| Electromagnetic casting(EMC) is a new type of casting technique that combines the magneto-hydrodynamics(MHD) theory and the developed casting engineering. It is one of the most promising ways to produce high quality aluminum slabs.The problems concern EMC slabs are the stability of the technology process and the shape defection.During EMC process, the control of suitable and stable liquid column is the crucial point to obtain ingot of desirable size and shape. The appropriate height, vertical side faces and stable surface are requested for the liquid column. In this paper, according to the theory of fuzzy model, a fuzzy neural network with self-learning and self-adapting characters is used to deal with the electrical signal of EMC parameters to forecast the height of liquid column quickly and precisely. The results show that the fuzzy neural network model has excellent pattern recognition and forecasting ability. With the use of the neural network to help control the EMC process, the fluctuation of height of liquid column is reduced greatly and the surface quality of the ingot is significantly improved.Because different sections are under different cooling conditions, the EMC slabs will deform and have various shape defection during different casting process. During the start phase of casting process, the butt of the slab will curl up, this makes the butt has to be cut away. The butt curl also greatly influences the subsequent stability of the casting process. During the stationary phase, the rolling faces of slabs will have non-uniform contraction, largest at the center but much smaller at the corner. So before rolling, the faces should be scalped to be flat and parallel. These shape defections of slabs not only add additional working procedures to subsequent mechanical process, but also waste the material, decrease the productivity and bring much waste in time and economy. So decreasing the butt curl defection and producing near-flat or flat slabs is now an impending and significant task.From the viewpoint of improving the metal getting and stabilizing the casting process, first the technologies to decrease the butt curl of slabs are studied.The understanding of deformation mechanisms is the key point to control the pull-in of the rolling faces of slab. In this paper, it is studied theoretically and experimentally. The influences of some casting parameters, such as pouring temperature, water flow rate and the distribution of jet hole on the flatness of rolling faces of slabs have been studied, especially emphasized is the relationship of temperature field and the pull-in formation process of the slab and the corresponding controlling method. It is found that the pouring temperature doesn't have great influence on the pull-in of rolling faces; while the width over thickness ratio, water flow rate and casting speed apparently affect the pull-in. Controlling the temperature field can directly control the deformation of the rolling faces of slabs. Rational cooling and solidification process can obtain almost rectangle cross section of slabs.Computer numerical simulation is very valuable in helping forecast the quality of ingot and optimize the casting conditions economically with less time and effort. In this paper, some unique characters hi numerical simulation of EMC temperature field, such as the enlarging calculating space, the simultaneous heat-transfer and mass-transfer, the effect of electromagnetic induced heat, and so on, are solved by the mathematical model in detail. The effects of some parameters on the temperature field have been simulated and discussed to supply excellent information for the design of casting parameters.Based on the deformation mechanism, the simplified models to calculate the deformation of slabs are put forward. And the influences of casting parameters on the butt curl and the pull-in of rolling faces have been calculated and discussed, which can provide guides hi practical industries.Multi-layer feedforward artificial neural network has excellent ability...
|
PDF Full Text Request