Papers by Author: Z.F. Chi

Abstract: Based on craft features analysis of vacuum casting, this paper concluded that the automation level and control precision of control system are key influencing factors for products quality of vacuum casting. So the purpose of this paper is to resolve these disadvantages of existing control system of vacuum casting equipments. This paper designed a new full automatic intelligent control system which adopts two-level distributed structure. The system is consisted of motor control subsystem, vacuum monitoring and control subsystem, and bubbles detecting and eliminating subsystem. According to the function requirement of control system, this paper described selection and configuration of hardware and function implement of software in detail. Finally, this paper designed and realized a new automatic detecting and eliminating method of bubbles based on vision sensor technology, and presented an intelligent controller model based on Wavelet Neural Network (WNN) for motors. The experimental prototype of vacuum casting has been developed and a large number of experiments have been carrying on. These new controlling algorithms have being tested. The control system integrates vision sensor, intelligent control, microcomputer interface and imaging processing technologies, so it possess high precision, efficiency and flexibility and has a good market application prospect and high economic value. It will be breakthrough for precise rapid tooling.

Abstract: Bubbles in vacuum casting process are dense, tiny and overlapped. The difference between bubbles and the background is so indistinct that makes bubble detection a great difficulty. Bubble elimination has become the main obstacle in the way of vacuum casting machine’s automation development. Bubble automatic detection is the base of bubble elimination. A machine vision bubble on-line detecting and eliminating platform is constructed, using industrial high-speed CCD camera and professional LED illumination. According to the features of vacuum casting bubbles, the edge pixel ratio algorithm is designed especially for vacuum casting bubble detection. The algorithm is realized using VC++ and Open CV. The integrated system proposed can detect and eliminate vacuum casting bubbles on-line automatically. As it’s confirmed, there is an obviously positive correlation between the edge pixel ratio and the bubble denseness. This detection system makes sense in vacuum casting bubble elimination.

Abstract: Differential pressure Vacuum Casting is a new kind of rapid manufacturing technology. The filling speed is the key process parameter which decides the filling quality. In order to ensure the stabilization of the flowing speed, qualitative model has been set up. Both CAE software and BP Neural Network Models have been used to find the relation between the inlet pressure and the forward speed. Using the inverse BP Model and giving a constant velocity of every section in the cavity, the expected pressure values have been got. In terms of the Filling Length Measure Method (FLMM) proposed by the researchers, the time points of flowing stage have been evaluated. The relation between the expected values of inlet pressure and time points has been concluded from this study.

Abstract: As one key technology of Rapid Tooling (RT), Indirect Nonmetal Rapid Tooling (INRT) has been used widely, so this paper analyzed the modeling process of INRT and concluded that the velocity of stirring and casting are key factors which affect products quality and production efficiency. In nature, just need to control the speeds of driving motors. In order to resolve problems of existing modeling equipments, such as strong dependency on operators experience and low control precision, this paper studied an intelligent controller model based on Fuzzy Wavelet Network (FWN) which integrated wavelet analysis and fuzzy neural network technologies. The structure and implementation algorithm of FWN are introduced in detail. After considering the advantages of Permanent Magnet Synchronous Motor (PMSM), the INRT equipments use it as driving motors. Because it is nonlinear, uncertain, coupled, sensitive and multivariable system, this paper made improvement on traditional IP controller, FWN acted as error compensator which can compensate for disturbance influence and trace expected speed in real time. The IP intelligent controller based on FWN can strengthen interference ability. The robustness of this control system has been verified by simulation experiment. This controller realized closed loop control of stirring and casting motor and improved the products quality and control precision of INRT.

Abstract: Bubbles in the manufacturing process are common. The bubbles often lead to the decrease of the product’s surface quality and internal performance. This paper summarized the published researches and applications of the detection and processing for bubble images, of which the advantages and disadvantages were also presented. Based on the above mentioned results, this paper then proposed a new bubble image processing algorithm for vacuum casting process, in which the characteristics of the bubbles in vacuum casting process and the problems possibly caused in detail were analyzed. According to the characteristics of bubbles in vacuum casting process, an image processing algorithms was designed using Matlab. The simulation result showed the efficiency of the proposed algorithm.

Abstract: This paper proposed a new method of automated eliminating bubbles based on vision sense technology because detection and elimination of bubbles of injection materials is key factor which influences the quality of products and hinders realization of full automatic control for the vacuum casting. This method uses vision sensor to replace artificial visual measurement, and adopts processing of the bubbles images to judge the effect of eliminating bubbles instead of experiences of operators. This paper developed a bubbles monitoring control system to realize the automated eliminating bubbles method. The system adopts two-level distributed structure which is consisted of host and slave computers. The slave computer is composed of image acquisition module, controller module and execution module. It is responsible for controlling operation devices to finish vacuum casting. The host computer is responsible for supervisory control which includes human-machine interface module, camera control module and images analysis and processing module. The function of every module is described in detail. The operator is just needed to set relative parameters, and then the system will implement automatically. In a word, the controlling system based on this new method possesses some advantages such as high precision, efficiency and flexibility.