Grinding wheel state determination device and grinding wheel state determination method

Abstract

To provide a technique capable of properly determining a grinding wheel state without being limited to clogging.SOLUTION: A grinding wheel state determination device that determines a state of a grinding wheel based on sparks generated when a workpiece is processed by the grinding wheel includes an acquisition unit that acquires a plurality of spark images in which the sparks are continuously imaged in time series, a light quantity detection unit that detects a light quantity of the sparks included in each of the plurality of acquired spark images, a calculation unit that calculates a function relating to periodicity of a time-series change of the light quantity based on each detected light quantity, and a determination unit that determines the state of the grinding wheel using the calculated function.SELECTED DRAWING: Figure 2

Description

[Technical Field]

[0001] The present disclosure relates to a grinding wheel condition determination device and a grinding wheel condition determination method. [Background technology]

[0002] Patent Document 1 discloses a technique for estimating the state of a grinding wheel using event data that captures sparks generated between the grinding wheel and a workpiece. [Prior art documents] [Patent documents]

[0003] [Patent Document 1] International Publication No. 2022 / 181098 Summary of the Invention [Problem to be solved by the invention]

[0004] The technology of Patent Document 1 can determine the occurrence of clogging of a grinding wheel. However, Patent Document 1 does not specifically consider determining the state of a grinding wheel other than clogging, such as abrasive grain spillage or dullness. Therefore, there is a demand for a technology that can appropriately determine the state of a grinding wheel other than clogging. [Means for solving the problem]

[0005] The present disclosure can be realized in the following forms.

[0006] (1) According to a first aspect of the present disclosure, there is provided a grinding wheel condition determination device that determines the condition of a grinding wheel based on sparks generated when a workpiece is machined by the grinding wheel. This grinding wheel condition determination device includes an acquisition unit that acquires a plurality of spark images in which the sparks are captured consecutively in time series, a light amount detection unit that detects the light amount of the sparks included in each of the acquired spark images, a calculation unit that calculates a function relating to the periodicity of the time-series change in the light amount based on each of the detected light amounts, and a determination unit that determines the condition of the grinding wheel using the calculated function. According to this embodiment, the periodicity of the time-series change in the amount of light from the sparks is quantified as a function, and the condition of the grinding wheel can be appropriately determined using the function. (2) In the above aspect, the function may be an autocorrelation function. According to this aspect, the periodicity of the sparks can be easily quantified as an autocorrelation function. (3) In the above-described embodiment, the calculation unit may extract a maximum value that appears at a lug number corresponding to an integer number of revolutions of the grinding wheel in the autocorrelation function, and the determination unit may determine the condition of the grinding wheel using the maximum value. According to this embodiment, the condition of the grinding wheel can be determined using the maximum value that adequately reflects the periodicity of sparks, thereby making it possible to more appropriately determine the condition of the grinding wheel. (4) In the above aspect, the calculation unit may calculate a plurality of the autocorrelation functions at predetermined time intervals and extract the maximum value for each of the calculated autocorrelation functions, and the determination unit may determine that the grinding wheel is in a dull state when the time-series change in the maximum value shows an increasing trend of at least a predetermined level. According to this aspect, the periodic increase in sparks in the dull state can be appropriately detected as a time-series increase in the maximum value, and the dull state can be more appropriately determined. (5) In the above embodiment, the calculation unit may calculate a plurality of the autocorrelation functions at predetermined time intervals and extract the maximum value for each of the calculated autocorrelation functions, and the determination unit may determine that the grinding wheel is in a state of overflow when the time series change in the maximum value shows a decreasing trend of at least a predetermined degree. According to this embodiment, a decrease in the periodicity of sparks in a state of overflow can be appropriately detected as a time series decrease in the maximum value, and the state of overflow can be more appropriately determined. (6) In the above embodiment, the determination unit may determine that the grinding wheel is in a dull state when the maximum value is equal to or greater than a predetermined threshold value. According to this embodiment, the dull state of the grinding wheel can be determined more easily. (7) In the above embodiment, the determination unit may determine that the grinding wheel is in a state of overflow when the maximum value is equal to or less than a predetermined threshold value. According to this embodiment, it is possible to more easily determine the state of overflow. (8) In the above embodiment, the determination unit may determine that the grinding wheel is in an abnormal state when a maximum value appears in the autocorrelation function at a lug number corresponding to a non-integer number of revolutions of the grinding wheel. According to this embodiment, it is possible to more easily determine that the grinding wheel is in an abnormal state. The present disclosure may be realized in various forms other than a grinding wheel condition determination device, such as a grinding wheel condition determination system, a grinding wheel condition determination method, a computer program for implementing the method, and a non-transitory recording medium on which the computer program is recorded. [Brief explanation of the drawings]

[0007] [Figure 1] FIG. 1 is an explanatory diagram showing the configuration of a grindstone condition determination system in a first embodiment. [Figure 2] FIG. 2 is an explanatory diagram showing the configuration of a CPU. [Figure 3] 10A and 10B are diagrams illustrating the state of the grindstone determined by the determining unit. [Figure 4] 4 is a flowchart of a determination process in the first embodiment. [Figure 5]10A and 10B are diagrams illustrating an example of a light amount detection process. [Figure 6] FIG. 10 is a diagram illustrating an example of a function calculation process. [Figure 7] 10A and 10B are diagrams illustrating examples of autocorrelation functions calculated for each grindstone state. [Figure 8] FIG. 10 is a diagram illustrating an example of time-series changes in the first maximum value. [Figure 9] 10 is a flowchart of a determination process in the second embodiment. [Figure 10] 10A and 10B are diagrams illustrating an example of determination of a grindstone state in the second embodiment. [Figure 11] 10 is a flowchart of a determination process in the third embodiment. DETAILED DESCRIPTION OF THE INVENTION

[0008] A. First embodiment: 1 is an explanatory diagram showing the configuration of a grinding wheel condition determination system 10 in a first embodiment. The grinding wheel condition determination system 10 includes an imaging device 12 and a grinding wheel condition determination device 14. In this embodiment, the grinding wheel condition determination system 10 is used when grinding a workpiece W with a grinding wheel 200 attached to a machining device such as a grinding machine or a machining center. The workpiece W is, for example, a steel material.

[0009] The grinding wheel condition determination device 14 determines the condition of the grinding wheel 200 based on sparks SP generated by friction from the grinding wheel 200 when grinding the workpiece W. More specifically, the sparks SP are generated by contact between the workpiece W and the grinding wheel 200 that is grinding the workpiece W. Hereinafter, the "condition of the grinding wheel 200" will also be referred to as the "grinding wheel condition."

[0010] The imaging device 12 generates multiple spark images by continuously capturing images of the sparks SP in a time series. The imaging device 12 is, for example, a camera. The imaging device 12 is configured to be able to communicate with the grinding wheel condition determination device 14 via wired or wireless communication. In this embodiment, the imaging device 12 continuously captures images of the sparks SP at very short time intervals. Here, the very short time interval refers to a time interval in which, when focusing on one of the streamlines that form the sparks SP, the process from the appearance to the disappearance of that streamline can be captured in multiple images. More specifically, in this embodiment, the spark images are captured at an imaging speed of 10,000 frames per second. Furthermore, in this embodiment, the spark images are color images. The resolution of the spark images is 1280 x 720 pixels.

[0011] The grinding wheel condition determination device 14 is configured by a computer having one or more CPUs 15, a memory 16 including ROM and RAM, an input / output interface 17, and an internal bus 18. The CPU 15, memory 16, and input / output interface 17 are connected via the internal bus 18 to enable bidirectional communication. The CPU 15 executes a program PG stored in the memory 16 to realize various functions, including the functions of various functional units described below and the function of executing determination processing.

[0012] In this embodiment, an output device 70 is connected to the grindstone condition determination device 14. The output device 70 includes, for example, a display device such as a display that outputs visual information, a speaker that outputs audio information, and the like.

[0013] 2 is an explanatory diagram showing the configuration of the CPU 15. The CPU 15 includes, as functional units, an acquisition unit 21, a brightness conversion processing unit 22, a binarization processing unit 23, a light amount detection unit 24, a calculation unit 25, a determination unit 26, a grindstone adjustment unit 32, and an output unit 33.

[0014] The acquisition unit 21 acquires a plurality of spark images generated by the imaging device 12 and stores them in the memory 16 .

[0015] The brightness conversion processing unit 22 converts each color spark image acquired by the acquisition unit 21 into a grayscale spark image, and stores each converted spark image in the memory 16. Hereinafter, a "grayscale spark image" will also be simply referred to as a "grayscale image." A grayscale conversion method is, for example, a method of converting the RGB values ​​of each pixel in a spark image into a brightness value using a predetermined conversion formula, and specifically, the average method, weighted average method, brightness method, etc. Note that the grayscale conversion method is not limited to the above methods.

[0016] The binarization processing unit 23 generates a binarized spark image by binarizing each grayscale image generated by the brightness conversion processing unit 22 using a predetermined threshold. Hereinafter, a "binarized spark image" will also be simply referred to as a "binarized image." The binarization processing unit 23 stores each generated binarized image in the memory 16. By performing such binarization, the shape of the sparks SP in the spark image can be made clearer, and the pixel regions representing the sparks SP in the spark image can be made more distinct. Note that in this embodiment, the binarization processing unit 23 generates a binarized image by converting pixels in the grayscale image before binarization that have a brightness value greater than a threshold into white pixels and pixels in the grayscale image before binarization that have a brightness value greater than the threshold into black pixels. Therefore, in the binarized image, the sparks SP are represented in white. In other embodiments, during binarization, for example, pixels in the image before binarization that have a brightness value greater than a threshold into black pixels and pixels in the image before binarization that have a brightness value greater than the threshold into white pixels may be converted.

[0017] The light amount detection unit 24 executes a light amount detection process. The light amount detection process is a process for detecting the light amount of the sparks SP in each spark image. In this embodiment, in the light amount detection process, the light amount detection unit 24 detects the number of pixels of the sparks SP in each spark image as the light amount of the sparks SP in each spark image. More specifically, the light amount detection unit 24 detects the number of pixels of the sparks SP in each binary image.

[0018] The calculation unit 25 executes a function calculation process. The function calculation process is a process for calculating a function relating to the periodicity of the time-series change in the light intensity of the spark SP based on each light intensity detected for each spark image by the light intensity detection unit 24. Hereinafter, the "periodicity of the time-series change in the light intensity of the spark SP" will also be simply referred to as "spark periodicity." In this embodiment, the calculation unit 25 calculates an autocorrelation function as a function relating to the spark periodicity.

[0019] In this embodiment, the calculation unit 25 further executes a feature extraction process, which is a process of extracting a predetermined feature from the autocorrelation function.

[0020] The determining unit 26 determines the state of the grindstone using the function relating to the spark periodicity calculated by the calculating unit 25.

[0021] The grindstone adjustment unit 32 executes a grindstone adjustment process. The grindstone adjustment process is a process for adjusting the grindstone 200 using the grindstone state determined by the determination unit 26. In this embodiment, "adjustment of the grindstone 200" includes at least one of adjusting the maintenance timing of the grindstone 200, adjusting the replacement timing of the grindstone 200, and adjusting the processing conditions of the grindstone 200. Specifically, the maintenance of the grindstone 200 includes truing and dressing.

[0022] The output unit 33 uses the output device 70 to output various information including the determination result of the grindstone condition to the user of the grindstone condition determination device 14.

[0023] FIG. 3 is a diagram illustrating the grinding wheel state determined by the determination unit 26 in this embodiment. FIG. 3 shows three grinding wheel states: a normal state, a dulled state, and a partially unfilled state. The normal state, the dulled state, and the partially unfilled state each relate to the state of the abrasive grains GR on the surface SR of the grinding wheel 200. For each grinding wheel state, FIG. 3 schematically shows the grinding wheel 200, the state of chips CH generated by grinding using the grinding wheel 200, and sparks SP generated from the grinding wheel 200. Note that in FIG. 3, the abrasive grains GR are hatched with diagonal lines. Also, in FIG. 3, the chips CH are hatched with dotted patterns. Also, in FIG. 3, the magnitude of the light intensity of the sparks SP is represented by the length of the white arrow.

[0024] More specifically, sparks SP are generated when chips CH are heated and oxidized. Therefore, the state of the sparks SP changes depending on the state of the chips CH. The state of the sparks SP here refers to, for example, the movement speed, movement direction, size, spatial distribution, and time distribution of the sparks SP. Furthermore, the state of the chips CH refers to, for example, the movement speed, movement direction, size, temperature, shape, and chemical composition of the chips CH. The state of the chips CH changes depending on the size and sharpness of the abrasive grains GR on the surface SR of the grinding wheel 200. As a result, there is a good correlation between the state of the sparks SP and the state of the grinding wheel.

[0025] In the normal state shown in FIG. 3 , abrasive grains GR of various sizes and sharpness are exposed on the surface SR of the grinding wheel 200. In the normal state, the self-sharpening action of the grinding wheel 200 operates normally. More specifically, in the normal state, the wear of the abrasive grains GR on the surface SR and the shedding of the worn abrasive grains GR from the surface SR occur in a balanced manner, so that the grinding wheel 200 is maintained in a state suitable for grinding even without truing or dressing. In the normal state, this self-sharpening action operates normally, so that the state of the abrasive grains GR on the surface SR changes slightly during one rotation of the grinding wheel 200. As a result, in the normal state, during grinding using the grinding wheel 200, a relatively high spark periodicity corresponding to one rotation of the grinding wheel 200 is exhibited.

[0026] The dulled and overgrown states are abnormal states in which the self-sharpening mechanism does not function properly, i.e., the balance between wear and shedding of GR abrasive grains on the surface SR is disrupted. The dulled state is a state in which the shape of the GR abrasive grains on the surface SR is flattened due to wear compared to the normal state. Hereinafter, the area of ​​the surface SR where the shape of the GR abrasive grains is flattened is also referred to as the "dulled region." The dulled state occurs when worn GR abrasive grains do not fall off at the appropriate time and remain on the surface SR. The overgrown state is a state in which GR abrasive grains fall off the surface SR more frequently than in the normal state. The overgrown state occurs when GR abrasive grains fall off the surface SR before they are worn down to the appropriate degree. Note that in Figure 3, the broken lines indicate GR abrasive grains that have fallen off the surface SR. The "-" in the overgrown state in Figure 3 indicates that chips CH and sparks SP, which would normally occur under normal conditions, were not generated due to the shedding of GR abrasive grains.

[0027] The inventors of the present application have found that spark periodicity increases in a dulled state compared to a normal state. In a dulled state, the shape of each abrasive grain GR is flattened in the dulled region, resulting in less variation in the shape of each abrasive grain GR on the surface SR. Therefore, in a dulled state, the shape of chips CH becomes more uniform, and the state of sparks SP becomes more uniform, compared to a normal state. Furthermore, self-sparking of the grinding wheel 200 is less likely to occur in the dulled region. As a result, in a worn state, sparks SP are generated based on the dulled region every time the grinding wheel 200 rotates, and spark periodicity increases compared to a normal state.

[0028] The inventors of the present application also found that spark periodicity is lower in a state of overflow compared to a normal state. In a state of overflow, the abrasive grains GR fall off, resulting in greater variation in the distribution of the abrasive grains GR on the surface SR and in the shape of each abrasive grain GR, compared to a normal state. Furthermore, in a state of overflow, the abrasive grains GR fall off the surface SR more frequently than in a normal state, which makes it more likely that the condition of the surface SR will vary with each rotation of the grinding wheel 200. As a result, in a state of overflow, spark periodicity is lower than in a normal state.

[0029] 4 is a flowchart of a determination process for realizing the grindstone condition determination method in this embodiment. The determination process is executed by the CPU 15 of the grindstone condition determination device 14. In this embodiment, the grindstone condition determination device 14 executes the determination process at predetermined time intervals while the workpiece W is being machined by the grindstone 200. The time interval at which the determination process is executed is, for example, one second.

[0030] 4, the acquisition unit 21 acquires a plurality of spark images and stores them in the memory 16. In step S2, the brightness conversion processing unit 22 converts the brightness of each spark image acquired in step S1 into a grayscale image consisting of brightness values ​​only, and stores them in the memory 16. In step S3, the binarization processing unit 23 converts each grayscale image generated in step S2 into a binary image and stores them in the memory 16.

[0031] In step S4, the light amount detection unit 24 executes a light amount detection process. In step S4, the light amount detection unit 24 detects the light amount of the sparks SP in each of the binary images generated in step S3, and stores the detected amount in the memory 16. More specifically, the light amount detection unit 24 counts the number of pixels representing the sparks SP in each of the binary images, i.e., the number of white pixels.

[0032] FIG. 5 is a diagram illustrating an example of light intensity detection processing. In FIG. 5, spark images Im1, Im2, and Im3 are shown as examples of spark images. Spark images Im1, Im2, and Im3 are spark images captured at times t1, t2, and t3, respectively. Time t2 is a time after time t1, and time t3 is a time after time t2. Each spark image in FIG. 5 is represented in a coordinate system having mutually orthogonal x- and y-axes as its coordinate axes. Note that in FIG. 5, for convenience of illustration, white images are represented in black, and black pixels are represented in white.

[0033] As shown in FIG. 5, spark image Im1 at time t1 includes spark SP1 and spark SP2. Thereafter, between time t1 and time t2, spark SP1 and spark SP2 each move in the +x direction and the +y direction without disappearing. Furthermore, a new spark SP3 occurs between time t1 and time t2. As a result, spark image Im2 at time t2 includes spark SP1, spark SP2, and spark SP3. Thereafter, between time t2 and time t3, spark SP1 and spark SP2 each disappear. Furthermore, between time t2 and time t3, spark SP3 moves in the +x direction and the +y direction without disappearing. As a result, spark image Im3 at time t3 includes only spark SP3.

[0034] In spark image Im1, the number of pixels of spark SP1 is 4, the number of pixels of spark SP2 is 7, and the total number of pixels of each spark SP is 11. Therefore, the light amount detection unit 24 detects the light amount of the spark SP in spark image Im1, i.e., the light amount of the spark SP at time t1, as "11." In substantially the same manner, the light amount detection unit 24 detects the light amounts of the spark SP in spark images Im2 and Im3, i.e., the light amounts of the spark SP at times t2 and t3, as "17" and "6," respectively.

[0035] 4, the calculation unit 25 executes a function calculation process. In step S5, the calculation unit 25 calculates an autocorrelation function related to spark periodicity based on the light intensity of the sparks SP in each spark image detected in step S4.

[0036] FIG. 6 is a diagram illustrating an example of the function calculation process. The light intensity in each spark image is detected by the light intensity calculation process described above, thereby obtaining time-series data TD representing the time-series change in the light intensity, as shown in the upper part of FIG. 6. The calculation unit 25 calculates an autocorrelation function AF based on the time-series data TD, as shown in the lower part of FIG. 6. More specifically, the calculation unit 25 repeatedly calculates the correlation coefficient between the time-series data TD and shift data obtained by shifting the time of the time-series data TD by the number of lags while changing the number of lags, and defines the correspondence between the calculated correlation coefficient and the number of lags as the autocorrelation function AF. The lower part of FIG. 6 shows the autocorrelation function AF calculated in this manner in the form of a graph, with the vertical axis representing the autocorrelation and the horizontal axis representing the number of lags. The "autocorrelation" in the autocorrelation function AF represents the correlation coefficient calculated for each number of lags. The number of lags has a time dimension.

[0037] In step S6 of FIG. 4, the calculation unit 25 executes a feature extraction process. In this embodiment, the calculation unit 25 extracts a first maximum value LM1 in the autocorrelation function AF as the feature value. The first maximum value LM1 is a maximum value that appears in the autocorrelation function AF at a lag number corresponding to an integer number of rotations of the grinding wheel 200. In other words, the first maximum value LM1 is a maximum value that appears in the autocorrelation function AF corresponding to the rotation period of the grinding wheel 200. FIG. 6 shows, as examples of the first maximum value LM1, a maximum value LM1a at lag number T1, a maximum value LM1b at lag number T2, and a maximum value LM1c at lag number T3. The lag number T1 is the lag number corresponding to one rotation of the grinding wheel 200. The lag number T2 is the lag number corresponding to two rotations of the grinding wheel 200. The lag number T3 is the lag number corresponding to three rotations of the grinding wheel 200.

[0038] FIG. 7 is a diagram illustrating examples of autocorrelation functions AF calculated for each grinding wheel state. FIG. 7 shows an autocorrelation function AF1 in the normal state, an autocorrelation function AF2 in the blemished state, and an autocorrelation function AF3 in the overflow state. As shown in FIG. 7, the autocorrelation function AF2 in the blemished state has larger maximum values ​​LM1a, LM1b, and LM1c than the autocorrelation function AF1 in the normal state. Specifically, the maximum values ​​LM1a, LM1b, and LM1c in the autocorrelation function AF2 are each equal to or greater than 0.3 and greater than 0.2. This is due to the higher spark periodicity in the blemished state compared to the normal state, as described above. Furthermore, the autocorrelation function AF3 in the overflow state has smaller values ​​of the first maximum values ​​LM1 than the autocorrelation function AF1. Specifically, the maximum value LM1a in the autocorrelation function AF3 is a value slightly exceeding 0.2, while the maximum values ​​LM1b and LM1c are each less than 0.2. This is because, as described above, the spark periodicity is lower in the overflow state compared to the normal state.

[0039] 7 shows a second maximum value LM2 in addition to the first maximum value LM1. The second maximum value LM2 will be described in a third embodiment below.

[0040] 4, the calculation unit 25 detects the trend of time-series changes in the first maximum value LM1 extracted in step S6. Hereinafter, the trend of time-series changes in the first maximum value LM1 will also be referred to as the time-series trend. In step S7 in this embodiment, the calculation unit 25 detects the trend of time-series changes in the maximum value LM1a as the time-series trend. Note that in other embodiments, the calculation unit 25 may detect the trend of time-series changes in other first maximum values ​​LM1, such as maximum values ​​LM1b and LM1c, as the time-series trend.

[0041] FIG. 8 is a diagram illustrating an example of a time series change in the local maximum value LM1a. In this embodiment, the determination process is repeatedly performed at predetermined time intervals to calculate multiple autocorrelation functions AF, and a local maximum value LM1a is extracted for each calculated autocorrelation function AF. FIG. 8 shows the time series change in the series of local maximum values ​​LM1a extracted in this manner in the form of a graph, with the local maximum value LM1a on the vertical axis and time on the horizontal axis. Note that time "0" in FIG. 8 represents the timing immediately after truing and dressing of the grinding wheel 200. That is, FIG. 8 shows the time series change in the local maximum value LM1a when the first determination process is performed immediately after maintenance of the grinding wheel 200, and then the determination process is repeatedly performed at predetermined time intervals. In other embodiments, the timing at which the first determination process is performed does not have to be immediately after maintenance of the grinding wheel 200.

[0042] In this embodiment, in step S7 of FIG. 4, the calculation unit 25 first calculates the latest moving average based on the time-series changes shown in FIG. 8. The latest moving average is the moving average for the latest maximum value LM1a. Specifically, the calculation unit 25 calculates the moving average of a predetermined number of maximum values ​​LM1a, including the latest maximum value LM1a, as the latest moving average. Next, the calculation unit 25 detects the difference between the calculated latest moving average and the previous moving average as the time-series trend. The previous moving average is the moving average for the maximum value LM1a immediately preceding the latest maximum value LM1a. In step S7 of FIG. 4, the calculation unit 25 may use, for example, the latest moving average calculated in the previous determination process as the previous moving average, or may calculate a new previous moving average. In other embodiments, the time-series trend may be detected, for example, as the slope of the change between the latest moving average and the previous moving average.

[0043] In step S8 of FIG. 4, the determination unit 26 determines whether the time series trend is an increasing trend of a predetermined first degree or more. Hereinafter, an "increasing trend of a first degree or more" will also be simply referred to as an "increasing trend." More specifically, in step S8 of this embodiment, the determination unit 26 determines whether the difference in the moving averages calculated in step S7 is a first degree or more. If the time series trend is an increasing trend in step S8, the determination unit 26 determines in step S9 that the grinding wheel 200 is in a dulled state. For example, in the period Pd2 shown in FIG. 8, the maximum value LM1a tends to increase, the time series trend is determined to be an increasing trend, and it can be determined that the grinding wheel 200 is in a dulled state.

[0044] If the time series trend is not an increasing trend in step S8 of FIG. 4, the determination unit 26 determines in step S10 whether the time series trend is a decreasing trend of a predetermined second degree or more. Hereinafter, a "decreasing trend of a second degree or more" will also be simply referred to as a "decreasing trend." More specifically, in step S10 of this embodiment, the determination unit 26 determines whether the difference in the moving averages calculated in step S7 is equal to or less than a second degree. If the time series trend is a decreasing trend in step S10, the determination unit 26 determines in step S11 that the grinding wheel 200 is in a non-uniform state. For example, in the period Pd3 shown in FIG. 8, the maximum value LM1a tends to decrease, the time series trend is determined to be a decreasing trend, and the grinding wheel 200 may be determined to be in a non-uniform state.

[0045] As shown in FIG. 4, if step S9 or S11 is performed, the grindstone adjustment unit 32 performs a grindstone adjustment process in step S20. In step S20, the grindstone adjustment unit 32 may, for example, advance the maintenance period of the grindstone 200 or the replacement period of the grindstone 200. Note that, for example, if the grindstone condition determination system 10 or the machining device is equipped with a device capable of performing maintenance or replacement of the grindstone 200, the grindstone adjustment unit 32 may send a control signal to the device so that maintenance or replacement is performed at the changed time. Also, in step S20, the grindstone adjustment unit 32 may change the processing conditions, for example, to improve the glazing or chipping state. More specifically, if the glazing state is determined to exist in step S9, the grindstone adjustment unit 32 may, for example, increase the feed rate of the grindstone 200, decrease the rotation speed of the grindstone 200, or increase the cutting depth of the grindstone 200. Furthermore, if it is determined in step S11 that there is overflow, the grinding wheel adjustment unit 32 may, for example, perform at least one of the following actions: reducing the feed speed of the grinding wheel 200, increasing the rotation speed of the grinding wheel 200, or reducing the cutting depth of the grinding wheel 200.

[0046] If the time series trend is not a decreasing trend in step S10, the determination unit 26 determines in step S30 that the grinding wheel condition is normal. That is, in this embodiment, the determination unit 26 determines that the grinding wheel condition is normal when the time series trend is neither an increasing trend nor a decreasing trend. For example, in the period Pd1 shown in FIG. 8, the maximum value LM1a fluctuates within a certain range, and it can be determined that the grinding wheel condition is normal.

[0047] 4, the output unit 33 outputs the determination result of the grinding wheel condition using the output device 70. In the determination process, the output unit 33 may output, using the output device 70, not only the determination result of the grinding wheel condition, but also various other information such as the autocorrelation function AF, the feature amount of the autocorrelation function AF, the tendency of time-series changes in the feature amount, the time-series tendency, and the adjustment details adjusted in the grinding wheel adjustment process.

[0048] According to the grinding wheel condition determination device 14 in this embodiment described above, a function relating to the spark periodicity is calculated based on the light intensity of each spark SP included in each of the plurality of spark images, and the calculated function is used to determine the condition of the grinding wheel 200. According to this embodiment, the spark periodicity is quantified as a function, and the function can be used to appropriately determine the condition of the grinding wheel 200.

[0049] In this embodiment, the autocorrelation function AF is calculated as a function related to the spark periodicity, so that the spark periodicity can be easily quantified as the autocorrelation function AF.

[0050] In this embodiment, the grinding wheel condition is determined using the first maximum value LM1. The first maximum value LM1 is a maximum value that appears in accordance with the rotation period of the grinding wheel 200, and therefore reflects the spark periodicity well. By determining the grinding wheel condition using such first maximum value LM1, the grinding wheel condition can be determined more appropriately.

[0051] In this embodiment, a blinded state is determined when the time-series change in the first maximum value LM1 tends to increase. As described above, in a blinded state, the spark periodicity increases compared to a normal state. In this embodiment, this increase in spark periodicity can be properly detected as a time-series increase in the first maximum value LM1, and the blinded state can be more properly determined.

[0052] In this embodiment, the overflow state is determined when the time-series change in the first maximum value LM1 is decreasing. As described above, in the overflow state, the spark periodicity decreases compared to the normal state. In this embodiment, such a decrease in the spark periodicity can be properly detected as a time-series decrease in the first maximum value LM1, and the overflow state can be more properly determined.

[0053] Furthermore, in this embodiment, the grindstone adjustment process is performed, so that the grindstone 200 can be appropriately adjusted in accordance with the determination result of the grindstone condition. As a result, the grindstone 200 can be used in a more appropriate state.

[0054] In this embodiment, the determination result of the grindstone condition is output. Therefore, the user can, for example, check the output determination result and adjust the grindstone 200 based on the checked determination result. Furthermore, if at least one piece of information among the autocorrelation function AF, the feature amount of the autocorrelation function AF, the trend of time-series changes in the feature amount, and the time-series trend is output as visual information, the user can visually check the spark periodicity.

[0055] B. Second embodiment: FIG. 9 is a flowchart of the determination process in the second embodiment. Unlike the first embodiment, in the second embodiment, the calculation unit 25 does not detect a time-series trend. Furthermore, the determination unit 26 determines the grinding wheel condition using a threshold value, as will be described later, without relying on a time-series trend. The grinding wheel condition determination system 10 and the grinding wheel condition determination device 14 in the second embodiment are the same as those in the first embodiment unless otherwise specifically described. Note that in FIG. 9, steps that are the same as those in FIG. 4 are assigned the same reference numerals as those in FIG. 4.

[0056] 9, in this embodiment, after step S6, in step S12, the determination unit 26 determines whether the first maximum value LM1 is equal to or greater than a predetermined threshold. Specifically, in step S12 in this embodiment, the determination unit 26 determines whether the maximum value LM1b is equal to or greater than a first threshold. If the maximum value LM1b is equal to or greater than the first threshold in step S12, the determination unit 26 determines in step S13 that the grinding wheel 200 is in a dull state.

[0057] If the maximum value LM1b is not equal to or greater than the first threshold in step S12, the determination unit 26 determines in step S14 whether the first maximum value LM1 is equal to or less than a predetermined threshold. Specifically, in step S14 in this embodiment, the determination unit 26 determines whether the maximum value LM1b is equal to or less than a second threshold. The second threshold is a value smaller than the first threshold. In this embodiment, the first threshold is 0.3, and the second threshold is 0.2. If the maximum value LM1b is equal to or less than the second threshold in step S14, the determination unit 26 determines in step S15 that the grinding wheel 200 is in a state of chipping.

[0058] FIG. 10 is a diagram illustrating an example of determining the grinding wheel condition in the second embodiment. FIG. 10 shows a graph in which the first maximum values ​​LM1 of the autocorrelation functions AF1, AF2, and AF3 shown in FIG. 7 are plotted. The number N of the first maximum value in FIG. 10 indicates that the first maximum value LM1 appears corresponding to N rotations of the grinding wheel 200. For example, the first maximum values ​​LM1 numbered 1, 2, and 3 correspond to the maximum values ​​LM1a, LM1b, and LM1c, respectively. As shown in FIG. 10, the maximum value LM1b in the autocorrelation function AF2 is equal to or greater than the first threshold value TH1. Therefore, when the autocorrelation function AF2 is calculated for the grinding wheel 200 in the function calculation process, the grinding wheel condition of the grinding wheel 200 is determined to be in a dull state in step S12 of FIG. 9. Furthermore, as shown in FIG. 10, the maximum value LM1b in the autocorrelation function AF3 is equal to or less than the second threshold value TH2. Therefore, when the autocorrelation function AF3 is calculated for the grinding wheel 200 in the function calculation process, the grinding wheel state of the grinding wheel 200 is determined to be a chipped state in step S14 of FIG.

[0059] In other embodiments, the grinding wheel 200 may be determined to be in a dull state when, for example, other first maximum values ​​LM1, such as the maximum values ​​LM1a and LM1c, are equal to or greater than a first threshold value. The grinding wheel 200 may also be determined to be in a dull state when two or more first maximum values ​​LM1 are each equal to or greater than a first threshold value, or when at least one of the two or more first maximum values ​​LM1 is equal to or greater than a first threshold value. In this case, the first threshold values ​​may be the same or different. In a similar manner, the grinding wheel 200 may be determined to be in a dull state when other first maximum values ​​LM1, such as the maximum values ​​LM1a and LM1c, are equal to or less than a second threshold value.

[0060] The first threshold value TH1 and the second threshold value TH2 may be determined based on, for example, experimental results so as to appropriately determine whether the eye is blinded or overflowing. Note that the term "experiment" as used herein includes simulated experiments. The first threshold value TH1 and the second threshold value TH2 may be determined based on, for example, the value of the first maximum value LM1 in a normal state. For example, the first threshold value TH1 may be a value obtained by adding a predetermined value to the first maximum value LM1 of the autocorrelation function AF calculated immediately after dressing. For example, in a configuration in which the blinded state is determined when the maximum value LM1b is equal to or greater than the first threshold value TH1, as in the present embodiment, the first threshold value TH1 may be a value obtained by adding a predetermined value to the maximum value LM1b of the autocorrelation function AF calculated immediately after dressing. In this case, the predetermined value may be determined based on, for example, experimental results. Similarly, the second threshold value TH2 may be a value obtained by subtracting a predetermined value from the first maximum value LM1 of the autocorrelation function AF calculated immediately after dressing.

[0061] When step S13 or S15 is executed, the grindstone adjustment unit 32 executes a grindstone adjustment process in step S20. Furthermore, when the first maximum value LM1 is not equal to or less than the second threshold value in step S14, the determination unit 26 determines that the grindstone condition is normal in step S30. That is, in this embodiment, the determination unit 26 determines that the grindstone condition is normal when the maximum value LM1b is not equal to or greater than the first threshold value TH1 and not equal to or less than the second threshold value TH2.

[0062] According to the grinding wheel condition determination device 14 in the second embodiment described above, when the first maximum value LM1 is equal to or greater than the threshold value, it is determined that the grinding wheel 200 is in a dull state. Therefore, the dull state can be determined more easily using the threshold value.

[0063] In this embodiment, when the first maximum value LM1 is equal to or less than the threshold value, it is determined that the grinding wheel 200 is in a state of overflow. Therefore, the use of the threshold value makes it possible to more easily determine the state of overflow.

[0064] In other embodiments, the determination of the grinding wheel condition using the time-series trend described in the first embodiment and the determination of the grinding wheel condition using the threshold value described in the second embodiment may be used in combination. For example, a dull state may be determined when the time-series trend is increasing and the first maximum value LM1 is equal to or greater than a threshold value. Alternatively, a dull state may be determined when at least one of the following conditions is met: the time-series trend is increasing and the first maximum value LM1 is equal to or greater than a threshold value. Similarly, a smear state may be determined when the time-series trend is decreasing and the first maximum value LM1 is equal to or less than a threshold value. Alternatively, a smear state may be determined when at least one of the following conditions is met: the time-series trend is decreasing and the first maximum value LM1 is equal to or less than a threshold value.

[0065] C. Third embodiment: Fig. 11 is a flowchart of the determination process in the third embodiment. In the third embodiment, unlike the first and second embodiments, the determination unit 26 determines the grinding wheel condition based on the second maximum value LM2 shown in Fig. 7, rather than the first maximum value LM1. The grinding wheel condition determination system 10 and the grinding wheel condition determination device 14 in the third embodiment are the same as those in the first embodiment unless otherwise specifically described. In Fig. 11, the same steps as those in Figs. 4 and 9 are denoted by the same reference numerals as those in Figs. 4 and 9.

[0066] In step S6b in Fig. 11, the calculation unit 25 extracts the second maximum value LM2 shown in Fig. 7 as a feature amount from the autocorrelation function AF. The second maximum value LM2 shown in Fig. 7 is a maximum value that appears at a lug number corresponding to a non-integer number of rotations of the grinding wheel 200 in the autocorrelation function AF.

[0067] The inventors of the present application have found that, as shown in Fig. 7, the second maximum value LM2 appears in the autocorrelation function AF2 in the glazed state and in the autocorrelation function AF3 in the overflow state, but does not appear in the autocorrelation function AF1 in the normal state. For example, in the autocorrelation functions AF2 and AF3, the second maximum value LM2 appears at the lag number HT1 corresponding to a half rotation of the grinding wheel 200, but in the autocorrelation function AF1, the second maximum value LM2 does not appear at the lag number HT1. In this embodiment, the judgment unit 26 uses this second maximum value LM2 to judge the grinding wheel condition.

[0068] After step S6b shown in FIG. 11, in step S16, the judgment unit 26 judges whether or not a second maximum value LM2 was extracted in step S6b. More specifically, in step S16 in this embodiment, the judgment unit 26 judges whether or not a maximum value LM2h was extracted in step S6b. If the maximum value LM2h was extracted, in step S17, the judgment unit 26 judges that the grinding wheel 200 is in an abnormal state. The abnormal state here means that the grinding wheel 200 is in at least one of a glazing state and a chipped state. If the maximum value LM2h was not extracted, in step S30, the judgment unit 26 judges that the grinding wheel 200 is in a normal state.

[0069] In this embodiment, the determination unit 26 does not determine whether the grinding wheel 200 is in a dulled state or a non-perforated state. Therefore, in this embodiment, in the grinding wheel adjustment process of step S20, the grinding wheel adjustment unit 32 preferably performs at least one of adjusting the maintenance timing and the replacement timing of the grinding wheel 200.

[0070] According to the grinding wheel condition determination device 14 in the third embodiment described above, when the second maximum value LM2 appears in the autocorrelation function AF, it is determined that the grinding wheel 200 is in an abnormal state. Therefore, when the second maximum value LM2 appears in the autocorrelation function AF, it is possible to immediately determine that the grinding wheel 200 is in an abnormal state. Conversely, when the second maximum value LM2 does not appear, it is possible to immediately determine that the grinding wheel 200 is in a normal state. In this way, in this embodiment, it is possible to more easily determine that the grinding wheel 200 is in an abnormal state, and for example, it is possible to further reduce the processing load of the grinding wheel condition determination device 14 in the determination process.

[0071] In other embodiments, at least one of the determination of the grinding wheel condition using the time-series trend described in the first embodiment and the determination of the grinding wheel condition using the threshold value described in the second embodiment and the determination of the grinding wheel condition using the second maximum value LM2 described in the third embodiment may be used in combination. For example, when an abnormal state is determined by the determination of the grinding wheel condition using the second maximum value LM2, at least one of the determination using the time-series trend and the determination using the threshold value may be performed. In this way, if the second maximum value LM2 does not appear, it is possible to immediately determine that the grinding wheel is in a normal state, and if the second maximum value LM2 appears, it is possible to identify whether the abnormal state is a glazed state or a non-existent state.

[0072] D. Other Embodiments: (D-1) In the above embodiments, an autocorrelation function is used as a function relating to spark periodicity, but this is not limiting. For example, a power spectrum may be used as a function relating to spark periodicity.

[0073] (D-2) In the first and second embodiments, both the blinded state and the overflowed state are determined, but it is also possible to determine only one of them.

[0074] (D-3) In each of the above embodiments, a normal state is determined, but a normal state does not have to be determined. More specifically, in the first and second embodiments, even if it is determined that neither the overflow state nor the blind state exists, step S30 in Fig. 4 or 9 may be omitted. Furthermore, in the third embodiment, even if it is determined that the abnormal state exists, step S30 in Fig. 11 may be omitted.

[0075] (D-4) In each of the above embodiments, the determination result of the grindstone condition does not have to be output. In this case, the grindstone condition determination device 14 does not have to be equipped with the output unit 33.

[0076] (D-5) In each of the above embodiments, grindstone adjustment may not be performed. In this case, the grindstone condition determination device 14 may not include the grindstone adjustment unit 32.

[0077] (D-6) In each of the above embodiments, grayscaling or binarization may not be performed. In this case, the grindstone condition determination device 14 may not include the brightness conversion processing unit 22 or the binarization processing unit 23.

[0078] The present disclosure is not limited to the above-described embodiments and can be realized in various configurations without departing from the spirit thereof. For example, the technical features in each embodiment corresponding to the technical features in the form described in the Summary of the Invention section can be appropriately replaced or combined to solve some or all of the above-described problems or achieve some or all of the above-described effects. Furthermore, if a technical feature is not described as essential in this specification, it can be appropriately deleted. [Explanation of symbols]

[0079] 10... Grinding wheel condition determination system, 12... Imaging device, 14... Grinding wheel condition determination device, 15... CPU, 16... Memory, 17... Input / output interface, 18... Internal bus, 21... Acquisition unit, 22... Brightness conversion processing unit, 23... Binarization processing unit, 24... Light amount detection unit, 25... Calculation unit, 26... Determination unit, 32... Grinding wheel adjustment unit, 33... Output unit, 70... Output device, 200... Grinding wheel

Claims

1. A grinding wheel condition determination device that determines the condition of a grinding wheel based on sparks generated when a workpiece is machined by the grinding wheel, an acquisition unit that acquires a plurality of spark images obtained by successively capturing images of the spark in time series; a light amount detection unit that detects the amount of light of the spark included in each of the acquired spark images; a calculation unit that calculates a function relating to the periodicity of the time-series change in the light amounts based on the detected light amounts; a determination unit that determines the state of the grindstone using the calculated function.

2. The grindstone condition determination device according to claim 1, The grinding wheel condition determination device, wherein the function is an autocorrelation function.

3. The grindstone condition determination device according to claim 2, the calculation unit extracts a maximum value that appears at a lug number corresponding to an integer number of rotations of the grinding wheel in the autocorrelation function, The determination unit determines the state of the grindstone using the maximum value.

4. The grindstone condition determination device according to claim 3, The calculation unit calculating a plurality of said autocorrelation functions at predetermined time intervals; extracting the maximum value for each of the calculated autocorrelation functions; The determination unit determines that the grinding wheel is in a dulled state when the time-series change in the maximum value shows an increasing trend of at least a predetermined level.

5. The grindstone condition determination device according to claim 3, The calculation unit calculating a plurality of said autocorrelation functions at predetermined time intervals; extracting the maximum value for each of the calculated autocorrelation functions; The determination unit determines that the grinding wheel is in a state of chipped grains when the time-series change in the maximum value shows a decreasing trend of at least a predetermined level.

6. The grindstone condition determination device according to claim 3, The determination unit determines that the grinding wheel is in a dulled state when the maximum value is equal to or greater than a predetermined threshold value.

7. The grindstone condition determination device according to claim 3, The determination unit determines that the grinding wheel is in a state of chipped grinding wheel when the maximum value is equal to or less than a predetermined threshold value.

8. The grindstone condition determination device according to claim 2, The determination unit determines that the grinding wheel is in an abnormal state when a maximum value appears in the autocorrelation function at a lug number corresponding to a non-integer number of revolutions of the grinding wheel.

9. A grinding wheel condition determination method for determining the condition of a grinding wheel based on the condition of sparks generated when a workpiece is machined by the grinding wheel, comprising: acquiring a plurality of spark images in which the sparks are successively captured in time series; Detecting the amount of light of the spark included in each of the acquired spark images; calculating a function relating to the periodicity of the time-series change in the light amounts based on the detected light amounts; The grinding wheel condition determination method determines the condition of the grinding wheel using the calculated function.

Images