Collapse Modes Research Articles

Prediction of erosion defects and buckling pressure of pipe bend based on bidirectional fluid-structure interaction method

In this study, a bidirectional fluid-structure interaction (FSI) method was proposed based on dynamic mesh technology to calculate the erosion defects of pipe bends under high pressure. In calculating the erosion rate of pipe bends, a stress erosion model was adopted, and the displacement and stress of the inner wall of pipe bend under high external pressure were considered. The influence of different parameters on the calculation results of erosion depth was discussed, and the buckling pressure and collapse mode of pipe bend under the erosion defect of liquid-solid flow were also studied. The proposed method can be used for predicting pipeline erosion defects under high pressure and evaluating the remaining limit load of pipelines with such erosion defects.

Advancements in Text-to-Image Generation through Generative AI

Text-to-image generation, a fascinating intersection of natural language processing and computer vision, has witnessed remarkable progress in recent years. This research paper provides a comprehensive review of the state-of-the-art techniques, challenges, and applications in the field of text-to-image generation. The paper aims to analyze various approaches, discuss their strengths and limitations, and highlight potential directions for future research. Generative Artificial Intelligence (Generative AI) has revolutionized the fusion of textual information and visual content, giving rise to sophisticated Text-to-Image generators. In this context we will discuss dynamic landscape of Generative AI-driven text-to-image synthesis, exploring the state-of-the-art models, underlying architectures, and the impact of training strategies on the quality of generated images. The paper provides a comprehensive overview of Training strategies, encompassing dataset selection and fine-tuning approaches, are scrutinized for their impact on model performance. Common challenges, such as handling ambiguous textual descriptions and ensuring the avoidance of mode collapse, are addressed, offering insights into potential avenues for improvement. Training strategies, encompassing dataset selection and fine-tuning approaches, are scrutinized for their impact on model performance. Common challenges, such as handling ambiguous textual descriptions and ensuring the avoidance of mode collapse, are addressed, offering insights into potential avenues for improvement. Applications of Generative AI text-to-image generators are explored, ranging from content creation to virtual environment design, highlighting their versatility and real-world utility. Ethical considerations surrounding potential misuse, including the creation of deepfakes, are examined to foster a balanced understanding of the technology's societal implications. The research paper concludes with a forward-looking exploration of future directions in Generative AI text-to-image generation. Emphasizing the transformative potential of this technology, the paper envisions advancements in model architectures, training methodologies, and emerging applications, inviting researchers and practitioners to contribute to the ongoing evolution of this exciting field.

STABLE DIFFUSION TEXT TO IMAGE USING AI

The Stable Diffusion Text-to-Image Generation Project is an innovative endeavor in the field of generative adversarial networks (GANs) and natural language processing (NLP). This project aims to bridge the semantic gap between textual descriptions and visual content by utilizing the Stable Diffusion training framework to generate highly realistic and coherent images from text prompts. The project leverages recent advancements in deep learning techniques to tackle the challenging task of text-to image synthesis. The project introduces an innovative approach at the crossroads of generative adversarial networks (GANs) and natural language processing (NLP). It aims to bridge the semantic gap between textual descriptions and visual content by utilizing the Stable Diffusion training framework. The key goal is to generate highly realistic and coherent images from text prompts, leveraging recent deep learning Stable Diffusion. The Stable Diffusion training framework plays a central role in this project. It’s a sophisticated training methodology for GANs, designed to stabilize the training process. GANs have exhibited great potential in generating images but often suffer from issues like mode collapse and training instability. Keywords: Stable Diffusion, Text-to-Image Generation, Image Synthesis, Natural Language Processing(NLP), Generative Adversarial Networks (GANs)

Collapse analysis of restrained concrete-filled steel tubular columns in fire conditions

This paper presents numerical analyses of the thermo-mechanical behavior of restrained square concrete-filled steel tubular (CFST) columns subjected to standard fire. It is found that the behavior is significantly influenced by the load ratio and axial restraint level. The collapse of restrained columns in fire conditions can be defined as the axial force in columns reaching 0 or the axial contraction deformation reaching a hundredth of column length. Three collapse modes are discovered, i.e., the expansion deformation collapse mode, contraction deformation collapse mode, and load-bearing capacity loss collapse mode. Under a low load ratio, three distinct modes of collapse are possible. Conversely, at a high load ratio, only the contraction deformation collapse mode and the load-bearing capacity loss collapse mode are observed. Through analytical and numerical approaches, a method for identifying the collapse mode based on geometric dimensions, axial restraint level, and load ratio has been established. This identification method can significantly contribute to optimizing fire protection strategies for CFST columns in structural designs.

SURVEY ON STABLE DIFFUSION TEXT TO IMAGE USING AI

The Fusion Nexus Text-to-Image Synthesis Initiative integrates cutting-edge Generative Adversarial Networks (GANs) with Natural Language Processing (NLP) techniques to narrow the semantic divide between textual input and visual output. Built upon the robust Stable Diffusion training paradigm, this initiative is engineered to produce immersive, true-to-life images based on descriptive text prompts. While GANs have exhibited potential in image generation, issues like mode collapse and training instability have impeded their effectiveness. The Fusion Nexus initiative circumvents these challenges by leveraging the Stable Diffusion framework, which furnishes a stable and reliable training methodology for GANs. By amalgamating recent advancements in deep learning, this project spearheads a novel approach to text-to-image synthesis. Its primary aim is to craft cohesive and highly realistic visual representations from textual descriptions, thereby bridging the gap between linguistic expression and visual perception. This ambitious undertaking marks a significant stride at the convergence of GANs and NLP, presenting a promising solution to the intricate task of text-to-image generation. Keywords: Adversarial, Diffusion, Framework, Fusion, GANs, Generation, Generation, Image, Language, Natural, NLP, Processing, Robust, Stable, Text-to-image, Training, Visuals.

Diff-IF: Multi-modality image fusion via diffusion model with fusion knowledge prior

Multi-modality image fusion (MMIF) aims to aggregate the complementary information from diverse source image domains. As the generative adversarial network-based methods have been the primary choice and demonstrated satisfactory fusion performance, they suffer from the unstable training and mode collapse. To tackle this challenge, we propose a novel diffusion model for MMIF incorporating fusion knowledge prior, termed as Diff-IF. Diff-IF proposes a trainable diffusion model paradigm for multi-modality image fusion, resolving the issue of lacking the ground truth for the diffusion model in image fusion tasks. It decomposes the diffusion-based image fusion method into conditional diffusion model and fusion knowledge prior with the targeted search to derive the prior distribution for the specific image fusion task. In particular, the forward diffusion process is guided by the fusion knowledge prior distribution through targeted search, while the reverse diffusion process is designed to generate high-quality fused images. Extensive experiments demonstrate that Diff-IF achieves outstanding performance, including exemplary visual preservation, and good preservation of weak textures, across various MMIF tasks such as infrared-visible image fusion and medical image fusion. The code will be available at https://github.com/XunpengYi/Diff-IF.

Text Feature Adversarial Learning for Text Generation With Knowledge Transfer From GPT2.

Text generation is a key component of many natural language tasks. Motivated by the success of generative adversarial networks (GANs) for image generation, many text-specific GANs have been proposed. However, due to the discrete nature of text, these text GANs often use reinforcement learning (RL) or continuous relaxations to calculate gradients during learning, leading to high-variance or biased estimation. Furthermore, the existing text GANs often suffer from mode collapse (i.e., they have limited generative diversity). To tackle these problems, we propose a new text GAN model named text feature GAN (TFGAN), where adversarial learning is performed in a continuous text feature space. In the adversarial game, GPT2 provides the "true" features, while the generator of TFGAN learns from them. TFGAN is trained by maximum likelihood estimation on text space and adversarial learning on text feature space, effectively combining them into a single objective, while alleviating mode collapse. TFGAN achieves appealing performance in text generation tasks, and it can also be used as a flexible framework for learning text representations.

Athermal pulsational mode dynamics with negative ions

The stability dynamics of the pulsational mode of gravitational collapse (PMGC), excitable in athermal (nonthermal) spherical complex dust molecular clouds (DMCs) in the presence of active role of negative ions, is semi-analytically investigated. The adopted partially ionized DMC model consists of five constitutive plasma species, well-coupled via the gravito-electrostatic Poisson formalism, depicting the corresponding potential-density correlations. The nonthermal uncorrelated lighter (inertialess) species are electrons, positive ions, and negative ions. The thermal correlated heavier (inertial) species are neutral and charged dust grains. The application of a standard spherical wave analysis results in a generalized linear cubic dispersion equation involving a new set of multi-factorial dispersion coefficients dependent on diverse equilibrium plasma parameters. A numerical illustrative platform is broadly created to analyze the DMC stabilization and destabilization factors relative to the non-local inward Newtonian self-gravity. It is interestingly seen that the DMC stabilizing agents are the radial cloud size, dust charge, neutral dust viscoelastic relaxation time, Tribeche thermostatistical parameters, etc. In contrast, the dust mass and equilibrium neutral dust density show destabilizing effects, and so forth. The results systematically explored here are fairly consistent and reliably correlated with the previous results on structure formation mechanism in the DMCs as heretofore reported extensively in the literature.

Hierarchical Clustering-Based Collapse Mode Identification and Design Optimization of Energy-Dissipation Braces Inspired by the Triangular Resch Pattern

Hierarchical Clustering-Based Collapse Mode Identification and Design Optimization of Energy-Dissipation Braces Inspired by the Triangular Resch Pattern

Comprehensive Overview of GAN Technology: Architecture, Training, and Applications

Abstract: Generative Adversarial Networks (GANs) is a groundbreaking artificial intelligence technology that transforms generative modeling through the implementation of a novel adversarial training framework. GANs are made up of two neural networks, the generator and the discriminator, which compete in a minimax game in order to generate counterfeit samples of data and distinguish between real and fake data. This adversarial training method results in the creation of highly proficient generative models capable of producing data that is identical to real-world samples. GANs have been shown to have significant outcomes in a variety of programs, consisting of synthesizing images, style transfer, medical visualization, and natural language processing. However, GANs encounter difficulties with problems like mode collapse and operational irregularities. Presently, research is aimed at tackling these problems and strengthening GAN frameworks and training approaches. GANs’ adaptability and perspective contribute to an intriguing technology for a wide range of sectors and artistic activities, with consequences for artificial intelligence growth and generative modeling breakthroughs. GANs incorporate the potential of competitive training with their capacity to generate extremely realistic data in a variety of domains. We will begin by looking at the fundamental concepts, and underlying principles that reinforce GANs, and their general architecture, investigate their different possible uses, and discuss the obstacles and potential developments in this instantly transforming field.

Image Generation from Text

Abstract: This work diverts from the usual by using Generative Adversarial Networks (GANs) to intentionally create flawed yet captivating images. We introduce a controlled chaos system, where a stalling generator network and a disoriented discriminator network co-exist. This strange training method encourages the development of peculiar irregularities within semi-realistic images. We welcome the unpredictability that arises from this chaos, purposely avoiding issues like mode collapse. Moreover, the project delves into a new evaluation paradigm for evaluating these deliberately flawed images. We go beyond typical metrics, exploring the use of faulty metrics and questionable visual inspection techniques. In addition, we investigate the impact of clumsy dataset selection, potentially employing unconventional or tampered datasets to train the GAN for specific types of errors. This work explores the potential of GANs for jumbled up art creation, data manipulation, and image confusion, opening a new frontier in GAN applications. While potentially causing confusion in the field, this research also reveals the unimaginable artistic and data exploration capabilities of GANs.

Self-clustered GAN for precipitation nowcasting

This paper proposes a novel GAN framework with self-clustering approach for precipitation nowcasting (ClusterCast). Previous studies have primarily captured the motion vector using only a single latent space, making the models difficult to adapt to disparate space-time distribution of precipitation. Environmental factors (e.g., regional characteristics and precipitation scale) have an impact on precipitation systems and can cause non-stationary distribution. To tackle this problem, our key idea is to train a generator network to predict future radar frames by learning a sub-network that automatically labels precipitation types from a generative model. The training process consists of (i) clustering the hierarchical features derived from the generator stem using a sub-network and (ii) predicting future radar frames according to the self-supervised labels, enabling heterogeneous latent representation. Additionally, we attempt an ensemble forecast that prescribes random perturbations to improve performance. With the flexibility of representation learning, ClusterCast enables the model to learn precipitation distribution more accurately. Results indicate that our method generates non-blurry future frames by preventing mode collapse, and the proposed method demonstrates robustness across various precipitation scenarios. Extensive experiments demonstrate that our method outperforms four benchmarks on a 2-h prediction basis with a mean squared error (MSE) of 8.9% on unseen datasets.

Effect of chord-cut defects on the collapse response of thick-walled pipelines

Corrosion defects can significantly reduce the ultimate bearing capacity of thick-walled pipelines in deep water. Uneven depth defects pose a serious threat to the integrity of pipeline structures. However, in the study of uneven depth defects, the influence of chord-cut defects is often overlooked. To study the impact of chord-cut corrosion defect on pipeline collapse response, a dynamic model is established. Two sets of full-scale pipeline experiments are conducted to verify the model. The validated model is used to obtain the pipeline collapse responses under different defect parameters, which include defect length, defect depth, defects’ location, and pipeline ovality. The changes in pipeline collapse pressure and collapse mode under different defect parameters are analyzed. The research shows the increase in defect length and the decrease in diameter-to-thickness ratio (D/t) will enhance the influence of defect angle θ on pipeline collapse configuration. And the change of collapse mode will lead to abnormal changes in collapse pressure. With the increase of internal defect length, the collapse mode variation rule is “∞” - “∩” - “P” - “U”. A collapse pressure empirical formula for pipeline with chord-cut corrosion defects is obtained. The empirical formula fits well with the numerical simulation results.

Image Enhancement Based on Dual-Branch Generative Adversarial Network Combining Spatial and Frequency Domain Information for Imbalanced Fault Diagnosis of Rolling Bearing

To address the problems of existing 2D image-based imbalanced fault diagnosis methods for rolling bearings, which generate images with inadequate texture details and color degradation, this paper proposes a novel image enhancement model based on a dual-branch generative adversarial network (GAN) combining spatial and frequency domain information for an imbalanced fault diagnosis of rolling bearing. Firstly, the original vibration signals are converted into 2D time–frequency (TF) images by a continuous wavelet transform, and a dual-branch GAN model with a symmetric structure is constructed. One branch utilizes an auxiliary classification GAN (ACGAN) to process the spatial information of the TF images, while the other employs a GAN with a frequency generator and a frequency discriminator to handle the frequency information of the input images after a fast Fourier transform. Then, a shuffle attention (SA) module based on an attention mechanism is integrated into the proposed model to improve the network’s expression ability and reduce the computational burden. Simultaneously, mean square error (MSE) is integrated into the loss functions of both generators to enhance the consistency of frequency information for the generated images. Additionally, a Wasserstein distance and gradient penalty are also incorporated into the losses of the two discriminators to prevent gradient vanishing and mode collapse. Under the supervision of the frequency WGAN-GP branch, an ACWGAN-GP can generate high-quality fault samples to balance the dataset. Finally, the balanced dataset is utilized to train the auxiliary classifier to achieve fault diagnosis. The effectiveness of the proposed method is validated by two rolling bearing datasets. When the imbalanced ratios of the four datasets are 0.5, 0.2, 0.1, and 0.05, respectively, their average classification accuracy reaches 99.35% on the CWRU bearing dataset. Meanwhile, the average classification accuracy reaches 96.62% on the MFS bearing dataset.

Shaking table tests on inclined collapse mechanism of RC frame structures subjected to earthquakes

Inclined collapse of structures under strong earthquakes threatens the adjacent infrastructures. The collapse mode and collapse direction should be considered with caution to curb the secondary disaster in the urban area. However, current studies on the inclined collapse mechanism are predominantly limited to the qualitative and intuitive understanding. To fill this gap, this study presents an insight into the inclined collapse mechanism of the reinforced concrete (RC) frame structures with two typical features subjected to earthquakes, i.e., the structures with slab opening on one side of the first floor and the structures with disparate axial compression ratios of the first story columns on the opposite sides. First, shaking table tests were performed on two 1/10-scale three-story RC frame structure models undergoing inclined collapse. Then the multi-scale finite element models were established, validated against the test results and used for parametric studies. In the parametric studies, two coefficients were considered to describe the typical features of the RC frame structures, i.e., the strong column-weak beam ratio η and differential rate of the axial compression ratios λ. The test and numerical results found that the inclined collapse mechanism was attributed to the different distributions and discrepant resistance deterioration of the plastic hinges among the first story columns of the RC frame structures with the two features. This causes the inconsistent damage levels among the first story columns, the dominant direction of the lateral displacement, and the inclined collapse of the structures. Finally, the collapse mode and collapse direction can be controlled by appropriately adjusting the two coefficients η and λ for the structures with the typical features concerned in this study.

Large deformation response of a novel triply periodic minimal surface skeletal-based lattice metamaterial with high stiffness and energy absorption

Triply periodic minimal surface (TPMS) lattice metamaterials present superior mechanical performance over traditional strut-based lattice metamaterials due to their unique structural characteristics. However, the high stiffness–stable plastic response trade-off dilemma is still the major challenge for skeletal-based metamaterials. Herein, a novel TPMS skeletal lattice metamaterial is proposed. Finite element simulations are conducted to reveal its elastic properties and plastic response under compression. Afterwards, validation experiments are performed on the lattice specimens fabricated by the most common Fused Deposition Modeling (FDM) process. The nominal stress–strain curves and collapse mode of the lattice materials are obtained accordingly. Both the numerical simulations and experiments demonstrate that the proposed TPMS lattice metamaterial exhibits high specific modulus, strength and energy absorption, together with a smooth and elongated stress plateau, thereby overcoming the strength-efficiency trade-off. Different from the rigid nodal region in traditional strut-based lattices, the strut connection parts in the novel TPMS lattices experience shear and twisting, which avoids the catastrophic failure of the struts and enhances the loading efficiency of the structure under compression. Meanwhile, the numerical results indicate that the stable plastic response of the designed architecture is insensitive to the plastic flow behavior of the bulk material, which is also supported by the experimental results. It is also demonstrated that the novel TPMS lattice metamaterial presents several times higher stiffness as well as energy absorption simultaneously than the current strut-based lattice materials, which can be potentially applied as load-bearing components and impact energy absorbers.

A Comprehensive guide to Generative Adversarial Networks (GANs) and application to individual electricity demand

Generative Adversarial Networks (GANs) have become a popular research topic in many research fields in the last decade. The fragmented body of knowledge on GANs drives researchers to a trial-and-error procedure while choosing the right GAN for a given task. This study provides a comprehensive guide on GANs, where we start with considering problems such as nonconvergence, mode collapse, vanishing gradient, and unstable training. We then compare various GANs concerning the application perspective, their outputs, and evaluation metrics. In addition, we suggest using a new evaluation metric to identify the best candidate of GANs for time series tasks, the so-called average coverage error (ACE). Finally, we discuss the application of the ACE using the electricity consumption of an individual industry customer and display significant selection criteria. The ACE is beneficial in identifying the best GANs by comparing outcomes and reduces the computational cost of searching related to the evaluation metrics.

A double-stage buckling restrained brace with progressive failure pattern: Experimental verification and numerical calibration

Buckling-restrained brace (BRB) is a widely used energy dissipator by the yielding of core steel plate. Single-core BRB (SBRB) only has one yielding point and a single failure pattern. Consequently, the SBRB has a risk of abrupt failure under extremely rare earthquakes, which may compromise the seismic safety and collapse mode of the building with SBRB. A new double-stage buckling-restrained brace (DSBRB) characterized by a progressive failure pattern is proposed. The progressive failure pattern of DSBRB is realized by setting up multiple core plates with different cross-sectional areas and lengths in the yielding part. A quasi-static test was conducted to investigate the difference in cyclic behavior and failure pattern between the SBRB and DSBRB. The test results showed that the yielding displacement and yielding force of the tested DSBRB were 8.74 % and 5.64 % smaller than those of SBRB respectively, resulting in a greater equivalent damping ratio for DSBRB under the small deformation. SBRB showed an abrupt failure pattern while the DSBRB was a progressive double-stage failure pattern at the cost of a reduced equivalent viscous damping ratio under the extreme deformation. The theory equations about the elastic stiffness, yielding force, and yielding displacement of DSBRB were derived and justified by the test results, i.e., differences between test results and theoretical derivation were all within ± 8.51 %. Two numerical models were established in ABAQUS and calibrated by the test results. The concept of double-stage BRB with a progressive failure pattern was verified.

Data enhancement method based on cyclic generation adversarial network

This study introduces a novel data augmentation technique employing Cycle Generative Adversarial Networks (CycleGAN) to mitigate the challenges posed by the paucity of image datasets in deep learning domains. Through the adept training of a CycleGAN model, this method substantially enriches image datasets, thereby enhancing the efficiency of deep learning models in target detection tasks. Distinct from conventional approaches, our strategy incorporates advanced activation functions, relative discriminators, and residual connections, which collectively foster greater image diversity and mitigate mode collapse, all while maintaining a low computational overhead. Evaluations conducted on diverse datasets, including MNIST, Synthetic Aperture Radar (SAR), and medical blood cell images, demonstrate the method's superior augmentation capabilities compared to traditional Deep Convolutional GAN (DCGAN) techniques, underscoring its efficacy and potential utility in preprocessing for deep learning applications.

Compressive properties of AlSi10Mg lattice structures with novel BCCZZ and FCCZZ configurations fabricated by selective laser melting

PurposeThis paper aims to examine the static compression characteristics of cell topologies in body-centered cubic with vertical struts (BCCZ) and face-centered cubic with vertical struts (FCCZ) along with novel BCCZZ and FCCZZ lattice structures.Design/methodology/approachThe newly developed structures were obtained by adding extra interior vertical struts into the BCCZ and FCCZ configurations. The samples, composed of the AlSi10Mg alloy, were fabricated using the selective laser melting (SLM) additive manufacturing technique. The specific compressive strength and failure behavior of the manufactured lattice structures were investigated, and comparative analysis among them was done.FindingsThe results revealed that the specific strength of BCCZZ and FCCZZ samples with 0.5 mm strut diameter exhibited approximately a 23% and 18% increase, respectively, compared with the BCCZ and FCCZ samples with identical strut diameters. Moreover, finite element analysis was carried out to simulate the compressive response of the lattice structures, which could be used to predict their strength and collapse mode. The findings showed that while the local buckling of lattice cells is the major failure mode, the samples subsequently collapsed along a diagonal shear band.Originality/valueAn original and systematic investigation was conducted to explore the compression properties of newly fabricated lattice structures using SLM. The results revealed that the novel FCCZZ and BCCZZ structures were found to possess significant potential for load-bearing applications.