Multi-source multi-class classification methods based on multi-class Support Vector Machines and data fusion strategies are proposed in this paper. The centralized and distributed fusion schemes are applied to combine...Multi-source multi-class classification methods based on multi-class Support Vector Machines and data fusion strategies are proposed in this paper. The centralized and distributed fusion schemes are applied to combine information from several data sources. In the centralized scheme, all information from several data sources is centralized to construct an input space. Then a multi-class Support Vector Machine classifier is trained. In the distributed schemes, the individual data sources are proc-essed separately and modelled by using the multi-class Support Vector Machine. Then new data fusion strategies are proposed to combine the information from the individual multi-class Support Vector Machine models. Our proposed fusion strategies take into account that an Support Vector Machine (SVM) classifier achieves classification by finding the optimal classification hyperplane with maximal margin. The proposed methods are applied for fault diagnosis of a diesel engine. The experimental results showed that almost all the proposed approaches can largely improve the diagnostic accuracy. The robustness of diagnosis is also improved because of the implementation of data fusion strategies. The proposed methods can also be applied in other fields.展开更多
Fuel injectors are considered as an important component of combustion engines. Operational weakness can possibly lead to the complete machine malfunction, decreasing reliability and leading to loss of production. To o...Fuel injectors are considered as an important component of combustion engines. Operational weakness can possibly lead to the complete machine malfunction, decreasing reliability and leading to loss of production. To overcome these circumstances, various condition monitoring techniques can be applied. The application of acoustic signals is common in the field of fault diagnosis of rotating machinery. Advanced signal processing is utilized for the construction of features that are specialized in detecting fuel injector faults. A performance comparison between novelty detection algorithms in the form of one-class classifiers is presented. The one-class classifiers that were tested included One-Class Support Vector Machine (OCSVM) and One-Class Self Organizing Map (OCSOM). The acoustic signals of fuel injectors in different operational conditions were processed for feature extraction. Features from all the signals were used as input to the one-class classifiers. The one-class classifiers were trained only with healthy fuel injector conditions and compared with new experimental data which belonged to different operational conditions that were not included in the training set so as to contribute to generalization. The results present the effectiveness of one-class classifiers for detecting faults in fuel injectors.展开更多
The recent deep prospecting results in the Jiaojia area of Eastern Shandong Province indicate that the Jiaojia ore field composed of several individual gold deposits as previously suggested is actually an ultra-large ...The recent deep prospecting results in the Jiaojia area of Eastern Shandong Province indicate that the Jiaojia ore field composed of several individual gold deposits as previously suggested is actually an ultra-large gold deposit.This deposit covers an area of ～40 km2,and shows a structural control by the Jiaojia fault and its secondary faults.Gold orebodies generally occur along the same mineralization-alteration belt,and the main orebodies intersect with each other or exhibit a parallel or overlapping distribution.This deposit＇s reserves are estimated to be 1,200t of gold,being the first gold deposit of more than 1000t gold reserves in China.The No.Ⅰ-1 orebody in the Shaling-deep Sizhuang ore blocks holds gold reserves greater than 350 t,or 29 percent of the total reserves,followed by the No.Ⅰ orebody in Matang-Jiaojia ore blocks with exceeding 150t gold reserves.This deposit mainly occurs in the footwall of the Jiaojia fault,and presents zoned patterns in mineralization,alteration and structures.The strongly mineralized zones agree with strongly altered and tectonically fractured zones.These orebodies display strataform-like,veinlike or lenticular shapes,and generally show characteristics of pinching out and reappearing,branching and converging,expanding and shrinking.The orebodies commonly occur along positions where the fault strike changes and in gentle locations with dips changing from steep to gentle.The main orebodies are parallel to the main plane of the orecontrolling fault,and tend to be gentle from the surface to the deep.The orebodies mainly plunge to the southwest,with plunge angle of 45°-606° Orebodies near the main plane of the ore-controlling fault have more gold resource than those away from main fault zone.The slant depth of orebodies is generally larger than the length along its strike direction; orebodies become thick and gold grades become low from the shallow area to the deep area.Ore-forming fluids are H2O-CO2-NaCl±CH4 type with medium-temperature and moderate to low salinity.Sulfur isotopic values （δ34SCDT） for gold ores range between 11.08‰ and 12.58‰,indicating mixed sulfur sources; hydrogen isotopic values （δDVSMOW） range from-83.68‰ to-116.95‰ and oxygen isotopic values （δ18OV-SMOW） range between 12.04‰ and 16.28‰.The hydrogen and oxygen isotopes suggest that ore-forming fluids originated from primary magma,and mixing with a large amount of atmospheric water during the late stage.The Eastern Shandong Province gold deposits are associated with magmatic activities which have mantle crust-mixed source,and also share some similarities with orgenic and epithermal hydrothermal golddeposits.Because Eastern Shandong Province gold deposits with unique metailogenic features and formation setting which are different from other gold deposit types in the world,we call it the Jiaojiatype gold deposits.The kiloton class Jiaojia gold deposit is related to fluid activities,extension and detachment resulted from thermal upweiling of magmas.The strong magmatic activities in the middle to late stage of early Cretaceous in Eastern Shandong Province lead to active fluids,and provided abundant ore-forming materials for gold depsoits.Moreover,many extensional structures resulting from crustal extension provided favourable space for orebody positioning.展开更多
针对行星滚柱丝杠(planetary roller screw mechanism,PRSM)在实际应用中故障机理不明和故障种类少,难以有效进行故障决策这一现存问题,提出采用单分类模型——深度支持向量数据描述(deep support vector data description,deep SVDD)...针对行星滚柱丝杠(planetary roller screw mechanism,PRSM)在实际应用中故障机理不明和故障种类少,难以有效进行故障决策这一现存问题,提出采用单分类模型——深度支持向量数据描述(deep support vector data description,deep SVDD)进行故障检测,判断PRSM是否处于正常状态。首先,在PRSM试验台上采集正常状态、润滑失效和滚柱一侧断齿3种状态的振动信号;其次,对数据进行归一化并通过窗口裁剪的方式进行数据增强,以扩充样本数量;然后,通过小波包变换对信号进行分解,以初步提取数据的特征;最后,利用deep SVDD实现PRSM故障检测,同时与单分类支持向量机(one-class support vector machine,OCSVM)和支持向量数据描述(support vector data description,SVDD)方法进行对比,结果表明,deep SVDD具有更好的分类能力和较高的训练效率,较为适合实现PRSM故障检测。展开更多
文摘Multi-source multi-class classification methods based on multi-class Support Vector Machines and data fusion strategies are proposed in this paper. The centralized and distributed fusion schemes are applied to combine information from several data sources. In the centralized scheme, all information from several data sources is centralized to construct an input space. Then a multi-class Support Vector Machine classifier is trained. In the distributed schemes, the individual data sources are proc-essed separately and modelled by using the multi-class Support Vector Machine. Then new data fusion strategies are proposed to combine the information from the individual multi-class Support Vector Machine models. Our proposed fusion strategies take into account that an Support Vector Machine (SVM) classifier achieves classification by finding the optimal classification hyperplane with maximal margin. The proposed methods are applied for fault diagnosis of a diesel engine. The experimental results showed that almost all the proposed approaches can largely improve the diagnostic accuracy. The robustness of diagnosis is also improved because of the implementation of data fusion strategies. The proposed methods can also be applied in other fields.
文摘Fuel injectors are considered as an important component of combustion engines. Operational weakness can possibly lead to the complete machine malfunction, decreasing reliability and leading to loss of production. To overcome these circumstances, various condition monitoring techniques can be applied. The application of acoustic signals is common in the field of fault diagnosis of rotating machinery. Advanced signal processing is utilized for the construction of features that are specialized in detecting fuel injector faults. A performance comparison between novelty detection algorithms in the form of one-class classifiers is presented. The one-class classifiers that were tested included One-Class Support Vector Machine (OCSVM) and One-Class Self Organizing Map (OCSOM). The acoustic signals of fuel injectors in different operational conditions were processed for feature extraction. Features from all the signals were used as input to the one-class classifiers. The one-class classifiers were trained only with healthy fuel injector conditions and compared with new experimental data which belonged to different operational conditions that were not included in the training set so as to contribute to generalization. The results present the effectiveness of one-class classifiers for detecting faults in fuel injectors.
基金financially supported by the National Natural Science Foundation of China (Grant No. 41230311)the National Science and Technology Support Program (Grant No.2011BAB04B09)
文摘The recent deep prospecting results in the Jiaojia area of Eastern Shandong Province indicate that the Jiaojia ore field composed of several individual gold deposits as previously suggested is actually an ultra-large gold deposit.This deposit covers an area of ～40 km2,and shows a structural control by the Jiaojia fault and its secondary faults.Gold orebodies generally occur along the same mineralization-alteration belt,and the main orebodies intersect with each other or exhibit a parallel or overlapping distribution.This deposit＇s reserves are estimated to be 1,200t of gold,being the first gold deposit of more than 1000t gold reserves in China.The No.Ⅰ-1 orebody in the Shaling-deep Sizhuang ore blocks holds gold reserves greater than 350 t,or 29 percent of the total reserves,followed by the No.Ⅰ orebody in Matang-Jiaojia ore blocks with exceeding 150t gold reserves.This deposit mainly occurs in the footwall of the Jiaojia fault,and presents zoned patterns in mineralization,alteration and structures.The strongly mineralized zones agree with strongly altered and tectonically fractured zones.These orebodies display strataform-like,veinlike or lenticular shapes,and generally show characteristics of pinching out and reappearing,branching and converging,expanding and shrinking.The orebodies commonly occur along positions where the fault strike changes and in gentle locations with dips changing from steep to gentle.The main orebodies are parallel to the main plane of the orecontrolling fault,and tend to be gentle from the surface to the deep.The orebodies mainly plunge to the southwest,with plunge angle of 45°-606° Orebodies near the main plane of the ore-controlling fault have more gold resource than those away from main fault zone.The slant depth of orebodies is generally larger than the length along its strike direction; orebodies become thick and gold grades become low from the shallow area to the deep area.Ore-forming fluids are H2O-CO2-NaCl±CH4 type with medium-temperature and moderate to low salinity.Sulfur isotopic values （δ34SCDT） for gold ores range between 11.08‰ and 12.58‰,indicating mixed sulfur sources; hydrogen isotopic values （δDVSMOW） range from-83.68‰ to-116.95‰ and oxygen isotopic values （δ18OV-SMOW） range between 12.04‰ and 16.28‰.The hydrogen and oxygen isotopes suggest that ore-forming fluids originated from primary magma,and mixing with a large amount of atmospheric water during the late stage.The Eastern Shandong Province gold deposits are associated with magmatic activities which have mantle crust-mixed source,and also share some similarities with orgenic and epithermal hydrothermal golddeposits.Because Eastern Shandong Province gold deposits with unique metailogenic features and formation setting which are different from other gold deposit types in the world,we call it the Jiaojiatype gold deposits.The kiloton class Jiaojia gold deposit is related to fluid activities,extension and detachment resulted from thermal upweiling of magmas.The strong magmatic activities in the middle to late stage of early Cretaceous in Eastern Shandong Province lead to active fluids,and provided abundant ore-forming materials for gold depsoits.Moreover,many extensional structures resulting from crustal extension provided favourable space for orebody positioning.
文摘针对行星滚柱丝杠(planetary roller screw mechanism,PRSM)在实际应用中故障机理不明和故障种类少,难以有效进行故障决策这一现存问题,提出采用单分类模型——深度支持向量数据描述(deep support vector data description,deep SVDD)进行故障检测,判断PRSM是否处于正常状态。首先,在PRSM试验台上采集正常状态、润滑失效和滚柱一侧断齿3种状态的振动信号;其次,对数据进行归一化并通过窗口裁剪的方式进行数据增强,以扩充样本数量;然后,通过小波包变换对信号进行分解,以初步提取数据的特征;最后,利用deep SVDD实现PRSM故障检测,同时与单分类支持向量机(one-class support vector machine,OCSVM)和支持向量数据描述(support vector data description,SVDD)方法进行对比,结果表明,deep SVDD具有更好的分类能力和较高的训练效率,较为适合实现PRSM故障检测。
文摘深度学习近年来在故障诊断领域受到广泛应用,但基于深度学习的故障诊断模型缺乏工程上的物理解释性,难以保证其故障诊断结果的稳定性。以轴承为例,建立了以小波时频图像为故障诊断依据的卷积神经网络模型(convolutional neural network,CNN),提出了一种基于梯度加权类激活热力图(gradient-weighted class activation map,Grad-CAM)的网络模型鲁棒性分析方法,并利用美国凯斯西储大学(Case Western Reserve University,CWRU)轴承数据集进行验证。首先,将故障直径轴承数据以不同方式混合并训练大、小多个模型。其次,利用Grad-CAM方法,建立时频区域与故障模式之间的联系。最后,利用其他工况下的轴承故障数据,以及含噪数据进行测试,并根据结果结合模型最注重的时频区域进行分析。结果表明,基于深度学习的轴承故障诊断模型在参数较少时更加注重低频区域,并能使其具有更好的鲁棒性。
