Journal of Mechatronics and Artificial Intelligence in Engineering
Not a member yet
1200 research outputs found
Sort by
Research on converter transformer state early warning system based on confidence ratio-EEMD and multi-cascade network
Get PDFAiming at the problems of poor prediction effect of non-stationary parameters and single warning rule of UHV converter transformer, this study proposes an intelligent warning method based on decomposition-multi-level cascade network and fuzzy set. Firstly, the integrated empirical modal decomposition technique is used to decompose the target parameter sequence into multiple sub-sequences, and the effective components are screened by the DPR-KLdiv confidence ratio, which is dynamically grouped and reconstructed to form a multilevel feature input; and the multilevel cascade network is constructed by combining multi-device parameters to make the time series prediction. The fuzzy function is further introduced to establish the parameter state mapping rules to expand the alarm triggering conditions. The experiments are validated by actual equipment data, and the local discharge signals of different defects are detected by ultra-high frequency method to enhance the generalization ability of the parameters. The results show that the average RMSE and MAE of this method are 23.21 and 18.47 respectively under the hours step prediction, and the accuracy of the warning is over 90 %, which effectively improves the accuracy of non-smooth parameter prediction and the flexibility of the warning decision
A multi-scale convolutional Siamese network for few-shot fault diagnosis of unmanned aerial vehicle rotor
Get PDFUnmanned Aerial Vehicles (UAVs) often face infrequent fault occurrences and high manual annotation costs, resulting in a critical shortage of valid fault samples for diagnostic research. Traditional fault diagnosis methods struggle with small sample sizes. This paper proposes a novel deep metric learning method, the Multi-Scale Convolutional Siamese Network (MSCSN), to address the few-shot learning problem in UAV rotor fault diagnosis. First, discrete wavelet transform (DWT) is used to compress and normalize the vibration signals, enhancing the prominence of signal features. Then, based on the multi-scale convolutional neural network (MS-CNN) model, the network automatically extracts multi-level features from rotor fault vibration signals, improving its adaptability to complex data. Finally, the Siamese network structure, with shared parameters and identical architecture, processes sample pairs and incorporates a small number of support samples for few-shot learning. Experimental results show that the proposed model achieves a highest accuracy of 94.42 % in few-shot tasks. In cross-domain transfer learning tests, the model achieves an average accuracy of 90.28 %, demonstrating its superior generalization ability and robustness across different environments. We also validated the model's stability using the publicly available MVS-UAV-BF dataset
Artificial intelligence-based stock market price prediction, a review
Get PDFThe stock market, a cornerstone of the global financial system, is characterized by its dynamic and volatile nature, which makes accurate price-trend prediction challenging. However, traditional statistical models often fail to capture this complexity. Recent advancements in Artificial Intelligence (AI), particularly Machine Learning (ML) and Deep Learning (DL), have transformed stock market forecasting by using diverse datasets and algorithms. This review examines recent studies on AI methodologies for stock market price trend prediction models by analyzing architectures, datasets, performance metrics, and limitations, with a focus on hybrid models, sentiment analysis, and dataset diversity. Hybrid approaches, including the Multi-Model Generative Adversarial Network Hybrid Prediction Algorithm (MMGAN-HPA), K-means long short-term memory (LSTM), and LSTM autoregressive output (LSTM-ARO), improve predictive accuracy by combining statistical methods with deep learning. Sentiment analysis models such as Stock Senti WordNet (SSWN) and Hybrid Quantum Neural Network (HQNN) integrate social media sentiment to capture market dynamics. Real-time frameworks that use stream processing show promise for high-frequency trading applications. This review addresses key challenges including data noise, nonstationarity, overfitting risks, and black-box model interpretability. Solutions include GAN-based synthetic data generation, transformer-based architectures such as SpectralGPT, and optimization techniques for computational efficiency. This review provides a taxonomy of AI-based approaches, while identifying gaps for future research. These findings highlight the potential of AI in financial forecasting while emphasizing the need for interdisciplinary collaboration to address its limitations in data quality, methodology, interpretability, and ethics
Analysis of modal and vibration response characteristics of high-pressure storage tanks
Get PDFAnalysis of dynamic characteristic was conducted focusing on the transportation of high-pressure storage tanks, covering two scenarios: independent transportation and mixed transportation. For independent transportation, analysis of free modal was carried out to obtain the first four orders of modal shapes. Additionally, the influence of two constraint methods on the modal characteristics and stress distribution was studied, including fixed at both ends and fixed at the cylinder body. Results show that when the tank was fixed at the cylinder body, it had a higher natural frequency and a lower stress level, making it safer. For mixed transportation, a finite element model was built for 6 high-pressure storage tanks, and analysis of random vibration was performed. The results showed that stress was mainly concentrated on the crossbeams and connection nodes, while the stress on the main body of the storage tanks was relatively low. The overall structure exhibited excellent fatigue performance and met the mechanical and safety requirements under random vibration conditions
High-speed roller/rail dynamics and thermodynamics considering surface roughness and revolution
Get PDFConsidering rotation and not considering rotation, a calculation was conducted on the friction and wear between the sliding pair in a high-speed rotating machine using a plane of Ra6.3 and Ra3.2, indicating that Ra3.2 has advantages. In higher firing rate Gatling guns, the guide rail should be processed more finely and have a smaller roughness. The results demonstrate that the stress increases a lot when the bolt is surface rough, which is 11.5 % higher than the flat condition. The temperature of Ra6.3 is about 100° higher than the flat condition. It plays an important role in improving the service life of friction surfaces
Structural damage detection by progressive continuous wavelet transform and singular value decomposition of noisy mode shapes
Get PDFFor decades, damage identification based on structural mode shapes has been a popular research topic. While mode shapes provide valuable spatial structural information, the sensitivity to localized damage remains limited. In contrast, modal curvature exhibits high sensitivity to local damage, enabling precise damage localization. However, its susceptibility to environmental noise poses a significant limitation. To this end, a novel damage identification method is proposed by integrating continuous wavelet transform (CWT) and singular value decomposition (SVD). First, the CWT is applied to structural mode shapes for generating continuous wavelet coefficients. Subsequently, the SVD is performed on these coefficients, yielding new damage indicator termed as the singular image of continuous wavelet coefficients (SICWC). The SICWC enhances damage sensitivity and localization accuracy by suppressing noise-induced global trends in structural mode shapes. The effectiveness of proposed method is validated through numerical simulations of a cantilever beam under noisy conditions, as well as experimental detection of a cracked beam using mode shapes acquired via a scanning laser vibrometer. The results demonstrate that SICWC effectively mitigates the limitations of traditional damage detection methods based on mode shape and curvature
Use of fragility curves to assess the seismic vulnerability of soft rock tunnels: a review
Get PDFDue to their distinct geotechnical and structural features, soft rock tunnels pose serious issues because of their seismic sensitivity. These tunnels, often constructed in formations with lower shear strength and higher deformability, are particularly susceptible to damage during earthquakes. Fragility curves, which graphically represent the probability that a structure may sustain damage up to or beyond a particular threshold as a function of seismic intensity, are essential tools for evaluating the seismic resilience of these infrastructures. This research looks closely at the use of fragility curves to assess the seismic vulnerability of soft rock tunnels. Exploring the fundamental concepts and methodologies involved in constructing fragility curves, including seismic hazard analysis, structural modeling, damage state definition, data collection and statistical analysis is looked at first. The review highlighted the integration of soft rock characteristics such as strength and deformation properties into the fragility assessment process. Key developments in the topic are covered such as how machine learning and Bayesian inference might improve the precision and usefulness of fragility curves. The paper identified key findings such as the high sensitivity of fragility curves to geotechnical properties and seismic intensity levels and emphasized the importance of accurate data collection and model calibration. Important gaps in seismic risk evaluations are filled by integrating cutting-edge methodologies, such as Bayesian inference and real-time machine learning models that clarify the seismic behaviour of soft rock tunnels in the real world. For the purpose of strengthening earthquake-resistant infrastructure in earthquake-prone areas, engineers, scholars and policymakers are given practical insights
Dynamic behaviors and double-frequency synchronization analysis of a dynamic vibration absorption system driven by three co-rotating exciters
Get PDFThe recovery efficiency of drilling fluid is directly affected by working performance of the vibration screen. Therefore, a newly dynamic vibration absorption system driven by different excitation frequencies is designed through double-frequency synchronization theory to improve the mechanical performance of screening equipment. Firstly, the differential equations of motion of vibration system are deduced by Lagrange method. Then, the theoretical conditions of the system implementing double-frequency synchronization are obtained based on asymptotic method, and stability criterion of the synchronization is revealed according to Routh-Hurwitz criterion. Subsequently, the effects of structure parameters on vibration isolation ability, synchronous state, and stability of synchronization are numerically discussed. Finally, the feasibility of the theoretical method and the obtained results is further verified by simulation and experiment. It is found that the vibration isolation and synchronization performance of the system is influenced by the motor parameters and system structure. The system has the best vibration isolation ability when ωm0= 157 rad/s, which is considered as the best operating frequency of the present vibration system. Meanwhile, when the mass ratio κ between the high-frequency co-rotating rotor and the low-frequency co-rotating rotor is smaller, the absolute value of the stability coefficient Si is larger, and the stability phase difference is smaller, and the system is more stable. The present work can provide theoretical direction for the design of new screening equipment
Improved CEEMD-based correction method for low-frequency shock response spectrum in large dual-wave shock tester devices
Get PDFThe shock response spectrum (SRS), calculated from a shock acceleration signal, is a critical indicator of shock environments. However, under intense loads, acceleration sensors are prone to trend term errors that can cause significant drift in the low-frequency spectral lines of large dual-wave shock tester devices. To address this issue, the complementary ensemble empirical mode decomposition (CEEMD) method was employed to decompose acceleration signals and restore the actual shock environment. Intrinsic mode functions (IMFs) were cross-correlated and compared to a predefined threshold to identify the effective IMF components required to reconstruct the signal. K-means clustering was employed to further validate the effectiveness of the IMFs for enhanced selection accuracy. Finally, the reconstructed acceleration signal was used to calculate a corrected SRS. The proposed approach demonstrated significant improvements over the traditional CEEMD algorithm. The corrected SRS exhibits a 5.6316 dB/oct slope in the low-frequency band, reflecting an equal displacement trend. The maximum error at the corresponding frequency was less than 6 % in comparison to the relative displacement response measured by low-frequency spring oscillators. This improved CEEMD correction method can effectively restore the actual shock environment of a dual-wave shock tester device, offering a valuable reference for evaluating shock resistance in onboard equipment
Stodola-Vianello iteration method for the free flexural vibration frequencies of Shimpi’s single variable shear deformable beams
Get PDFThe natural vibration frequency analysis of beams is vital for their design against resonance failures because such failures occur when the excitation load frequencies of vibration coincide with such natural frequencies. This work presents a single variable shear deformable beam equation formulated using Shimpi’s displacement field assumptions. This results in a quadratic shear stress profile over the depth and a satisfaction of the transverse shear stress-free boundary conditions. The governing equation is obtained using a first principles consideration and equilibrium method as a partial differential equation (PDE) which is non-homogenous for forced vibrations and homogeneous for free vibrations. The study then used the Stodola-Vianello iteration method to solve the resulting homogeneous PDE for simply supported boundary conditions and harmonic response. The problem reduced to an iterative problem of algebra involving the computation of an (n+1)th vibratory modal shape function from an nth shape function that satisfies the boundary conditions. This work used a sinusoidal shape function which is exact for the simply supported boundary condition investigated. The use of boundary conditions solved the integration constants involved. Application of the convergence rule led to the eigenequation from which the eigenvalues were found. The eigenvalues were presented for the first four modes of vibration and for a rectangular beam. It was found that for l/h varying from 5 to 100, the natural vibration frequencies were identical with the ωn values obtained using Navier method for other thick beam vibration problems. It was also found that ωnwas close to the exact values for all vibration modes and for all values of l/h between 5 and 100. For all vibration modes and all considered l/h values negligible differences, were observed between the ωn obtained using SVIM and the exact values obtained by previous researchers