论文部分内容阅读
Aiming at some problems of slag detection system (SDS) based on vibration detection method,such as many interference source,high signal separation difficulty and control time delay etc,a kind of steel water continuous casting SDS realization method based on pattern recognition was put forward.By deep investigation for continuous casting production process and its operation characteristic,continuous casting production status model was established,and dynamic recognition theory oriented to slag detection was discussed.According to analysis outcome of field measurement data,the WP+VQ slag pattern recognition method was brought forward.Using wavelet package (WP) analysis technology,large quantity of preprocessed field experiment data was extracted to constitute the characteristic vector,system knowledge study and rule training were realized to judge the status of steel water by vector quantization (VQ) technology.Combined the analysis of vibration signal variation tendency of slag physical process,its yielding generality was handled by parameter optimization method,and the systems recognition scope and effectivity were guaranteed.Application instance experiments prove that this system can successfully judge some kinds of typical casting status,has the capability of advance warning,and can satisfy the continuous casting production demand.