Dissertation > Excellent graduate degree dissertation topics show

Study on the Straddle-type Monorail Cast Steel Pedestal Health Status Detection Based on Vibration Test Technology

Author: LiuJiaZuo
Tutor: WangTongQing
School: Chongqing University
Course: Instrument Science and Technology
Keywords: BEARING A class support vector machine Kernel Principal Component Analysis Fault Diagnosis
CLC: U232
Type: Master's thesis
Year: 2011
Downloads: 23
Quote: 0
Read: Download Dissertation

Abstract

BEARING thesis Chongqing monorail transport system as the research object. BEARING is the connection track beams and pier key force parts, plays a vital role in the health status of the safe operation of the entire monorail transport system. Due to manufacturing defects, corrosion of the metal components of the climate and environment, long-serving and complexity of the forces and other factors will have a huge negative impact on the health status of the BEARING. The monorail transportation security operations must BEARING health status monitoring. BEARING, however, generally elevated above the outdoor track pier in a unique position, and its larger, complex shape, using conventional ray detection, ultrasonic detection, microwave detection and non-destructive testing methods are difficult to carry out their health status monitoring. Pillar Program of the National Science and Technology the monorail transportation equipment R \u0026 D \state of health. First IPC, hammer, vibration sensor and data acquisition card and write the software constitutes the vibration data acquisition system suitable for the actual situation of the project site; then several experiments to determine the vibration sensor above BEARING the arrangement and excitation positions; the manual force Finally BEARING hammering vibrator and vibration impulse response signal acquisition BEARING. Lack of fault samples for the the BEARING system fault diagnosis problem, analysis of the vibration pulse collected BEARING response signal was first proposed, the paper-based class support vector machine (One-class SVM) fault diagnostic methods. The method requires only the normal data samples can establish a single value classifier to distinguish normal samples and non-normal samples. Test kernel principal component analysis (KPCA) the collected vibration pulse response signal decomposition, will be extracted to the winner of the element characteristic input to the One-class SVM classifier for training and testing, the classifier in the case of the appropriate sacrifice injustice detection rate ensure no undetected circumstances, diagnostic systems, the accuracy was 98%. The experimental results show that the One-class SVM classifier can accurately identify a fault samples, while the use of the extracted characteristic KPCA exploded fault information can be effectively concentrated, the One-class SVM classifier classified better computational efficiency. The research results show that the system uses the programs and methods used in the detection system is feasible and correct the performance and indicators have reached the expected requirements. The above methods proposed for the health detection BEARING provide scientific reference, has a certain reference value.

Study on the Straddle-type Monorail Cast Steel Pedestal Health Status Detection Based on Vibration Test Technology

Abstract

Related Dissertations