This study proposes the utilization of an optical fiber vibration sensor for detecting the superposition of extremely close frequencies in vibration signals. Integration of deep neural networks (DNN) proves to be meaningful and efficient, eliminating the need for signal analysis methods involving complex mathematical calculations and longer computation times. Simulation results of the proposed model demonstrate the remarkable capability to accurately distinguish frequencies below 1 Hz. This underscores the effectiveness of the proposed image-based vibration signal recognition system embedded in DNN as a streamlined yet highly accurate method for vibration signal detection, applicable across various vibration sensors. Both simulation and experimental evaluations substantiate the practical applicability of this integrated approach, thereby enhancing electric motor vibration monitoring techniques.
30 January 2023 in Elinvo (Electronics, Informatics, and Vocational Education)
This paper presents a digital implementation of a brushless direct current motor (BLDCM) drive with a six-step pulse width modulation (PWM) using a programmable simplified C-block based on the STM32 microcontroller. The implementation is conducted through the PSIM simulation platform, which is commonly used for power electronics and motor control. This approach combines the benefits of using a programmable simplified C-block for precise and flexible programming with the PWM concepts of the STM32 microcontroller. The PWM method used on the BLDCM drive is the unipolar upper PWM technique (H~PWM_L~ON). The performance of the PWM implementation is analyzed in detail, including the accuracy of the PWM generation using a Fast Fourier Transform (FFT), the gating of the IGBTs in the three-phase inverter, and the effect of the duty cycle on the BLDCM's speed, phase voltage, and phase current.
This study introduces a novel deep neural network (DNN) framework tailored to breaking the sampling limit for high-frequency vibration recognition using fiber Bragg grating (FBG) sensors in conjunction with low-power, low-sampling-rate FBG interrogators. These interrogators, while energy-efficient, are inherently limited by constrained acquisition rates, leading to severe undersampling and the obfuscation of fine spectral details essential for accurate vibration analysis. The proposed method circumvents this limitation by operating solely on raw time-domain signals, learning to recognize high-frequency and extremely close vibrational components accurately. Extensive validation using the combination of simulated and experimental datasets demonstrates the model’s superiority in frequency discrimination across a broad vibrational spectrum. This approach is expected to be a significant advancement in intelligent optical vibration sensing and compact, low-power condition monitoring solutions in complex environments.
Inexpensive, lightweight, sensitive, compact, easy-to-install, and mass-producible vibration sensors are the future wave. This paper validates the use of a 3D-printed doublet bracket that turns out to be a scissor-like X-structure with scissor-like kinematics when vibrating. The main concept is simple: an intact standard single-mode fiber (SMF) is folded into a droplet-bending shape and embedded in an arch bridge-type assembly connected to an X-shaped structure. When regular vibrations are transmitted to the X-structure, the continuous scissor motion of the X-structure drives the arch bridge-type assembly to regularly extend and shrink, thereby changing the size of the fiber bending, creating variations in the optical loss, and allowing for the measurement of the vibration. In addition, the sensitivity of the vibration measurement can be adjusted by altering the torque by varying the intersection angle of the doublet bracket in the X-structure or the fiber loss by varying the number of turns of fiber bending. In this way, the sensitivity can be adjusted according to the magnitude of the vibration amplitude of the object being measured in various measurement scenarios. For instance, if the vibration amplitude is weak, a high-sensitivity frame will be utilized for the measurement, and vice versa. Furthermore, this 3D-printed-based X-structure and fiber-bending-based vibration sensor can be integrated with free-space optics (FSO) for more versatile measurements in various applications. In summary, this scheme showcases an affordable, lightweight, variable-sensitivity, simple, and easily mass-producible fiber-optic vibration sensing technology.
27 May 2024 in Control Systems and Optimization Letters
Path planning is a crucial aspect of mobile robot navigation, ensuring that robots can safely travel from their initial position to the goal. In real-world applications, path planning is essential for autonomous vehicles, drones, warehouse robots, and rescue robots to navigate complex environments efficiently and safely. One effective method for path planning is the Rapidly-exploring Random Tree (RRT) algorithm, which is particularly practical in maze-like environments. The performance of RRT depends on the sampling methods used to explore the maze. Sampling methods are important because they determine how the algorithm explores the search space, affecting the efficiency and success of finding an optimal path. Poor sampling can lead to suboptimal or infeasible paths. In this study, we investigate different sampling strategies for RRT, specifically focusing on uniform sampling, Gaussian sampling, and the Motion Planning Network (MPNet) sampling. MPNet leverages a neural network trained on past environments, allowing it to predict promising regions of the search space quickly, unlike traditional methods like RRT that rely on random exploration without prior knowledge. This makes MPNet much faster and more efficient, especially in complex or high-dimensional spaces. Through a benchmarking analysis, we compare these methods in terms of their effectiveness in generating feasible paths. The results indicate that while all three methods are effective, MPNet sampling outperforms uniform and Gaussian sampling, particularly in terms of path length. The mean path length generated, based on a sample size of 30, is 13.115 meters for MPNet, which is shorter compared to uniform and Gaussian sampling, which are 18.27 meters and 18.088 meters, respectively. These findings highlight the potential to enhance path planning algorithms using learning-based sampling methods.
2 September 2026 in Electrical Engineering & Electromechanics
Introduction. Bearing faults in induction motors are one of the primary causes of performance degradation and unexpected failures in industrial systems. Early fault detection remains challenging because conventional protection systems generally respond only after severe damage occurs. In addition, motor current signals exhibit nonlinear and complex characteristics, requiring advanced analysis techniques for accurate fault identification. Problem. Existing fault diagnosis methods often suffer from limited classification accuracy, dependency on specific operating conditions, and insufficient integration between spectral feature extraction and adaptive classification techniques. Goal. To develop a non-invasive bearing fault classification method based on current spectrum analysis and artificial neural network (ANN) for induction motor condition monitoring. Methodology. The proposed method utilizes fast Fourier transform (FFT) to transform motor current signals from the time domain into the frequency domain for spectral feature extraction. The extracted features are then processed using principal component analysis (PCA) for dimensionality reduction before being used as inputs to the ANN classifier. Experimental testing is conducted under 3 bearing conditions, namely normal, 7-ball fault, and 6-ball fault conditions, using 50 datasets for each condition. Results. The results demonstrate that the proposed method successfully identifies bearing conditions with high classification accuracy and strong separation characteristics in the PCA space. FFT analysis also reveals consistent spectral changes corresponding to fault severity, particularly in sideband components and energy distribution patterns. Scientific novelty. This work integrates FFT-based current spectrum analysis, PCA-based feature reduction, and ANN classification into a unified non-invasive diagnosis framework for bearing fault detection. Practical value. The proposed approach provides a simple, adaptive, and reliable solution for early bearing fault detection without requiring additional mechanical sensors, making it suitable for industrial condition monitoring applications. References 32, tables 4, figures 7.