TWI794027B - Turning tool collapse detection system and method thereof - Google Patents

Abstract

A turning tool collapse detection system and a method thereof are proposed. The turning tool collapse detection system includes a turning-milling machine, a three-axis accelerometer and a processor. The turning-milling machine includes a turning tool. The three-axis accelerometer is disposed on the turning-milling machine and configured to measure the turning tool to obtain a first time-domain vibration signal and a second time-domain vibration signal. The processor is electrically connected to the three-axis accelerometer. The processor receives the first time-domain vibration signal and the second time-domain vibration signal and is configured to implement a pre-processing step, a model building step, a vibration predicting step and a turning tool collapse detecting step. The pre-processing step is performed to convert the first time-domain vibration signal into a plurality of first frequency band energies. The model building step is performed to calculate the first frequency band energies to build a prediction model according to a self-organizing map method and a minimum quantization error method and obtain a best warning minimum quantization error threshold set. The vibration predicting step is performed to convert the second time-domain vibration signal into a plurality of second frequency band energies, and then input the second frequency band energies into the prediction model to obtain a predicted minimum quantization error value. The turning tool collapse detecting step is performed to compare the predicted minimum quantization error value and the best warning minimum quantization error threshold set to obtain a compared result, and detect whether the turning tool is damaged according to the compared result. Therefore, the turning tool collapse detection system of the present disclosure can improve the accuracy of replacement timing of the turning tool.

Description

Turning tool chipping detection system and method

The present invention relates to a detection system and its method, in particular to a turning tool chipping detection system and its method.

Turning is a traditional processing method. Lathe operators usually judge whether the tool is worn out based on the rolling chips, current characteristics, sound or visual experience of the tool during turning. However, when the operator notices it, the tool is usually damaged, which may easily cause damage to the workpiece being processed. In order to avoid damage to the workpiece, the operator usually replaces the tool early, which can reduce the damage rate of the workpiece, but increases the time of downtime for replacement and the cost of the tool. It can be seen that there is currently a lack of a turning tool chipping detection system and method that can improve the accuracy of tool replacement timing in the market, so related companies are all looking for a solution.

Therefore, the purpose of the present invention is to provide a turning tool chipping detection system and method thereof, which uses artificial intelligence to detect tool chipping in the cutting state of turning, can more accurately predict that the tool is about to break down, and then stop the machine immediately for replacement The cutting tool is not only convenient for the operator to use, but also ensures that the workpiece in processing will not be damaged due to the breakage of the cutting tool, so it can solve the problem caused by the different manual judgment standards in the conventional technology.

One embodiment of the structural aspect according to the present invention provides a chipping detection system for turning, which includes a turning-milling compound machine, a three-axis accelerometer, and an arithmetic processor. The mill-turn machine includes a tool. The three-axis accelerometer is set on the turning-milling machine and is used to measure the tool to obtain the first time-domain vibration signal and the second time-domain vibration signal. The computing processor is electrically connected to the three-axis accelerometer, and the computing processor receives the first time-domain vibration signal and the second time-domain vibration signal and is configured to perform operations including the following steps: pre-processing step, model building step, vibration prediction step And the detection step of the broken knife. The preprocessing step is converting the first time-domain vibration signal into complex first frequency band energy. The model building step is to calculate the energy of the first frequency band according to the self-organizing map network algorithm and the minimum quantization error method to establish a prediction model, and obtain the optimal early warning minimum quantization error threshold group. The vibration prediction step is converting the second time-domain vibration signal into complex second frequency band energies, and then inputting the second frequency band energies into a prediction model to obtain a predicted minimum quantization error value. The tool chipping detection step is to compare the predicted minimum quantization error value with the best early warning minimum quantization error threshold value group to obtain a comparison result, and detect whether the tool is damaged according to the comparison result.

Thereby, the turning tool chipping detection system of the present invention uses artificial intelligence to detect the tool in the cutting state in the turning process, and can more accurately predict that the tool is about to collapse, and then stop the machine immediately to replace the tool, which is not only convenient for the operator to use, but also It can ensure that the workpiece in processing will not be damaged due to tool breakage.

Other examples of the foregoing embodiments are as follows: the foregoing preprocessing steps may include a first conversion step and a second conversion step. The first converting step is converting the first time-domain vibration signal into a frequency-domain vibration signal according to a Fourier transform. The second conversion step is to divide the vibration signal in the frequency domain into complex segment spectrum signals, and convert the segment spectrum signals into the first frequency band energies according to square root mean conversion.

Other examples of the above-mentioned embodiment are as follows: in the above-mentioned second conversion step, the number of these segment spectrum signals may be 13, and the spectrum range of at least one of these segment spectrum signals is 1000 Hz. The segment spectrum signal has a plurality of feature points, and the number of these feature points is 39. In addition, in the model building step, the operation processor adjusts the self-organizing map network algorithm to find a winning neuron, and then calculates these first frequency band energies according to the winning neuron and the minimum quantization error method to establish a prediction model, and obtain Get the best early warning minimum quantization error threshold group.

Other examples of the above-mentioned embodiment are as follows: the above-mentioned turning-milling machine may further include a warning light and a programmable logic controller. The programmable logic controller is electrically connected to the operation processor and the warning light. The operation processor determines whether to generate an early warning signal according to the comparison result, and the programmable logic controller controls the warning light according to the early warning signal. The optimal early warning minimum quantization error threshold group includes a first early warning threshold and a second early warning threshold, and the first early warning threshold is smaller than the second early warning threshold. When the comparison result is that the predicted minimum quantization error value is less than or equal to the first warning threshold, the arithmetic processor decides not to generate a warning signal, and the warning light turns green. When the comparison result is that the predicted minimum quantization error value is greater than the first warning threshold and less than or equal to the second warning threshold, the arithmetic processor determines to generate a warning signal and transmits the warning signal to the programmable logic controller, and the warning light turns yellow. . When the comparison result is that the predicted minimum quantization error value is greater than the second warning threshold, the arithmetic processor determines to generate a warning signal and transmits the warning signal to the programmable logic controller, the warning light turns red and the cutting tool stops.

Other embodiments of the above-mentioned embodiment are as follows: the above-mentioned turning tool chipping detection system may further include a graphical user interface, which is electrically connected to the arithmetic processor and receives the energy of the second frequency band and predicts the minimum quantization error value, the graphical user interface The interface displays the total amount of energy in these second frequency bands in the first area, and displays the predicted minimum quantization error value in the second area. The second area includes a first sub-area, a second sub-area and a third sub-area, and the colors of the first sub-area, the second sub-area and the third sub-area are green, yellow and red respectively. When the comparison result is that the predicted minimum quantization error value is less than or equal to the first warning threshold, the predicted minimum quantization error value is displayed in the first sub-area. When the comparison result is that the predicted minimum quantization error value is greater than the first warning threshold and less than or equal to the second warning threshold, the predicted minimum quantization error value is displayed in the second sub-area. When the comparison result is that the predicted minimum quantization error value is greater than the second warning threshold, the predicted minimum quantization error value is displayed in the third sub-area.

According to one embodiment of the method aspect of the present invention, a method for detecting tool chipping in turning is provided, which includes the following steps: a vibration measurement step, a preprocessing step, a model building step, a vibration prediction step, and a tool chipping detection step. The vibration measurement step is to drive the three-axis accelerometer to measure the tool of the turning-milling compound machine to obtain the first time-domain vibration signal. The preprocessing step is to drive the arithmetic processor to convert the first time-domain vibration signal into complex first frequency band energy. The model building step is to drive the computing processor to calculate the energy of the first frequency band according to the self-organizing map network algorithm and the minimum quantization error method to establish a prediction model, and obtain the optimal early warning minimum quantization error threshold set. The vibration prediction step is to drive the three-axis accelerometer to measure the tool to obtain the second time domain vibration signal, and drive the arithmetic processor to convert the second time domain vibration signal into complex second frequency band energy, and then drive the arithmetic processor to convert the second time domain vibration signal The two-band energy is input to the prediction model to obtain the predicted minimum quantization error value. The chipping detection step is to drive the arithmetic processor to compare the predicted minimum quantization error value with the optimal early warning minimum quantization error threshold value group to obtain a comparison result, and detect whether the tool is damaged according to the comparison result.

Thereby, the turning tool chipping detection method of the present invention utilizes artificial intelligence to detect the tool in the cutting state in the turning process, which can more accurately predict that the tool is about to collapse, and then stop the machine immediately to replace the tool, which is not only convenient for the operator to use, but also It can ensure that the workpiece in processing will not be damaged due to tool breakage, so it can solve the problem caused by the different standards of manual judgment in the conventional technology.

Other examples of the foregoing embodiments are as follows: the foregoing preprocessing steps may include a first conversion step and a second conversion step. The first converting step is converting the first time-domain vibration signal into a frequency-domain vibration signal according to Fourier transform. The second conversion step is to divide the vibration signal in the frequency domain into complex segment spectrum signals, and convert the segment spectrum signals into the first frequency band energies according to root-mean-square conversion.

Other examples of the above-mentioned embodiment are as follows: in the above-mentioned second conversion step, the number of these segment spectrum signals can be 13, the spectrum range of at least one of these segment spectrum signals is 1000 Hz, these segments The spectrum signal has a plurality of feature points, and the number of these feature points is 39. In the model building step, the operation processor adjusts the self-organizing map network algorithm to find the winning neuron, and then calculates these first frequency band energies according to the winning neuron and the minimum quantization error method to establish a prediction model and obtain the best Early warning minimum quantization error threshold group.

Other examples of the above-mentioned embodiment are as follows: the above-mentioned turning tool chipping detection method may further include a warning signal generating step and a warning step. The step of generating the warning signal is to drive the arithmetic processor to decide whether to generate the warning signal according to the comparison result. The warning step is to drive the programmable logic controller of the turning-milling compound machine to control the warning light of the turning-milling compound machine according to the early warning signal. The optimal early warning minimum quantization error threshold group includes a first early warning threshold and a second early warning threshold, and the first early warning threshold is smaller than the second early warning threshold. When the comparison result is that the predicted minimum quantization error value is less than or equal to the first warning threshold, the arithmetic processor decides not to generate a warning signal, and the warning light turns green. When the comparison result is that the predicted minimum quantization error value is greater than the first warning threshold and less than or equal to the second warning threshold, the arithmetic processor determines to generate a warning signal and transmits the warning signal to the programmable logic controller, and the warning light turns yellow. . When the comparison result is that the predicted minimum quantization error value is greater than the second warning threshold, the arithmetic processor determines to generate a warning signal and transmits the warning signal to the programmable logic controller, the warning light turns red and the cutting tool stops.

Other examples of the above-mentioned embodiment are as follows: the above-mentioned turning tool chipping detection method may further include an interface display step, which is to drive the graphical user interface to display the total amount of energy in the second frequency band in the first area, and to display the total amount of energy in the second frequency band in the second area. area shows the predicted minimum quantization error value. The second area includes a first sub-area, a second sub-area and a third sub-area, and the colors of the first sub-area, the second sub-area and the third sub-area are green, yellow and red respectively. When the comparison result is that the predicted minimum quantization error value is less than or equal to the first warning threshold, the predicted minimum quantization error value is displayed in the first sub-area. When the comparison result is that the predicted minimum quantization error value is greater than the first warning threshold and less than or equal to the second warning threshold, the predicted minimum quantization error value is displayed in the second sub-area. When the comparison result is that the predicted minimum quantization error value is greater than the second warning threshold, the predicted minimum quantization error value is displayed in the third sub-area.

Several embodiments of the present invention will be described below with reference to the drawings. For the sake of clarity, many practical details are included in the following narrative. It should be understood, however, that these practical details should not be used to limit the invention. That is, in some embodiments of the present invention, these practical details are unnecessary. In addition, for the sake of simplifying the drawings, some commonly used structures and elements will be shown in a simple and schematic way in the drawings; and repeated elements may be denoted by the same reference numerals.

In addition, when a certain element (or unit or module, etc.) is "connected" to another element herein, it may mean that the element is directly connected to another element, or it may mean that a certain element is indirectly connected to another element , that is, there are other elements interposed between the element and another element. And when it is stated that an element is "directly connected" to another element, it means that there is no other element interposed between the element and another element. The terms first, second, third, etc. are used to describe different components, and have no limitation on the components themselves. Therefore, the first component can also be called the second component. Moreover, the combination of components/units/circuits in this article is not a combination that is generally known, conventional or conventional in this field. Whether the components/units/circuits themselves are known or not can be used to determine whether the combination relationship is easily recognized by those in the technical field. Usually knowledgeable people do it easily.

Please refer to FIG. 1 and FIG. 2 together, wherein FIG. 1 is a schematic diagram of a turning tool chipping detection system 100 according to the first embodiment of the present invention; and FIG. 2 is a schematic diagram of a second implementation of the present invention A schematic flowchart of a method 700 for detecting tool chipping in turning. The tool chipping detection system 100 is configured to implement the tool chipping detection method 700 and includes a milling machine 200 , a three-axis accelerometer 300 , a computing processor 400 , a GUI 500 and a data acquisition card 600 .

The turn-mill machine 200 includes a cutting tool 210 , a programmable logic controller (PLC) 220 and a warning light 230 . The tool 210 is used for turning and milling workpieces. The programmable logic controller 220 is electrically connected to the arithmetic processor 400 and the warning light 230 , and the programmable logic controller 220 combines the warning lights generated by the warning light 230 to monitor the current processing status of the machine. The three-axis accelerometer 300 is installed on the turning-milling machine 200 and is electrically connected to the data acquisition card 600. The three-axis accelerometer 300 is used to measure the tool 210 and is captured by the data acquisition card 600 to obtain the first time-domain vibration signal. and the second time domain vibration signal. Any one of the first time-domain vibration signal and the second time-domain vibration signal includes an X-axis signal, a Y-axis signal, and a Z-axis signal. The computing processor 400 is electrically connected to the three-axis accelerometer 300 through the data acquisition card 600, and receives the first time-domain vibration signal and the second time-domain vibration signal. The computing processor 400 is used for computing and analyzing various data. The graphical user interface 500 is electrically connected to the computing processor 400 and used for visualizing data. In this way, the turning tool chipping detection system 100 of the present invention uses artificial intelligence to detect the cutting tool 210 in the cutting state in the turning process, and can more accurately predict that the tool 210 is about to collapse, and then stop the machine immediately to replace the tool 210, which not only facilitates the operation The use of personnel can also ensure that the workpiece in processing will not be damaged due to the collapse of the tool 210, so it can solve the problem caused by the different manual judgment standards in the prior art.

In one embodiment, the turning-milling compound machine 200 can be a numerical control (Computer Numerical Control; CNC) turning-milling compound machine; the arithmetic processor 400 can be an industrial computer; the workpiece can be alloy steel SCM440 before quenching and tempering, but the present invention This is not the limit.

As can be seen from FIG. 1 and FIG. 2 , the method 700 for detecting tool chipping in turning includes a vibration measurement step S02 , a preprocessing step S04 , a model building step S06 , a vibration prediction step S08 and a tool chipping detection step S10 . The vibration measurement step S02 is to drive the three-axis accelerometer 300 to measure the tool 210 of the turning-milling machine 200 to obtain a first time-domain vibration signal. The preprocessing step S04 is to drive the arithmetic processor 400 to convert the first time-domain vibration signal into a complex number of first frequency band energies. The model building step S06 is to drive the operation processor 400 to calculate the energy of the first frequency band according to the self-organizing map network algorithm (Self-Organizing Map; SOM) and the minimum quantization error method (Minimum Quantization Error; MQE) to establish a prediction model, And obtain the best early warning minimum quantization error threshold group. The vibration prediction step S08 is to drive the three-axis accelerometer 300 to measure the tool 210 to obtain the second time-domain vibration signal, and drive the arithmetic processor 400 to convert the second time-domain vibration signal into complex second frequency band energy, and then drive the arithmetic processor 400 Input the energy of the second frequency band into the prediction model to obtain the predicted minimum quantization error value. The chipping detection step S10 is to drive the arithmetic processor 400 to compare the predicted minimum quantization error value with the best early warning minimum quantization error threshold to obtain a comparison result, and to detect whether the tool 210 is damaged according to the comparison result. In this way, the method 700 for detecting chipping of the turning tool of the present invention uses artificial intelligence (such as SOM and MQE) to detect the tool 210 in the cutting state during the turning process, and can more accurately predict that the tool 210 is about to collapse, and then stop the machine immediately to replace the tool 210 is not only convenient for operators to use, but also ensures that the workpiece in processing will not be damaged due to the collapse of the tool 210, so it can solve the problem caused by the different manual judgment standards in the prior art. The following is a detailed embodiment to illustrate the details of the above steps.

Please refer to Figures 1 to 5 together, in which Figure 3 is a schematic diagram of processing step S04 prior to the detection method 700 of turning tool chipping in Figure 2; Figure 4 is a schematic diagram of turning tool chipping in Figure 2 The interface display step S16 of the detection method 700 displays the relationship between the total energy Acc-X, Acc-Y, Acc-Z and time of the three-axis frequency band energy of the second frequency band energy; and Figure 5 shows the turning of Figure 2 The interface display step S16 of the chipping detection method 700 displays a relationship graph of the predicted minimum quantization error value and time. The preprocessing step S04 includes a first converting step S042 and a second converting step S044. The first converting step S042 is converting the first time-domain vibration signal 110 into a frequency-domain vibration signal 120 according to a Fourier transform. The second conversion step S044 is to divide the frequency-domain vibration signal 120 into complex segment spectrum signals, and convert the segment spectrum signals into the first frequency band energies 130 according to Root Mean Square (RMS). In the second converting step S044 , the number of these segmental spectrum signals is 13, and a spectrum range of at least one of these segmental spectral signals is 1000 Hz. These segment spectrum signals have a plurality of feature points, and the number of these feature points is 39. In the model building step S06, the computing processor 400 adjusts the self-organizing map network algorithm to find a winning neuron (Best Matching Unit; BMU), and then calculates the energy of the first frequency band according to the winning neuron and the minimum quantization error method 130 to establish a prediction model, and obtain the best early warning minimum quantization error threshold group. In one embodiment, both the first time-domain vibration signal 110 and the second time-domain vibration signal collect 76800 sampling points per second, and the amplitude (g) of the first time-domain vibration signal 110 ranges from -15 g to 15 g The Fourier transform can be Fast Fourier Transform (FFT); the frequency (Hz) range of the frequency domain vibration signal 120 can be 0 Hz to 12800 Hz, and its intensity (g-rms) can be between 0 g-rms to between 0.15 g-rms; the spectrum range of the first twelve segments of these segment spectrum signals is 1000 Hz, and the spectrum range of the last segment (12000 Hz to 12800 Hz) is 800 Hz, but the present invention does not This is the limit. The first frequency band energy 130 is represented by frequency band (Hz) and energy (mg-rms). It is also worth mentioning that the present invention reduces the number of feature points from the original 38400 to 39 through the second conversion step S044, which can achieve the effect of dimensionality reduction.

As can be seen from FIG. 2 , the method 700 for detecting tool chipping in turning further includes an early warning signal generating step S12 , an alerting step S14 and an interface displaying step S16 . The early warning signal generating step S12 is to drive the arithmetic processor 400 to determine whether to generate an early warning signal according to the comparison result. The warning step S14 is to drive the programmable logic controller 220 of the turning-milling machine 200 to control the warning light 230 of the turning-milling machine 200 according to the warning signal. The optimal early warning minimum quantization error threshold group includes a first early warning threshold MQEV1 and a second early warning threshold MQEV2 , and the first early warning threshold MQEV1 is smaller than the second early warning threshold MQEV2 . When the comparison result shows that the predicted minimum quantization error value is less than or equal to the first warning threshold MQEV1 , the arithmetic processor 400 decides not to generate a warning signal, and the warning light 230 turns green. When the comparison result is that the predicted minimum quantization error value is greater than the first warning threshold MQEV1 and less than or equal to the second warning threshold MQEV2, the arithmetic processor 400 determines to generate a warning signal and transmits the warning signal to the programmable logic controller 220, The warning light 230 lights up yellow. When the comparison result shows that the predicted minimum quantization error value is greater than the second warning threshold MQEV2, the arithmetic processor 400 determines to generate a warning signal and transmits the warning signal to the programmable logic controller 220, the warning light 230 lights up in red and stops the cutting tool. 210 actions. Furthermore, the interface display step S16 is to drive the graphical user interface 500 to display the total frequency band energy Acc-X, Acc-Y, and Acc-Z of these second frequency band energies in the first area 510 on the three axes respectively, and in the first area 510 The second field 520 displays the predicted minimum quantization error value. "Acc-X", "Acc-Y", and "Acc-Z" respectively represent the total energy of the X-axis frequency band, the total energy of the Y-axis frequency band, and the total energy of the Z-axis frequency band. The second region 520 includes a first subregion 522, a second subregion 524 and a third subregion 526, the colors of the first subregion 522, the second subregion 524 and the third subregion 526 are green, yellow and red respectively, And represent normal (Good), warning (Worn) and stop (Bad) respectively. When the comparison result is that the predicted minimum quantization error value is less than or equal to the first warning threshold MQEV1, the predicted minimum quantization error value is displayed in the first sub-area 522; when the comparison result is that the predicted minimum quantization error value is greater than the first warning threshold MQEV1 And when it is less than or equal to the second warning threshold MQEV2, the predicted minimum quantization error value is displayed in the second sub-area 524; when the comparison result is that the predicted minimum quantization error value is greater than the second warning threshold MQEV2, the predicted minimum quantization error value is displayed in The third sub-region 526 . In this way, the method 700 for detecting tool chipping in turning of the present invention utilizes the early warning signal generation step S12, the warning step S14 and the interface display step S16 to facilitate the user to monitor the predicted minimum quantization error value of the current machining state of the machine by visualizing the data, It can not only ensure that the state of the tool 210 is controllable, but also ensure that when the three-axis accelerometer 300 is abnormal (damaged), it can be found immediately without affecting the quality of the workpiece.

In other embodiments, the computing processor 400 can be electrically connected to a network card to transmit data to the network or the cloud, thereby realizing the Internet of Things or big data analysis.

It can be seen from the above embodiments that the present invention has the following advantages: First, the use of artificial intelligence to detect the cutting state of the cutting tool in the turning process can accurately predict that the cutting tool is about to collapse, and then stop the machine immediately to replace the cutting tool, which is not only convenient for the operator Use can also ensure that the workpiece in processing will not be damaged due to tool breakage, so it can solve the problem caused by the different manual judgment standards in the prior art. Second, the number of feature points can be reduced from the original 38,400 to 39 through the second conversion step, thereby achieving the effect of dimensionality reduction. Third, using warning lights, programmable logic controllers, graphical user interfaces, early warning signal generation steps, warning steps, and interface display steps to facilitate users to monitor the current processing status of the machine through visual presentation of the predicted minimum quantization error The value can not only ensure the controllable state of the tool, but also ensure that the abnormality (damage) of the three-axis accelerometer can be found immediately without affecting the quality of the workpiece.

Although the present invention has been disclosed above in terms of implementation, it is not intended to limit the present invention. Anyone skilled in this art can make various changes and modifications without departing from the spirit and scope of the present invention. Therefore, the protection of the present invention The scope shall be defined by the appended patent application scope.

100: Turning tool chipping detection system 110: The first time domain vibration signal 120: frequency domain vibration signal 130: First frequency band energy 200: Turning and milling compound machine 210: Knife 220: Programmable logic controller 230: warning lights 300: Three-axis accelerometer 400: arithmetic processor 500: Graphical User Interface 510: the first area 520: second area 522: The first sub-area 524: Second sub-area 526: The third sub-area 600: data acquisition card 700: Detection method of tool chipping in turning S02: Vibration measurement steps S04: Preprocessing steps S042: First conversion step S044: Second conversion step S06: Model building steps S08: Vibration prediction step S10: Steps for detection of knife collapse S12: Early warning signal generation step S14: Warning step S16: Interface display steps Acc-X, Acc-Y, Acc-Z: the total amount of frequency band energy FFT: Fast Fourier Transform MQEV1: The first early warning threshold MQEV2: The second warning threshold RMS: root mean square conversion

Fig. 1 is a schematic diagram of a turning tool chipping detection system according to the first embodiment of the present invention; Figure 2 is a schematic flow chart showing the method for detecting chipping in turning according to the second embodiment of the present invention; Fig. 3 is a schematic diagram showing the processing steps before the method for detecting chipping in turning in Fig. 2; Fig. 4 is a diagram showing the relationship between the total amount of triaxial frequency band energy and time of the second frequency band energy in the interface display step of the turning tool chipping detection method in Fig. 2; and Fig. 5 is a graph showing the relationship between the predicted minimum quantization error value and time in the interface display steps of the turning tool chipping detection method in Fig. 2.

100: Turning tool chipping detection system

200: Turning and milling compound machine

210: Knife

220: Programmable logic controller

230: warning lights

300: Three-axis accelerometer

400: arithmetic processor

500: Graphical User Interface

600: data acquisition card

Claims (9)

A tool chipping detection system for turning, comprising: a turning-milling compound machine including a tool; a three-axis accelerometer set on the turning-milling compound machine and used to measure the tool to obtain a first time-domain vibration signal and a second time-domain vibration signal; and an arithmetic processor electrically connected to the three-axis accelerometer, the arithmetic processor receiving the first time-domain vibration signal and the second time-domain vibration signal and configured to implement the following steps: The operation of the steps: a preprocessing step, which is to convert the first time domain vibration signal into complex first frequency band energy; a model building step, which is to calculate these first frequency bands according to a self-organizing map network algorithm and a minimum quantization error method A frequency band energy is used to establish a prediction model, and obtain an optimal early warning minimum quantization error threshold group; a vibration prediction step is to convert the second time domain vibration signal into a complex number of second frequency band energies, and then use these first The energy of the two frequency bands is input to the prediction model to obtain a predicted minimum quantization error value; and a knife collapse detection step is to compare the predicted minimum quantization error value with the best early warning minimum quantization error threshold value group to obtain a comparison result, and according to the comparison result, it is detected whether the tool is damaged; wherein, the preprocessing step includes: a first conversion step, which is to convert the first time-domain vibration signal into a frequency-domain vibration signal according to a Fourier transform; and a second conversion step, which is to divide the frequency-domain vibration signal into complex segment spectrum signals, and convert these segment spectrum signals according to a root-mean-square conversion Change to these first frequency band energies. The turning tool chipping detection system as described in claim 1, wherein, in the second conversion step, the number of these segmental spectrum signals is 13, and a spectrum range of at least one of these segmental spectral signals is 1000Hz, the spectral signals of these sections have plural feature points, and the number of these feature points is 39; and in the model building step, the arithmetic processor adjusts the self-organizing map network algorithm to find a winning neuron , and then calculate the first frequency band energies according to the winning neuron and the minimum quantization error method to establish the prediction model, and obtain the optimal early warning minimum quantization error threshold set. The turning chipping tool detection system as described in claim 1, wherein the turning-milling compound machine further includes: a warning light; and a programmable logic controller electrically connected to the operation processor and the warning light, the calculation The processor determines whether to generate an early warning signal according to the comparison result, and the programmable logic controller controls the warning light according to the early warning signal; wherein, the optimal early warning minimum quantization error threshold set includes a first early warning threshold and A second early warning threshold, the first early warning threshold is smaller than the second early warning threshold; wherein, when the comparison result is that the predicted minimum quantization error value is less than or equal to the first early warning threshold, the arithmetic processor determines Do not generate this warning message , the warning light turns green; wherein, when the comparison result shows that the predicted minimum quantization error value is greater than the first warning threshold and less than or equal to the second warning threshold, the arithmetic processor decides to generate the warning signal and The warning signal is sent to the programmable logic controller, and the warning light turns yellow; wherein, when the comparison result shows that the predicted minimum quantization error value is greater than the second warning threshold, the arithmetic processor decides to generate the warning signal and send the warning signal to the programmable logic controller, the warning light turns red and stops the tool from moving. The turning tool chipping detection system as described in claim 3 further includes: a graphical user interface electrically connected to the arithmetic processor and receiving the second frequency band energies and the predicted minimum quantization error value, the graphical user interface The interface displays the total amount of energy in the second frequency bands in a first area, and displays the predicted minimum quantization error value in a second area; wherein, the second area includes a first sub-area, a second sub-area and A third sub-region, the colors of the first sub-region, the second sub-region and the third sub-region are green, yellow and red respectively; wherein, when the comparison result is that the predicted minimum quantization error value is less than or equal to the When the first warning threshold value is reached, the predicted minimum quantization error value is displayed in the first sub-area; wherein, when the comparison result is that the predicted minimum quantization error value is greater than the first warning threshold value and less than or equal to the second warning threshold value value, the predicted minimum amount The quantization error value is displayed in the second sub-area; wherein, when the comparison result is that the predicted minimum quantization error value is greater than the second warning threshold, the predicted minimum quantization error value is displayed in the third sub-area. A method for detecting tool chipping in turning, comprising the following steps: a vibration measurement step of driving a three-axis accelerometer to measure a tool of a turning-milling compound machine to obtain a first time-domain vibration signal; a preprocessing step, It is to drive an operation processor to convert the first time-domain vibration signal into complex first frequency band energy; a model building step is to drive the operation processor to calculate these according to a self-organizing map network algorithm and a minimum quantization error method The energy of the first frequency band is used to establish a prediction model, and obtain an optimal early warning minimum quantization error threshold group; a vibration prediction step is to drive the three-axis accelerometer to measure the tool to obtain a second time-domain vibration signal, and driving the arithmetic processor to convert the second time-domain vibration signal into complex second frequency band energies, and then driving the arithmetic processor to input the second frequency band energies into the prediction model to obtain a predicted minimum quantization error value; and A tool breaking detection step is to drive the arithmetic processor to compare the predicted minimum quantization error value with the optimal early warning minimum quantization error threshold group to obtain a comparison result, and detect whether the tool is based on the comparison result damage. The method for detecting chipping in turning as described in Claim 5, wherein The preprocessing step includes: a first conversion step, converting the first time domain vibration signal into a frequency domain vibration signal according to a Fourier transform; and a second conversion step, dividing the frequency domain vibration signal into complex regions segment spectrum signals, and convert the segment spectrum signals into the energy of the first frequency bands according to the root-mean-square conversion. The method for detecting chipping in turning as described in Claim 6, wherein, in the second conversion step, the number of these segmental spectrum signals is 13, and a spectrum range of at least one of these segmental spectral signals is 1000Hz, the spectral signals of these sections have plural feature points, and the number of these feature points is 39; and in the model building step, the arithmetic processor adjusts the self-organizing map network algorithm to find a winning neuron , and then calculate the first frequency band energies according to the winning neuron and the minimum quantization error method to establish the prediction model, and obtain the optimal early warning minimum quantization error threshold set. The method for detecting chipping of turning tools as described in claim 5 further includes: an early warning signal generation step, which is to drive the arithmetic processor to determine whether to generate an early warning signal according to the comparison result; and a warning step, which is to drive the car A programmable logic controller of the milling compound machine controls a warning light of the turning milling compound machine according to the early warning signal; wherein, the optimal early warning minimum quantization error threshold group includes a first early warning Threshold and a second early warning threshold, the first early warning threshold is less than the second early warning threshold; wherein, when the comparison result is that the predicted minimum quantization error value is less than or equal to the first early warning threshold, the operation The processor decides not to generate the warning signal, and the warning light turns green; wherein, when the comparison result is that the predicted minimum quantization error value is greater than the first warning threshold and less than or equal to the second warning threshold, the arithmetic processor decide to generate the warning signal and transmit the warning signal to the programmable logic controller, and the warning light turns yellow; wherein, when the comparison result is that the predicted minimum quantization error value is greater than the second warning threshold value, the The arithmetic processor determines to generate the warning signal and transmits the warning signal to the programmable logic controller, the warning light turns red and stops the tool from moving. The method for detecting chipping in turning as described in Claim 8 further includes: an interface display step, which is to drive a graphical user interface to display the total amount of energy in the second frequency bands in a first area, and to display in a second area displays the predicted minimum quantization error value; wherein, the second area includes a first sub-area, a second sub-area and a third sub-area, the first sub-area, the second sub-area and the third sub-area The colors of the areas are green, yellow and red respectively; wherein, when the comparison result is that the predicted minimum quantization error value is less than or equal to the first warning threshold value, the predicted minimum quantization error value is displayed on the first warning threshold value. A sub-area; wherein, when the comparison result is that the predicted minimum quantization error value is greater than the first warning threshold and less than or equal to the second warning threshold, the predicted minimum quantization error value is displayed in the second sub-area; wherein , when the comparison result is that the predicted minimum quantization error value is greater than the second warning threshold, the predicted minimum quantization error value is displayed in the third sub-area.

Images