Research On The Intelligent Process Decision Technique For High-speed Milling Of Die

As a new-emerging modern technology, high-speed milling technology is known as the great characters of rapid machining speed, high feeding rate, excellent surface quality and dimension precision. It has been introduced into the manufacture of the aerospace, aviation, Die, etc, and shows its particular technical advantages in these fields. For most of the Die parts have complicated freeform surfaces and they are difficult to be machined, the high-speed milling technology has much more advantages in advancing the manufacture economic efficiency than the traditional electrical discharge machining (EDM) to performance its potential in making Die. But it is not very long to enter into the actual high-speed milling of Die and it has the distinct machining mechanism and process with the conventional milling methods. So there are little exiting process data which fit for the high-speed milling of die. Therefore, building the process database for the high-speed milling of Die, and solving the problems of lacking process data will face in the actual productions according to the artificial intelligent technologies are very significate. The intelligent process decision techniques for high-speed milling of Die are just researched in this paper.1. Total Function Model DesignOn the basis of the analyzing of the intelligent process decision requirements, the actual process decision problems for high-speed milling of die and the reasonable artificial intelligence techniques are combined in the paper, the modular modeling idea is chosen to solve the process decision problems. The total function model using the system modeling method-IDEF0 is built in the paper, which shows as Fig.1. The intelligent process decision for high-speed milling of die mainly includes four modules: The Process Database, Multi-objective Optimize of High- speed Milling Parameters, Integrated Judge & Fuzzy Clustering for Die Material and Expert System Based on the Neural Network, which show in the Fig.1. 2. The Process Database Design for High-speed Milling of DieAccording to the process information requirements of the process decision, the required process informations are analyzed and disposed before the database is built. According to the database designing standard, the E-R conception model and logical structure of the database are designed. At last, the database is built by Microsoft SQL Server 2000, the database includes the process parameters of the die material, milling tool, milling machine, example data, and so on, which mainly provide the data supporting platform for the intelligent technique decision modules The management application program of the database is also realized using Visual C++ 6.0 in the paper.3. The Multi-objective Optimal Selection of Milling Parameters for High- speed Milling of DieIn order to pursue the higher productivity, the better surface quality and the lower production cost is the primary goal for the high-speed milling. By combining the actual high-speed milling process, the multi-objective optimal mathematical model of milling parameters for high-speed milling of die is built in the paper. The expression of the mathematical model shows as follows: The decision variables: milling speed v c, feed per tooth f z, axial milling depth a p, radial milling spacing a e, in the condition of satisfying the constraint conditions: the restriction of milling speed restriction of machining center, feed per tooth, milling axial depth, milling space, torsion moment of machining center, power of machining center, stability of high-speed milling process, to seek the object functions: the optimum of the least machining times, the minimum surface roughness, and the lowest machining cost at the same time.For the different kinds of high-speed milling working conditions, the different object functions and constraint conditions which include the decision variables are built in the paper. By analyzing the characteristic of the optimal mathematical model, the Non-Dominated Sorting Genetic Algorithm-II (NSGA-II) algorithm is selected to solving the optimal mathematical model which is constraint handled by the Punishment Function and designs the program combining the example using Matlab. The optimal and the orginal results are shown in the Tables 1.It can summarize that the optimal mathematical model and the solution arithmetic can achieve the goal of multi-objective optimal milling parameters selection, which can be used to guide the practices. It provides a new thought for optimal milling parameters selection in the detail working conditions and also reduces the actual experiment quantity in some extent.4. The Integrated Judge & Fuzzy Clustering Analysis for High-speed Milling Capability of Die MaterialsTo solve the problem of process decision for the Die materials which is not contained in the database or just discovered and haven't known the working condition, the characteristic of the materials which have the similar machining capabilities will have the similar technological capabilities is considered in the paper. First, the material's capable indexes which affect its high-speed milling capability are analyzed, and the capable indexes are used to judge the high-speed milling capability of Die materials in the paper. Then, the capability of the materials which have the same high-speed milling capabilities are analyzed with the method of fuzzy clustering, and the most similar material to the judged material is found out. We can use the material's process to appreciatively instead the process of the machined material. The integrated judge & fuzzy clustering arithmetic program is designed using Visual C++ 6.0. The program flow chart shows as Fig.2. The arithmetic program is validated by the example, from which it can summarize that the module can excellently achieve the goal of the most similar die materials selection. It can also advances the using efficiency of the database and provides a preferable approach to settle the problem of process data selection for the Die materials which is not contained in the database or just discovered.5. The Process Decision Based On the Neural Network-Expert System How to use the former examples'process data to guide the future production is the linchpin of the intelligent process decision. Expert system has the predominance in solving such problems. The process factors which affect the process decision are analyzed in the paper, and expert system model based on the neural network is constructed shows as Fig.3. According to the learning and forecasting of experiential machining data by the BP algorithm to realize the knowledge acquisition and reasoning of expert system. At last, the BP algorithm program is designed using Matlab for the network and the reasoning capability of the network is validated by the example. By contrasting the machining goal and the actual result, it indicates that the expert system model based on the neural network which is constructed in the paper has the ability to recommend the optimal process for high-speed milling of Die.