Abstract In this paper, the possibility and feasibility of computer simulation technology in the EDM process modeling, process parameter optimization, and processing mechanism research are elaborated in detail. The EDM process simulation system is established to realize the prediction of processing effects and the processing parameters. optimization. The simulation system was also used to find out the method of large thickness processing and achieved stable processing of 805 mm thick workpieces. At the same time, the paper also uses the computer simulation technology in the EDM heat conduction model to study, for the pulse power supply parameters and waveform design to find the theoretical basis. Computer Simulation Application for EDM Machining Lou Leming Xiao Wenfang Li Minghui Peng Yinghong Abstract In this paper,an EDM simulation system is presented,which has been used to optimize the process parameter and predict the erosion rate and surface roughness seccessfully.According to this system,the way to machine the thick wordpiece is found, and workpiece with 805 Mm thickness is machined stably.The simulation technology is also used to study the thermal model of EDM. In recent years, with the improvement of pulse power technology, the introduction of numerical control systems and the adoption of artificial intelligence technology, the processing capability of EDM has been greatly improved, and the operability of EDM machines has also been greatly enhanced, but its processing mechanism has been greatly improved. However, the research has not been able to make great progress. The reason is that the EDM processing mechanism is too complex and involves various theories such as electromagnetic fields, high-voltage electrostatics, fluid mechanics, dielectric physics, plasma physics, and underwater explosions. The theory itself needs to be re-recognized, so that the mechanism of EDM processing has not been able to make breakthroughs in recent years. On the other hand, the backwardness of research methods is also one of the important reasons. The increasing performance of computer technology makes the application of simulation technology more and more widely in industry. It also brings new ideas to the research of EDM theory, which makes it possible to overcome the limitations of traditional research methods and further reveal the microscopic conditions of the discharge process. And to establish the model that reflects the objective laws of discharge, and provide a reliable theoretical basis for machine tool design and process control. 1 Computer Simulation Technology and Its Application in EDM Because the EDM processing mechanism is too complex, so far it has not been able to truly solve the relationship between various processing parameters on the processing performance, the empirical formula is too simple and often results in a far cry from the expected results and the actual, which led to the process database The emergence of process databases and system development on this basis have greatly facilitated the development of EDM technology and facilitated operators so that they can start the machine tool and complete machining requirements without having much expertise. However, actually the data of the processing data sheet is obtained by the experience of skilled workers. However, whether or not the empirical data is the optimal data depends entirely on the machine tool production unit, so the processing performance under certain conditions is optimized, and How to expand the processing range and optimize the parameters by processing the data already in the database also requires deeper research on the processing mechanism. Simulation technology is more suitable for studying the processing mechanism like EDM. It is not completely clear, but it already has some experience and knowledge system. Simulation technology has also been widely used in the EDM field. Figure 1 shows the basic principles of computer simulation. Fig.1 Basic principle diagram of computer simulation From Figure 1, we can see that the main meaning of computer simulation is mainly reflected in: 1) Perfect system model: EDM system is a nonlinear system, only knows the input and output of the system, and does not know the internal structure and structure of the system. parameter. However, due to some existing knowledge, it is possible to construct a simple system concept model with hypotheses, and then use this concept model to perform simulation experiments. In the experiment, the structure and parameters of the model are corrected according to a certain method, so that it finally meets Known input and output relations of the system, thus establishing a mathematical model similar to the system characteristics, can use simulation technology to achieve EDM processing process modeling; 2) Prediction: Using computer simulation can avoid direct experiments, and can predict the characteristics of the system , further study the EDM process rules, and deepen the understanding of the EDM mechanism; 3) optimize the design: based on the established model, using certain optimization techniques can be achieved in the computer on the specified processing requirements of the most process parameters Optimization, at the same time, in the design process of the EDM pulse power supply and control system, the performance prediction and parameter optimization of the system are needed so that the designed system can reach the optimal target. The system design also needs to be repeated frequently, so it is necessary to use simulation technology to assist the design; 4) Repeated tests: Repeated tests are a very natural requirement in system analysis and design, so that a real system is repeated in exactly the same environment It is difficult to implement physically repeatable tests for a system with a certain randomness in EDM, but it is even more difficult to use a simulation technique to ensure the same output under the same parameter settings and input conditions. Economical and safety: Testing on EDM machines is not only time-consuming, but also costly. Therefore, it is difficult to systematically study the process rules. In addition, some process rules may not necessarily be studied on the machine tool, such as research on large-thickness machining, deep holes, or arcing that is easily destructive. In contrast, simulation tests are not only cost-effective but also safe. reliable. 2 EDM processing speed, surface roughness and other processing performance forecast and optimization To predict the machining speed and surface roughness and to optimize the machining parameters, first establish a model of the machining speed surface roughness. Modeling methods General linear regression, differential equations, neural networks and other methods. The commonly used EDM modeling method is the use of regression theory to establish the EDM speed model or surface roughness model, such as Anand Srinivas [3], M.kishi [4], YSLiao [8] and others who use multi-objective optimization and regression analysis. The relationship between the WEDM processing speed, surface roughness and other parameters and the electrode wire diameter, workpiece thickness and pulse width and other parameters. The advantage of regression analysis is that the calculation speed is fast and a definite mathematical formula can be drawn. In order to improve the regression accuracy of the model, it is generally necessary to add the high-order variables and the combination factors between the independent variables. Therefore, when the number of independent variables is large, the form of the regression equation is difficult to determine. Figure 2 EDM forming neural network model Forecast results and raw data tables Note: Processing Conditions: Cu(+)-Steel(-) From the data in the table, it can be seen that the data predicted by the model is compared with the experimental data in the processing data table, and most of the errors are within 10%, indicating that the model can accurately predict the processing speed. Figure 3 Wire Cutting Simulation System During the simulation study of WEDM processing using this simulation system, we also found that an interesting phenomenon is that under a certain processing current, when the workpiece reaches a certain thickness, the processing speed begins to drop significantly, indicating that the processing is no longer available at this time. Stable, but if the peak current is increased at this time, the processing can be performed normally. Based on this, we speculate that to cut a large thickness workpiece must have a corresponding high peak current, according to this speculation we successfully cut a workpiece with a thickness of 805 mm. 3 Using Simulation Technology to Study Electro-machining Mechanism EDM involves a variety of complex theories, it is difficult to establish a perfect theoretical model, in addition, due to the gap between the poles in the electrical processing process is generally only tens of microns, the discharge state between the poles is difficult to observe using conventional means, etc. The study of electromachining mechanisms has brought difficulties. Computer simulation techniques may help us to understand the discharge mechanism. Based on the existing understanding, a simulation model can be established, and then the simulation model can be used to study the effects of current waveform, pulse width, pulse frequency, and electrode material on processing efficiency and processing performance, or to study the influence of other factors on processing performance. . The successful application of simulation technology in many aspects and its initial application in the field of WEDM and EDM have shown that it is an effective tool for studying the EDM mechanism and can be used in many areas of EDM processing. Where: α is the thermal diffusivity, α(t) is the heat source radius, and T0 is the initial temperature. Where: q is the heat density. Based on this, the relationship between the pulse width and the relative electrode loss is shown in Figure 4. Figure 4 Relationship between Pulse Width and Relative Electrode Loss Utsumi [11] believes that under the vacuum arc discharge, the temperature of the cathode discharge point is the boiling point Tb of the cathode material, then the formula for the change of discharge channel radius can be obtained according to the above formula: The relationship between the discharge time and the radius of the heat source when the copper is processed in this way is shown in FIG. 5 . Figure 5 Relationship between discharge time and heat source radius G. Spur et al. [12] studied the variation of the channel radius under the action of a delta current during WEDM processing, and calculated the relationship between the radius of the discharge mark, the ratio of the depth to the diameter of the discharge trace, the pulse width, and the current intensity using the finite element method. It is concluded that in the future power supply design, the falling edge of the current should be reduced as much as possible, because it will help improve the pulse utilization rate, and it also helps to increase the processing speed. 4 Conclusion Based on the characteristics of EDM and the advantages of simulation technology in system modeling, prediction and optimization, this paper establishes EDM process prediction and optimization system. The prediction made by this system shows that the prediction error is basically within 10%, which means that the system model truly reflects the process law of the machine tool, and can fully use this system to predict the process effect and optimize the processing parameters, thus making it an EDM system. The direction of intellectualization laid the foundation. At the same time, we also used this system to study the process law and discovered the process rule of high-thickness wire-cutting. When high-thickness wire-cutting processes are used, high peak currents are necessary. According to this rule, we successfully cut 805 Mm thick workpiece. We also used simulation techniques to study the heat propagation model of EDM and achieved preliminary results. Camera Modules Cleaning Machine
ROC Precision Intelligent Technology Co., Ltd. was established in 2013 and is located at No. 3, Tengfeng Road, Fenghuang Community, Fuyong Street, Baoan District, Shenzhen. Focus on electronic manufacturing cleaning equipment . This Camera modules cleaning machine are used to Camera modules manufacturing production line. Using this kind of Camera modules cleaning machine can increase the quality and quantity of products.CCM modules in PCBA also can use cleaning machine ,CCM camera module process use cleaning machine Remove the dust.
Automatic Pcba Cleaning Machine,Pcba Cleaning Drying Machine,After Dip Pcba Cleaner Machine,Online Pcba Cleaning Equipment Shenzhen KunPeng Precision Intelligent Technology Co., Ltd , https://www.rocairseparations.com
Keywords computer simulation, EDM, neural network optimization
Department of Plasticity Engineering, Shanghai Jiaotong University
Keywords Computer simulation EDM machining Neural entwork Optimize 
Since computer simulation has a series of advantages mentioned above, we have conducted some research on the prediction of EDM process effects, the optimization of process parameters, and how to use simulation techniques to study the mechanism of EDM.
Because of the above shortcomings in regression analysis, the theoretical model of EDM is difficult to establish, and people have turned their attention to modern modeling methods such as neural networks. According to the basic theory of neural networks, any continuous function can be simulated with a three-tier network structure. As long as a proper network structure is given, the network model can be established after determining the weight relationship between the network layer units by learning the sample data. If new sample data is available later, the network can be used to relearn the network function. Adapt to new samples without rebuilding the model. Compared with linear regression, neural network not only has self-learning function, but also is insensitive to the number of independent variables, and there is no problem of determining the order of independent variables and the combination of independent variables in linear regression, so the independent variables It is more appropriate to use neural network modeling in more cases, and EDM is the case. Compared with EDM forming processing, WEDM has much less influence parameters and is easy to study using various methods. Using neural networks and other theories to study the influence of various influence parameters on the processing performance in WEDM processing has become a current one. Hot spot. For example, YSTarng [6] and TASpedding et al [5,7] used neural network to establish the processing speed and surface roughness model for WEDM processing, and then simulated annealing algorithm to optimize the processing parameters. The simulation results were in good agreement with the experimental data.
The impact parameters of EDM forming process are more than WEDM, and the difficulty is even greater. At present, the process modeling research for EDM forming process is also very rare. We analyzed that the main factors influencing the performance of EDM forming process are: peak current Ip, pulse width Ton, pulse-to-pulse Tpff, processing time Tdn, and tooling-up time Tup, while the processing performance is mainly reflected in the processing speed and surface roughness. The EDM forming process model was established as shown in Fig. 2. The following table shows the prediction results obtained by using this model and the comparison table of experimental data. 
I/A pulse width
t/μs pulse
t/μs lifting
t/μs processing
t/s Ra/μm MRR(mm3/min) Experimental value prediction error
(%) Experimental value prediction error
(%) 2.5 20 15 3 4 6 5.39 10.2 0.25 0.28 12.0 3.5 60 40 3 4 8 7.78 2.7 0.75 0.73 2.7 6.0 80 40 3 4 12 11.75 2.1 2.60 2.44 6.2 6.5 80 40 3 4 13 12.53 3.6 3.20 3.10 3.1 6.5 20 15 3 4 10 10.11 11.0 2.80 3.00 7.1 69.5 250 60 2 5 75 76.62 2.16 230.00 228.59 0.6 93.5 6 24 2 5 18 18.06 0.3 16.02 16.26 1.5 115.2 6 24 2 5 16 16.08 0.5 33.22 33.14 0.2
Using this model can not only predict processing speed, surface roughness, but also can be used to optimize machining parameters. This will pave the way for the development of intelligent machine tools for EDM machines, which will allow novices in electrical processing to produce products that meet the requirements without too much training. For this purpose, we have established a simulation system for wire cutting processing using neural networks. The basic block diagram is shown in Figure 3. 
According to the current level of understanding, it is generally believed that the essence of EDM is the thermal effect. According to the basic theory of heat conduction and the actual conditions of EDM, after certain assumptions, the heat conduction model of the electrode can be obtained as: 
If it is assumed that the radius of the discharge channel remains unchanged, the temperature field distribution of the heat source with a constant radius on the basis of the heat source theory is: 
China’s Chu Zhenbin also studied the propagation model of heat sources with a constant radius in 1978, and explained the conditions of low-loss processing qualitatively and coarsely. It lays a theoretical foundation for low-loss processing.
It can be said that the research on the heat transfer model of EDM has achieved certain results. However, all the above models have been greatly simplified in the process of solving, especially when solving the temperature field by analytical method, the simplification is more, and some are even not reasonable. With the continuous improvement of the computer level, the computer's ability to solve various complex models is also growing. A variety of symbolic computing software functions are becoming more and more powerful. Consider solving the temperature field with less simplification. The finite element analysis software, which is now more and more sophisticated, can also be used for numerical simulation under less simplified conditions to obtain the distribution of the temperature field. At the same time, the previous analysis mainly focused on the temperature field when the power waveform is a rectangular wave. When the current intensity changes with time, most of the above analysis is not suitable. We will use the finite element analysis method to study the surface heat source heat transfer model of transient heat flux, transient radius, hoping to provide theoretical support for future pulse power supply design and low-loss machining and even EDM milling processing.
Simulation technology can be used not only to study the heat transfer model, but also has been successfully applied in the research of the oil flushing method. Masuzawa et al. simulated the distribution of galvanic corrosion products under different modes of oil rushing, and concluded that moving rush oil is better than static rushing oil, and that the mirror edging of the workpiece is achieved by moving the oil rush method.
In summary, the simulation techniques are used in the EDM machining speed modeling, processing speed and surface roughness prediction, processing parameter optimization, EDM mechanism research such as the peak current, pulse width, and material properties on the theoretical analysis of the impact of processing performance and current Waveform effects on processing performance can play a role. And lay a good foundation for the future development of electric spark artificial intelligence system.