CN115041750A - Circular saw blade state evaluation method applied to automatic gear grinding machine - Google Patents
Circular saw blade state evaluation method applied to automatic gear grinding machine Download PDFInfo
- Publication number
- CN115041750A CN115041750A CN202210812441.6A CN202210812441A CN115041750A CN 115041750 A CN115041750 A CN 115041750A CN 202210812441 A CN202210812441 A CN 202210812441A CN 115041750 A CN115041750 A CN 115041750A
- Authority
- CN
- China
- Prior art keywords
- saw blade
- bulge
- value
- matrix
- circular saw
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Images
Classifications
-
- B—PERFORMING OPERATIONS; TRANSPORTING
- B23—MACHINE TOOLS; METAL-WORKING NOT OTHERWISE PROVIDED FOR
- B23D—PLANING; SLOTTING; SHEARING; BROACHING; SAWING; FILING; SCRAPING; LIKE OPERATIONS FOR WORKING METAL BY REMOVING MATERIAL, NOT OTHERWISE PROVIDED FOR
- B23D63/00—Dressing the tools of sawing machines or sawing devices for use in cutting any kind of material, e.g. in the manufacture of sawing tools
- B23D63/08—Sharpening the cutting edges of saw teeth
- B23D63/12—Sharpening the cutting edges of saw teeth by grinding
- B23D63/14—Sharpening circular saw blades
Abstract
The invention relates to a circular saw blade state evaluation method applied to an automatic gear grinding machine, which comprises six steps of saw blade point location scanning, point location data conversion, scalable matrix establishment, sawtooth wear analysis, model classification training and model use and optimization; the automatic tooth grinding machine is additionally arranged on the automatic tooth grinding machine, can automatically and intelligently complete the state evaluation of the circular saw blade in the tooth punching or tooth repairing process of the circular saw blade, and provides suggestions for a user of the circular saw blade, and the number of cutters can be cut at most before the next gear repairing process in the normal wear state of the saw teeth, so that the use safety of the circular saw blade is ensured, and safety accidents such as tooth cracking, tooth breaking, blade explosion and the like are avoided. The method provided by the invention has the advantages of obvious practical significance, lower economic cost and higher intelligent degree, thereby having better popularization value, being applicable to circular saw blade manufacturers and circular saw blade grinding service providers, and also being applicable to circular saw blade use units.
Description
Technical Field
The invention relates to a circular saw blade state evaluation method, in particular to a circular saw blade state evaluation method applied to an automatic gear grinding machine.
Background
Circular saw blades (simply "saw blades") are common tools used in cutting processes: the cutting process refers to cutting the shapes of materials such as pipes, stones, plates, woods, plastics and the like; the cutting process is a fundamental operation in the manufacturing industry, as is the turning, milling, planing, grinding, clamping, and the like. The circular saw blade comprises central base member and marginal sawtooth two parts, and when carrying out cutting work, the circular saw blade clamping drives the high-speed rotation of circular saw blade through equipment on equipment such as circular sawing machine, pipe cutting machine, finally realizes the cutting to the material by the sawtooth that is located the saw bit edge.
An automatic gear grinding machine is a machine tool device which utilizes a grinding wheel as a grinding tool to grind sawteeth, can fully automatically complete actions such as tool setting, tool feeding, gear grinding, chamfering and the like, and grinding precision does not depend on skill and experience of workers, so that the automatic gear grinding machine gradually replaces a manual gear grinding machine to become mainstream gear grinding equipment in the industry. Automatic gear grinding machines are important equipment accompanying the whole life cycle of a circular saw blade: when the circular saw blade is produced, the last procedure is to grind the outer edge of the circular saw blade into sawteeth, and the process is called as 'tooth punching' and needs to be finished on an automatic tooth grinding machine; every time a circular saw blade performs a certain amount of cutting operation (commonly referred to in the industry as "cutting a plurality of knives"), the saw teeth gradually become "dull", and at the same time, the saw teeth need to be reground to maintain the cutting performance of the circular saw blade, and the process is called "tooth trimming", and is also finished on an automatic tooth grinding machine.
With the awareness of safety in production, more and more manufacturers are aware that after the sharpening of the saw teeth is finished and before the next round of cutting operation is carried out, a "physical examination" needs to be carried out on the circular saw blade, namely, the number of knives which can be cut by the saw blade before the saw teeth become dull in the next round of operation is evaluated. The evaluation is vital because if the saw teeth are still cut continuously after being dulled, the potential safety hazard of production is easily caused, so that accidents such as saw tooth pulling crack, saw tooth broken teeth, matrix pulling crack, saw blade burst and the like occur in the high-speed rotation process of the circular saw blade in the cutting operation, the normal production operation is interrupted if the accidents are light, and the personal safety is threatened even if the accidents are heavy. For a user of a circular saw blade, it is not possible to wait until the appearance of cracks, twists, teeth, etc., which are visible to the naked eye, because it is already at a very dangerous place to wait until these phenomena appear. A circular saw blade user often desires to assess the state of the circular saw blade before these noticeable phenomena occur.
At present, the evaluation of the state of the circular saw blade in the industry mainly depends on the experience of workers, and the state of the circular saw blade is evaluated perceptually by three modes of sensing the integral rigidity of the circular saw blade by slightly breaking the circular saw blade with hands when the circular saw blade is static, sensing the feed feedback force of the circular saw blade when the circular saw blade is cut, sensing the deflection vibration degree of the circular saw blade when the circular saw blade is cut and the like. Along with the improvement of the production process level of the domestic circular saw blade and the improvement of the working rotating speed of the domestic circular saw blade caused by the improvement, the disadvantage of manually evaluating the safety state of the circular saw blade is more and more prominent, and an automatic device for quantitatively and objectively evaluating the state of the circular saw blade is urgently needed in the industry so as to ensure safe and accident-free processing and manufacturing production.
Disclosure of Invention
In order to overcome the defects of the prior art, the invention provides a circular saw blade state evaluation method applied to an automatic gear grinding machine. The invention is based on artificial intelligence algorithm, is additionally arranged on an automatic gear grinding machine, can automatically and intelligently complete the state evaluation of the circular saw blade in the process of tooth punching or tooth trimming of the circular saw blade, and provides suggestions for a user of the circular saw blade: under normal wear conditions, the circular saw blade can cut a certain number of knives in the next round.
The technical scheme adopted by the invention is as follows:
a circular saw blade state evaluation method applied to an automatic gear grinding machine comprises the following steps:
(1) scanning of saw blade point locations
Scanning from the outer edge of a central fixing hole of the saw blade to the lowest point of the sawtooth gullet along a radius line connecting the center point of the saw blade and the lowest point of the sawtooth gullet by using a laser sensor, wherein the scanning step distance ranges from one third to two thirds of the thickness of the sawtooth gullet; according to the method, each sawtooth is scanned for one line, the scanning line number is equal to the sawtooth number, and the point position data matrix shown as follows is obtained:
in the above equation, Lij represents the distance between the laser sensor and the saw blade scanned by the laser sensor, where: the index j indicates that this is the jth line of the scan, and the index i indicates that this is the ith point of the scan;
(2) point location data conversion
Firstly, subtracting the mounting distance between the laser sensor and the outer edge of the central fixed hole of the saw blade from the data obtained by scanning the laser sensor, and converting to obtain the actual size data of the saw blade:
in the above formula, Hij represents a scanned value of the saw blade at the ith point of the jth line obtained by scanning, and Hij is equal to Lij-h, where h represents a mounting distance between the laser sensor and the outer edge of the central fixing hole of the saw blade;
and then subtracting the actual size data of the saw blade from the theoretical curve data of the profile of the saw blade to obtain a saw blade size deviation value matrix:
in the above formula, Dij represents a saw blade size deviation value between a saw blade scanning value of the ith point of the jth line and a saw blade theoretical value, Dij is Hij-Fij, wherein Fij represents a design theoretical value of the ith point of the jth line of the circular saw blade;
finally, compensating the size deviation value of the saw blade to obtain a matrix as shown in the specification:
in the above formula, Zij is the size deviation value of the compensated saw blade;
(3) scalable matrix building
Searching local maximum values in the compensated matrix size deviation value matrix, and establishing a scalable matrix with the following form according to the local maximum values and subscripts thereof:
[ saw blade radius, bulge diameter, bulge center position, bulge plus and minus, and relation with previous bulge center position ];
(4) sawtooth wear analysis
Firstly, measuring and calculating a distance value between the addendum point of the saw tooth before cutting and the addendum point of the saw tooth after cutting by using an image superposition mode, representing the wear state of the saw tooth in the cutting of the wheel by using the distance value, and then calculating the ratio of the cutting tool number/the wear state of the saw tooth of the wheel so as to realize the normalization of the cutting tool number and the wear state;
(5) model classification training
According to the number of columns in the obtained extensible row matrix, dividing the saw blade into a plurality of types of unevenness, then using the corresponding extensible row matrix as a training input parameter for each type, and using the ratio of the number of cutting knives to the wear state of the saw blade as a training output parameter, thereby forming a data set to train a support vector machine (SVR) shown as the following:
f(x)＝w T x+b
in the formula, x represents a scalable matrix, f (x) represents a ratio of the number of cutting knives to the wear state, w represents a classification hyperplane normal vector of SVR, b represents an offset value, and w and b are obtained by data set training;
(6) model usage and optimization
Firstly, completing the steps of scanning the point positions of the saw blade, converting point position data, establishing a scalable matrix and training model classification in a tooth punching or tooth trimming stage so as to calculate the predicted value of the number of cutting knives/the abrasion state of the saw blade; secondly, after the saw blade is used, when entering the next gear trimming stage, analyzing the saw tooth wear state according to a saw tooth wear analysis method, and simultaneously requiring a user to input the actual number of cutting knives of the previous gear so as to calculate the actual value of the number of cutting knives/the wear state of the saw blade; then, judging whether the error between the predicted value and the actual value of the cutting tool number/the wear state is within an allowable range, if the error exceeds the allowable range, adding the obtained scalable matrix and the actual value of the cutting tool number/the wear state into an SVR training set, and retraining the SVR model; finally, returning to the initial step to continue a new round of work.
Preferably, the laser sensor only scans the lowest point of the gullet of the saw blade at the outer edge of the saw blade during scanning, i.e. the laser sensor does not scan the saw blade itself.
Preferably, Zij is calculated as follows:
in the above formula, err1 is the maximum measurement deviation of the laser sensor, and is a positive value; err2 is the maximum deflection allowed for the blade, and is positive.
Preferably, in the scalable matrix, the convex protrusions or the concave protrusions connected with the bulged finger saw blades can be obtained by searching local maximum values in the Zij matrix in a searching mode that:
in the Zij matrix, the majority of numbers in the matrix are 0 through searching, but the number in the region from Z31 to Z53 is not 0 continuously, so that the region is a bump; then carrying out bubble sorting on each number in the areas from Z31 to Z53, and finding that the numerical value of Z42 is the largest, so that Z42 is taken as a value of the center position of the bump in the scalable matrix, and then a distance value from Z31 to Z42 is taken as a value of the diameter of the bump in the scalable matrix; in the scalable matrix, positive and negative values of the bulge are determined according to the positive and negative values of the Z42, if the Z42 number is a negative number, the bulge is concave, and if the Z42 number is a positive number, the bulge is convex; in the telescopic matrix, a relation value with the center position of the previous bulge is obtained, if the saw blade has only one bulge, the relation value is 0, and if the saw blade has a plurality of bulges, the relation value represents an arc value between two adjacent bulges; in addition, in the above scalable matrix, the saw blade radius refers to the actual radius value of the saw blade when the laser sensor scans.
Preferably, if there is only one bulge in the blade, then the form of the telescopic matrix is as follows:
[ saw blade radius, bulge diameter, bulge center position, bulge plus or minus ]
If there are two bumps in the saw blade, then the scalable matrix is of the form:
[ saw blade radius, first bulge diameter, first bulge center position, first bulge positive and negative, second bulge diameter, second bulge center position, second bulge positive and negative, and camber value with first bulge ]
If there are three bumps in the saw blade, then the scalable matrix is of the form:
[ saw blade radius, first bulge diameter, first bulge center position, first bulge positive and negative, second bulge diameter, second bulge center position, second bulge positive and negative, camber value with first bulge, third bulge diameter, third bulge center position, third bulge positive and negative, camber value with second bulge ].
Preferably, in the sawtooth wear analysis, after each tooth trimming of the sawtooth is finished, the industrial camera takes a picture of the sawtooth which is sharp and stores the sawtooth, and when the sawtooth is taken for trimming again next time, the industrial camera takes a picture of the sawtooth which is passivated and stores the sawtooth again; then, the two illuminants are overlapped; finally, the distance value between the tooth tips of the saw teeth in the two pictures can be used for representing the abrasion loss of the saw teeth in the cutting of the round.
Preferably, in the model classification training, other artificial intelligence type prediction models including a BP neural network and a random forest can also be adopted.
The invention has the beneficial effects that:
1. the method provided by the invention can change the current situation that the state of the circular saw blade is evaluated only by manual experience in the prior art, thereby greatly improving the intelligent degree of the state evaluation of the circular saw blade, enabling a large number of circular saw blade users lacking relevant experience to evaluate by themselves and avoiding the generation of safety accidents;
2. after the method is popularized and used in a large scale, a large amount of original data of the circular saw blade can be accumulated, valuable quantitative data is provided for the optimization of the aspects of theoretical design, production process, inspection, detection and the like of the circular saw blade, and a data basis is also provided for the basic theory of the circular saw blade;
3. the method provided by the invention has the advantages of obvious practical significance, lower economic cost and good popularization value, and can be widely applied to circular saw blade manufacturers and circular saw blade grinding service providers and can also be applied to circular saw blade use units.
Drawings
FIG. 1 is a flow chart of the present invention model usage;
FIG. 2 is a schematic diagram of an internal blade spot location scanning method of the present invention;
FIG. 3 is a view of different types of circular saw blades corresponding to different telescopic matrices of the present invention;
FIG. 4 is a schematic diagram of a sawtooth wear state analysis method according to the present invention.
Detailed Description
The present invention is further described with reference to the following specific examples, which are not intended to be limiting, but are intended to be exemplary in nature and not to be limiting, and all equivalent modifications and equivalents of the known art that are within the spirit and scope of the present invention are intended to be protected by the present invention.
Referring to fig. 1, the invention provides a method for evaluating the state of a circular saw blade applied to an automatic gear grinding machine, which is used for proposing suggestions to a user of the circular saw blade, and under the normal wear state of the saw teeth, the maximum number of cutters can be cut before the next gear trimming, so that the use safety of the circular saw blade is ensured, and safety accidents such as tooth cracking, tooth breaking, blade burst and the like are avoided. The method requires a controller, a laser sensor (a laser distance sensor or a laser displacement sensor) and an industrial camera to be additionally arranged on an automatic gear grinding machine on the hardware, and a guide rail and a driving motor thereof which can drive the laser distance sensor to carry out parallel scanning on the section of the circular saw blade are additionally arranged. In addition, because the method needs to calculate the SVR, the controller can be a controller with an ARM-Linux architecture or an X86-Linux architecture, such as Raspberry pie 4, Saiyang Industrial control small host, and the like.
The method provided by the invention comprises six steps of saw blade point location scanning, point location data conversion, scalable matrix establishment, sawtooth wear analysis, model classification training, model use and optimization and the like, and the following steps are sequentially explained in detail.
Fig. 2 is a schematic diagram of the method for scanning the saw blade point position in the first step of the present invention, i.e., the step of scanning the saw blade point position. As shown in FIG. 1, a laser sensor is used to scan from the outer edge of the central fixed hole of the saw blade to the lowest point of the sawtooth gully along a radius line connecting the center point of the saw blade and the lowest point of the sawtooth gully. According to the method, each sawtooth is scanned by one line, the scanning line number is equal to the sawtooth number, and the point position data matrix shown as follows is obtained:
in the above equation, Lij represents the distance between the laser sensor and the saw blade scanned by the laser sensor, where: the subscript j indicates that this is the jth line of the scan, and any one line can be taken as the 1 st line; the index i indicates that this is the ith point of the sweep, and the first sweep point of the blade from the center of the line is point 1.
During scanning, the scanning step distance (i.e. the distance value between two scanning points) ranges from one third to two thirds of the thickness of the sawtooth (the thickness of the sawtooth is the distance value between the sawtooth peak and the lowest point of the sawtooth gullet). The determination of the scanning step distance value is a balance value obtained by considering the scanning speed and the scanning quality: if the scanning step distance is too large, unevenness of the circular saw blade may be missed; if the scanning step is too small, the scanning time is too long, and the circular saw blade does not have small point-like irregularities in general.
In the scanning process, laser sensor only scans the sawtooth gullet minimum of saw bit outer fringe, laser sensor does not scan sawtooth itself promptly, this is according to experience and reachs, unevenness degree can not appear in sawtooth itself under the normal conditions, even unevenness degree appears in certain sawtooth, gear grinding machine also can in time discover at the gear repair in-process, then this saw bit will be made useless, consequently, the meaning of scanning sawtooth itself is little, do not scan sawtooth itself simultaneously and can also reduce laser sensor's scanning trajectory control degree of difficulty by a wide margin.
The second major step of the method, namely the point location data conversion step, is to firstly subtract the mounting distance between the laser sensor and the outer edge of the central fixed hole of the saw blade from the data obtained by scanning the laser sensor and convert the data to obtain the actual size data of the saw blade;
in the above formula, Hij represents a scanned value of the saw blade at the ith point of the jth line obtained by scanning, and Hij is equal to Lij-h, where h represents a mounting distance between the laser sensor and the outer edge of the central fixing hole of the saw blade;
and then subtracting the actual size data of the saw blade from the theoretical curve data of the profile of the saw blade to obtain a saw blade size deviation value matrix:
in the above formula, Dij represents the deviation value of the saw blade size between the i-th point of the j-th line and the theoretical value of the saw blade, and Dij is Hij-Fij, wherein Fij represents the theoretical value of the design of the i-th point of the j-th line of the circular saw blade;
finally, compensating the size deviation value of the saw blade to obtain a matrix as shown in the specification:
in the above formula, Zij is a compensated deviation value of the saw blade size, and the calculation method of Zij is shown as the following formula:
in the above formula, err1 is the maximum measurement deviation of the laser sensor, and is a positive value; err2 is the maximum allowable deflection of the saw blade, and is positive, typically on the order of microns. The specific values of err2 will depend on the type of saw blade, for example, a normal high speed steel circular saw blade will have a maximum allowable deviation of 10 wire (+ -0.1 mm, i.e., an err2 value of 0.1 mm), and a high gauge (e.g., for precision cutting) high speed steel circular saw blade will have a maximum allowable deviation of 5 wire (+ -0.05 mm, i.e., an err2 value of 0.05 mm).
In the above equation, Zij is most widely compensated, and theoretically, each term in the Zij matrix of an acceptable circular saw blade should be 0. However, in the actual production and application process of the saw blade, especially after the saw blade passes through several rounds of trimming, the values in the Zij matrix are difficult to be kept all at 0, and at this time, theoretically, a user should replace a new saw blade, but under the normal actual condition, the user still continues to use the original saw blade, so that the method provided by the invention has the actual use value.
The third step of the method, namely the step of establishing the telescopic matrix, is to search a local maximum value in the compensated Zij matrix size deviation value matrix and establish the telescopic matrix with the following form according to the local maximum value and the subscript thereof:
[ saw blade radius, bulge diameter, bulge center position, bulge positive and negative, and relation with previous bulge center position ]
In the above scalable matrix, "bulge" refers to a protrusion or a recess connected to a saw blade, and can be obtained by searching for a local maximum in the Zij matrix, and the search mode is explained by taking the following Zij matrix as an example:
in the Zij matrix, most numbers in the matrix are 0 through searching, but in a region from Z31 to Z53, the number is not 0 continuously, so that the region is a bump. It should be noted that "not 0 continuously" is required to be counted as the same bump. Then, the bubble sorting is performed on each number in the region from Z31 to Z53, and the value of Z42 is found to be the largest, so that Z42 is taken as the value of "bump center position" in the scalable matrix, and then the value of the distance from Z31 to Z42 is taken as the value of "bump diameter" in the scalable matrix. In the scalable matrix, the positive and negative values of the "bulge" are determined according to the positive and negative values of the Z42, if the Z42 number is a negative number, the bulge is concave, and if the Z42 number is a positive number, the bulge is convex. In the above-mentioned scalable matrix, the value "relation with the center position of the previous bump" is 0 if the saw blade has only one bump, and the value represents the camber value between two adjacent bumps if the saw blade has a plurality of bumps. In addition, in the above-mentioned scalable matrix, "the saw blade radius" refers to the actual radius value of the saw blade during the scanning of the laser sensor, because the saw blade will be reduced by one turn after each tooth trimming, the saw blade radius value needs to be considered when evaluating the saw blade state.
The step creates a scalable matrix whose length varies with the type of blade. Fig. 3 shows different types of circular saw blades corresponding to different telescopic matrixes of the invention. As shown in fig. 3a, there is only one bulge in the blade, and the form of the telescopic matrix is as follows:
[ saw blade radius, bulge diameter, bulge center position, bulge plus or minus ]
As shown in fig. 3b and 3c, two bulges are arranged in the saw blade, both bulges are convex in fig. 3b, one bulge is convex and the other bulge is concave in fig. 3c, and the form of the telescopic matrix is as follows:
[ saw blade radius, first bulge diameter, first bulge center position, first bulge positive and negative, second bulge diameter, second bulge center position, second bulge positive and negative, and camber value with first bulge ]
As shown in fig. 3d, there are three internal drums of the saw blade, and the scalable matrix is formed as follows:
[ saw blade radius, first bulge diameter, first bulge center position, first bulge positive and negative, second bulge diameter, second bulge center position, second bulge positive and negative, camber value with first bulge, third bulge diameter, third bulge center position, third bulge positive and negative, camber value with second bulge ]
Generally, a circular saw blade rarely has three or more bulges, but the method provided by the invention is still applicable to the case of a plurality of bulges because of the establishment of a scalable matrix.
The fourth major step of the method of the present invention, namely the step of analyzing the wear of the saw teeth, is to measure and calculate the distance value between the top point of the saw teeth before cutting and the top point of the saw teeth after cutting by using an image superposition mode, characterize the wear state of the saw teeth in the current round of cutting by using the distance value, and then calculate the ratio of the cutting tool number/wear state of the saw teeth in the current round of cutting, so as to realize the normalization of the cutting tool number and the wear state.
FIG. 4 is a schematic diagram of a sawtooth wear state analysis method according to the present invention. In the method, after each tooth trimming of the sawteeth, the industrial camera can photograph and store the sharp sawteeth, and when the next time of taking the sawteeth for trimming the teeth again, the industrial camera can photograph and store the passivated sawteeth again; then, the two photographs are superimposed as shown in fig. 4; because the position of the industrial camera is fixed, and the positions of the sawteeth and the gear trimming grinding wheels of the sawteeth are also fixed, the sawteeth photos shot in two times can be quickly superposed; finally, as shown in FIG. 4, the distance between the tips of the saw teeth in the two pictures can be used to characterize the amount of wear of the saw teeth in the present round of cutting.
It should be noted that the normalization is realized by dividing the number of the cutting knives by the wear state, because the usage habits of each saw blade user on the saw blade are different, the saw blade state cannot be analyzed by simply comparing the number of the cutting knives of each round. According to the method, a comparability quantization index is provided for the subsequent support vector machine SVR through the ratio of the number of the cutting knives to the abrasion state of the sawtooth.
The fifth major step of the method of the present invention, i.e., the step of model classification training, is to classify the saw blade into a plurality of types of unevenness according to the number of columns in the scalable row matrix obtained in the above steps, then each type uses the corresponding scalable row matrix as a training input parameter, and uses the ratio of the cutting tool number/wear state of the saw blade as a training output parameter, thereby forming a data set to train the support vector machine SVR shown below:
f(x)＝w T x+b
in the above formula, x represents a scalable matrix, f (x) represents a ratio of "number of cutters/wear state", w represents a classification hyperplane normal vector of SVR, b represents an offset value, and w and b are obtained by data set training.
In the invention, the number of the bulges is different, so that the scalable matrixes with different lengths are obtained, and the different scalable matrixes correspond to different SVR models. However, when the number of the bulges is the same, and parameters such as the diameter of the bulge, the positive and negative (concave and convex) of the bulge, the central position of the bulge, the camber value of the center of the bulge and the like are different, the parameters only correspond to the scalable matrices with the same length, and the same SVR model is trained at the moment. After the SVR model is trained, the ratio of the number of cutting tools to the wear state can be predicted by inputting the scalable matrix, and then the number of the cutting tools of the next round of the user is suggested according to the ratio.
It should be added that, the invention adopts the support vector machine SVR model, which is a prediction algorithm required for reducing the economic cost of the controller and using lower hardware resources to operate. Under the condition of sufficient hardware resources, other artificial intelligence prediction models such as BP neural network and random forest can also replace the SVR model in the support vector machine.
The sixth step of the method, namely the model using and optimizing step, specifically comprises the following steps: firstly, completing the steps of scanning the point position of a saw blade, converting point position data, establishing a scalable matrix, calculating an SVR model and the like in a tooth punching or tooth trimming (gear grinding) stage so as to calculate the predicted value of the cutting tool number/wear state of the saw blade; secondly, after the saw blade is used, when entering the next gear trimming stage, analyzing the wear state of the saw teeth according to the method in the fourth step, and simultaneously requiring the user to input the actual number of cutting tools of the previous gear so as to calculate the actual value of the cutting tool number/wear state of the saw blade; then, judging whether the error between the predicted value and the actual value of the cutting tool number/wear state is within an allowable range, if the error exceeds the allowable range, adding the obtained telescopic matrix and the actual value of the cutting tool number/wear state into an SVR training set, and retraining the SVR model; finally, returning to the initial step to continue a new round of saw blade point location scanning and the like.
The method comprises six steps of saw bit point location scanning, point location data conversion, scalable matrix establishment, sawtooth wear analysis, model classification training and model use and optimization; the invention is additionally arranged on an automatic gear grinding machine (attached to the automatic gear grinding machine of the circular saw blade), can automatically and intelligently complete the state evaluation of the circular saw blade in the process of tooth punching or tooth repairing of the circular saw blade, and provides suggestions for a user of the circular saw blade, and the maximum number of cutters can be cut before the next gear repairing under the normal abrasion state of the saw teeth, thereby ensuring the use safety of the circular saw blade and avoiding the safety accidents of tooth cracking, tooth breaking, blade explosion and the like. The method provided by the invention has the advantages of obvious practical significance, lower economic cost and higher intelligent degree, thereby having better popularization value, being applicable to circular saw blade manufacturers and circular saw blade grinding service providers, and also being applicable to circular saw blade use units.
The above detailed description is intended to illustrate the present invention, not to limit the present invention, and any modifications and changes made within the spirit of the present invention and the scope of the claims fall within the scope of the present invention.
Claims (7)
1. A circular saw blade state evaluation method applied to an automatic gear grinding machine is characterized by comprising the following steps:
(1) scanning of saw blade point locations
Scanning from the outer edge of a central fixing hole of the saw blade to the lowest point of the sawtooth gullet along a radius line connecting the center point of the saw blade and the lowest point of the sawtooth gullet by using a laser sensor, wherein the scanning step distance ranges from one third to two thirds of the thickness of the sawtooth gullet; according to the method, each sawtooth is scanned by one line, the scanning line number is equal to the sawtooth number, and the point position data matrix shown as follows is obtained:
in the above equation, Lij represents the distance between the laser sensor and the saw blade scanned by the laser sensor, where: the index j indicates that this is the jth line of the scan, and the index i indicates that this is the ith point of the scan;
(2) point location data conversion
Firstly, subtracting the mounting distance between the laser sensor and the outer edge of the central fixed hole of the saw blade from the data obtained by scanning the laser sensor, and converting to obtain the actual size data of the saw blade:
in the above formula, Hij represents a scanned value of the saw blade at the ith point of the jth line obtained by scanning, and Hij is equal to Lij-h, where h represents a mounting distance between the laser sensor and the outer edge of the central fixing hole of the saw blade;
and then subtracting the actual size data of the saw blade from the theoretical curve data of the profile of the saw blade to obtain a saw blade size deviation value matrix:
in the above formula, Dij represents a saw blade size deviation value between a saw blade scanning value of the ith point of the jth line and a saw blade theoretical value, Dij is Hij-Fij, wherein Fij represents a design theoretical value of the ith point of the jth line of the circular saw blade; finally, compensating the size deviation value of the saw blade to obtain a matrix as shown in the specification:
in the above formula, Zij is the size deviation value of the compensated saw blade;
(3) scalable matrix building
Searching local maximum values in the compensated matrix size deviation value matrix, and establishing a scalable matrix with the following form according to the local maximum values and subscripts thereof:
[ saw blade radius, bulge diameter, bulge center position, bulge plus and minus, and relation with previous bulge center position ];
(4) sawtooth wear analysis
Firstly, measuring and calculating a distance value between the top point of a saw tooth before cutting and the top point of the saw tooth after cutting by using an image superposition mode, representing the wear state of the saw tooth in the cutting of the current round by using the distance value, and then calculating the ratio of the cutting tool number/the wear state of the saw tooth of the current round so as to realize the normalization of the cutting tool number and the wear state;
(5) model classification training
According to the number of columns in the obtained extensible row matrix, dividing the saw blade into a plurality of types of unevenness, using the corresponding extensible row matrix of each type as a training input parameter, and using the ratio of the number of cutting knives to the wear state of the saw blade as a training output parameter, thereby forming a data set to train the support vector machine SVR shown as follows:
f(x)＝w T x+b
in the formula, x represents a scalable matrix, f (x) represents a ratio of the number of cutting blades to the wear state, w represents a classification hyperplane normal vector of SVR, and b represents an offset value;
(6) model usage and optimization
Firstly, completing the steps of scanning the point positions of the saw blade, converting point position data, establishing a scalable matrix and training model classification in a tooth punching or tooth trimming stage so as to calculate the predicted value of the number of cutting knives/the abrasion state of the saw blade; secondly, after the saw blade is used, when entering the next gear trimming stage, analyzing the saw tooth wear state according to a saw tooth wear analysis method, and simultaneously requiring a user to input the actual number of cutting knives of the previous gear so as to calculate the actual value of the number of cutting knives/the wear state of the saw blade; then, judging whether the error between the predicted value and the actual value of the number of the cutting blades/the wear state is within an allowable range, if the error exceeds the allowable range, adding the obtained scalable matrix and the actual value of the number of the cutting blades/the wear state into an SVR training set, and retraining the SVR model; finally, returning to the initial step to continue a new round of work.
2. The circular saw blade condition evaluation method for an automatic gear grinding machine as set forth in claim 1, wherein: in the scanning process, the laser sensor only scans the lowest point of the sawtooth gullet at the outer edge of the saw blade, namely the laser sensor does not scan the sawtooth.
3. The circular saw blade condition evaluation method for an automatic gear grinding machine as set forth in claim 1, wherein:
zij is calculated as follows:
in the above formula, err1 is the maximum measurement deviation of the laser sensor, and is a positive value; err2 is the maximum deflection allowed for the blade, and is positive.
4. The circular saw blade condition evaluation method for an automatic gear grinding machine as set forth in claim 3, wherein: in the scalable matrix, the bulges connected with the saw blades of the bulge fingers are convex or concave, and the local maximum value can be searched in the Zij matrix, wherein the searching mode is as follows:
in the Zij matrix, the majority of numbers in the matrix are 0 through searching, but the number in the region from Z31 to Z53 is not 0 continuously, so that the region is a bump; then carrying out bubble sorting on each number in the areas from Z31 to Z53, and finding that the numerical value of Z42 is the largest, so that Z42 is taken as a value of the center position of the bump in the scalable matrix, and then a distance value from Z31 to Z42 is taken as a value of the diameter of the bump in the scalable matrix; in the scalable matrix, positive and negative values of the bulge are determined according to the positive and negative values of the Z42, if the Z42 number is a negative number, the bulge is concave, and if the Z42 number is a positive number, the bulge is convex; in the telescopic matrix, the relation value with the center position of the previous bulge is 0 if the saw blade has only one bulge, and the value represents the radian value between two adjacent bulges if the saw blade has a plurality of bulges; in addition, in the above scalable matrix, the saw blade radius refers to the actual radius value of the saw blade when the laser sensor scans.
5. The circular saw blade condition evaluation method for an automatic gear grinding machine as set forth in claim 4, wherein: if there is only one bulge in the blade, the form of the telescopic matrix is as follows:
[ saw blade radius, bulge diameter, bulge center position, bulge plus or minus ]
If there are two bumps in the saw blade, then the scalable matrix is of the form:
[ saw blade radius, first bulge diameter, first bulge center position, first bulge positive and negative, second bulge diameter, second bulge center position, second bulge positive and negative, and camber value between first bulge ]
If there are three bumps in the saw blade, then the scalable matrix is of the form:
[ saw blade radius, first bulge diameter, first bulge center position, first bulge positive and negative, second bulge diameter, second bulge center position, second bulge positive and negative, camber value with first bulge, third bulge diameter, third bulge center position, third bulge positive and negative, camber value with second bulge ].
6. The circular saw blade condition evaluation method for an automatic gear grinding machine as set forth in claim 1, wherein: in the sawtooth wear analysis, after the sawtooth is trimmed, the industrial camera can photograph and store the sharp sawtooth every time, and when the sawtooth is trimmed again next time, the industrial camera can photograph and store the passivated sawtooth again; then, overlapping the two illuminants; finally, the distance value between the tooth tips of the saw teeth in the two pictures can be used for representing the abrasion loss of the saw teeth in the cutting of the round.
7. The circular saw blade condition evaluation method for an automatic gear grinding machine as set forth in claim 1, wherein: in the model classification training, other artificial intelligence prediction models including BP neural network and random forest can also be adopted.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210812441.6A CN115041750A (en) | 2022-07-11 | 2022-07-11 | Circular saw blade state evaluation method applied to automatic gear grinding machine |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210812441.6A CN115041750A (en) | 2022-07-11 | 2022-07-11 | Circular saw blade state evaluation method applied to automatic gear grinding machine |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN115041750A true CN115041750A (en) | 2022-09-13 |
Family
ID=83164756
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202210812441.6A Pending CN115041750A (en) | 2022-07-11 | 2022-07-11 | Circular saw blade state evaluation method applied to automatic gear grinding machine |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN115041750A (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117182195A (en) * | 2023-08-02 | 2023-12-08 | 杭州博野精密工具有限公司 | High-precision hole grinding and tooth repairing integrated machine for saw blade machining |
| CN117808798A (en) * | 2024-02-29 | 2024-04-02 | 山东万利精密机械制造有限公司 | Visual acquisition and analysis method for intelligent manufacturing production data of circular sawing machine |
-
2022
- 2022-07-11 CN CN202210812441.6A patent/CN115041750A/en active Pending
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117182195A (en) * | 2023-08-02 | 2023-12-08 | 杭州博野精密工具有限公司 | High-precision hole grinding and tooth repairing integrated machine for saw blade machining |
| CN117182195B (en) * | 2023-08-02 | 2024-04-12 | 杭州博野精密工具有限公司 | High-precision hole grinding and tooth repairing integrated machine for saw blade machining |
| CN117808798A (en) * | 2024-02-29 | 2024-04-02 | 山东万利精密机械制造有限公司 | Visual acquisition and analysis method for intelligent manufacturing production data of circular sawing machine |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN115041750A (en) | Circular saw blade state evaluation method applied to automatic gear grinding machine | |
| CN112867580B (en) | Method and device for machining a workpiece | |
| Armarego et al. | Computerized end-milling force predictions with cutting models allowing for eccentricity and cutter deflections | |
| CN109063326B (en) | Gear accurate modeling method considering microscopic shape correction and actual machining errors | |
| Yan et al. | Sustainability assessment of machining process based on extension theory and entropy weight approach | |
| Pal et al. | Optimization of quality characteristics parameters in a pulsed metal inert gas welding process using grey-based Taguchi method | |
| JP6005498B2 (en) | Lens processing system, tool change time detection method, and spectacle lens manufacturing method | |
| CN101041198A (en) | Method of for cutting un-orthogonal crossed big connecting pipe hole by the digital controlled fire on the thick cylinder | |
| CN110083967A (en) | A kind of sbrasive belt grinding process parameter optimizing and evaluation index mathematical model modeling method | |
| CN106296679A (en) | A kind of method determining ERW welding quality | |
| CN1598534A (en) | Soft investigating method for overflow grain index of ore grinding system based on case inference | |
| CN101893430A (en) | Processing method of abnormal measured values based on CNC gear measuring center | |
| CN107186287B (en) | A method of it reducing the internal tooth strength gear honing top gem of a girdle-pendant and cuts radial load size | |
| CN110146375B (en) | Method for determining mapping relation between fatigue crack initiation position of part and surface integrity | |
| JP2023507178A (en) | Method and apparatus for determining cutting parameters of a laser cutting machine | |
| Kurukulasuriya et al. | Sustainable machining: Assessment of environmental performance of milling | |
| Brecher et al. | Analysis of abrasive grit cutting for generating gear grinding | |
| JP2006102843A (en) | Optimum machining apparatus and optimum machining method | |
| US7035761B2 (en) | Method for analyzing waviness of a surface | |
| CN113626953A (en) | High-energy-efficiency milling error dynamic distribution characteristic identification method | |
| Kundrak et al. | Experimental study on surface roughness of face milled parts with round insert at various feed rates | |
| CN111408811A (en) | Welding bead root process inspection ruler for X-shaped groove and manufacturing method and using method thereof | |
| CN112036661A (en) | Ceramic cutter reliability prediction method based on random distribution of mechanical properties of cutter | |
| Gindy | Selection of drilling conditions for glass fibre reinforced plastics | |
| CN114264772B (en) | Part surface scratch contrast detection device and manufacturing method |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination |