JP3249111B2 - Tool life sensor - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION [Industrial applications]

The present invention relates to a tool life sensing device, and in particular, amplifies a cutting sound or a grinding sound obtained by removing an environmental sound from a mechanical system, frequency-analyzes the amplified cutting sound or grinding sound, and determines a tool life from a recognition pattern thereof. Is fed back to the tool system.

[Prior art]

Conventionally, as a method of determining the tool life, for example, a method of directly measuring a vibration change due to tool wear, a method of recognizing a pattern of the vibration change and determining the same, and electrically converting a cutting resistance or a grinding resistance. Thus, there are known a method of judging from the change state, a method of indirectly judging from the temperature of the processed portion or the surface accuracy of the processed portion, and the like. Also, recently, for example, JP-A-60-186353,
As seen in JP-A-61-192450, JP-A-2-15954, etc., a solid-state AE sensor processes an elastic wave generated at the time of processing to generate an abnormal signal having a specific amplitude or more. There is disclosed a tool breakage prediction device that detects a failure. Further, in Japanese Unexamined Patent Publication No. 1-277299, a cutting sound of a drill or the like is picked up by an acoustic microphone, a detection signal thereof is analog-amplified by an amplifier, and then the amplified signal is divided into a plurality of frequency bands. Further, there is disclosed an acoustic signal identification device which determines a tool life by extracting an energy component of the natural frequency. However, in this apparatus, the subject to be examined is a voice, an elastic wave (AE), and a fricative sound, and an acoustic signal relating to the characteristic sound is identified. Also, JP-A-61-72310 and JP-A-61-295771
In the detection device disclosed in Japanese Patent Application Laid-Open Publication No. H11-157, it is disclosed that an array sensor is used and Fourier transform is performed in order to eliminate unnecessary information.
This is a light receiving element array such as a CCD, and the latter transmits image information from an optical fiber with time.

[Problems to be solved by the invention]

However, the conventional methods described above have a problem that the tool life cannot be determined in accordance with a complicated change in tool wear, and none of them has been put to practical use. This is mainly because the vibration information is obtained by solid-state transmission and the simple detection signal is analyzed even in the air. Further, the AE sensor according to the first three publications described above has a major drawback in that since solid-state transmission is performed, high-frequency components are absorbed and particle size defects constituting a tool edge cannot be detected. In contrast, Japanese Unexamined Patent Publication No.
In the acoustic signal identification device disclosed in Japanese Patent No. 277299, in the identification of the degree of wear of a drill or the like, the above-described friction sound is extracted as a characteristic sound. Therefore, in the actual cutting sound or grinding sound, besides the friction sound, a batter of a substance other than the lack of the particles constituting the tool edge, the vibration sound of the motor, the vibration sound emitted from the entire machine or a unique part, and the like are superimposed. Therefore, there is a problem that it is difficult to identify the degree of wear of the drill or the like from true information. Further, the conventional device using the array sensor is not provided for frequency analysis of cutting sound or grinding sound, and is completely different from the present invention in its use, purpose, and receiving principle. On the other hand, from the viewpoint of the future of automation, labor saving and speeding up of mechanical devices, highly reliable tool life determining means has become increasingly important, and its development has been demanded. In view of the above, the present invention provides a tool life sensing device with higher reliability so that a tool defect can be predicted in advance.

[Means for Solving the Problems]

The present invention has been made in view of the above points, and a sound collector for sensing a cutting sound or a grinding sound, and at least 10 KHZ to 100
A sound collector having a KHZ high-frequency sensing ability, an amplifier for amplifying the collected cutting sound or grinding sound, a frequency analyzer for analyzing the sound pressure obtained by the amplifier, and a frequency analyzer The present invention has been improved on the premise of a tool life sensing device comprising a centralized controller for processing obtained information. That is, the sound collector is at least 10 KHZ to 100 KHZ.
Has a sensing ability in a high frequency range, and environmental noise from a mechanical system is removed, and the frequency analyzer is connected in parallel with an array sensor adapted to the sensing ability of the sound collector. In addition, the sound pressure obtained from the amplifier is fast Fourier transformed and analyzed to confirm the dominant frequency, check the frequency shift, and calculate the wave packet period. According to the array sensor, it is possible to effectively remove all invalid environmental sounds, and depending on the processing conditions, it becomes possible to select and rank a plurality of predominant functions, and also to determine the wave packet of the effective sound. This is because highly accurate information can be provided for analysis. Further, the centralized controller may be configured such that the artificial intelligence system built in the neuron computer has at least a frequency shift and a modulation of a wave packet period obtained from the frequency analyzer as information corresponding to tool wear from normal wear of the tool. Is selected based on the analysis data, fuzzy control is performed based on the analysis data, the tool life is determined based on the obtained frequency pattern recognition, and the result is fed back to the tool system. The reason for this application is to cope with complicated changes in tool wear, and the present inventors have found that in frequency shift and wave packet cycle modulation, information corresponding to tool loss from tool normal wear exists. Based on this, it was determined that fuzzy control was suitable.

[Action]

The tool life sensing device in the present invention is a sound collector,
A cutting sound or a grinding sound from which the environmental sound from the mechanical system is removed is sensed. In this case, since a specific frequency corresponding to the tool damage exists in the high frequency region of the cutting sound or the grinding sound, the pattern can be easily recognized. The amplifier amplifies the sensed cutting sound or grinding sound. Next, in the frequency analyzer, Fourier transform and analysis are performed by the array sensor to check the dominant frequency, check the frequency, calculate the wave packet period, and the like. Further, the centralized controller is constituted by a neuron computer, in which an artificial intelligence system incorporates at least a frequency shift and a wave packet obtained from the frequency analyzer as information corresponding to a defect from normal wear of the tool. The fuzzy control is performed based on the analysis data on the modulation of the period, the life is determined from the pattern recognition obtained based on the fuzzy control, and the result is fed back to the tool system.

【Example】

Hereinafter, an embodiment of the tool life sensing device of the present invention will be described with reference to the drawings. FIG. 1 is a flowchart showing an outline of the present invention, in which a cutting sound or a grinding sound from a work material (not shown) is collected by a sound collector 1 composed of, for example, a microphone having a sensing ability in a high frequency region, Next, the collected cutting sound or grinding sound is amplified by the amplifier 2 and recorded by the recorder 3. In this case, the sound collector 1
Has a sound collecting ability of at least 10 KHZ to 100 KHZ, and an electrostatic or piezoelectric microphone is preferable.
As the electrostatic method, a method using a Ti alloy film or the like is applied. As the piezoelectric method, PZT ceramic or PVF is used.
Is effective. In addition, it is preferable that the sound collector 1 considers removal of environmental sound from a mechanical system by a blow tube, for example. The amplifier 2 is necessary because the sound collection of the sound collector 1 alone may not be sufficient in some cases.
Is used when it is necessary to keep a record, but the frequency analyzer 4 described later is used for comprehensive sound analysis during preliminary idle rotation, and the background is removed. Further, the amplified cutting sound or grinding sound is subjected to pattern analysis by the frequency analyzer 4, and life determination is determined by the centralized controller 5 based on the analysis. This determination result is fed back to the tool system to continue or stop the machining operation. In addition, the frequency analyzer 4 converts the effective sound obtained by removing the environmental sound such as the machine-specific vibration sound, chip vibration sound, coolant splashing sound, etc. from the complete recording to the sine function by FFT (fast Fourier transform) analysis. Convert. Then, confirmation of the dominant frequency, check of the frequency shift, modulation calculation of the wave packet period, and the like are performed. Therefore, the frequency analyzer 4 has a built-in system for connecting in parallel an array sensor consisting of 10 to 100 input lines having a predetermined frequency sensitivity of 10 KHZ to 100 KHZ, and selecting valid information. According to this array sensor, it is possible to effectively remove all invalid environmental sounds, and depending on the processing conditions, it is possible to select and rank a plurality of predominant functions. In addition, it provides highly accurate information for the analysis of the effective sound wave packet. The centralized controller 5 determines the tool life.
Information processing is performed by fuzzy control based on each frequency analysis result. Therefore, an artificial intelligence system for creating rules for control, simulation, and evaluation is built in. In the conventional vibration sensing pattern recognition method, the analog binary logic method is used, which is different from that in which effective tool life control cannot be performed. This is because the fuzzy theory determination method in which the ambiguity is quantified is optimal for the tool life determination based on the ambiguous characteristic of tool wear. In particular, a neuron computer equipped with a neural network exhibits its true value in a determination that requires learning, perception, and determination based on data obtained under various processing conditions. Further, the result determined by the centralized controller 5 is fed back to the tool system to continue or stop the machining operation. In this case, the tool system means a tool holder or the like attached to a lathe, and the tool edge is replaced within an allowable safety range based on the tool life determined by the centralized controller 5. The lateral flank wear width, which is a reference for the tool edge life, is 1 KHZ to 10 K by various cutting tests and grinding tests.
In most cases within the range of 0 KHZ, it was confirmed that it corresponds to the height and frequency shift of the predominant sound pressure peak appearing within the range of 10 KHZ to 100 KHZ. Therefore, the centralized controller 5 can also issue a command for accurate tool change before cutting edge chipping.

【The invention's effect】

The present invention amplifies the cutting sound or the grinding sound collected by the sound collector 1 as described above, analyzes the frequency, and performs fuzzy control of the centralized controller 5 with built-in artificial intelligence to extend the tool life. The judgment and the judgment result are fed back to the tool system. Therefore, the frequency pattern corresponding to the tool wear is accurately recognized and identified, and the tool life can be reliably determined. As a result, the reliability of the command to the tool system is increased, which is suitable for automation of machinery and labor saving.

[Brief description of the drawings]

FIG. 1 is a flowchart showing an embodiment of a tool life sensing device according to the present invention.

[Explanation of symbols]

 1… Sound collector 2… Amplifier 3… Recorder 4… Frequency analyzer

────────────────────────────────────────────────── ─── Continuation of the front page (56) References JP-A-1-277299 (JP, A) JP-A-2-15954 (JP, A) (58) Fields investigated (Int. Cl. ⁷ , DB name) B23Q 17/09 G01M 13/00 G01M 19/00

Claims (1)

(57) [Claims] A sound collector for sensing a cutting sound or a grinding sound;
An amplifier for amplifying the cutting sound or the grinding sound received by the sound collector, a frequency analyzer for analyzing the sound pressure amplified by the amplifier, and a centralized controller for processing information obtained from the frequency analyzer. In the tool life sensing device, the sound collector has a sensing ability in a high frequency range of at least 10 KHZ to 100 KHZ, and environmental sound from a mechanical system is removed, and the frequency analyzer is An array sensor adapted to the sensitivity of the sound collector is connected in parallel, and the sound pressure obtained from the amplifier is fast Fourier transformed and analyzed to confirm the dominant frequency, check the frequency shift, calculate the wave packet period. The centralized controller is composed of a neuron computer, and the built-in artificial intelligence system responds to normal wear of the tool and loss of the tool. As information, analysis data on at least the frequency shift and the modulation of the wave packet period obtained from the high-frequency analyzer are selected, fuzzy control is performed based on the selected data, and the tool life is determined based on the pattern recognition of the obtained frequency. A tool life sensing device characterized in that the result is fed back to a tool system.