speed. This is a graph illustrating comparative examples and examples of threshold-based determination.

[0009] <First Embodiment> Figure 1 is a block diagram showing an NC machine tool 1 and a control device 2. The NC machine tool 1 is a well-known type that performs machining such as cutting on a workpiece by moving a tool mounted on a rotating spindle and a workpiece held by a jig such as a table relatively in three orthogonal axial directions.

[0010] The illustrated NC machine tool 1 includes a numerical control unit 1A, a sequence control unit 1B, a power amplification unit 1C for servo motors M1 and M2, a power amplification unit 1D for spindle motor M3, a communication control unit 1E for sending and receiving data with the control device 2, and various sensors, etc. The control device 2 also includes a communication control unit 2A for sending and receiving data with the communication control unit 1E of the NC machine tool 1, a data acquisition unit 2B, a main control unit (calculation unit) 2C, and a determination notification unit 2D, etc.

[0011] The tool defect detection method according to the present invention is a program input to a control device 2, which determines tool defects during machining by an NC machine tool 1, and sets threshold values ​​based on the cumulative value of the load applied to the tool in the immediately preceding predetermined section and the difference (deviation) from the average value of the load.The tool defect detection method then determines whether or not there is a tool abnormality based on the cumulative value of the load over the Nth time and the difference (deviation) from the average value of the load, by comparing it with the set threshold values.

[0012] The above tool defect detection method, when performing cutting operations, inputs the rotational speed and torque of servo motors M1 and M2 and spindle motor M3 to control device 2 during the machining process of NC machine tool 1. The control device 2 then calculates the moving average (integral value of the load, standard deviation of the load) of the cutting load generated on the tool in the preceding predetermined section, calculates a threshold for judgment based on the moving average, and determines whether or not there is a tool defect based on that threshold and the cutting load calculated in the current operation.

[0013] The specific steps of the tool defect detection method described above will be explained below using the flowchart shown in Figure 2. Figure 2 illustrates the case where cutting is performed on NC machine tool 1.

[0014] In other words, in the tool defect detection method, when control is started, the previous value (moving average) is read in step S1 in Figure 2, and then the start signal is read in step S2. At this time, the tool defect detection method uses macro variables used in the program of the NC machine tool 1, with 1 indicating start and 0 indicating end, thereby starting and ending the reading of the data to be judged for the judgment process in step S3.

[0015] The above tool defect detection method determines whether or not to start monitoring in step S3. If it is determined not to start monitoring (No), the process returns to step S2. If it is determined to start monitoring (Yes), the process proceeds to step S4.

[0016] Next, in step S4, the load on the spindle is read, in step S5 the spindle speed is read, in step S6 the spindle load is stored, and in step S7 the spindle speed is stored. At this time, the control device 2 stores the actual values ​​of the load and speed, and also counts up the number of reads.

[0017] Then, in step S8, the tool defect detection method reads the termination signal. At this time, similar to step S2, the tool defect detection method uses macro variables used in the program of the NC machine tool 1, with 1 indicating the start and 0 indicating the end, thereby starting and ending the reading of the data to be judged for the judgment process in step S9.

[0018] In the tool defect detection method described above, as shown in Figure 3, the predetermined section immediately preceding one machining cycle is defined as the section in which the rotational speed of the spindle during machining in the NC machine tool 1 is maintained, and the changes in the loads occurring in the rotational direction and axial direction of the spindle are monitored and set accordingly.

[0019] Figure 3 is a graph showing the relationship between the cutting load generated on the spindle and the spindle rotation speed. During machining, the rotation speed is kept constant (for example, 900 rpm) from the start point S to the end point E. In the tool defect detection method, the period from the start point S to the end point E is defined as a predetermined interval, and the macro variable is set to 1 (start) in step S2 and to 0 (end) in step S8.

[0020] The above tool defect detection method determines in step S9 whether monitoring has ended or not. If it is determined that monitoring has not ended (No), the process returns to step S4. If it is determined that monitoring has ended (Yes), the process returns to step S10 to determine whether the spindle speed is at a predetermined value. In this case, the predetermined value is set to, for example, ±20 rpm if the spindle speed (cutting speed) is 1000 rpm.

[0021] Then, in step S10, if it is determined that the rotational speed of the spindle is not a predetermined value (No), step S10 is repeated. If it is determined that the rotational speed is a predetermined value (Yes), the cumulative value is calculated in step S11, and then the moving average (cumulative) is calculated in step S12.

[0022] Next, in step S13, the tool defect detection method calculates a threshold for judgment based on the moving average of the tool. After calculating the standard deviation in step S14, the moving average is calculated in step S15, the threshold for judgment is calculated from the moving average in step S16, and a comparison is made with the previous value (moving average) in step S17.

[0023] In other words, the above tool defect detection method calculates two thresholds in steps S13 to S17: a threshold for the cumulative value of the load applied to the tool and a threshold for the difference (deviation) from the average value of the load. It then compares the threshold for the cumulative value of the load in the predetermined interval with the cumulative value of the load in the current (Nth) test, and also compares the threshold for the difference from the average value of the load in the predetermined interval with the difference from the average value of the load in the current (Nth) test.

[0024] In a more preferred embodiment, the above tool defect detection method may detect sudden changes and unstable states of the load by calculating the standard deviation of the instantaneous value of the load applied to the tool. Furthermore, in addition to setting the threshold, the method may also record the difference between the cumulative load value detected in the previous machining cycle and the average load value, and detect sudden changes by comparing the absolute value of the difference between the cumulative load value detected in the current machining cycle and the average load value with a preset reference value.

[0025] The above tool defect detection method determines in step S18 whether both the cumulative value and the standard deviation are greater than a threshold, in step S19 whether either the cumulative value or the standard deviation is greater than a threshold, and in step S20 whether both the cumulative value and the standard deviation are less than a threshold.

[0026] In step S18, if both the cumulative value and the standard deviation are greater than the threshold (Yes), the judgment result in step S21 is determined to be abnormal; otherwise, the process proceeds to step S19.

[0027] In step S19, if either the cumulative value or the standard deviation is greater than the threshold (Yes), an alarm is output in step S22 indicating a problem with the judgment result; otherwise, the process proceeds to step S23.

[0028] Then, in step S20, if both the cumulative value and the standard deviation are smaller than the threshold (Yes), the judgment result in step S23 is determined to be normal. In step S20, there is no negative judgment (No) because steps S18 and S19 precede it.

[0029] Figure 4 is a graph showing the determination based on a threshold. On the left side of the figure, the cutting load applied to the tool is shown as raw data waveform (instantaneous value), and the cumulative value of the load in the predetermined section explained in Figure 3 is shown, with an example where the cumulative threshold is 2800. When tool defects are detected using only this cumulative threshold, the results are as shown in Comparative Examples 1 to 4. In Comparative Example 1, there was no abrupt change in the cutting load during the machining cycle and the cutting of the predetermined section was completed, so there was no abrupt change in the cumulative value, and the cumulative value exceeds the threshold (2800), so it is determined to be normal. In Comparative Examples 2 and 3, there was abrupt change in the cutting load during the machining cycle and the cutting of the predetermined section could not be completed, and the cumulative value is below the threshold (2800), so there is a possibility that the tool broke, and it is determined to be abnormal.

[0030] Furthermore, in Comparative Example 4, although there is a sudden change in cutting load near the end of the machining cycle, the cutting of the predetermined section is almost completed, so the cumulative value exceeds the threshold (2800). In this case, as in the conventional example, there is a risk of misjudging the tool as normal even though an abnormality has occurred.

[0031] In contrast, the tool defect detection method according to the present invention, as shown in the right figure of Figure 4, uses the same cumulative threshold and the difference (deviation / standard deviation) threshold from the average value of the load as shown in the left figure, with the case where the standard deviation threshold is 1.2 being illustrated. In this case, in Embodiment 4 shown in the right figure of Figure 4, although the cumulative value is normal relative to the threshold, both the standard deviation and the cumulative value exceed their respective thresholds (1.2) (Yes in step S19), so it is possible to grasp the sudden change in cutting load that could not be determined from the cumulative value alone, and it is clearly determined that there is an abnormality in the tool. The judgment conditions and judgment results in S18 to S23 can be increased in number by setting the thresholds in stages.

[0032] Next, in the tool defect detection method, after the comparison determination described above, the number of repetitions is counted up in step S24, and in step S25, it is determined whether the number of repetitions has reached the number of reads. If it has not reached the number (No), the process returns to step S10. If it has reached the number (Yes), the current value (moving average) is written in step S26, and then the determination result is written in step S27 before the control process ends.

[0033] Thus, the above tool defect detection method can improve the accuracy of tool defect detection by increasing the accuracy of calculating the load generated on the tool during machining. For example, even if an abnormality occurs in the tool near the end of the machining section, that abnormality can be reliably detected.

[0034] Incidentally, when detecting tool defects in NC machine tools, it is extremely important to determine the cutting section of a machining cycle in order to calculate the cumulative value and standard deviation. If one tries to determine the start and end timing of the cutting section by considering the magnitude and rise and fall of the cutting load on the spindle and the thrust load of the motor that controls the axis feed, the cutting load becomes unstable when the tool bites into the workpiece and when it is withdrawn from the workpiece, causing wild fluctuations in the load. As a result, it becomes difficult to correctly determine the cutting section, and the area of ​​the cutting waveform (cumulative load) cannot be calculated correctly, resulting in an inaccurate judgment and false detection.

[0035] Furthermore, when detecting defects in NC machine tools, a method is known to monitor sudden anomalies during cutting by calculating the derivative value from the instantaneous value of the cutting load. However, there are various cases in which the derivative value rises or falls sharply, such as chip jamming, the formation and detachment of built-up edges, and the tool biting into the cutting edge when the core is misaligned. While some cases may lead to tool abnormalities such as breakage, in many cases the cutting process ends as normal. For this reason, the derivative value of the cutting load cannot correctly determine tool abnormalities and results in false detections.

[0036] Furthermore, when detecting defects in NC machine tools, collecting instantaneous cutting load data used for judgment requires high-frequency sampling of less than a few milliseconds. Even if a single machining operation lasts only a few seconds, this results in a massive amount of data—thousands of entries. Calculating and judging irrelevant data takes a long time, causing processing delays. Therefore, it becomes necessary to remove extraneous data before calculating and judging the cutting load. Additionally, various conditions are required to judge this extraneous data, increasing the burden on data processing and hindering accurate judgment.

[0037] In contrast to these, the above-described tool defect detection method sets thresholds based on the cumulative load applied to the tool in the immediately preceding predetermined section and the difference between that and the average load. By comparing these thresholds, the presence or absence of a tool abnormality is determined based on the cumulative load of the Nth load and the difference between that and the average load. This improves the accuracy of calculating the load applied to the tool during machining, allowing for accurate determination of tool abnormalities and eliminating the aforementioned factors that hinder false detection and correct judgment.

[0038] Furthermore, the above-described tool defect detection method is set by monitoring the changes in loads occurring in the rotational direction and axial direction of the spindle, while ensuring that the preceding predetermined section is a section in which the rotational speed of the spindle during machining is maintained. As a result, the above-described tool defect detection method can improve the accuracy of the predetermined section and achieve further improvement in the accuracy of calculating the cutting load (integrated load value, standard deviation of load) for one machining cycle.

[0039] Furthermore, the above tool defect detection method detects rapid changes and unstable states of load by calculating the standard deviation of the instantaneous load on the tool. This eliminates situations where large fluctuations in the differential value of the cutting load may be misinterpreted as tool abnormalities. Moreover, by incorporating the differential value into the judgment criteria depending on the manufacturing method and processing conditions, further improvements in the accuracy of judgment regarding rapid changes in cutting load can be achieved.

[0040] Furthermore, the above tool defect detection method uses macro variables in the NC machine tool program, with 1 indicating the start and 0 indicating the end, thereby starting and ending the reading of data to be judged for the judgment process. As a result, the above tool defect detection method can execute this notification trigger before and after the actual machining in the NC program, suppressing the collection of unnecessary data, reducing judgment processing on unnecessary data, and thus reducing the judgment processing time.

[0041] Furthermore, the above tool defect detection method, in addition to setting a threshold, records the cumulative load value detected in the previous machining cycle and the difference between that value and the average load value. It then compares the absolute value of the cumulative load value detected in the current machining cycle and the difference between that value and the average load value with a preset reference value to detect sudden changes. As a result, the accuracy of tool abnormality detection is further improved by this tool defect detection method, eliminating false detections and over-detections.

[0042] The tool defect detection method according to the present invention is not limited to the above-described embodiment in terms of its specific configuration or applicable NC machine tools, and can be modified as appropriate without departing from the spirit of the present invention.

[0043] 1. NC machine tool 2. Control device

Claims

1. A method for determining tool defects during machining using an NC machine tool, characterized by setting threshold values ​​based on the difference between the cumulative load applied to the tool in the immediately preceding predetermined section and the average load, and determining whether or not there is an abnormality in the tool based on the difference between the cumulative load of the Nth load and the average load by comparing with each of the threshold values.

2. The tool defect detection method according to claim 1, characterized in that the predetermined section is a section in which the rotational speed of the spindle during machining in the NC machine tool is maintained, and is set by monitoring changes in the respective loads occurring in the rotational direction and axial direction of the spindle.

3. The tool defect detection method according to claim 1, characterized in that it detects sudden changes in load and unstable conditions by calculating the standard deviation of the instantaneous value of the load generated on the tool.

4. The tool defect detection method according to claim 1, characterized in that a macro variable used in the program of the NC machine tool is used to start and end the reading of data to be judged for the judgment process, with 1 indicating the start and 0 indicating the end.

5. The tool defect detection method according to claim 1, characterized in that, in addition to setting a threshold based on the difference between the cumulative load value and the average load value in the immediately preceding predetermined section, the method also records the difference between the cumulative load value and the average load value detected in the immediately preceding machining cycle, and detects abrupt changes by comparing the absolute value of the difference between the cumulative load value and the average load value detected in the current machining cycle with a predetermined reference value.

Images