Sort
Technical CommunicationPreface
In the previous articles of the special topic of tool monitoring technology, we introduced the commonly used tool monitoring methods, including power monitoring method, vibration monitoring and acoustic emission monitoring method. In this article, we introduced a relatively small cutting monitoring method -- visual inspection method.
The author selected "Research on Tool Wear Detection System Based on Machine Vision", which proposed a non-contact tool wear detection system based on machine vision, discussed the commonly used lathe tool, and used digital image processing method to measure tool wear with visual aid system, which solved the safety shortage of traditional manual detection. The precision and quality of the product are avoided due to the shape error caused by tool wear.
First, visual performance of tool wear
The main factor leading to tool wear is cutting temperature.
When cutting, due to the cutting temperature force is quite high, resulting in the hardness of the tool reduction, deformation, gradually make the tool lose the cutting ability or make the edge shape change, resulting in cutting size change, so that the actual cutting quantity is smaller than the predetermined cutting quantity, and the processing surface will be all or part of the smooth surface into rough surface, The cutting force increases, the cutting temperature increases, the cutting performance changes, the chip range increases and the power consumption increases. The tool wear mode is mainly friction type, melting type, diffusion type, chemical type and electrolytic type wear.
Tool wear performance can be classified as normal and abnormal.
The end of the service life of traditional cutting tools can be attributed to the gradual increase of tool wear, resulting in tool damage or sudden cutting edge damage.
The damage of the cutting tool will reduce the surface quality of the cutting workpiece, using direct or indirect sensing technology to monitor the surface quality of the cutting tool, can judge whether the cutting tool is coming to the end of its service life. Tool wear performance is shown below.
As can be seen from the figure, the geometric appearance of tool surface wear can be divided into A, B and C areas.
According to the ISO definition of tool life, if the tool belly wear average wear value VB>0.3mm, or maximum wear VBmax> 0.6mm, then the tool life has exceeded the tool life. Dented wear occurs on the upper top surface of the tool. This wear is caused by high temperature during cutting.
Second, the predictability of tool wear
The variation of the width of the wear zone of the general turning tool insert during cutting is shown in the figure below.
The curve can be divided into three sections, in which most of the insert wear is concentrated in the tip part.
Section AB is the initial wear zone, sharp edge will be quickly friction and small wear zone.
The BC section is a uniform wear zone, and the wear width increases approximately linearly and uniformly until the tool is nearly damaged.
The CD section is a rapid wear area, the width and speed of wear increase, because of the increase of tool temperature, this section pretended to be the tool wear sensitive reaction area, at this time, the tool wear area has accounted for a large proportion of the whole area.
Three, the principle of machine vision detection
Based on machine vision, the non-contact tool wear judgment auxiliary system is based on the tool wear to determine the time to replace the tool.
In this system, the image of the assumed tool is taken first. Then it is divided into two stages. The first stage is image processing, and the second stage is abrasion determination. Image processing stage includes image capture drive and control and image processing.
In the phase of determination of wear, the calculation and classification of wear are carried out directly according to the positioning results, and a complete set of auxiliary system is designed.
As long as the outline of the tool is obtained, the LED ring light source is used to take the image of the tool on the simulation workbench. The direction of the scanning image is from top to bottom, from left to right, and the feature range is searched in sequence. The transverse scanning search mode is shown as the figure below.
Four, the determination of tool wear
The actual image of tool wear is divided into normal wear and abnormal wear, as shown in the figure below.
As far as normal wear is concerned, after image processing, it can be clearly seen that the wear area turns black. The paper will search for its characteristics and further calculate the wear amount.
Abnormal wear is often caused by tool breaks and defects, and the area of wear is irregular and cannot be judged by the black area.
When calculating the amount of wear, black pixels are numbered for the tool with normal wear, that is, the total number of black pixels is used as the determination of the amount of wear. For the amount of abnormal wear, the judgment of the tool on the verge of being scrapped is based on the pixel area size after frame selection. If it is larger than the pixel area, the tool is judged to be scrapped. Otherwise, it is judged that the tool is not scrapped. Judging whether the tool is scrapped or not, the results can be used to further modify the operating procedure.
If the tool has reached the standard of scrapping, replace it with a new tool immediately; On the contrary, it will be in accordance with the degree of wear as the basis of the operation process to adjust the size of the cutter feed.
Summary and prospect
Machine vision method is used to further test the turning tool. After filtering image processing, the range of wear is selected to calculate the final amount of wear and wear area, and determine whether the tool is scrapped by the results of the amount of wear and wear area, so as to achieve the purpose of inspection.
The original article only explores "visual detection" theoretically, and summarizes the core technologies in the application and promotion process as follows:
Efficient testing process
According to the conditions of visual inspection, the image environment and light source are set. Due to the reduction of external uncontrollable factors, the tool wear in the image is the same as that in the image. Therefore, the image processing can be analyzed in a specific range to reduce the size of the image and effectively shorten the total processing time of the whole process.
Quantified wear criteria
General normal wear of the tool wear amount is not clear, the use of precision measuring instruments will increase a lot of costs, has been used for a certain time of the tool to do inspection, accurately find the tool wear position, and according to the box selection range results, calculate the amount of pixels.
Stable testing environment
The image taken by the original text was completed in the laboratory, the environment of the shooting was fixed, and it was not easy to have other interference factors. But if it is actually applied to the lathe machine, it should try to overcome the influence caused by shaking, and pay attention to whether the light source is consistent when shooting, to avoid the gray value of image processing is not the same because of the inconsistent light intensity.
Precise monitoring range
In order to make the system constructed in the original text more complete and have more new directions of thinking, when selecting the lens, considering the small range of tool wear applied, the image of the whole tool cannot be clearly seen after fixing the focus with the optimal aperture value, which will cause the parts beyond the tip of the tool cannot be recognized in the image processing, and more comparison should be made in the selection of the lens. It will be of great help to the complete inspection of the tool.
Optimal image Angle
Since pixel size and abrasion area are used as the basis to determine scrap, the image Angle is a direct factor affecting the size of abrasion area. Therefore, in the aspect of image taking Angle testing, more tests can be conducted by experimental design and other methods in the future to obtain the best image taking Angle.