CN114083358A

CN114083358A - Industrial robot polishing process optimization method

Info

Publication number: CN114083358A
Application number: CN202210058614.XA
Authority: CN
Inventors: 韩旭; 段书用; 李本旺; 陶友瑞; 赵赢
Original assignee: Chongqing Dexin Robot Detection Center Co ltd; Hebei University of Technology
Current assignee: Chongqing Kairui Robot Technology Co ltd; Hebei University of Technology
Priority date: 2022-01-19
Filing date: 2022-01-19
Publication date: 2022-02-25
Anticipated expiration: 2042-01-19
Also published as: CN114083358B

Abstract

The invention provides an industrial robot polishing process optimization method, which comprises the following steps: mounting a polishing robot on a polishing experiment platform; setting a first grinding parameter; driving a polishing robot to polish a workpiece according to a first polishing parameter; acquiring a first processing parameter of the tail end of the robot in the polishing process; acquiring a second processing parameter of the surface of the workpiece after polishing; constructing a prediction model by using the first grinding parameter, the first processing parameter and the second processing parameter; inputting the second polishing parameter into the prediction model to obtain a prediction result; and judging a prediction result, updating the second polishing parameter and driving the robot to polish the workpiece. According to the polishing process optimization method, the prediction model is built, the prediction machining result is analyzed, the optimal machining mode is selected, the quality detection of the surface of the workpiece is accelerated, and the efficiency of automatic production is improved.

Description

Industrial robot polishing process optimization method

Technical Field

The invention belongs to the technical field of polishing processes, and particularly relates to an optimization method of a polishing process of an industrial robot.

Background

With the proposal of industry 4.0, the application of the robot in the production and manufacturing fields is more and more extensive. For example, CN113245988A discloses a grinding adaptive control platform, in which a robot grinds a workpiece. However, the robot realizes multiple degrees of freedom, and adopts a mode of connecting 3-6 joints in series, and the robot is a serial mechanism, so that the rigidity of the robot is insufficient. Vibration can be generated due to insufficient rigidity in the machining process, and the quality of a machined part and the service life of the industrial robot can be influenced by serious vibration. In the field of polishing, an actuator such as a polishing head at the end of a robot is in greater contact with the surface of a workpiece, so that the robot is subjected to a greater load. At present, in order to polish a specific surface roughness, a general empirical formula is adopted, but the method is not suitable for machining of a robot with weak rigidity, and meanwhile, manual operation and a complicated process are needed in the traditional methods for detecting the surface quality such as a needle touch method, a contrast method and a die method. Therefore, it is not suitable for the present automated production.

Disclosure of Invention

In view of the defects or shortcomings in the prior art, the invention aims to provide an industrial robot polishing process optimization method, which improves the polishing process on the existing polishing platform, constructs a prediction model, predicts a processing result, selects an optimal processing mode, accelerates the quality detection of the surface of a workpiece, improves the efficiency of automatic production, and reduces the labor cost.

In order to achieve the above purpose, the embodiment of the invention adopts the following technical scheme:

an industrial robot polishing process optimization method comprises the following steps:

s1, mounting the polishing robot on a polishing experiment platform;

s2, setting first grinding parameters, wherein the first grinding parameters comprise grinding depth, feeding speed and grinding path;

s3, driving the grinding robot to grind the workpiece according to the first grinding parameter;

s4, acquiring a first processing parameter of the tail end of the robot in the grinding process, wherein the first processing parameter is as follows: root mean square value of robot vibration signal data;

s5, obtaining a second processing parameter of the surface of the workpiece after polishing, wherein the second processing parameter is as follows: the number of vibration lines per unit area on the surface of the workpiece;

s6, constructing a prediction model by the first grinding parameter, the first processing parameter and the second processing parameter;

s7, inputting a second polishing parameter into the prediction model to obtain a prediction result, wherein the second polishing parameter comprises a second polishing depth, a second feeding speed and a second polishing path, and the prediction result comprises a first prediction parameter and a second prediction parameter;

s8, determining whether the first prediction parameter does not exceed the first preset value and the second prediction parameter does not exceed the second preset value,

if the first prediction parameter exceeds the first preset value and/or the second prediction parameter exceeds the second preset value, updating the second polishing parameter, and repeating the steps S7-S8;

and if the first prediction parameter does not exceed the first preset value and the second prediction parameter does not exceed the second preset value, driving the robot to polish the workpiece by using the second polishing parameter.

According to the technical scheme provided by the embodiment of the application, the step of setting the first grinding parameter comprises the following steps:

s21, selecting a polishing depth range; selecting a feed speed range; designing a polishing path;

s22, arranging the polishing depth ranges into i polishing depths at equal intervals, arranging the feeding speed ranges into j feeding speeds at equal intervals, and selecting h polishing paths;

s23, sequentially arranging i polishing depths and j feeding speeds in the order from large to small or from small to large, listing all experimental combinations as follows:

wherein D represents the sanding depth, E represents the feed speed, and F represents the sanding path;

s24, selecting an experiment combination sequence with a preset experiment combination number by adopting an orthogonal experiment design method;

and S25, wherein the first grinding parameter is the first experimental combination in the experimental combination sequence.

According to the technical scheme provided by the embodiment of the application, the step of obtaining the first processing parameter comprises the following steps:

s41, measuring the vibration signal value when the robot processes;

s42, acquiring the number of vibration signals generated during robot machining;

s43, solving the root mean square value of the vibration signal to obtain a first processing parameter, wherein the formula is as follows:

；

in the formula:y _rmsroot mean square;nthe number of vibration signals;xis the vibration signal value.

According to the technical scheme provided by the embodiment of the application, the step of obtaining the second processing parameter comprises the following steps:

s51, acquiring an image of the surface of the processed workpiece;

s52, identifying the vibration lines on the surface of the workpiece;

s53, acquiring the number of the identified vibration marks on the surface of the workpiece;

and S54, calculating the number of the vibration lines on the unit area of the surface of the workpiece to obtain a second machining parameter.

According to the technical scheme provided by the embodiment of the application, the step of identifying the vibration lines on the surface of the workpiece comprises the following steps:

s521, acquiring a workpiece surface image, and adjusting the image contrast to obtain a first image;

s522, marking the vibration lines on the first image to obtain a second image;

s523, setting model parameters, wherein the model parameters comprise img-size and Epoch.

S524, constructing a training model according to the set model parameters, the acquired workpiece surface image and the second image;

s525, inputting the third image into the training model;

s526, judging whether the training model identifies the vibration line of the third image;

if not, updating the model parameters, and repeating the steps S524-S526 until the vibration lines are identified by the training model;

and if so, starting the work of automatically detecting the vibration lines.

According to the technical scheme provided by the embodiment of the application, the step of updating the second grinding parameter comprises the following steps:

s81, if the first prediction parameter exceeds the first preset value and the second prediction parameter exceeds the second preset value, then

S811, reducing the second grinding depth;

s812, updating the second polishing path, and selecting the polishing path with less reciprocating times at the fixed processing point;

s813, reducing the second feeding speed and the machining speed of the polishing head;

s82, if the first prediction parameter exceeds the first preset value and the second prediction parameter does not exceed the second preset value, then

S821, reducing the second grinding depth;

s822, updating the second polishing path, and selecting the polishing path with less reciprocating times at the fixed processing point;

s83, if the second prediction parameter exceeds the second preset value and the first prediction parameter does not exceed the first preset value, then

S831, reducing the second feeding speed and the processing speed of the polishing head;

s832, the second polishing route is updated, and a polishing route having a small number of reciprocations at a fixed machining point is selected.

The invention has the following beneficial effects:

the grinding process optimization method considers the influence of three grinding parameters including grinding depth, feeding speed and grinding path on the surface of a workpiece and the intensity of vibration, establishes a prediction model of robot grinding processing based on depth learning, firstly inputs a second grinding parameter into the prediction model before processing, outputs two prediction parameters including a root mean square value of a vibration signal and the number of vibration lines on the surface unit area of the workpiece, and uses the two prediction parameters as evaluation indexes to judge whether the quality of the processed surface and the intensity of vibration reflected by the two prediction parameters exceed preset values or not, and only carries out next processing and manufacturing if the quality and the intensity of vibration exceed the preset values, and if the quality and the intensity of vibration exceed the preset values, the second grinding parameter is updated, so that the workpiece is ground by the best grinding parameter, and the processing efficiency of the robot is effectively improved. And because the grinding robot generally has 3-6 joints connected in series in order to realize multiple degrees of freedom, the grinding robot has weaker rigidity and is easy to vibrate in the processing process, so that vibration lines are generated on the processing surface, and the processing quality of the robot is reduced. In the preprocessing analysis method, the number of the vibration lines on the unit area of the surface of the workpiece is used as one of the evaluation indexes, the workpiece is polished by the polishing parameters with the small number of the vibration lines, meanwhile, an image processing model is set and a training model is established in the process of evaluating the quality of the machined workpiece, the training model can rapidly identify the vibration lines, and compared with the traditional detection modes such as a needle touch method and a contrast method, manual operation is reduced, and the detection efficiency is higher.

Drawings

Other features, objects and advantages of the invention will become more apparent upon reading of the detailed description of non-limiting embodiments made with reference to the following drawings:

FIG. 1 is a flow chart of a method for optimizing a polishing process according to the present application;

FIG. 2 is a flow chart of identifying a chatter mark as described herein;

FIG. 3 is a schematic view of a lateral buffing path according to the present application;

FIG. 4 is a schematic view of a longitudinal sanding path as described herein;

FIG. 5 is a schematic view of a transverse zig-zag sanding path according to the present application;

FIG. 6 is a schematic view of a longitudinal zig-zag sanding path according to the present application;

FIG. 7 is a path A vibration detection view of a lateral buffing path;

FIG. 8 is a path B vibration detection view of a longitudinal buffing path;

FIG. 9 is a diagram showing the results of training models using 4 models.

Detailed Description

The present invention will be described in further detail with reference to the accompanying drawings and examples. It is to be understood that the specific embodiments described herein are merely illustrative of the relevant invention and not restrictive of the invention. It should be noted that, for convenience of description, only the portions related to the present invention are shown in the drawings.

It should be noted that: like reference numbers and letters refer to like items in the following figures, and thus, once an item is defined in one figure, it need not be further defined and explained in subsequent figures.

In the description of the present disclosure, it should be noted that the terms "center", "upper", "lower", "left", "right", "vertical", "horizontal", "inner", "outer", "clockwise", "counterclockwise", "front", "rear", "side", and the like indicate orientations or positional relationships based on the orientations or positional relationships shown in the drawings, or orientations or positional relationships that are conventionally laid out when the disclosed products are used, and are only for convenience of describing and simplifying the present disclosure, but do not indicate or imply that the referred device or element must have a specific orientation, be constructed in a specific orientation, and be operated, and thus should not be construed as limiting the present disclosure. Furthermore, the terms "first," "second," "third," and the like are used solely to distinguish one from another and are not to be construed as indicating or implying relative importance.

Furthermore, the terms "horizontal", "vertical", "overhang" and the like do not imply that the components are required to be absolutely horizontal or overhang, but may be slightly inclined. For example, "horizontal" merely means that the direction is more horizontal than "vertical" and does not mean that the structure must be perfectly horizontal, but may be slightly inclined.

In the description of the present disclosure, it should also be noted that, unless otherwise explicitly specified or limited, the terms "disposed," "mounted," "connected," and "butted" are to be construed broadly, e.g., as being fixedly connected, detachably connected, or integrally connected; can be mechanically or electrically connected; they may be connected directly or indirectly through intervening media, or they may be interconnected between two elements. The specific meaning of the above terms in the present disclosure can be understood in specific instances by those of ordinary skill in the art.

It should be noted that the embodiments and features of the embodiments may be combined with each other without conflict. The present invention will be described in detail below with reference to the embodiments with reference to the attached drawings.

s1, mounting the polishing robot on a polishing experiment platform;

s4, acquiring a first processing parameter of the tail end of the robot in the grinding process, wherein the first processing parameter is as follows: root mean square value of the robot vibration signal;

Wherein:

the polishing robot is installed on a polishing experiment platform. The end of the robot joint is provided with an actuator, and the actuator is generally a polishing head or a turning machine for polishing work and the like.

The first grinding parameter of the robot is directly set through a controller of the robot. The first grinding parameters include grinding depth, feed rate and grinding path.

After the first polishing parameter is set, the polishing head of the polishing robot is driven by the first polishing parameter to polish the workpiece.

The first processing parameter is a root mean square value of the vibration data signal and reflects the vibration intensity of the robot in the polishing processing process. Because the joint of the robot is a serial mechanism, the rigidity is weak, and the robot can vibrate in the polishing process. Generally, the larger the amplitude of the vibration, the stronger the vibration, which affects the machining effect and also affects the life of the robot; the smaller the amplitude is, the less the abrasion to the robot is, and the service life of the robot can be prolonged.

In order to obtain first processing parameter, the experiment platform of polishing still includes vibration sensor, vibration sensor installs on the executor of robot end, vibration sensor is three-dimensional vibration sensor. Therefore, the vibration sensor collects vibration signal data of the actuator in the workpiece polishing process, namely displacement acceleration signal data of the actuator; and stopping collecting after polishing. And solving the root mean square value of vibration signal data to obtain a first processing parameter, wherein the vibration signal data reflects the vibration magnitude of the robot in the processing process and can also be used as an evaluation index of the robot processing.

The second processing parameter is the number of the vibration lines on the surface of the processed workpiece in unit area, and reflects the processing quality of the robot. The robot produces the vibration at the in-process of polishing, causes the machined part surface to produce the chatter mark, can know and evaluate processingquality through the chatter mark quantity of unit area, and generally the chatter mark is less, and processingquality is higher.

In order to obtain the second processing parameter, the polishing experiment platform further comprises an interconnection image acquisition device. The image acquisition device is generally a camera, is positioned above the workpiece, acquires images of the processed workpiece, and transmits the acquired images to the image recognition model. The image recognition model recognizes the number of the vibration lines of the image and calculates the number of the vibration lines in unit area, so as to obtain a second processing parameter.

And constructing a prediction model by using the set first grinding parameter and the first processing parameter and the second processing parameter which are acquired when the robot is driven by the first grinding parameter to process and after the robot is processed as sample data. Theoretically, the more sample data of the prediction model, the more reliable the evaluation index output by the prediction model is, and the more accurate the prediction result is, so that a plurality of first grinding parameters can be set as required, and the prediction model can be constructed by correspondingly obtaining a plurality of first processing parameters and second processing parameters.

And before formal machining, inputting the second polishing parameter into the prediction model to obtain a first prediction parameter and a second prediction parameter of a prediction result. The second grinding parameters comprise a second grinding depth, a second feeding speed and a second grinding path, and the second grinding depth, the second feeding speed and the second grinding path are the grinding depth, the feeding speed and the grinding path for planning to carry out actual grinding on the workpiece; the first prediction parameter is the root mean square value of the predicted vibration data signal, and the second prediction parameter is the number of vibration lines per unit area of the surface of the workpiece to be processed. Only if the two predicted parameters do not exceed the preset values, the polishing work can be started according to the second polishing parameter; if only one predicted parameter exceeds the preset value, updating and adjusting the second polishing parameter, and only starting polishing work according to the updated second polishing parameter until the two predicted parameters predicted by the updated second polishing parameter do not exceed the preset values. The updated second dressing parameter includes an updated second dressing depth, an updated second feed speed, and an updated second dressing path.

Referring to fig. 1, after the polishing process adopts the optimization method described in the present application, the processing quality and the vibration intensity after processing can be known in advance according to the predicted first prediction parameter and the predicted second prediction parameter, the processing effect is evaluated, and then the optimal processing mode is selected, so that the processing efficiency is improved, the loss of the robot is reduced, the service life of the robot is prolonged, and the cost is reduced; the image acquisition device and the image recognition model are adopted to extract and recognize vibration pattern data of the surface of the machined workpiece, compared with the traditional method that the machined workpiece needs to be manually taken down and then the surface of the machined workpiece is detected, after the machining of the robot is completed, vibration patterns are automatically recognized through the image recognition model, the workpiece is clamped and polished through the mechanical arm, the automation is realized in the whole process, the manual operation cost is reduced, the automation degree is improved, and the detection efficiency is also improved.

Further, the step of setting the first grinding parameter comprises:

Wherein:

the polishing depth range is selected according to the effective polishing numerical range, effective polishing can be achieved only by effectively polishing the workpiece above 0.2mm, but the depth cannot exceed 1mm to the maximum extent.

The feed speed range is determined by the maximum speed and the minimum speed which can be reached when the robot processes, and an appropriate feed speed range is selected between the maximum speed and the minimum speed. The grinding speed is too high, the surface of a workpiece is not sufficient, the grinding is too slow, and the temperature of a contact surface is increased, so that the workpiece is deformed. The grinding speed is related to the grinding machine, and is generally 3mm/s in order to ensure the minimum processing efficiency.

The grinding path of design processing, in order to raise the efficiency, the area of contact of the grinding face and the work piece is the biggest when letting the robot grind at every turn as far as possible, specifically can be through the angle of polishing of adjustment grinding head to reduce the number of times of making a round trip of grinding, avoid local repetition to polish.

The selected polishing depth range and the selected feeding speed range are equally divided into i and j data, and the uniform dispersion of the experiment is ensured. Wherein i and j may be the same or different.

Because the polishing experiment is carried out on all the combinations of the three parameters, the number of experimental combinations is huge, a plurality of optimal experimental combination schemes are selected by adopting an orthogonal experiment method, and only the optimal realization combination is carried out, so that the number of experiments is reduced, and the workload is greatly reduced. For example, a three-factor, three-level experiment requires experiments with 3^3=27 combinations as required by the overall experiment, and the number of repetitions of each combination has not been considered. If the experiment is arranged according to the L9(3^4) orthogonal table, only 9 times are needed, and 15 times of experiments are carried out according to the L15(3^7) orthogonal table.

For example, in one experiment, the first grinding parameter to build a predictive model is exemplified by grinding the 3C cell phone case.

The selected polishing depth range is 0.2-0.6 mm; 3 sanding depths, 0.2mm, 0.4mm, 0.6mm, were selected equally spaced from the sanding depth range.

The selected feed speed range is selected according to the percentage of the maximum grinding speed which can be realized by the robot teach pendant, namely a controller, and the 3 feed speeds which are selected at equal intervals are respectively 10%, 30% and 50% of the maximum speed.

A, B, C three grinding paths are designed. Specifically, the sanding path generally has a transverse direction, a longitudinal direction, a transverse zigzag, a longitudinal zigzag, and the like, as shown in fig. 3-6.

The three polishing parameters are labeled, the experimental process is carried out according to an orthogonal experiment table shown in table 1, and the 3-factor orthogonal experiment only needs to be carried out for 9 times.

After 9 groups of experimental combination sequences are obtained, the first experimental combination in the experimental combination sequences is the first grinding parameter, and then the first grinding parameter in the experimental combination sequences is extracted in an iterative manner according to the sequence to drive the grinding robot to grind the workpiece, so that a corresponding first processing parameter and a corresponding second processing parameter are obtained.

Further, the step of obtaining the first processing parameter includes:

s41, measuring the vibration signal value when the robot processes;

；

Wherein:

the vibration signal value, i.e., the displacement acceleration signal value, is measured by a three-way vibration sensor mounted on the robot end effector. The three directions of the three-direction vibration sensor refer to the three directions of xyz after a coordinate system is established by the tail end of the robot. The method comprises the steps of measuring vibration signal data of the tail end of the robot by using a three-way vibration sensor, substituting a root mean square formula after obtaining the quantity of vibration signals, calculating a root mean square value of the root mean square formula to obtain vibration signal data, and expressing the amplitude of vibration of the robot tail end actuator, namely the intensity of vibration by using the root mean square value of the vibration signals.

Specifically, the robot is weak in rigidity and prone to generate vibration, and the vibration affects the service life of the robot. And vibration sensor gathers vibration signal data at the robot in-process of polishing, adopts behind the root mean square, becomes vibration signal's continuous data discrete data, makes things convenient for later stage comparison and judges. The root mean square value of the vibration signal data can effectively evaluate the vibration degree of the robot actuator, and the root mean square value is used as an evaluation standard for evaluating the vibration degree of the vibration actuator, so that the second polishing parameter for reducing the vibration amplitude of the robot is selected, and the service life of the robot can be effectively prolonged.

For example, in a specific embodiment of the present application, data collected by a three-way sensor is processed, and a vibration amplitude signal of an actuator during the working process of the six-degree-of-freedom grinding robot is plotted, wherein fig. 7 is a vibration detection diagram of a path a of a transverse grinding path, and fig. 8 is a vibration detection diagram of a path B of a longitudinal grinding path.

As can be seen from fig. 7 and 8, the vibration of the actuator of the robot along the path a is greater than that of the path B, and the trend of the vibration of the path B changes more smoothly, which is beneficial to the stable operation of the actuator. Fig. 7 and 8 illustrate data as an image, and the data should be actually subjected to root mean square processing.

Further, the step of obtaining the second processing parameter includes the following steps:

s51, acquiring an image of the surface of the processed workpiece;

s52, identifying the vibration lines on the surface of the workpiece;

Wherein:

the image acquisition device positioned above the workpiece acquires an image of the surface of the workpiece, and the image acquisition device sends the acquired image to the image recognition model; the image recognition model recognizes the collected image to obtain the number of the vibration lines on the surface of the workpiece, the number of the vibration lines in unit area is further calculated to obtain a second processing parameter, and the second processing parameter is used as an evaluation index for evaluating the quality of the processed workpiece.

Particularly, in the polishing process of the robot, the robot has insufficient rigidity, so that the phenomenon of vibration can be caused, vibration lines are easy to generate on the surface of a workpiece, and the processing quality is seriously influenced. In order to improve the processing quality, improve the quality detection efficiency of the surface of a workpiece, improve the efficiency of automatic production and reduce the labor cost, the vibration lines are identified in a visual mode and the number of the vibration lines on the surface of the workpiece is extracted quickly.

Specifically, the image acquisition device, namely a camera, is controlled to move to a position to be detected, and the extraction of the vibration lines on the surface of the workpiece is carried out. The quality of the processed surface quality can be effectively judged according to the number of the vibration lines in the unit area, generally, the less the number of the vibration lines is, the higher the processing quality is, and the detection efficiency can be effectively improved by adopting the image recognition model for recognition.

Further, the step of identifying the vibration lines on the surface of the workpiece comprises the following steps:

s522, marking the vibration lines on the first image to obtain a second image;

s525, inputting the third image into the training model;

and if so, starting the work of automatically detecting the vibration lines.

Wherein:

as shown in fig. 2, the image recognition model obtains a flow of the processed workpiece surface image collected by the image collecting device.

The first image is an image after the contrast of the image is adjusted, and the contrast is adjusted to enable the vibration lines of the acquired image to be displayed more clearly and ensure the acquisition effect; the problem that the shot surface image of the workpiece is not clear enough due to light, brightness and the like and influences on the collection quantity and the collection quality is avoided.

And the second image is the marked image, the first image is marked manually by adopting labellmg software, and the vibration lines are marked out and used as the input of the training model, so that the image recognition model can recognize and learn the vibration lines.

And setting model parameters to adjust the speed and accuracy of the recognition of the training model, thereby ensuring the accuracy of the vibration line recognition. The model parameters img-size may be understood as the image resolution, typically 640 dpi. The Epoch represents that the model is completely trained once by using all data of the training set, which is called as 'first generation training', the model is gradually learned through multiple Epoch training, and is gradually stabilized after 150 epochs, and the number of training rounds can also be understood as 300 by default, and needs to be specified. Generally, the higher the image resolution, the more accurate the number of training rounds, but the longer the training time.

And constructing a training model by using the acquired workpiece surface image and the labeled second image in a YOLOv5 neural network deep learning mode, so that the image recognition device can recognize the vibration lines. The YLOLv5 model has four models of s, m, l and x, the processing speed and the processing accuracy are different, and an optimal model needs to be selected for training in model training.

For example, the results after training with 4 models are shown in FIG. 9. As can be seen from fig. 9, the training models of s have the characteristics of small number and fast calculation, and training by using the s model can be considered.

The third image is an image of the surface of the workpiece after actual processing, and whether the training model can recognize the vibration lines or not needs to be tested in order to detect the training result. And inputting a third image, namely the actually processed workpiece surface image, into the training model, and judging whether the training model identifies the vibration lines or not, if the vibration lines cannot be identified correctly, continuing to adjust model parameters and training the model again until the image identification model can automatically identify the vibration lines and then putting the vibration lines into the work of automatically detecting the vibration lines.

Specifically, the number of the vibration lines recognized by the training model accounts for more than 90% of the total number of the vibration lines, or the recognition rate of the vibration lines reaches more than 90%, that is, the vibration lines can be recognized accurately by the training model.

Further, the step of updating the second grinding parameter comprises:

S811, reducing the second grinding depth;

S821, reducing the second grinding depth;

s832, the second polishing route is updated, and the polishing route with a small number of reciprocations at the fixed processing point is selected.

Wherein:

setting: the second grinding parameters that are not updated include a second grinding depth, a second feed speed, and a second grinding path.

(1) If the first prediction parameter exceeds the first preset value, the second prediction parameter exceeds the second preset value.

The second sanding depth is reduced first. When the second sanding depth is reduced, the second sanding depth is reduced in a fixed interval decreasing mode. For example, when the current second grinding depth is 0.4mm, the second grinding depth is decreased by 0.01mm each time. And forming an updated second polishing parameter by the updated second polishing depth, the updated second feeding speed and the updated second polishing path, inputting the updated second polishing parameter into the prediction model to obtain a first prediction parameter and a second prediction parameter, and continuously judging:

and if the first predicted parameter does not exceed the first preset value and the second predicted parameter does not exceed the second preset value, polishing the workpiece by the updated second polishing parameter.

And if the first prediction parameter exceeds the first preset value and the second prediction parameter exceeds the second preset value, continuing to decrease in a decreasing mode at fixed intervals. And forming an updated second polishing parameter by the updated second polishing depth, the updated second feeding speed and the updated second polishing path, inputting the updated second polishing parameter into the prediction model to obtain the first prediction parameter and the second prediction parameter, continuously judging, performing reciprocating operation until the limit value of the polishing depth range is reduced to the minimum value, and updating the second polishing path.

Specifically, updating the second grinding path and optimizing the second grinding path may be performed as follows: the second grinding path is composed of a plurality of track lines, the number of grinding vibration marks of each track line is recorded, the vibration mark with the minimum vibration mark is defined as the optimal track line, each processing point has the corresponding optimal processing track line, and the optimal track lines of all points to be processed are combined when the whole surface is processed to obtain the optimal grinding path.

And forming an updated second polishing parameter by using the minimum value of the first polishing depth, the second feeding speed and the updated second polishing path, inputting the updated second polishing parameter into the prediction model to obtain a first prediction parameter and a second prediction parameter, and continuously judging:

And if the first prediction parameter exceeds the first preset value and the second prediction parameter exceeds the second preset value, reducing the second feeding speed. When the second feed speed is reduced, the feed speed is reduced in a stepwise manner at fixed intervals. For example, when the current feed speed is 50% of the maximum speed, the feed speed is gradually decreased by 10% of the maximum speed each time. And (3) forming a second polishing parameter after updating by using the second polishing depth minimum value, the second feeding speed after updating and the second polishing path after updating, inputting the second polishing parameter after updating into a prediction model to obtain a first prediction parameter and a second prediction parameter, and continuously judging:

and if the first prediction parameter exceeds the first preset value and the second prediction parameter exceeds the second preset value, continuing to reduce the feeding speed until the first prediction parameter does not exceed the first preset value and the second prediction parameter does not exceed the second preset value, and polishing the workpiece by using the updated second polishing parameter.

(2) And if the first prediction parameter exceeds the first preset value and the second prediction parameter does not exceed the second preset value, reducing the second polishing depth at equal intervals, and updating the second polishing path until the first prediction parameter does not exceed the first preset value and the second prediction parameter does not exceed the second preset value, so that the workpiece is polished by the updated second polishing parameter.

(3) And if the second predicted parameter exceeds the second preset value and the first predicted parameter does not exceed the first preset value, reducing the second feeding speed at equal intervals, and updating the second polishing path until the first predicted parameter does not exceed the first preset value and the second predicted parameter does not exceed the second preset value, so that the workpiece is polished by the updated second polishing parameter.

It can be understood that the fixed interval setting cannot be too large, which may cause the updated data to deviate too much and miss the most suitable polishing parameters; too small, the update will be less efficient. It is preferable that: the reduction interval of the sanding depth was 0.02mm, and the reduction interval of the feed rate was 5% of the maximum rate.

The first prediction parameter is the root mean square value of vibration signal data predicted by the prediction model and reflects the vibration intensity, namely the amplitude; the second prediction parameter is the number of the vibration lines of the unit area of the surface of the workpiece predicted by the prediction model, and reflects the processing quality.

(1) When the first prediction parameter exceeds a first preset value and the second prediction parameter exceeds a second preset value, namely the amplitude is too large or the vibration lines are too large, the amplitude is preferentially reduced, namely the second grinding depth is preferentially reduced, the grinding depth of the robot affects the contact force of the robot, and the deeper the grinding depth is, the larger the contact force is, the larger the vibration amplitude is; updating the second grinding path, selecting a grinding path with less reciprocating times at a fixed processing point, reducing the reciprocating times of the grinding path, increasing the reciprocating times to increase the vibration amplitude, grinding one processing point for the second time, and influencing the processing quality due to inconsistent force of the first contact and the second contact; finally, the second feeding speed is reduced, the too high feeding speed affects the vibration stability of the robot, and the higher the speed, the worse the stability, the vibration lines are increased, and the processing quality is affected; too fast a feed rate may also result in incomplete machining of the workpiece.

(2) When the first prediction parameter exceeds a first preset value, the second prediction parameter does not exceed a second preset value, namely the amplitude is too large, and the number of the vibration grains is moderate, the second polishing depth is reduced firstly, the polishing depth of the robot influences the contact force of the robot, and the deeper the polishing depth, the larger the contact force is, the larger the vibration amplitude is; and updating the second polishing path, and selecting the polishing path with less reciprocating times at the fixed processing point, wherein the vibration amplitude is influenced by the increase of the reciprocating times.

(3) When the second prediction parameter exceeds a second preset value, the first prediction parameter does not exceed the preset value, namely the number of the chatter marks is too large, the amplitude is moderate, the second feeding speed, namely the grinding head processing speed is preferentially reduced, because the grinding head moving speed is too high, the vibration stability of the robot is influenced, the higher the speed is, the lower the stability is, the higher the chatter marks are, and the processing quality is influenced; too fast a feed rate may also result in incomplete machining of the workpiece; and secondly, updating the second grinding path to avoid the overlapping of the grinding paths as much as possible, grinding one machining point for the second time, wherein the forces of the first contact and the second contact are inconsistent and the machining quality is influenced.

The foregoing description is only exemplary of the preferred embodiments of the invention and is illustrative of the principles of the technology employed. It will be appreciated by those skilled in the art that the scope of the invention herein disclosed is not limited to the particular combination of features described above, but also encompasses other arrangements formed by any combination of the above features or their equivalents without departing from the spirit of the invention. For example, the above features and (but not limited to) features having similar functions disclosed in the present invention are mutually replaced to form the technical solution.

Claims

1. A grinding process optimization method for an industrial robot is characterized by comprising the following steps:

s1, mounting the polishing robot on a polishing experiment platform;

and if the first predicted parameter does not exceed the first preset value and the second predicted parameter does not exceed the second preset value, driving the robot to polish the workpiece by using the second polishing parameter.

2. A method for optimizing a grinding process of an industrial robot according to claim 1, characterized in that the step of setting a first grinding parameter comprises:

3. An industrial robot sanding process optimization method according to claim 1, characterized in that the obtaining of the first machining parameters comprises the steps of:

s41, measuring the vibration signal value when the robot processes;

；

4. An industrial robot sanding process optimization method according to claim 1, characterized in that the obtaining of the second machining parameters comprises the steps of:

s51, acquiring an image of the surface of the processed workpiece;

s52, identifying the vibration lines on the surface of the workpiece;

5. An industrial robot grinding process optimization method according to claim 4, characterized in that identifying workpiece surface chatter marks comprises the steps of:

s522, marking the vibration lines on the first image to obtain a second image;

s523, setting model parameters, wherein the model parameters comprise img-size and Epoch;

s525, inputting the third image into the training model;

and if so, starting the work of automatically detecting the vibration lines.

6. A method for optimizing a grinding process of an industrial robot according to claim 1, characterized in that the step of updating the second grinding parameter comprises:

S811, reducing the second grinding depth;

S821, reducing the second grinding depth;