Contrary to the common assumption that software is "environmentally friendly" simply because it virtual, processes and methods used develop, maintain deploy do have an environmental social impact normally not accounted for by development practices. For example, e-waste could be greatly minimized if vendors would take into consideration lifetime of old hardware. Like so, dependable minimizes waste resources support system. This paper introduces a set engineering metrics can assess economic,...

Accurate fault diagnosis is essential for the safe operation of rotating machinery. Recently, traditional deep learning-based have achieved promising results. However, most these methods focus only on supervised learning and tend to use small convolution kernels non-effectively extract features that are not controllable poor interpretability. To this end, study proposes an innovative semi-supervised method bearing diagnosis. Firstly, multi-scale dilated squeeze-and-excitation residual blocks...

Bearing pitting, one of the common faults in mechanical systems, is a research hotspot both academia and industry. Traditional fault diagnosis methods for bearings are based on manual experience with low diagnostic efficiency. This study proposes novel bearing method deep separable convolution spatial dropout regularization. Deep extracts features from raw vibration signals, during which 3 × 1 convolutional kernel one-step size selects effective by adjusting its weights. The similarity...

With the rapid development of aviation industry, videoscope inspection aeroengines has become crucial for ensuring aircraft flight safety. Recently, deep learning, particularly convolutional neural networks (CNNs), have shown remarkable efficacy in tasks. However, these methods usually require large-scale training labels with accurate annotation. In images, defects are often present at micrometer level and manual labeling by professional personnel, posing further challenges to model...

Deep supervised learning-based fault diagnosis methods require a large amount of labeled data, which frequently contradicts the typical engineering scenario, in numerous samples are available but only small portion labeled. To conduct this case, unsupervised and semi-supervised representation learning have recently been proposed. However, most these designed without much consideration for downstream classification tasks thus insufficiently relevant providing sufficient targeted assistance....

In remote sensing image analyses, the extraction of lake water bodies has been emphasized owing to its pivotal role in interpreting aquatic ecosystems, assessing hydrological trends, and detecting environmental changes. Although deep-learning-based techniques have effectively deployed for this task, several challenges persist, including mis-segmentation low-contrast regions, insufficient delineation fuzzy boundaries, over-segmentation micro-regions. To address these issues, an innovative,...

SaaS (Software as a Service) deliver software service over the Internet, eliminating need to install and run application on customers' own computers simplifying maintenance support. Access control is an important information security mechanism, according user identity attribution of predefined group users restrict access certain items, limit use functions. In view features multi-tenant, if we apply existing methods systems directly, following problems will appear: (1) role name conflicts (2)...

In aeroengine maintenance, endoscopic imaging serves as a crucial tool for detecting blade defects and evolves toward intelligence driven by computer vision technology. Currently, supervised-learning-based defect segmentation methods mainly rely on extensive pixel-level annotations, making it laborious time consuming. This article shifts focus to the abundant unlabeled data in real-world scenarios introduces an innovative semisupervised method termed SAIT. Within this framework, three...

In recent years, meta-learning-based methods have been widely used in cross-domain fault diagnosis and promising results can be obtained even with limited target training data. However, data scarcity problems exist not only the domain but also source domain, which puts a damper on meta-knowledge learning process since provide sufficient tasks. this study, novel adversarial one-shot network named AOCN for is proposed, requires few samples as low one labeled sample per class. The main idea of...

The performance of many components in intelligent transportation systems depends heavily on the quality traffic forecasting. After analyzing deficiency existing algorithm and methods forecasting, we develop a new forecasting model based logic reasoning this paper, describe details each part model. Finally through an example, introduce working order

SaaS (software as a service) is new software development mode. It has drawn many IT workers' attentions. Its core concept that transform from developing softwares into services in the process of development. And applied mode developed system can provide users with greater operating space, and it be more meet needs users. In application system, multi-tenants share single instance effectively reduce costs on hardware maintenance maximize scale application. Although architecture implements true...

Convolutional neural networks(CNNs) show significant potential for bearing fault diagnosis. However, tra-ditional CNNs face challenges such as poor noise resistance, high computational complexity, reliance on extensive samples, and limited generalizability. As a result, this paper proposes WDSC-Net, lightweight, multiscale feature fusion method, focusing labeled samples. Initially, wide kernel convo-lutional is employed, aiming to reduce parameters complexity. Next, features are fed into 1×1...

X-ray nondestructive testing means are widely used in the inspection process of internal defects parts. In practical inspection, generally determined and rated by manual based on images, which is inefficient cannot meet requirements high-volume automatic inspection. This paper proposes an defect classification model improved U-Net. First, a classifier added behind encoder The connected series to form main branch complete task. Second, decoder U-Net adding attention module. auxiliary model,...