WEKO3
アイテム
{"_buckets": {"deposit": "8deb4267-465a-4375-8c00-4d4df43faec4"}, "_deposit": {"created_by": 16, "id": "10683", "owners": [16], "pid": {"revision_id": 0, "type": "depid", "value": "10683"}, "status": "published"}, "_oai": {"id": "oai:mie-u.repo.nii.ac.jp:00010683", "sets": ["712"]}, "author_link": [], "item_7_biblio_info_6": {"attribute_name": "書誌情報", "attribute_value_mlt": [{"bibliographicIssueDates": {"bibliographicIssueDate": "2014-01-01", "bibliographicIssueDateType": "Issued"}}]}, "item_7_date_granted_59": {"attribute_name": "学位授与年月日", "attribute_value_mlt": [{"subitem_dategranted": "2014-03-25"}]}, "item_7_degree_grantor_57": {"attribute_name": "学位授与機関", "attribute_value_mlt": [{"subitem_degreegrantor": [{"subitem_degreegrantor_name": "三重大学"}], "subitem_degreegrantor_identifier": [{"subitem_degreegrantor_identifier_name": "14101", "subitem_degreegrantor_identifier_scheme": "kakenhi"}]}]}, "item_7_degree_name_56": {"attribute_name": "学位名", "attribute_value_mlt": [{"subitem_degreename": "博士(学術)"}]}, "item_7_description_14": {"attribute_name": "フォーマット", "attribute_value_mlt": [{"subitem_description": "application/pdf", "subitem_description_type": "Other"}]}, "item_7_description_4": {"attribute_name": "抄録", "attribute_value_mlt": [{"subitem_description": "Condition diagnosis plays a significant role in modern equipment management. Establishment of an intelligent system of condition diagnosis not only improves the productivity, but also reduces the cost of maintenance and the risk of unexpected failures. In the field of condition diagnosis of the plant machinery, particularly rotating machinery, vibration information is widely used to detect a fault and identify the fault types. Certainly, condition diagnosis based on vibration information depends largely on feature extraction. Only when the features of vibration information are sensitively extracted in any condition of a machine can condition diagnosis be effective. However, plant machines are operating under unsteady conditions, even if the machines are in the normal state, rotating speed and operating load can vary. These can influence the spectrum feature of the vibration information measured. Moreover, vibration information often contain strong noise, especially at an early stage of a fault. Therefore, it is difficult to extract the features of vibration information. In addition, when building an intelligent system fordiagnosing the condition of plant machinery, symptom parameters (SPs) and artificial intelligence (AI) are required. A high sensitive SP can express features information of machine conditions. However, in most cases of condition diagnosis for plant machinery, the sensitivity of some SPs is not high. The main reasons can be explained as follow: (1) the pure features are ambiguous in vibration information, because a fault is at an early stage, or the measure point is farther from failure part; (2) previous work namely feature extraction is unsatisfied; (3) the selected SPs cannot sensitively reflect the conditions of monitoring machine. In the case of AI, neural network (NN), genetic algorithm (GA), support vector machine (SVM), etc. have some special advantages as well as some disadvantages. For example, NN and GA will never converge when the first-layer data has the same values in different states, and SVM is only a two-class classifier. Moreover, these methods cannot deal with ambiguous classification problems. In order to extract effectively the features of vibration information, improve the ambiguous relationship between SPs and machine conditions, strengthen the sensitivity of SPs, and increase the efficiency of condition diagnosis at an early stage, this thesis has focused on studying condition diagnosis based on vibration information and support vector machine for plant machinery. This thesis proposes statistic filter is performed to extract fault features from vibration information measured. Statistic filter is a method of signal processing by statistical tests of spectrums between normal information and measured information. In the field of machinery diagnosis, statistic filter is used to smooth away noise, what’s more, leave fault features unchanged. This thesis defines many dimensional symptom parameters (DSPs) and non-dimensional symptom parameters (NDSPs) in time domain and in frequency domain for automatic diagnosis, which reflect the features of vibration information measured in plant machinery. Moreover, optimal composition of symptom parameters (OCSPs) that is a SPs’ group from multiple directions is proposed to improve the ambiguous relationship between SPs and machine conditions. In addition, distinctive frequency components (DFCs) and relative ratio symptom parameters (RRSPs) are proposed to exclusively detect and identify structural faults of rotating machinery. DFCs are the values of the spectrum features of structural faults, and RRSPs are a new type of SPs defined by the features of structural faults. Finally, discrimination index (DI) and principal component analysis (PCA) are employed to evaluate the sensitivity of DSPs, NDSPs or RRSPs, respectively. Each type of SPs has its good points. In the case of the diagnosis of structural faults of rotating machinery, the sensitivity of RRSPs is much higher, as the capability of DFCs is more effective. This thesis proposes a sequential diagnosis method for monitoring and diagnosing the operating condition of rotating machinery, and the diagnosis approach is constructed on the basis of vibration information and support vector machines (SVMs). DSPs, NDSPs, RRSPs and DFCs are regarded as the objects of a follow-on process, respectively. SVMs are used to establish the sequential diagnosis system for faults detection and the identification of fault types. This thesis presents a fuzzy diagnosis method based on SVMs and possibility theory (PT). Soft margin SVMs are used to merge NDSPs or OCSPs into synthetic symptom parameters (SSPs) aiming to increase the diagnosis’ efficiency. PT is used to convert the probability distribution function of a SSP into a possibility function, and then the possibility function is regarded as the membership function for fuzzy inference. Many practical examples of fault detection and the identification of fault types in rotating machinery are provided to verify that all of the proposed methods are effective.", "subitem_description_language": "en", "subitem_description_type": "Abstract"}]}, "item_7_description_5": {"attribute_name": "内容記述", "attribute_value_mlt": [{"subitem_description": "本文 / Graduate School of Bioresources Mie University", "subitem_description_type": "Other"}, {"subitem_description": "5, 3, 154p", "subitem_description_type": "Other"}]}, "item_7_dissertation_number_60": {"attribute_name": "学位授与番号", "attribute_value_mlt": [{"subitem_dissertationnumber": "甲学術第1677号"}]}, "item_7_publisher_30": {"attribute_name": "出版者", "attribute_value_mlt": [{"subitem_publisher": "三重大学"}]}, "item_7_text_65": {"attribute_name": "資源タイプ(三重大)", "attribute_value_mlt": [{"subitem_text_value": "Doctoral Dissertation / 博士論文"}]}, "item_7_version_type_15": {"attribute_name": "著者版フラグ", "attribute_value_mlt": [{"subitem_version_resource": "http://purl.org/coar/version/c_970fb48d4fbd8a85", "subitem_version_type": "VoR"}]}, "item_access_right": {"attribute_name": "アクセス権", "attribute_value_mlt": [{"subitem_access_right": "open access", "subitem_access_right_uri": "http://purl.org/coar/access_right/c_abf2"}]}, "item_creator": {"attribute_name": "著者", "attribute_type": "creator", "attribute_value_mlt": [{"creatorNames": [{"creatorName": "XUE, HONGTAO", "creatorNameLang": "en"}, {"creatorName": "薛, 紅涛", "creatorNameLang": "ja"}]}]}, "item_files": {"attribute_name": "ファイル情報", "attribute_type": "file", "attribute_value_mlt": [{"accessrole": "open_date", "date": [{"dateType": "Available", "dateValue": "2017-02-20"}], "displaytype": "detail", "download_preview_message": "", "file_order": 0, "filename": "2013DB007.pdf", "filesize": [{"value": "2.6 MB"}], "format": "application/pdf", "future_date_message": "", "is_thumbnail": false, "mimetype": "application/pdf", "size": 2600000.0, "url": {"label": "2013DB007.pdf", "objectType": "fulltext", "url": "https://mie-u.repo.nii.ac.jp/record/10683/files/2013DB007.pdf"}, "version_id": "ec49e3ca-683f-4b79-a2f0-b1578d2e54c3"}]}, "item_language": {"attribute_name": "言語", "attribute_value_mlt": [{"subitem_language": "eng"}]}, "item_resource_type": {"attribute_name": "資源タイプ", "attribute_value_mlt": [{"resourcetype": "doctoral thesis", "resourceuri": "http://purl.org/coar/resource_type/c_db06"}]}, "item_title": "Study on Intelligent Condition Diagnosis Based on Vibration", "item_titles": {"attribute_name": "タイトル", "attribute_value_mlt": [{"subitem_title": "Study on Intelligent Condition Diagnosis Based on Vibration", "subitem_title_language": "en"}]}, "item_type_id": "7", "owner": "16", "path": ["712"], "permalink_uri": "http://hdl.handle.net/10076/13915", "pubdate": {"attribute_name": "PubDate", "attribute_value": "2014-06-18"}, "publish_date": "2014-06-18", "publish_status": "0", "recid": "10683", "relation": {}, "relation_version_is_last": true, "title": ["Study on Intelligent Condition Diagnosis Based on Vibration"], "weko_shared_id": -1}
Study on Intelligent Condition Diagnosis Based on Vibration
http://hdl.handle.net/10076/13915
http://hdl.handle.net/10076/13915ed87de4a-467d-4b7d-9957-d11e29ef7588
名前 / ファイル | ライセンス | アクション |
---|---|---|
2013DB007.pdf (2.6 MB)
|
|
Item type | 学位論文 / Thesis or Dissertation(1) | |||||||||
---|---|---|---|---|---|---|---|---|---|---|
公開日 | 2014-06-18 | |||||||||
タイトル | ||||||||||
言語 | en | |||||||||
タイトル | Study on Intelligent Condition Diagnosis Based on Vibration | |||||||||
言語 | ||||||||||
言語 | eng | |||||||||
資源タイプ | ||||||||||
資源タイプ識別子 | http://purl.org/coar/resource_type/c_db06 | |||||||||
資源タイプ | doctoral thesis | |||||||||
アクセス権 | ||||||||||
アクセス権 | open access | |||||||||
アクセス権URI | http://purl.org/coar/access_right/c_abf2 | |||||||||
著者 |
薛, 紅涛
× 薛, 紅涛
|
|||||||||
抄録 | ||||||||||
内容記述タイプ | Abstract | |||||||||
内容記述 | Condition diagnosis plays a significant role in modern equipment management. Establishment of an intelligent system of condition diagnosis not only improves the productivity, but also reduces the cost of maintenance and the risk of unexpected failures. In the field of condition diagnosis of the plant machinery, particularly rotating machinery, vibration information is widely used to detect a fault and identify the fault types. Certainly, condition diagnosis based on vibration information depends largely on feature extraction. Only when the features of vibration information are sensitively extracted in any condition of a machine can condition diagnosis be effective. However, plant machines are operating under unsteady conditions, even if the machines are in the normal state, rotating speed and operating load can vary. These can influence the spectrum feature of the vibration information measured. Moreover, vibration information often contain strong noise, especially at an early stage of a fault. Therefore, it is difficult to extract the features of vibration information. In addition, when building an intelligent system fordiagnosing the condition of plant machinery, symptom parameters (SPs) and artificial intelligence (AI) are required. A high sensitive SP can express features information of machine conditions. However, in most cases of condition diagnosis for plant machinery, the sensitivity of some SPs is not high. The main reasons can be explained as follow: (1) the pure features are ambiguous in vibration information, because a fault is at an early stage, or the measure point is farther from failure part; (2) previous work namely feature extraction is unsatisfied; (3) the selected SPs cannot sensitively reflect the conditions of monitoring machine. In the case of AI, neural network (NN), genetic algorithm (GA), support vector machine (SVM), etc. have some special advantages as well as some disadvantages. For example, NN and GA will never converge when the first-layer data has the same values in different states, and SVM is only a two-class classifier. Moreover, these methods cannot deal with ambiguous classification problems. In order to extract effectively the features of vibration information, improve the ambiguous relationship between SPs and machine conditions, strengthen the sensitivity of SPs, and increase the efficiency of condition diagnosis at an early stage, this thesis has focused on studying condition diagnosis based on vibration information and support vector machine for plant machinery. This thesis proposes statistic filter is performed to extract fault features from vibration information measured. Statistic filter is a method of signal processing by statistical tests of spectrums between normal information and measured information. In the field of machinery diagnosis, statistic filter is used to smooth away noise, what’s more, leave fault features unchanged. This thesis defines many dimensional symptom parameters (DSPs) and non-dimensional symptom parameters (NDSPs) in time domain and in frequency domain for automatic diagnosis, which reflect the features of vibration information measured in plant machinery. Moreover, optimal composition of symptom parameters (OCSPs) that is a SPs’ group from multiple directions is proposed to improve the ambiguous relationship between SPs and machine conditions. In addition, distinctive frequency components (DFCs) and relative ratio symptom parameters (RRSPs) are proposed to exclusively detect and identify structural faults of rotating machinery. DFCs are the values of the spectrum features of structural faults, and RRSPs are a new type of SPs defined by the features of structural faults. Finally, discrimination index (DI) and principal component analysis (PCA) are employed to evaluate the sensitivity of DSPs, NDSPs or RRSPs, respectively. Each type of SPs has its good points. In the case of the diagnosis of structural faults of rotating machinery, the sensitivity of RRSPs is much higher, as the capability of DFCs is more effective. This thesis proposes a sequential diagnosis method for monitoring and diagnosing the operating condition of rotating machinery, and the diagnosis approach is constructed on the basis of vibration information and support vector machines (SVMs). DSPs, NDSPs, RRSPs and DFCs are regarded as the objects of a follow-on process, respectively. SVMs are used to establish the sequential diagnosis system for faults detection and the identification of fault types. This thesis presents a fuzzy diagnosis method based on SVMs and possibility theory (PT). Soft margin SVMs are used to merge NDSPs or OCSPs into synthetic symptom parameters (SSPs) aiming to increase the diagnosis’ efficiency. PT is used to convert the probability distribution function of a SSP into a possibility function, and then the possibility function is regarded as the membership function for fuzzy inference. Many practical examples of fault detection and the identification of fault types in rotating machinery are provided to verify that all of the proposed methods are effective. | |||||||||
言語 | en | |||||||||
内容記述 | ||||||||||
内容記述タイプ | Other | |||||||||
内容記述 | 本文 / Graduate School of Bioresources Mie University | |||||||||
内容記述 | ||||||||||
内容記述タイプ | Other | |||||||||
内容記述 | 5, 3, 154p | |||||||||
書誌情報 |
発行日 2014-01-01 |
|||||||||
フォーマット | ||||||||||
内容記述タイプ | Other | |||||||||
内容記述 | application/pdf | |||||||||
著者版フラグ | ||||||||||
出版タイプ | VoR | |||||||||
出版タイプResource | http://purl.org/coar/version/c_970fb48d4fbd8a85 | |||||||||
出版者 | ||||||||||
出版者 | 三重大学 | |||||||||
学位名 | ||||||||||
学位名 | 博士(学術) | |||||||||
学位授与機関 | ||||||||||
学位授与機関識別子Scheme | kakenhi | |||||||||
学位授与機関識別子 | 14101 | |||||||||
学位授与機関名 | 三重大学 | |||||||||
学位授与年月日 | ||||||||||
学位授与年月日 | 2014-03-25 | |||||||||
学位授与番号 | ||||||||||
学位授与番号 | 甲学術第1677号 | |||||||||
資源タイプ(三重大) | ||||||||||
Doctoral Dissertation / 博士論文 |