This paper starts from the sensor and divides the intelligent technology of CNC machine tools into intelligent sensors, intelligent functions, intelligent components, intelligent systems and other parts. It sums up the intelligent technology, points out the inadequacies, reveals the direction of development, and looks forward to the future. .
Intelligent machines first appeared in the first monograph "Manufacturing Intelligence" in the area of smart manufacturing research published by P. W. Wright and D. A. Bourne in 1998. Due to its important role in advanced manufacturing, smart technology has attracted the attention of various countries. The United States launched the SMPI; Europe implemented the “NextGenerationProductionSystem” study; Germany introduced the “Industry 4.0” program; China’s mid-to-long-term technological development put forward an urgent need for “digital intelligent manufacturing technology” and developed a corresponding The "Thirteenth Five-Year Plan" development plan; at the 2006 IMTS 2006, the first intelligent machine tool named "Intelligent Machine" launched by Japan's Mazak Company and the one named "thinc" launched by Japan's Okuma Corporation. The intelligent numerical control system has opened the era of intelligentized numerical control machine tools. This paper starts from the sensor and divides the intelligent technology of CNC machine tools into intelligent sensors, intelligent functions, intelligent components, intelligent systems and other parts. It sums up the intelligent technology, points out the inadequacies, reveals the direction of development, and looks forward to the future. .
Smart Sensors The CNC machine tool manufacturing system consisting of machine tools, cutting tools, and workpieces, in the course of machining, with the removal of materials, accompanied by a variety of complex physical phenomena, implies a wealth of information. In this kind of dynamic, non-linear, time-varying and non-deterministic environment, the perception technology of the CNC machine tool itself is the basic condition for achieving intelligence. In order to realize intelligence, CNC machine tools need various sensors to collect external environment and internal state information to approximate the function of the five senses of human beings to perceive environmental changes, as shown in Table 1. For humans, the eye is the most important sensory organ in the five senses, and can obtain more than 90% of environmental information, but the application of visual sensors in CNC machine tools is still relatively small. With the increasing level of automation and intelligence, visual functions will play an increasingly important role in CNC machine tools. Table 1 Available sensors for CNC machine tools:
With the development of technologies such as MEMS (Micro Electro Mechanical Systems) technology, embedded technology, and smart materials and structures, sensors tend to be miniaturized. The mature applications of micro-sensors such as MEMS microsensors, thin-film sensors, and fiber optic sensors lay the foundation for the integration of sensors into CNC machine tools. Due to unpredictable or unpredictable complex phenomena and strange problems in the manufacturing process, and the timeliness, accuracy, and integrity of the information being monitored, the sensors are required to have the intelligence of analysis, reasoning, and learning. This requires the sensor to have a high-performance smart processor to act as a "brain." Qualcomm is developing an artificial intelligence system microprocessor that can simulate human brain work. In the future, through semiconductor integration technology, high-performance artificial intelligence system microprocessors and sensors, signal processing circuits, and I/O interfaces can be integrated on the same chip to form a large scale integrated circuit smart sensor that not only has detection, recognition, and memory. , analysis and other functions, but also have self-study and even thinking ability. It is believed that with the continuous advancement of computer technology, signal processing technology, MEMS technology, high-tech materials technology, and wireless communication technology, smart sensors will bring about new changes in the intelligent sensing of CNC machine tools.
The intelligent function CNC machine tool develops to high speed, high efficiency, high precision, require the numerical control machine bed to have the intelligent function such as thermal compensation, vibration monitoring, wear monitoring, condition monitoring and fault diagnosis. Combine several or several kinds of intelligent sensors, adopt artificial intelligence method, realize the intelligent function of the numerical control machine tool through identifying, analyzing, judging and reasoning, lay the foundation for the realization of the intelligent component. Errors in CNC machine tools include geometric errors, thermal (deformation) errors, force (deformation) errors, and assembly errors. Research shows that geometric errors and thermal errors account for more than 50% of the total machine tool error and are the key factors affecting the machining accuracy of the machine tool, as shown in Figure 1. Among them, geometric error is the error caused by the machine tool structure itself during manufacturing and assembly. It does not change much with time and belongs to static error. The error prediction model is relatively simple and can be effectively controlled by the system's compensation function. Thermal error is The time changes greatly, it belongs to dynamic error, and the error prediction model is complex. It is a difficult and hot spot in international research.
The heat sources of CNC machine tools in the machining process include bearings, ball screws, motors, gear boxes, guide rails, and tools. The temperature rise of these parts will cause changes such as spindle extension, coordinate changes, tool elongation, etc., resulting in an increase in machine tool errors. Due to the large number of temperature-sensitive points and wide distribution, the optimal design of temperature test point location is very important. The main methods include genetic algorithm, neural network, fuzzy clustering, rough set, information theory, and gray system [6]. On the basis of determining the temperature measurement points, commonly used neural networks, genetic algorithms, fuzzy logic, gray systems, support vector machines, etc. are used for error prediction and compensation.
In the aerospace field, with the wide application of difficult-to-machine materials such as titanium alloys, nickel alloys, and high-strength steels, and under high-speed cutting conditions, the increase in cutting volume has made it easy for tools and workpieces to vibrate and seriously affect the workpieces. Processing accuracy and surface quality. Since the cutting force is the original characteristic signal of the cutting process, it can best reflect the dynamic characteristics of the machining process. Therefore, the vibration monitoring can be performed with the cutting force monitoring and prediction. With force meters, force sensors, feed motor currents, etc., particle swarm optimization, fuzzy theory, genetic algorithms, and gray theory are used to model and predict cutting forces. Taking into account the main causes of machine tool vibration spindle, screw, bearing and other components, you can also collect these components vibration, cutting force, acoustic emission and other signals, the use of neural networks, fuzzy logic, support vector machines and other intelligent methods to directly vibration monitor.
The cutter is installed on the front end of the spindle and comes into contact with the machined workpiece. It directly cuts the surface of the workpiece. The effect on the machining quality is the most direct and critical. Tool wear, breakage and other anomalies affect machining accuracy and work safety. In view of the disadvantages of the direct measurement method that require off-line detection, one or more indirect signals such as current, cutting force, vibration, power, and temperature are often collected, and intelligent algorithms such as RBF neural network, fuzzy neural network, wavelet neural network, and support vector machine are used. Intelligent monitoring of tool wear conditions. With the increase in the degree of automation, CNC machine tools integrate more and more functions, and their complexity has been continuously improved. In order to operate efficiently, it is necessary to monitor and evaluate the internal state of CNC machine tools and to make early warning and diagnosis of faults. Because the failure mode is not very reproducible and sample collection is difficult, intelligent methods such as BP neural networks that require many samples are not suitable for this occasion. State monitoring and fault diagnosis often use intelligent methods such as SOM neural network, fuzzy logic, support vector machine, expert system and multi-agent. Researchers continue to explore and study new methods or hybrids of intelligent functions. However, most of them focus on the laboratory environment. They lack methods with high real-time and strong on-line capabilities. They need to be further developed in a simple, rapid, and adaptable manner. Smart method.
Intelligent parts CNC machine parts mainly include support structure parts, main transmission parts, feed transmission parts, tools and other parts, involving the bed, column, spindle, tools, screws and guides, and rotating shafts and other components. These components can be integrated with one or more intelligent functions of smart sensors to form smart components of CNC machine tools.
The main shaft is the main transmission component. As the core component, it directly relates to the workpiece machining accuracy. Due to the high spindle speed, especially the spindle, heat, wear, and vibration have a great influence on the machining quality. Therefore, more and more smart sensors are integrated into the spindle to realize functions such as monitoring, early warning, and compensation of working conditions. . The "smart spindle" developed by Japan's Yamazaki Mazak is equipped with various sensors such as temperature, vibration, displacement, and distance. It not only has temperature, vibration, and fixture life monitoring and protection functions, but also can intelligently coordinate processing based on temperature and vibration conditions. parameter. Electro-spindles manufactured by Step-Tec and IBAG in Switzerland are equipped with sensors such as temperature, acceleration, and axial displacement. As shown in Fig. 3, they can perform thermal compensation and vibration monitoring.
