An Incorrect Data Detection Method for Big Data Cleaning of Machinery Condition Monitoring

Condition based maintenance (CBM) improves decision-making performances for a maintenance program through machinery condition monitoring. Therefore, it is a key step to trace machinery health condition for CBM. In this paper, a novel method is proposed to establish a health evaluation index named automatic evaluation index (AEI) and its corresponding dynamic threshold using Wavelet Packet Transform (WPT) and Hidden Markolv Model (HMM). In this process, WPT is used to decompose signal into detail signals and exhibits prominent gear fault features. In addition, HMM employed here is to recognize two concerned states of gear in the whole life validation, including normal gear state and early gear fault state. It is also important to build a dynamic threshold to differentiate the two states automatically. The proposed dynamic threshold not only renews by itself according to the history values of AEI but also easily and automatically detects occurrence of gear early fault. Finally, a set of whole life time data ending in gear failure is used to verify the proposed method effectively. Further, some related parameters included in this method are discussed and the obtained results show that condition monitoring performance of the proposed method is excellent in detection of gear failure.

Download Full-text

A Novel Big Data Cleaning Algorithm Based On Edge Computing In Industrial Internet of Things

10.1109/ctisc52352.2021.00043 ◽

2021 ◽

Author(s):

Mao Chen ◽

Wen Chen ◽

Yongqi Zhu

Keyword(s):

Big Data ◽

Internet Of Things ◽

Data Cleaning ◽

Edge Computing ◽

Industrial Internet Of Things ◽

Industrial Internet

Download Full-text

IoT Based Decision Making System to Improve Veracity of Big Data

International Journal of Engineering & Technology ◽

10.14419/ijet.v7i3.1.16799 ◽

2018 ◽

Vol 7 (3.1) ◽

pp. 63 ◽

Cited By ~ 1

Author(s):

R Revathy ◽

R Aroul Canessane

Keyword(s):

Decision Making ◽

Big Data ◽

Internet Of Things ◽

Data Streams ◽

Data Cleaning ◽

Data Handling ◽

Secure Data ◽

Decision Making System ◽

Data Investigation ◽

Impact Data

Data are vital to help decision making. On the off chance that data have low veracity, choices are not liable to be sound. Internet of Things (IoT) quality rates big data with error, irregularity, deficiency, trickery, and model guess. Improving data veracity is critical to address these difficulties. In this article, we condense the key qualities and difficulties of IoT, which impact data handling and decision making. We audit the scene of estimating and upgrading data veracity and mining indeterminate data streams. Also, we propose five suggestions for future advancement of veracious big IoT data investigation that are identified with the heterogeneous and appropriated nature of IoT data, self-governing basic leadership, setting mindful and area streamlined philosophies, data cleaning and handling procedures for IoT edge gadgets, and protection safeguarding, customized, and secure data administration.

Download Full-text