Search Results (3)

Traditional fault diagnosis methods for rolling bearings rely on nemerous labeled samples, which are difficult to obtain in engineering applications. Moreover, when unseen fault categories appear in the test set, these models fail to achieve accurate diagnoses, as the fault categories are not represented in the training data. To address these challenges, a zero-shot fault diagnosis model for rolling bearings is proposed, which realizes knowledge transfer from seen to unseen categories by constructing attribute information, thereby reducing the dependence on labeled samples. First, an attribute method Discrete Label Embedding Method (DLEM) based on word embedding and envelope analysis is designed to generate fault attributes. Fault features are extracted using the Swin Transformer model. Then, the attributes and features are input into the constructed model Distribution Consistency and Multi-modal Cross Alignment Variational Autoencoder (DCMCA-VAE), which is built on Convolutional Residual SE-Attention Variational Autoencoder (CRS-VAE). CRS-VAE replaces fully connected layers with convolutional layers and incorporates residual connections with the Squeeze-and-Excitation Joint Attention Mechanism (SE-JAM) for improved feature extraction. The DCMCA-VAE also incorporates a reconstruction alignment module with the proposed distribution consistency loss $L_{WT}$ and multi-modal cross alignment loss function $L_{MCA}$. The reconstruction alignment module is used to generate high-quality features with distinguishing information between different categories for classification. In the face of multiple noisy datasets, this model can effectively distinguish unseen categories and has stronger robustness than other models. The model can achieve 100% classification accuracy on the SQ dataset, and more than 85% on the CWRU dataset when unseen and seen categories appear simultaneously with noise interference. Full article

The clinical application of different relative biological effectiveness (RBE) models for carbon ion RBE-weighted dose calculation hinders a global consensus in defining normal tissue constraints. This work aims to update the local effect model (LEM)-based constraints for the rectum using microdosimetric kinetic model (mMKM)-defined values, relying on RBE translation and the analysis of long-term clinical outcomes. LEM-optimized plans of treated patients, having suffered from prostate adenocarcinoma (n = 22) and sacral chordoma (n = 41), were recalculated with the mMKM using an in-house developed tool. The relation between rectum dose-volume points in the two RBE systems (D_LEM|v and D_MKM|v) was fitted to translate new LEM-based constraints. Normal tissue complication probability (NTCP) values, predicting late rectal toxicity, were obtained by applying published parameters. No late rectal toxicity events were reported within the patient cohort. The rectal toxicity outcome was confirmed using dosimetric analysis: D_MKMVHs lay largely below original constraints; the translated D_LEM|v values were 4.5%, 8.3%, 18.5%, and 35.4% higher than the nominal D_MKM|v of the rectum volume, v—1%, 5%, 10% and 20%. The average NTCP value ranged from 5% for the prostate adenocarcinoma, to 0% for the sacral chordoma group. The redefined constraints, to be confirmed prospectively with clinical data, are D_LEM|5cc ≤ 61 Gy(RBE) and D_LEM|1cc ≤ 66 Gy(RBE). Full article

Deltas and estuaries built by the Mississippi/Atchafalaya River (MAR) in the United States and the Usumacinta/Grijalva River (UGR) in Mexico account for 80 percent of all Gulf of Mexico (GoM) coastal wetlands outside of Cuba. They rank first and second in freshwater discharge to the GoM and owe their natural resilience to a modular geomorphology that spreads risk across the coast-scape while providing ecosystem connectivity through shelf plumes that connect estuaries. Both river systems generate large plumes that strongly influence fisheries production over large areas of the northern and southern GoM continental shelves. Recent watershed process simulations (DLEM, MAPSS) driven by CMIP3 General Circulation Model (GCM) output indicate that the two systems face diverging futures, with the mean annual discharge of the MAR predicted to increase 11 to 63 percent, and that of the UGR to decline as much as 80 percent in the 21st century. MAR delta subsidence rates are the highest in North America, making it particularly susceptible to channel training interventions that have curtailed a natural propensity to shift course and deliver sediment to new areas, or to refurbish zones of high wetland loss. Undoing these restrictions in a controlled way has become the focus of a multi-billion-dollar effort to restore the MAR delta internally, while releasing fine-grained sediments trapped behind dams in the Great Plains has become an external goal. The UGR is, from an internal vulnerability standpoint, most threatened by land use changes that interfere with a deltaic architecture that is naturally resilient to sea level rise. This recognition has led to successful efforts in Mexico to protect still intact coastal systems against further anthropogenic impacts, as evidenced by establishment of the Centla Wetland Biosphere Preserve and the Terminos Lagoon Protected Area. The greatest threat to the UGR system, however, is an external one that will be imposed by the severe drying predicted for the entire Mesoamerican “climate change hot-spot”, a change that will necessitate much greater international involvement to protect threatened communities and lifeways as well as rare habitats and species. Full article