Kybernetika a robotika

Browse

Now showing 1 - 5 of 45

A Dead-Time Compensation Strategy Based on an Online Learned Artificial Neural Network
(IEEE, 2025-04-03) Buchta, Luděk; Kozovský, Matúš; Blaha, Petr
This article presents an innovative approach to mitigate the harmonic distortion of the phase currents of a permanent magnet synchronous motor (PMSM) controlled by a field-oriented control (FOC) algorithm. The issue of phase current harmonic distortion is often a consequence of the output voltage deformation caused by the non-linearities of the voltage source inverter (VSI). The relationship between the disturbance voltages of the inverter and the phase currents of the motor is non-linear. Therefore, we used an artificial neural network (ANN) to identify the compensation voltages. The topology is designed to allow the neural network to solve complex problems with the limited computing resources available on the AURIX TC397 microcontroller. The input vector is assembled from quantities available in the PMSM FOC algorithm. The online learning process based on the back-propagation algorithm is adapted to operate directly on the microcontroller. The proposed strategy with ANN is verified on a real PMSM. The results show the excellent ability of the proposed ANN to suppress the harmonic distortion of the PMSM phase currents without knowledge of the VSI parameters.
Validated low-cost standardized VICON configuration as a practical approach to estimating the minimal accuracy of a specific setup
(NATURE PORTFOLIO, 2025-07-02) Chromý, Adam; Šopák, Petr; Cígler, Hynek
Motion Capture (MoCap) is rapidly growing in the sports, biomechanics, healthcare, and medicine segments, where accuracy is crucial. Current research studies are concurrently confirming that the accuracy can be determined only for the specific analyzed configuration and thus recommending performing your own accuracy verification on your specific setup. However, it is often hard to perform since it requires significant effort, time, knowledge of statistical data analysis and often equipment and tools that are not commonly available. This paper deals with this by creating a standardized setup with carefully evaluated accuracy, substituting the on-site validation process (in case of using such a setup) or providing the worst-case accuracy (when a more advanced setup is used). The setup is designed to be low-cost, easily reproducible and cover a wide range of applications - thus VICON setup with five VERO v1.3 cameras is used. The accuracy was evaluated using the robotic manipulator EPSON C3, determining that the absolute positioning accuracy of such a standardized setup is 0.65 mm on average (SD = 0.48, with maximal error of 2.47 mm) and rotation accuracy 0.40 degrees (SD = 0.35, with maximal error of 2.0 degrees), which is negligible considering the experimental diameter of 1.4 m and full angular span. The major source of error was specific to particular spatial and rotational positions; other systematic and other random errors were noticeably smaller. If the standardized setup is used and all its requirements are met, a similar accuracy as validated above can be expected without the need to explicitly validate the specific configuration, which is time-consuming and resource-intensive.
Analyzing the impact of mechanical damage in the piezoelectric ceramics elements of knock sensors on the frequency characteristic
(The International Institute of Acoustics and Vibration, 2024-10-14) Klusáček, Stanislav; Fialka, Jiří; Havránek, Zdeněk; Skalský, Michal; Pikula, Stanislav; Šedivá, Soňa; Beneš, Petr
Piezoelectric ceramics have been used on a long-term basis as the active element in multiple types of sensors, in examining or performing vibration, knock effects, acoustic emission, and defectoscopy, and in ultrasonic classes. Mechanical damage, including cracks, breakage, and partial separation of the piezoceramics, can arise from the normal use of sensors in industrial applications. The defects, however, are not inspectable optically when the piezoelectric element is encapsulated inside the sensor. Accelerometers, for instance, are calibrated immediately after manufacturing and then periodically in the frequency range used. Other sensors, such as those for knock detec-tion, which are permanently fixed to a machine or engine block, are only verified before ship-ment from the manufacturer. This paper focuses on the impacts exerted by the mechanical damage or breaking of an element on the rel-evant frequency characteristics and functionality verification, also highlighting the possibilities for detecting such adverse issues. In the investigation, we examined various categories of piezo-electric elements, invariably with the most common mechanical damage that occurs during the industrial use. The defects on the piezoelements were observed in detail using an optical scan-ning microscope, and their impacts were evaluated with the frequency and the direct measure-ment methods via a d33 meter. In this context, the paper aims to define a measurement method applicable in quantifying the level of damage on an active piezoelectric element. In this context, it is also determined how the actual damage affects the output parameters of the sensor.
Simulation-Based Diagnosis for Cyber-Physical Systems - A General Approach and Case Study on a Dual Three-Phase E-Machine
(Schloss Dagstuhl – Leibniz-Zentrum für Informatik, 2024-11-26) Kaufmann, David; Kozovský, Matúš; Wotawa, Franz
This paper presents a simulation-based approach for fault diagnosis in cyber-physical systems. We utilize simulation models to generate training data for machine learning classifiers to detect faults and identify the root cause. The presented processing pipeline includes simulation model validation, training data generation, data preprocessing, and the implementation of a diagnosis method. A case study with a dual three-phase e-machine highlights the results and challenges of the simulation-based diagnosis approach. The e-machine simulation model provides a complex and robust system representation, including the capability to inject inter-turn short-circuit faults. The introduced validation procedures of the simulation model revealed limitations in signal similarity and distinguishability compared to real system behavior. Based on the discovered limitations, the overall best results are achieved by applying an Autoencoder model for anomaly detection, followed by a Random Forest classifier to identify the specific anomalies. Further, the focus is on identifying the affected e-machine phase rather than the exact number of faulty winding turns. The paper shows the challenges when applying a simulation-based diagnosis approach to time-series data and underlines the required analysis of simulation models. In addition, the flexible adaption in the diagnosis strategies enhances the efficient utilization of cyber-physical system models in fault diagnosis and root cause identification.
PMSM fault detection using unsupervised learning methods based on conditional convolution autoencoder
(IEEE, 2024-11-03) Kozovský, Matúš; Buchta, Luděk; Blaha, Petr
The challenges of fault detection and condition monitoring in powertrain systems have become increasingly prominent, particularly with the widespread adoption of failoperational systems. These systems are pivotal in diverse sectors, including the robotics, automotive industry, and various industrial applications. A critical attribute of such systems lies in their capability to identify non-standard behaviour of the system. This study describes a inovative conditional convolutional autoencoder-based fault detection algorithm for the permanent magnet synchronous motor. The study compares a train process of conditional convolutional autoencoder with a classical convolutional autoencoder. The presented autoencoder structure was designed to be implementable into the target microcontroller AURIX TC397 while providing sufficient recognition capabilities of the interturn short-circuit. Autoencoders are trained on data obtained during healthy motor operation and subsequently used to detect interturn short-circuit faults on the experimental dual three-phase permanent magnet synchronous motor with the possibility of emulating an interturn short-circuit fault. The paper provides insights into the achieved autoencoder inference times and the sensitivity in detecting the fault.

Browse

Recent Submissions