Sort
Technical CommunicationFirst, tool monitoring application background
With the improvement of automation and intelligence of production, higher requirements are put forward for on-line monitoring of tool damage.
In the process of cutting, the cutting edge of the cutting tool is not only under high cutting force and cutting temperature, but also because of the hard point contained in the processing material will give impact to the high-speed running tool, these complex conditions are easy to cause damage to the tool in the process of processing.
If the damaged tool in the process of processing is not replaced in time and continue to process will lead to cutting force and cutting temperature rise sharply, the surface roughness and size of the workpiece out of machining, serious even lead to machine damage and endanger the operator. Therefore, on-line monitoring of tool damage is very necessary for mass production of production line.
In recent years, with the increase of production labor cost, manufacturing enterprises have higher and higher requirements for the efficiency of mechanical processing process. This kind of market requirements make more and more machinery manufacturing enterprises will invest more funds in the application field of CNC production line with higher degree of automation and intelligence.
Although these advanced production modes greatly save labor costs and improve processing efficiency, they put forward high requirements for real-time monitoring of process system.
The monitoring of the tool is the most critical in the process system. The state of the tool not only affects the rhythm of the production line, but also relates to the processing efficiency of the production line and the safety of the machine tool and other process systems.
In the traditional machining mode, the state of the tool is generally judged whether the tool is damaged by the observation and experience of the machine tool operator. This judgment method has a strong subjectivity.
With the greatly improved degree of automation and intelligence in the production workshop, most of the production lines in the workshop are managed by few people or even unmanned production lines with higher intelligence. Obviously, the traditional relying on the personal experience of machine tool operators to judge whether the cutting tool needs to be replaced in the cutting process has been unable to meet the current advanced manufacturing mode.
At present, many production lines are statistical tool current processing time or processing pieces as the basis for tool change, and this method as the basis for tool change is often more conservative when setting the number of processing pieces or processing time, it is likely to lead to the waste of tool life, and if the sudden damage of the tool is unpredictable within the tool life.
Two, two types of tool damage forms
The main forms of tool failure are tool wear and tool damage.
In the process of machining tool wear is a gradual change process, as long as the moving tool and workpiece contact will occur tool wear.
Generally, the tool has a blunt standard, the tool in the cutting process does not meet the blunt standard will lose the ability to cut, this phenomenon is tool damage.
According to the time when the tool damage occurs in the cutting process, the tool damage can be divided into early damage and late damage.
At the beginning of cutting, there is no obvious wear on the cutting edge and tool face, but the tool has lost its cutting ability. This phenomenon is called early damage of the tool.
Late tool damage is the loss of cutting ability caused by the tool after a period of cutting.
Another division of tool damage is divided into plastic damage and brittle damage according to the type of damage.
2.1 Plastic damage of the tool
In the process of cutting, the cutting edge and the tool face are under very high temperature and pressure, resulting in very large plastic deformation of the contact parts of the cutting edge and the tool face and the machining surface, and the large plastic deformation of the cutting edge and the tool face exceeds the blunt standard of the tool, the tool will lose the ability of cutting. The tool failure caused by the plastic deformation of the blade and the tool face is the tool plastic failure. The plastic damage of tool is mainly related to the hardness and plasticity of tool material.
2.2 Brittle damage of the tool
The brittle damage of the tool refers to the phenomenon that when the tool made of brittle material is used for cutting, the blade is usually broken, the blade is broken, and the front and back tool face is cracked and spalling. Ceramic, hard alloy and other brittle, high hardness materials made of cutting tools usually brittle damage.
The characteristics of the breaking edge is that there is a small size gap on the cutting edge of the cutting tool, and the size of the gap is not much different from the size of the feed. The breaking edge of the cutting tool of the brittle material usually occurs in the early stage of the cutting process, such as the cutting tool of the ceramic and hard alloy material, often occurs in this early damage phenomenon.
Fracture is usually a damage to the cutting edge of a tool. It is typically characterized by a large break in the cutting edge or a break in the entire cutting edge. Once the tool breaks, the tool is completely scrapped, and the possibility of use after grinding is very small. The fracture of the blade usually occurs in the middle and late cutting process, and often occurs more frequently in intermittent machining.
The fragmentation and fracture of the blade are different, usually there is a large size gap on the cutting edge, the fragmentation of the tool can not continue to cut, but the fragmentation of the blade through grinding can be used for a period of time, fragmentation appears more in the early processing, the form of damage is usually early damage.
Spalling is relative to the tool face, usually refers to the front and back of the tool face in the process of spalling a layer of small debris, this tool face flaking usually reduces the strength of the cutting edge, so that the tool loses the cutting ability.
The spalling of the front and back tool surfaces usually occurs in the early stage of intermittent cutting. The accumulation of chip nodules in the machining process will accelerate the spalling of the tool. Cracks are fatigue expansion caused by the impact of high temperature and high pressure on the front and back tool surfaces, leading to the loss of cutting ability of the tool. Cracks usually appear in the late stage of the tool intermittent machining.
Three. Research status of tool monitoring
On-line tool damage monitoring has always been a hot topic in the field of machining, and scholars at home and abroad have made a lot of research achievements in recent years.
Generally speaking, the tool damage monitoring system includes signal acquisition, signal processing, damage feature extraction and damage recognition processes.
In the process of processing, there are many signals that can reflect the state of the tool, such as cutting force signal, machine tool spindle power signal, current signal, vibration signal, acoustic emission (AE) signal, etc. According to monitoring strategy and monitoring signal, it can be divided into direct monitoring method and indirect monitoring method.
3.1 Direct monitoring method
Direct monitoring method is to use some measurement methods to determine whether the volume or weight of the tool is reduced directly. This method is usually to directly detect the position, shape and other parameters of the blade or observe whether the blade has a notch.
But the direct monitoring method can only detect the tool in the non-cutting state of the tool, need to stop the detection, occupy the production time so that the use of machine tool efficiency is reduced, in addition, the direct monitoring method can not monitor the tool in the process of sudden damage, so there are few direct monitoring method of research. However, the author still found a "machine vision based tool wear detection system research" paper, the follow-up will be introduced for everyone, please pay attention to.
3.2 Indirect monitoring method
Indirect monitoring method is to detect the parameters that can reflect the cutting state of the tool, and compare these parameters with a set of standard processing parameters, so as to indirectly judge the state of the tool.
Indirect monitoring method can avoid the direct monitoring method can not real-time on-line monitoring and can not detect the tool in the process of sudden damage shortcomings, so indirect monitoring method has been the focus of research.
Tool in the cutting process has a lot of measurement parameters can reflect the damage of the tool, such as cutting force, spindle power or torque, spindle current, cutting in the process of sound signal, vibration signal, workpiece surface roughness and workpiece size sudden change, but these available detection signals, considering the signal collection is convenient or not, monitoring strategy is difficult, the number of signal interference source, Commonly used monitoring signals in actual production include cutting force, spindle power or torque, spindle current, acoustic emission (AE), vibration signal, etc.
The spindle power or torque signal is convenient to collect, no need to change the mechanical structure of the machining system, the sensor is simple, and the spindle power signal is rarely interfered by the cutting fluid, oil smoke and other workshop environment in the processing process, so the spindle power is a very practical monitoring signal for tool damage monitoring.
Cutting force signal can be a more intuitive reaction of the state of the cutting tool in the process of processing, but the installation of cutting force sensor needs to change the structure of the machine tool, which is the biggest obstacle to its application in production. In production, the use of cutting force signal monitoring is generally installed in the shank force sensor, made of force measuring shank, this application also has great limitations. Vibration signal due to more signal interference sources, there are few corresponding monitoring system.
However, many universities in the study of wear theory and data processing algorithm, often take cutting force as the data for analysis, such as "based on cutting parameters and tool state of the turning force model", "based on fuzzy neural network tool wear identification" and "based on tool state of the cutting force model research" and other classic papers, the author will be introduced in the follow-up for you.
In recent years, there are many researches on acoustic emission sensor, which is a promising tool damage monitoring method. However, the high cost of acoustic emission sensor is also a difficult problem to be overcome in its practical application.
In view of the study of acoustic emission signals, the author quoted "Automatic identification of tool wear and damage state in the turning process" as an introduction.
Various monitoring methods are shown in the following table:
Four, commonly used tool monitoring methods
4.1 Power Monitoring Method
The use of spindle power as monitoring signal has been studied by scholars for a long time. Lu Xiaoying et al. from Tsinghua University studied the relationship between spindle power and tool damage during tool processing, and proposed that sudden tool damage during tool processing would cause a sudden change in spindle power.
Based on this principle, "self-learning tool monitoring" is proposed, and "machining state recognition based on machine tool velocity vector" algorithm is adopted to realize data synchronization and comparison. This method is very effective for tool damage monitoring in batch production, but not for single or small batch production.
Ouyang Huibin from University of South China has proposed a tool damage monitoring system based on power change rate, which collects the spindle power value in real time and differentiates it. If its differential value exceeds specified threshold value, the tool damage can be judged. This monitoring strategy can judge the tool damage to a certain extent, but because the alarm criterion condition is very sensitive to the setting of threshold, threshold setting is difficult and easy to cause false positives.
Some foreign scholars also collect the spindle power signal and process the signal by using the autoregressive (AR) model. It is concluded that when the order of the model is low, the peak signal of tool breakage can be seen from the residual sequence.
Wan Jun from Tsinghua University processed the collected spindle power signals by discrete autoregressive (AR) model, and introduced normalized deviation to treat the power signals. The monitoring strategy was proved to be effective through experiments.
4.2 Acoustic emission monitoring method
Acoustic emission (AE) as tool damage monitoring signal is the most studied in recent years, this kind of monitoring signal is considered to be the most valuable research. Acoustic emission (AE) is a physical phenomenon in which transient elastic stress waves are generated by the rapid release of energy within an object or material.
The sudden damage of the tool in the process of processing will produce this elastic stress wave, usually the stress wave is high frequency energy wave, because of the high frequency characteristics of the tool damage AE signal to avoid the noise and vibration in the environment in the process of processing serious pollution of the low frequency area, high sensitivity in the high frequency area, strong anti-interference ability.
Wang Haili et al. from Shanghai Jiao Tong University proposed that there are four AE signals in the cutting process, that is, normal cutting signal, tool damage signal, chip breaking signal and other random signals, and proposed using time-frequency analysis to take energy signal as the characteristic quantity of tool damage monitoring.
AE sensors are difficult to be used in practical production mainly because of the expensive price of AE sensors, the difficulty of sensor installation and the problem of signal processing acquired by sensors.
There are many sound sources in the process of processing. How to choose the signal processing method to remove the noise is always a difficult point. In the process of machining, the tool movement, the workpiece frequent replacement, the cutting fluid splashing and so on to the sensor installation has brought problems.
The existing researches on AE signal are all based on laboratory research, and there is no commercialized tool damage monitoring system based on AE sensor. In order to solve the installation problem of AE sensor, Wang Xinyi et al. studied a new type of sensor that uses cutting fluid as a medium to transmit acoustic emission signals, which has a high sensitivity.
4.3 Vibration monitoring method
Serious wear or damage in the process of cutting tool will cause serious vibration in the process system, so vibration signals are often used in tool damage monitoring.
The vibration acceleration signal of spindle bearing and the X and Y direction vibration signals of tool bar are collected by the beam accumulation medium. The power spectrum of spindle bearing acceleration signal and the frequency band energy ratio of tool bar vibration signal are used as characteristic quantities, and the artificial neural network is used to make judgment decisions to judge the state of tool processing.
Chen Quntao et al. collected sound signals and vibration signals. Based on the two signals, they used empirical mode decomposition to obtain the set of eigenmode functions of the signals. Finally, the obtained eigenmode function components were analyzed by power spectrum to realize the identification of tool damage within the frequency domain.
4.4 Multi-sensor Monitoring Method
In order to strengthen the reliability and accuracy of the tool damage monitoring system, synthesize the advantages of various monitoring methods and expand the application scope of the monitoring system, many scholars put forward the method of multi-sensor comprehensive monitoring.
Liu Zhiyan et al. from Yanshan University used acoustic emission and motor current signals, used spectrum analysis for acoustic emission signals, and used threshold comparison for current signals to comprehensively monitor tool state.
Liu Xiaoming et al. made comprehensive use of acoustic emission, cutting force and power differential monitoring tools. The monitoring system first judged the working condition. For fine machining, acoustic emission signals and cutting force signals were comprehensively used for monitoring; for rough machining, power differential and cutting force signals were comprehensively used for discrimination.
Five, tool monitoring application background
With the continuous development of computer technology and sensor technology, and the continuous improvement of processing automation, the development of tool damage monitoring system is also more and more attention by manufacturers and scholars. The future development trend of tool damage monitoring is as follows:
The development of special intelligent sensors is in urgent need of high quality signal acquisition in complex processing environment, and the installation of sensors is convenient, practical and low cost.
Most of the existing tool monitoring involves a "sample learning" process, and with the change of processing conditions, the learning sample also has to change, the study of efficient and intelligent signal processing technology, reduce the monitoring model due to the change of processing parameters and frequent acquisition of learning samples.
Tool damage monitoring system is not only limited to laboratory research stage, it will be applied. In recent years, it is a trend to develop a convenient and practical monitoring system. Tool damage monitoring system will not exist alone. It may be integrated with production line management system, workshop ERP management system and other production management systems into a highly intelligent and practical plant management system.
Six, processing signal acquisition
Taking FANUC CNC machine tools as an example, the reason why CNC machine tools can realize automatic and intelligent processing is closely related to the sensors in CNC machine tools. The commonly used built-in sensors in CNC machine tools can be roughly divided into the following types:
Position sensor
The position sensor is mainly used to detect the position of the current tool rest and to sense whether the tool rest moves to the limit position in the process of movement. Position sensors are divided into contact type and proximity type. The commonly used built-in sensors in CNC machine tools include travel switch, proximity switch, magnetic switch, etc.
Displacement sensor
Displacement sensor is used to detect the specific position of the current tool in the machine coordinate system. The commonly used displacement sensors include linear grating, induction synchronizer, rotary transformer, etc.
Speed sensor
The speed sensor is used to detect the speed of each shaft and the linear motion speed of the saddle in the CNC machine tool. The commonly used speed sensor in the CNC machine tool has the speed measuring generator, pulse encoder and so on.
Power sensor
The power sensor is used to detect the power of each shaft in the CNC machine tool, including the spindle power and feed shaft power. The commonly used power sensors in NC machine tools are Hall sensor, thermistor and so on.
Reference: Design and Development of Tool Damage Monitoring System for Production Line