- Research Article
- 10.1016/j.jallcom.2026.186943
Positive temperature coefficient of magnetization in AlNiCo magnets with Dy / Gd / Er addition
- Mar 01, 2026
- Journal of Alloys and Compounds
- Tianyu Hu + 7 more +7
Publications from 2021 to 2026
Showing 10 of 76 papers
Positive temperature coefficient of magnetization in AlNiCo magnets with Dy / Gd / Er addition
Prediction of Lost Circulation Using Artificial Neural Networks in the Al Jeribe Formation of Omar Field
Lost circulation is a significant challenge in drilling operations, leading to fluid losses, increased non-productive time, and well instability. This paper develops a predictive model to quantify lost circulation in the Al Jeribe Formation in the Al-Omar field, one of the largest oilfields in the Middle East. Lost circulation is especially prevalent when drilling through the Al Jeribe formation due to the presence of vugs and caves. However, current models for predicting lost circulation often suffer from limited accuracy and efficiency due to the complexity of geological formations and the variability of drilling conditions, leading to unreliable predictions in challenging environments. This research aims to overcome these limitations by developing a more accurate and efficient predictive model tailored to the Al Jeribe Formation, providing valuable insights to mitigate fluid loss and improve drilling efficiency. This paper introduces a novel predictive model for lost circulation in the Al Jeribe Formation, utilizing artificial neural networks (ANNs) trained on extensive field data from over 100 wells. The model incorporates key drilling parameters such as mud weight (MW), yield point (Yp), equivalent circulation density (ECD), rate of penetration (ROP), revolutions per minute (RPM), strokes per minute (SPM), Plastic viscosity (PV), and weight on bit (WOB) as input parameters. The ANN achieved excellent predictive performance, with Training R2 = 0.99 and Testing R2 = 0.99. Error metrics also confirmed strong generalization, with RMSE = 1.70% (training) and 1.40% (testing), and AAPE = 11.0% (training) and 10.2% (testing). In addition, the model identified the most critical parameters influencing lost circulation and provided optimized parameter ranges to mitigate fluid loss during drilling operations. This study focuses on lost circulation prediction in the Al Jeribe Formation, identifying key drilling parameters and providing optimized ranges to reduce losses and improve wellbore stability. It offers insights not covered in previous research, specifically targeting the Al Jeribe Formation. The model predicts lost circulation and suggests practical adjustments to drilling parameter values. The findings are expected to enhance drilling efficiency and minimize downtime in the Al-Omar field. This methodology can also be applied to similar geological formations worldwide to reduce lost circulation in oil fields.
Read morePrediction of remaining useful life for electronic equipment based on online PINN.
As critical components of spacecraft systems, electronic equipment requires accurate remaining useful life (RUL) prediction to ensure reliable operations. Traditional physics-based and data-driven methods are limited by poor generalization and substantial data demands, respectively. While Physics-Informed Neural Networks (PINNs) integrate physical laws to enhance accuracy, they remain susceptible to performance degradation under shifting operational conditions due to catastrophic forgetting. This paper proposes a novel online PINN framework integrated with continual learning to dynamically adapt to new degradation patterns. Validated on NASA's IGBT dataset, our method reduces RUL prediction error to 27.2% of a traditional LSTM model and achieves an online update time of less than 800ms/cycle. These results demonstrate a significant improvement in robustness and accuracy, providing a superior solution for RUL estimation of aerospace power electronics under extreme environments.
Read moreAn aeromagnetic compensation method based on the extended Tolles Lawson model
Abstract Aeromagnetic survey is widely applied in mineral exploration, shipwreck salvage, and unexploded ordnance detection. To maintain optimal system detection performance, it is essential to compensate for magnetic interference originating from the airborne platform and external environments. The conventional TLG compensation model can only suppress platform-induced dynamic magnetic interference and geomagnetic gradient interference, but fails to account for magnetic interference caused by geological anomalies within the survey area, which limits further improvements in system performance. To address this limitation, this paper proposes a magnetic interference compensation method specifically designed for airborne detection in regions with complex geological magnetic anomalies. First, a high-precision spatial reconstruction of the geological magnetic anomaly distribution is achieved using an ensemble machine learning-based spatial field reconstruction approach. Subsequently, the conventional TLG compensation model is extended by incorporating the reconstructed anomaly distribution, resulting in a novel TLGF compensation model. Finally, by utilizing the platform’s positional information in conjunction with the TLGF model, simultaneous suppression of platform-induced dynamic interference, geomagnetic gradient interference, and regional geological magnetic interference is realized. Experimental results demonstrate that, compared to conventional methods, the proposed approach reduces the standard deviation of residual magnetic fields by 52.4-62.1%, significantly improving the system’s detection robustness in complex geological environments.
Read moreA Lightweight Neural Network Approach for MOSFET Remaining Useful Life Prediction
To meet the health management requirements of aerospace missions and considering the operating characteristics of Metal-Oxide-Semiconductor Field-Effect Transistors (MOSFETs) under computational resource constraints, this paper proposes a remaining useful life (RUL) prediction method based on the integration of a Gated Recurrent Unit (GRU) and a Variational Autoencoder (VAE). The GRU is employed to model the temporal evolution of device operating features and predict degradation trends, while the VAE performs latent-space modeling and reconstruction-error quantification of degradation states, thereby enabling both quantitative RUL estimation and uncertainty characterization. In degradation trend prediction, the GRU is compared with the Extended Kalman Filter (EKF) and Long Short-Term Memory (LSTM) networks. The results show that the GRU significantly outperforms EKF in accuracy and achieves performance comparable to LSTM, while requiring only about three-quarters of the parameters and slightly higher inference speed, highlighting its lightweight nature. By combining historical operating data with GRU predictions, the VAE effectively captures latent features of the degradation process and quantifies the degree of degradation through reconstruction error, thus enabling robust RUL estimation with uncertainty analysis. Experimental results demonstrate that the proposed lightweight neural network achieves a root-mean-square error below 0.003 for degradation prediction and an average error of less than 10% for RUL estimation. The model consistently tracks the overall degradation trend and maintains high accuracy with minimal computational overhead, indicating strong potential for deployment in resource-constrained end devices.
Read moreA C-Modelica co-simulation-based method for constructing switch driver functional digital model
Abstract This study addresses the technical limitations of the Modelica-based simulation mode in the dynamic interaction characteristics of simulation systems within measurement and control systems. It proposes the C-Modelica co-simulation architecture, which establishes a parameter read/write channel for the Modelica static model through the shared memory middleware. By constructing a dynamic compilation interface for control logic using C code blocks, the architecture allows dynamic modification of digital model parameters during the simulation process, enabling external injection of user control behaviors. The proposed C-Modelica co-simulation architecture’s dynamic behavior control capabilities during the simulation process are verified by using commonly used switch driver functional modules in measurement and control systems. This provides theoretical and practical support for digital model construction and behavior validation in measurement and control systems.
Read moreTactical-Grade Three-Axis Interferometric Fiber-Optic Gyroscope With a Silica Integrated Optical Chip
As an important component in inertial navigation system (INS), the fiber-optic gyroscope (FOG) is developing toward integration and miniaturization. In this article, an integrated photonic chip comprised of a light source, three photodiodes, and four 3-dB couplers for three-axis FOG based on silica waveguide is proposed. The interferometric FOG (IFOG) assembly process is simplified, the cost is reduced, and production efficiency is improved. The overall size of the three-axis IFOG is reduced significantly to only <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"> <tex-math notation="LaTeX">$50\times 50\times 30$ </tex-math></inline-formula> mm, and a bias stability is better than 0.5°/h for a sensing coil with a diameter of 26 mm and can satisfy the requirements of tactical-grade compact INSs. To the best of our knowledge, this is the first reported to date for the three-axis silica integrated optical chip-based IFOG.
Read more"Comparative analysis of rate of penetration prediction and optimization in deep wells using real-time continuous stacked generalization ensemble learning: A case study in Shunbei"
A digital-twin-based online monitoring method for aging of thermal interface material of power converter
This paper proposes a digital-twin-based aging state online monitoring method of thermal interface material (TIM) on the bottom of an insulated gate bipolar translator (IGBT) module, The proposed method uses voltage, current, and temperature data which is measured by the sensors in power converter as the input data of digital twin model, and digital twin module continuously modifies its thermal conductivity parameter to follow the thermal status of power converter, and the thermal resistance of thermal interface material is identified after multiple consecutive iterations, which is a good characterization parameter of TIM aging state. For verification, the digital twin simulation models and test platform are built to verify the accuracy. It is demonstrated that the proposed method can monitor the aging status of thermal interface material in real-time.
Read moreDesign and Trajectory Planning of 6-DOF Reconfiguration Parallel Platform