Search

The deterioration of power generation facilities built during the early stages of plant operation is becoming increasingly severe, raising concerns about potential socioeconomic harm from pipe leaks. Consequently, there is a pressing need for rapid leak detection and proactive responses. Prior research primarily relied on various signal processing techniques and supervised learning for leak detection. However, these approaches struggle with accurate detection amid environments with diverse background noises and weak leak signals, exacerbated by challenges in gathering sufficient real-world leakage data, which can lead to overfitting during model learning. Therefore, in this paper, an adaptable leak detection model suitable for various environments was proposed to ensure precise leak detection. Frequency spectrum feature extraction and transfer learning were utilized to achieve accurate leak detection, even with limited data. In addition, an unsupervised learning-based autoencoder model is employed to identify leaks accurately by learning general patterns, even when leakage data is limited. Experimental results demonstrate that the proposed model, integrating feature extraction techniques using the Uniform Manifold Approximation and Projection (UMAP) algorithm and employing transfer learning, achieved a higher accuracy performance with 6.35 percentage points (%p) compared to the model lacking these techniques. In addition, these findings confirm a slight decrease in accuracy performance even when using minimal learning data. Moreover, the leak detection performance was superior to the existing models considered in this study, achieving a high accuracy rate of 99.19%.

Background: Osteoarthritis (OA) is a complex disease marked by the degradation of articular cartilage. Objective: This study aimed to explore the relationship between cartilage volume/thickness and clinical outcomes in knee OA patients treated with intra-articular injections over one year. Methods: Twenty-four patients with mild-to-moderate OA were retrospectively analyzed using knee MRI. OA features were assessed semiquantitatively with the Whole-Organ Magnetic Resonance Imaging Score (WORMS), while cartilage thickness and volume in the medial femoral condyle (MFC) and medial tibial plateau (MTP) were measured. T1ρ and T2 values for MFC cartilage were also recorded. Clinical outcomes were evaluated using the Korean Western Ontario and McMaster Universities (K-WOMAC) and Knee Injury Osteoarthritis Outcomes (KOOS) scores. Spearman’s rank test assessed the associations between imaging changes and clinical outcomes. Results: The baseline MTP and MFC cartilage thickness and MTP cartilage volume showed significant correlations with clinical outcomes. Additionally, less progressive cartilage loss in the medial femorotibial joint (MFTJ) and overall joint was linked to a better clinical response over 12 months. Conclusions: In conclusion, thicker baseline MFTJ cartilage and minimal cartilage loss were associated with favorable clinical outcomes in knee OA patients receiving intra-articular injections.

Creating an effective deep learning technique for accurately diagnosing leak signals across diverse environments is crucial for integrating artificial intelligence (AI) into the power plant industry. We propose an automatic weight redistribution ensemble model based on transfer learning (TL) for detecting leaks in diverse power plant environments, overcoming the challenges of site-specific AI methods. This innovative model processes time series acoustic data collected from multiple homogeneous sensors located at different positions into three-dimensional root-mean-square (RMS) and frequency volume features, enabling accurate leak detection. Utilizing a TL-driven, two-stage learning process, we first train residual-network-based models for each domain using these preprocessed features. Subsequently, these models are retrained in an ensemble for comprehensive leak detection across domains, with control weight ratios finely adjusted through a softmax score-based approach. The experiment results demonstrate that the proposed method effectively distinguishes low-level leaks and noise compared to existing techniques, even when the data available for model training are very limited.

The conventional Weinberg converter, owing to its high efficiency, continuous input and output current, and the soft switching of its switches and diodes, is essentially utilized in battery discharge regulators (BDR) – components that make up the overall electrical power system for geostationary satellites. However, this converter faces a crucial drawback in that the output current ripple is considerably high during the commutating operation of its output diodes. Consequently, the output root mean square (RMS) current and output voltage ripple are large. This may directly influence the lifetime and overall capacitance of the bus capacitors, ultimately affecting the longevity and size of the geostationary satellite. Therefore, reducing the BDR’s output current ripple is essential in designing a geostationary satellite’s electrical power system. As a result, this paper proposes an improved Weinberg converter to reduce the output current ripple and output voltage ripple. Specifically, the proposed converter not only has a lower output current and voltage ripples compared to the conventional converter but also reduces the magnetizing current offset of the coupled inductor. Furthermore, similar to the conventional converter, it is capable of soft switching operation of all switches and achieving high efficiency without the reverse recovery problem of diodes. Overall, the proposed converter is verified with experimental results from a 750 W-rated prototype.

The arc-fault phenomenon in photovoltaic (PV) systems has emerged as a major problem in recent years. Existing studies on arc-fault detection in conventional PV systems primarily focus on detecting typical stable arc-faults. Low-energy arc-faults are more challenging to detect than stable arc-faults because of their low current distortions, short durations, and nonlinear properties. These low-energy arc-faults, which are precursors to stable arc-faults, could even inflict serious damage on the system components. Here, a transfer learning-based low-energy arc-fault detection network (TL–LEDarcNet) using a two-stage training method is proposed to proactively detect series DC arc-faults by considering low-energy arc-faults. A one-layer long short-term memory network combined with a lightweight one-dimensional convolutional neural network was developed to detect low-energy arc-faults by only using the sensed current information. The results of offline and online experiments conducted with a commercial grid-connected PV inverter indicate that the proposed method can perform real-time operations on a single-board computer and detect low-energy arc-faults with an accuracy of 95.8%, which is higher than previous methods considered in this study.

We propose a two‐dimensional (2D) scattering‐center‐extraction (SCE) method using sparse recovery based on the compressive‐sensing theory, even with data missing from the received radar cross‐section (RCS) dataset. First, using the proposed method, we generate a 2D grid via adaptive discretization that has a considerably smaller size than a fully sampled fine grid. Subsequently, the coarse estimation of 2D scattering centers is performed using both the method of iteratively reweighted least square and a general peak‐finding algorithm. Finally, the fine estimation of 2D scattering centers is performed using the orthogonal matching pursuit (OMP) procedure from an adaptively sampled Fourier dictionary. The measured RCS data, as well as simulation data using the point‐scatterer model, are used to evaluate the 2D SCE accuracy of the proposed method. The results indicate that the proposed method can achieve higher SCE accuracy for an incomplete RCS dataset with missing data than that achieved by the conventional OMP, basis pursuit, smoothed L0, and existing discrete spectral estimation techniques.

Abstract Background The Multicenter Orthopaedic Outcomes Network (MOON) group recently reported that medial meniscus (MM) repairs are associated with more frequent re-operations when compared to lateral meniscus (LM) repairs. The purpose of this study was to compare the meniscal healing and the incidence of subsequent re-operation of medial and lateral meniscal tears that occurred concurrently with anterior cruciate ligament (ACL) injuries. Methods We retrospectively reviewed patients who underwent second-look arthroscopy after primary ACL reconstruction (ACLR) between June 2005 to December 2016. The healing of meniscal tears following repair or left in situ, and re-tear following partial meniscectomy, were evaluated via second-look arthroscopy and compared between medial and lateral meniscus. Moreover, the incidence of subsequent meniscal re-operation after the index ACLR were investigated and compared between medial and lateral meniscus. Subsequent meniscal re-operation was performed in cases of the following three symptomatic meniscus tears: re-tears at the meniscectomy site; new tears; and failed healing of repaired or left in situ meniscus. Results There were 148 meniscal tears in 121 patients at index ACLR. There were 62 MM tears, 38 LM tears, and 24 bilateral meniscus tears. At second-look arthroscopy, the “successful healing” rate for tears following repair was higher in LM tears (91.2%) compared to MM tears (80.0%), although it was not statistically significant (p > 0.05). No significant differences were observed in the healing of left in situ tears or re-tear of meniscectomy site between medial and lateral meniscus. Patients with MM tears combined with ACL injuries had a higher incidence of subsequent meniscal re-operation compared to patients with LM tears (25.6% vs 16.1%, p = 0.025). Conclusions There was a trend for the successful healing rate to be higher in LM repairs than MM repairs. Subsequent meniscal re-operations after ACLR were more frequent in patients with medial meniscal tears concurrently with ACL injuries in comparison to patients with lateral meniscal tears. Level of study Level IV, retrospective case series.

Background: Although the use of adjustable-loop suspensory fixation has increased in recent years, the influence of the shortcomings of suspensory fixation, such as the bungee-cord or windshield-wiper effects, on tunnel widening remains to be clarified. Hypothesis/Purpose: The purpose of this study was to compare adjustable-loop femoral cortical suspensory fixation and interference screw fixation in terms of tunnel widening and clinical outcomes after anterior cruciate ligament reconstruction (ACLR). We hypothesized that tunnel widening in the adjustable-loop femoral cortical suspensory fixation (AL) group would be comparable to that in the interference screw fixation (IF) group. Methods: This study evaluated patients who underwent primary ACLR at our institution between March 2015 and June 2019. The femoral and tibial tunnel diameters were measured using plain radiographs in the immediate postoperative period and 2 years after ACLR. Tunnel widening and clinical outcomes (Lysholm score, 2000 International Knee Documentation Committee subjective score, and Tegner activity level) were compared between the two groups. Results: There were 48 patients (mean age, 29.8 ± 12.0 years) in the AL group and 44 patients (mean age, 26.0 ± 9.5 years) in the IF group. Tunnel widening was significantly greater in the AL group than that in the IF group at the tibia anteroposterior (AP) middle (2.03 mm vs. 1.32 mm, p = 0.017), tibia AP distal (1.52 mm vs. 0.84 mm, p = 0.012), tibia lateral proximal (1.85 mm vs. 1.00 mm, p = 0.001), tibia lateral middle (2.36 mm vs. 1.03 mm, p < 0.001), and tibia lateral distal (2.34 mm vs. 0.85 mm, p < 0.001) levels. There were no significant differences between the two groups with respect to femoral tunnel widening and clinical outcomes. Conclusions: Tibial tunnel widening was significantly greater in the AL group than in the IF group at 2 years after primary ACLR. However, the clinical outcomes in the two groups were comparable at 2 years.

We devise a layer‐wise hint training method to improve the existing hint‐based knowledge distillation (KD) training approach, which is employed for knowledge transfer in a teacher‐student framework using a residual network (ResNet). To achieve this objective, the proposed method first iteratively trains the student ResNet and incrementally employs hint‐based information extracted from the pretrained teacher ResNet containing several hint and guided layers. Next, typical softening factor‐based KD training is performed using the previously estimated hint‐based information. We compare the recognition accuracy of the proposed approach with that of KD training without hints, hint‐based KD training, and ResNet‐based layer‐wise pretraining using reliable datasets, including CIFAR‐10, CIFAR‐100, and MNIST. When using the selected multiple hint‐based information items and their layer‐wise transfer in the proposed method, the trained student ResNet more accurately reflects the pretrained teacher ResNet's rich information than the baseline training methods, for all the benchmark datasets we consider in this study.

A deep-learning technology for knowledge transfer is necessary to advance and optimize efficient knowledge distillation. Here, we aim to develop a new adversarial optimization-based knowledge transfer method involved with a layer-wise dense flow that is distilled from a pre-trained deep neural network (DNN). Knowledge distillation transferred to another target DNN based on adversarial loss functions has multiple flow-based knowledge items that are densely extracted by overlapping them from a pre-trained DNN to enhance the existing knowledge. We propose a semi-supervised learning-based knowledge transfer with multiple items of dense flow-based knowledge extracted from the pre-trained DNN. The proposed loss function would comprise a supervised cross-entropy loss for a typical classification, an adversarial training loss for the target DNN and discriminators, and Euclidean distance-based loss in terms of dense flow. For both pre-trained and target DNNs considered in this study, we adopt a residual network (ResNet) architecture. We propose methods of (1) the adversarial-based knowledge optimization, (2) the extended and flow-based knowledge transfer scheme, and (3) the combined layer-wise dense flow in an adversarial network. The results show that it provides higher accuracy performance in the improved target ResNet compared to the prior knowledge transfer methods.

Background: Although several studies have compared the clinical outcomes of septic and aseptic revision total knee arthroplasty (TKA), their results have been controversial. Therefore, this study aimed to compare clinical outcomes and complications of septic and aseptic revision TKA through a systematic review and meta-analysis. Methods: The PubMed (MEDLINE) and Embase databases were searched for studies evaluating the clinical outcomes and complications of 2-stage septic revision and aseptic revision TKAs. A systematic review of clinical outcomes (Knee Society Knee and Function Scores and range of motion) and complications (reoperation, infection, and failure rates) was conducted. Results: Thirteen studies were included in the systematic review. The mean MINORS (Methodological Index for NOn- Randomized Studies) score of the included studies was 20.5 (range, 18 to 22). The meta-analysis revealed higher reoperation (risk ratio [RR], 1.98; 95% confidence interval [CI], 1.50 to 2.62; p < 0.00001), infection (RR, 4.08; 95% CI, 2.94 to 5.64; p < 0.00001), and failure rates (RR, 2.88; 95% CI, 1.38 to 6.03; p = 0.005) in septic revision TKAs than in aseptic revision TKAs. Moreover, septic revision TKAs showed lower Knee Society Knee Scores compared with aseptic TKAs (mean difference [MD], 26.86; 95% CI, 211.80 to 21.92; p = 0.006). However, the Knee Society Function Score (MD, 21.84; 95% CI, 27.84 to 3.80; p = 0.52) and range of motion (MD, 26.96_; 95% CI, 216.23_ to 2.31_; p = 0.14) were not significantly different between septic and aseptic revision TKAs. Conclusions: Despite the heterogeneity of prosthesis designs and surgical protocols used in septic and aseptic revision TKAs, the results of this systematic review suggest that 2-stage septic revision TKAs have poorer clinical outcomes and higher complication rates than aseptic revision TKAs do. Level of Evidence: Therapeutic Level III. See Instructions for Authors for a complete description of levels of evidence. [ABSTRACT FROM AUTHOR]