Research And Development On Monitoring And Optimizing Of Processing For NC Machines

In the processing of machining, the efficiency of processing and the quality of products will be badly influenced by any faults of machine tools and equipments and the cost of operation will be added. Using the monitoring and optimization to the processing of machining, faults will be avoided and high value of in economic and social area will be shown, especially in the area of improving the efficiency of NC tools.The object of this paper is realizing the monitoring and control of operation and the optimization of the parameters of processing. The theories of monitoring and optimization are discussed and the model of the relationship between the parameters of monitoring and processing is promoted in this paper. The system of on-line condition monitoring based on Artificial Immune Theory and Neural Network is consisted of the following four parts: DAQ module, signal analysis module, monitoring module and optimization module.In signal analysis module, multi-signals processing methods including wavelet analysis are used to extract and integrate the features of the signals of every sensor from DAQ module. For example, power spectrum is used in the frequency domain and variance estimate in the time domain to the vibration signal. Features of sensors are provided for the following monitoring module.Some modification has been modified in the Negative-Selection Algorithm (NSA) in the monitoring module. Clone & Evolution Algorithm based on Genetic Algorithm are applied to NSA. The modified NSA can efficiently collect the recent information and carry out the pattern recognition in the dispersed form, so that faults can be found and transacted with related measures in time. This monitoring module is consists of two parts: one is self-studying, including training and modeling, the other is on-line monitoring.Artificial Neural Network (ANN) is applied to propose theparameters using in machining. In order to satisfy requirements for individual operations, fuzzy ARTMAP neural network is used to implement the continuous learning based on samples of cutting conditions that obtained from previously learning and recommended from a machining data handbook. Those samples could be substituted by better cutting conditions by a proposed replacement algorithm. For the continual improvement of the cutting conditions, the system, which is capable of the evolutionary learning of cutting conditions, is proposed. This is a key component of the operation optimization system. Monitoring and optimization modules for machining are carried out on line simultaneously. Not only the faults monitoring and the faults avoiding, but also the efficiency and effectiveness of machining are considered.The modules are designed and developed in Matlab6.5 and Virtual Instrument-LabVIEW. Experiments are implemented the ordinary lathe and NC lathe and expect results were achieved.