Existing methods for transformer fault diagnosis either train a classifier to fit the relationship between dissolved gas and fault type or find some similar cases with unknown samples by calculating the similarity met...Existing methods for transformer fault diagnosis either train a classifier to fit the relationship between dissolved gas and fault type or find some similar cases with unknown samples by calculating the similarity metrics.Their accuracy is limited,since they are hard to learn from other algorithms to improve their own performance.To improve the accuracy of transformer fault diagnosis,a novel method for transformer fault diagnosis based on graph convolutional network(GCN)is proposed.The proposed method has the advantages of two kinds of existing methods.Specifically,the adjacency matrix of GCN is utilized to fully represent the similarity metrics between unknown samples and labeled samples.Furthermore,the graph convolutional layers with strong feature extraction ability are used as a classifier to find the complex nonlinear relationship between dissolved gas and fault type.The back propagation algorithm is used to complete the training process of GCN.The simulation results show that the performance of GCN is better than that of the existing methods such as convolutional neural network,multi-layer perceptron,support vector machine,extreme gradient boosting tree,k-nearest neighbors and Siamese network in different input features and data volumes,which can effectively meet the needs of diagnostic accuracy.展开更多
A procedure to recognize individual discontinuities in rock mass from measurement while drilling(MWD)technology is developed,using the binary pattern of structural rock characteristics obtained from in-hole images for...A procedure to recognize individual discontinuities in rock mass from measurement while drilling(MWD)technology is developed,using the binary pattern of structural rock characteristics obtained from in-hole images for calibration.Data from two underground operations with different drilling technology and different rock mass characteristics are considered,which generalizes the application of the methodology to different sites and ensures the full operational integration of MWD data analysis.Two approaches are followed for site-specific structural model building:a discontinuity index(DI)built from variations in MWD parameters,and a machine learning(ML)classifier as function of the drilling parameters and their variability.The prediction ability of the models is quantitatively assessed as the rate of recognition of discontinuities observed in borehole logs.Differences between the parameters involved in the models for each site,and differences in their weights,highlight the site-dependence of the resulting models.The ML approach offers better performance than the classical DI,with recognition rates in the range 89%to 96%.However,the simpler DI still yields fairly accurate results,with recognition rates 70%to 90%.These results validate the adaptive MWD-based methodology as an engineering solution to predict rock structural condition in underground mining operations.展开更多
在分析现有程序代码抄袭检测系统的特点及局限性的基础上,提出一种综合文本分析、结构度量和属性计数技术的混合式程序抄袭检测方法。应用文档指纹技术和Winnowing算法计算程序的文本相似度;将程序代码表示成动态控制结构树(Dynamic Con...在分析现有程序代码抄袭检测系统的特点及局限性的基础上,提出一种综合文本分析、结构度量和属性计数技术的混合式程序抄袭检测方法。应用文档指纹技术和Winnowing算法计算程序的文本相似度;将程序代码表示成动态控制结构树(Dynamic Control Structure tree,DCS),运用Winnowing算法计算DCS树相似度,从而得到结构相似度;收集并统计程序中的每个变量信息,应用变量相似度算法分析变量信息节点获取变量相似度;分别赋予文本相似度、结构相似度和变量相似度一个权值,计算得到总体的代码相似度。实验结果表明,所提出的方法能够有效检测出各种抄袭行为。针对不同的抄袭门槛值,使用该方法的检测结果准确度和查全率高于JPLAG系统。特别对于结构简单的程序组,此方法和JPLAG系统检测结果的平均准确度分别为82.5%和69.5%,说明所提的方法更加有效。展开更多
基金This manuscript is supported by the China Scholarship Council.
文摘Existing methods for transformer fault diagnosis either train a classifier to fit the relationship between dissolved gas and fault type or find some similar cases with unknown samples by calculating the similarity metrics.Their accuracy is limited,since they are hard to learn from other algorithms to improve their own performance.To improve the accuracy of transformer fault diagnosis,a novel method for transformer fault diagnosis based on graph convolutional network(GCN)is proposed.The proposed method has the advantages of two kinds of existing methods.Specifically,the adjacency matrix of GCN is utilized to fully represent the similarity metrics between unknown samples and labeled samples.Furthermore,the graph convolutional layers with strong feature extraction ability are used as a classifier to find the complex nonlinear relationship between dissolved gas and fault type.The back propagation algorithm is used to complete the training process of GCN.The simulation results show that the performance of GCN is better than that of the existing methods such as convolutional neural network,multi-layer perceptron,support vector machine,extreme gradient boosting tree,k-nearest neighbors and Siamese network in different input features and data volumes,which can effectively meet the needs of diagnostic accuracy.
基金conducted under the illu MINEation project, funded by the European Union’s Horizon 2020 research and innovation program under grant agreement (No. 869379)supported by the China Scholarship Council (No. 202006370006)
文摘A procedure to recognize individual discontinuities in rock mass from measurement while drilling(MWD)technology is developed,using the binary pattern of structural rock characteristics obtained from in-hole images for calibration.Data from two underground operations with different drilling technology and different rock mass characteristics are considered,which generalizes the application of the methodology to different sites and ensures the full operational integration of MWD data analysis.Two approaches are followed for site-specific structural model building:a discontinuity index(DI)built from variations in MWD parameters,and a machine learning(ML)classifier as function of the drilling parameters and their variability.The prediction ability of the models is quantitatively assessed as the rate of recognition of discontinuities observed in borehole logs.Differences between the parameters involved in the models for each site,and differences in their weights,highlight the site-dependence of the resulting models.The ML approach offers better performance than the classical DI,with recognition rates in the range 89%to 96%.However,the simpler DI still yields fairly accurate results,with recognition rates 70%to 90%.These results validate the adaptive MWD-based methodology as an engineering solution to predict rock structural condition in underground mining operations.
文摘在分析现有程序代码抄袭检测系统的特点及局限性的基础上,提出一种综合文本分析、结构度量和属性计数技术的混合式程序抄袭检测方法。应用文档指纹技术和Winnowing算法计算程序的文本相似度;将程序代码表示成动态控制结构树(Dynamic Control Structure tree,DCS),运用Winnowing算法计算DCS树相似度,从而得到结构相似度;收集并统计程序中的每个变量信息,应用变量相似度算法分析变量信息节点获取变量相似度;分别赋予文本相似度、结构相似度和变量相似度一个权值,计算得到总体的代码相似度。实验结果表明,所提出的方法能够有效检测出各种抄袭行为。针对不同的抄袭门槛值,使用该方法的检测结果准确度和查全率高于JPLAG系统。特别对于结构简单的程序组,此方法和JPLAG系统检测结果的平均准确度分别为82.5%和69.5%,说明所提的方法更加有效。
