|
|
Comparison of methods for identification of rolling bearing failures
Kokeš, Miroslav ; Hnidka,, Jakub (referee) ; Klusáček, Stanislav (advisor)
The aim of this master thesis is the comparison of selected methods and parameters for roller bearings diagnostics. Selected statistical parameters are kurtosis, crest factor, and parameter K(t). The other selected methods are envelope analysis, cepstral analysis, and ACEP method. These methods are implemented in LabVIEW software and compared based on noise resistance, computation speed, and overall capability of identifying roller bearing faults.
|
|
|
Implementation and practical verification of methods for predictive identification of rolling bearings failures
Bár, Martin ; Havránek, Zdeněk (referee) ; Klusáček, Stanislav (advisor)
The aim of this thesis is to identify and classify rolling bearing failures. The first part of the thesis deals with the diagnosis of bearings using in-house measured data. Faults were introduced into two bearings. The first one was deformed and corroded. In the second bearing, the outer raceway was damaged. Vibration data was collected at regular intervals and processed in MATLAB. The values of the statistical features indicated faults in both bearings. Envelope analysis showed that both bearings had developed a fault on the outer raceway and a gradual fault on the cage. In the second part of this thesis, machine learning methods were used to classify defective bearings using the CWRU data set. The accelerometer data were divided into blocks in two ways. Vibration images were created from these blocks for convolutional neural networks (CNNs). The best prediction accuracy was achieved by 1D convolutional neural network (1DCNN) (99.2 %), followed by neural network (94.6 %) and SVM (94.4 %). Random Forest and SVM are the best methods when the training set is reduced, and among CNNs, MATLAB architecture and 1DCNN are the best. The most noise resistant method is Random Forest and neural network, and among CNNs, 1DCNN is the best. Methods using statistical features perform better than CNNs on extremely noisy data. Convolutional networks do not achieve good accuracy, which is probably due to the conversion of raw accelerometer data into vibration images.
|
|
|
Comparative analyses of muscle activation lower extremities during running at different surface
Král, David ; Bačáková, Radka (advisor) ; Hojka, Vladimír (referee)
Title: Comparative analyses of muscle activation lower extremities during running at different surface. Objectives: The aim of this bachelor's thesis is to compare the level of activation of selected muscles of the lower extremities and the relative timing of these activations. Realize measurements on three types of surface: tartan, grass, sand and find out the differences in activations of selected muscles according to the surface. Methods: In this thesis, I used the method of analysis and the method of comparison. I applied the analysis method in the analysis of the measured signals for the running step and the comparison method in the section comparing average running step cycles from different surface types. Results: I found out that running on the tartan, acitvates all the monitored muscles in more than 75% of the average cycles within 10% of the running step time period. For softer surfaces - grass and sand, my research did not support the hypothesis, that the softer the surface is, the greater are the differences in the activation of individual muscles versus tartan. Further I found out that on sandy and grassy surfaces, the average activation time of the muscles which are more involved in stabilizing the ankle and foot, i.e. tibialis anterior and peroneus longus, increases. On a sandy...
|
|
|
Implementation and practical verification of methods for predictive identification of rolling bearings failures
Bár, Martin ; Havránek, Zdeněk (referee) ; Klusáček, Stanislav (advisor)
The aim of this thesis is to identify and classify rolling bearing failures. The first part of the thesis deals with the diagnosis of bearings using in-house measured data. Faults were introduced into two bearings. The first one was deformed and corroded. In the second bearing, the outer raceway was damaged. Vibration data was collected at regular intervals and processed in MATLAB. The values of the statistical features indicated faults in both bearings. Envelope analysis showed that both bearings had developed a fault on the outer raceway and a gradual fault on the cage. In the second part of this thesis, machine learning methods were used to classify defective bearings using the CWRU data set. The accelerometer data were divided into blocks in two ways. Vibration images were created from these blocks for convolutional neural networks (CNNs). The best prediction accuracy was achieved by 1D convolutional neural network (1DCNN) (99.2 %), followed by neural network (94.6 %) and SVM (94.4 %). Random Forest and SVM are the best methods when the training set is reduced, and among CNNs, MATLAB architecture and 1DCNN are the best. The most noise resistant method is Random Forest and neural network, and among CNNs, 1DCNN is the best. Methods using statistical features perform better than CNNs on extremely noisy data. Convolutional networks do not achieve good accuracy, which is probably due to the conversion of raw accelerometer data into vibration images.
|
|
|
Comparison of methods for identification of rolling bearing failures
Kokeš, Miroslav ; Hnidka,, Jakub (referee) ; Klusáček, Stanislav (advisor)
The aim of this master thesis is the comparison of selected methods and parameters for roller bearings diagnostics. Selected statistical parameters are kurtosis, crest factor, and parameter K(t). The other selected methods are envelope analysis, cepstral analysis, and ACEP method. These methods are implemented in LabVIEW software and compared based on noise resistance, computation speed, and overall capability of identifying roller bearing faults.
|
guest ::
