Basic Research On Process Optimization For CNC Mechanical Polishing Aluminum Alloy Wheel

Automobile industry is aiming at lightweight, high-speed, modernization. As a result, Aluminum alloy wheels are being used widely all over the world. In order to improve the reliability and durability of the wheels to run on the road in all-weather conditions and meet the increasing aesthetic needs of customers, it is necessary to plating or coating wheels. And polishing is one of the most steps during plating or coating aluminum alloy wheels. Wheels' appearance depends on polishing technology 70 percent and the others 30 percent. To get rid of the current situation that migrant workers polish wheels by handy air-driven or electric sanding tools, improve the productivity of polishing wheels and ease the shortage of workers, etc., automatically polishing aluminum alloy wheels technology is imperative. The CNC mechanical polishing machine tool prototype is developed and a great deal of basic research work has been done, however, quantitatively assessing the relation between process parameters and material removal depth, and surface roughness is still difficult. Besides, determining process parameters when material removal depth is constant and surface rough is limited is not easy. To solve these problems, this paper carried out a more systematic study.Modeling the quantitative relation between process parameters and material removal depth, the algorithm, which is based on Preston equation and assuming the contact pressure between the polishing tool and the wheel surface meets Hertzian distribution premise, is proposed. The model form the algorithm was macroscopic and focused on the relation between material removal depth and the normal polishing force, the geometric physical properties (such as elastic modulus, Poisson's ratio and the relative principle curvature), the motion property (such as feed rate and polishing tool tangential velocity).Besides, another model on abrasive grits and material removal depth is constructed, which is based on probability statistics, the contact elastic mechanics, the contact plastic mechanics and the abrasive cutting theory. Specially, the effect of the grain size on material removal depth is considered. The geometry property of abrasive grain (such as maximum diameter, average diameter, grain ratio and so on) was charactered by the grit number and the structure number, and the relation between grits and material removal depth was revealed. Algorithm about cutter location data for cylindrical coordinate machine tool polishing axissymmetric surface of wheel is discussed. Tool path is multi end spiral.Multi-axis polishing is confined to the constant polishing area of workpiece.The generatrix is interpolated two times.One is for discrete the generatrix with line segment,and the other is for interpolating the profile of circular truncated cone.Modeling the relation between surface roughness and process parameters, a PSO-based multilayer perceptron algorithm is used, which is based on PSO training the connection weights and the parameters of the Sigmoid activation function. The advantage of PSO and MLP is combined then the relation between process parameters (such as normal polishing force, equivalent radius, feed rate, polishing tool's tangent velocity) is modeled.To determine process parameters using expert system when material removal depth is constant and surface roughness is limited between two known numbers, the expert system architecture is built, which is based on modeling polishing process. The system is made of global database, process database, inference engine, knowledge database, and so on. Acquiring polishing process knowledge, representing them, principles of inference engine, designing all kinds of tables and function modules are discussed. Finally, width-first algorithm and forward reasoning method is used when process parameters are needed adjusting.