Diesel Engine Gearbox Fault Diagnosis Based on Multi-features Extracted from Vibration Signals

The vibration-based feature extraction is the main approach for mechanical fault diagnosis,whereas,in some conditions vibration signal is not easily measured because of its contact-measuring.Acoustic-based diagnosis(ABD) can overcome this disadvantage.However,for traditional ABD it is hard to choose proper measuring positions and the acoustic signals acquired based on single channel measurement can be used only for local analysis.Based on near-field acoustic holography(NAH),a new feature extraction method by using sound field distribution for rolling bearing fault diagnosis is presented.Firstly,sound fields in different bearing conditions are reconstructed by NAH.Using gray level co-occurrence matrix(GLCM) features extracted from acoustic images,the inner relationship between bearing conditions and sound fields is established.These features are fed into support vector machine(SVM) classifier for fault diagnosis.The effectiveness of our proposed method is demonstrated on the experimental investigation.The method provides a new reference for mechanical fault diagnosis.

Diagnosing Rolling Bearing Faults Using Spatial Distribution Features of Sound Field

The invention discloses a method for detecting a non-roundness performance of a vehicle of an urban rail vehicle The method comprises: a vibration acceleration sensor is installed at the outer side of a track below an orbital plane to collect a vibration signal of a vehicle wheel; an envelope entropy is selected as a fitness function and multi-times iteration is carried out based on a particle swarm algorithm to obtain optimal values of a component number K and a penalty parameter alpha; the optimal K and alpha are substituted into a VMD algorithm and a vibration signal corresponding to a single wheel is decomposed to obtain an IMF component; three IMF components with the smallest envelope entropies are selected to calculate five feature quantities: a peakedness value, a root-mean-squarevalue, an energy value, an energy moment and an information entropy; with a Gaussian kernel function and a directed acyclic graph classification method, a weight coefficient of the Gaussian kernel function is calculated by using kernel polarization and a multi-core LS-SVM model is constructed; and wheel sample classification is carried out based on the feature quantity of the vibration signal, sothat a non-roundness result of the vehicle wheel is obtained According to the invention, online non-contact detection is realized; and the method has advantages of high speed, high accuracy and strong anti-interference

Method for detecting non-roundness performance of vehicle of urban rail vehicle

Aiming at the problem of fault diagnosis and safety region boundary estimation of rolling bearing in metro vehicle, A method combined empirical mode decomposition (EMD) with support vector machine (SVM) was proposed. Firstly, The intrinsic mode functions (IMF) containing fault information was obtained by EMD decomposition based on the collected vibration data. Then, RMS value, energy, Shannon entropy and energy moments were calculated as characteristic index of roller bearing. Next the safety region boundary was demarcated by LSSVM and fault classification of bearings were identified by DAGSVM. The experiment results indicated that the proposed method could identify the working status and fault type of metro roller bearings accurately and effectively.

Fault diagnosis and safety region estimation based on SVM and EMD of roller bearing in metro vehicle

In this paper, based on the correlation matrix, the fault detection method of controller is studied based on the fault diagnosis characteristics of environment control and life support system(ECLSS). Combined with the characteristics of actual system fault diagnosis, the reasons for the existence of the feedback loop are analyzed, the different diagnostic effects caused by voltage and current detection methods are compared, and the detection method of breaking the feedback loop by manual testing is proposed. Through modeling and analysis, complete design ideas and methods are provided for improving testability design of controller products, which can help guide Built-In-Test(BIT) design of controllers. 

Wei Dong

Papers

Fault diagnosis and safety region estimation based on SVM and EMD of roller bearing in metro vehicle

Research on a system controller fault detection method based on correlation matrix