基于归一化小波能量系数的管道安全识别算法

针对管道泄漏检测系统存在一定误报率的问题，提出一种基于归一化小波能量系数的油气管道安全识别算法．首先提出最佳母小波选择标准，以现场数据为基础选取最佳母小波，然后提取归一化小波能量系数的最大值作为特征量，最后利用Fisher分类器对压力数据进行识别，判断管道是否处于安全状况．利用现场数据对该方法进行验证分析，结果表明该算法识别正确率较高，实用性好，且可以降低误报率．

Recognition Algorithm Based on Normalized Wavelet Energy Coefficients for the Safety of Pipeline

A recognition algorithm based on normalized wavelet energy coefficients for the safety of pipe- line was proposed to solve the problem of false alarm rate existing in the pipeline leak detection system. First, the best mother wavelet selection criteria were put forward. Then the best mother wavelet was se- lected based on field data, and the maximum normalized wavelet energy coefficient was extracted as fea- ture vector. Finally, the feature vector was identified by the Fisher classifier to determine whether the pipeline is in a safe condition. Experimental data obtained from the oil field were used to evaluate the method. The results show that it is an effective recognition algorithm for reducing false alarm rate, with high accuracy and good real-time performance.