الفهرس | Only 14 pages are availabe for public view |
Abstract ding processes is one of the most popular means in the machining of parts, particularly in fabricating products from brittle materials such as ceramics and composites. When fine surface finish and high geometrical accuracy are required, grinding appears to be the most practical and economical means. In the finishing machining operations, grinding determines the major portion of the machining cost. This research introduces an artificial intelligent system for monitoring and controlling surface grinding and dressing operations. The aim of this system is to achieve the desired workpiece surface roughness and to reduce the production cycle under grinding wheel surface topography variations. Multi sensor system is used for monitoring. A high response pressure sensor (HRPS) is used for monitoring of wheel topography condition. A cutting force dynamometer is used for monitoring grinding force components. The core of the system consists of two multi-layer feed forward artificial neural networks based on back error propagation learning algorithm. The first one is used for process design to achieve the desired surface roughness. It extracts the suitable process variables such as grinding wheel speed and feed rate. The second one monitors the cutting operation using the sensors readings. It extracts the different controlling decisions; accept the process, redesign the process and start dressing operation. According to these decisions, a PC master control program generates the appropriate control codes and sends them to the machine controllers to take the required actions. The neural network efficiency can be enhanced by continuous learning, adding new learning features. A special hybrid neural network is established for detecting wheel-workpiece contact event. In addition, an automatic dressing system is developed to conduct the dressing operation whenever it is needed. The system testing assessment shows its ability to keep the workpiece surface roughness within the desired range by selecting different proper process variables during the grinding wheel life. The desired workpiece surface roughness is achieved with an acceptable tolerance. The accuracy of wheel-workpiece contact detection has been improved. In addition, the grinding cycle time has been decreased due to using the wheel direct approach rather than the conventional reciprocating approach. The automated dressing system decreases the total process cycle time and increases the role of automation in the system. The developed system can be used for conducting researches for different types of grinding and dressing operations. The neural networks efficiency can be enhanced by continuous learning for new examples. |