Search

Fuel load significantly affects fire development in compartments, and its design value derives from the statistical results of numerous surveys. This work enhances the fuel load database by employing a digitized survey method to assess 27 university and commercial offices in Hong Kong and Mainland, China. The results revealed significant differences in fuel load densities: university offices averaged 382 MJ/m2 (SD: 297 MJ/m2) and commercial offices averaged 1804 MJ/m2 (SD: 1319 MJ/m2). Moreover, it proposed an online questionnaire method to overcome onsite access limitations. Furthermore, it identified higher fuel load densities in commercial offices due to higher paper-made content and greater occupancy density than university offices. The fuel load of university offices was lower than that in previous surveys and design codes, while commercial fuel load was higher. Notably, there is an increasing tendency of fuel load density and plastic combustible composition over the years. Additionally, it considers that Gompertz distribution better fits cumulative probabilities of fuel load density data with a small sample size. Overall, it provides a valuable database for future fire scenario design, fire codes edition, and fire safety assessment and discusses future collaboration with AI applications. • Conducted a digitized fuel load survey in 27 university and commercial offices. • Revealed differences in fuel load density and combustible composition between two office types. • Statistical analyzed the variation among this survey, past literature and design codes. • Revealed Gompertz distribution's superiority for fitting fuel load cumulative probability. • Provided a valuable digitized fuel load database for future fire safety assessment. [ABSTRACT FROM AUTHOR]

Open-vocabulary learning can identify categories marked during training (seen categories) and generalize to categories not annotated in the training set (unseen categories). It could theoretically extend segmentation systems to more universal applications. However, current open-vocabulary segmentation frameworks are primarily suited for specific tasks or require retraining according to the task, and they significantly underperform in inferring seen categories compared to fully supervised frameworks. Therefore, we introduce a universal open-vocabulary segmentation framework based on the latent diffusion process (SegLD), which requires only a single training session on a panoptic dataset to achieve inference across all open-vocabulary segmentation tasks, and reaches SOTA segmentation performance for both seen and unseen categories in every task. Specifically, SegLD comprises two stages: in the first stage, we deploy two parallel latent diffusion processes to deeply fuse the text (image caption or category labels) and image information, further aggregating the multi-scale features output from both latent diffusion processes on a scale basis. In the second stage, we introduce text queries, text list queries, and task queries, facilitating the learning of inter-category and inter-task differences through the computation of contrastive losses between them. Text queries are then further fed into a Transformer Decoder to obtain category-agnostic segmentation masks. Then we establish classification loss functions for the type of text input during training, whether image captions or category labels, to help assign a category label from the open vocabulary to each predicted binary mask. Experimental results show that, with just a single training session, SegLD significantly outperforms other contemporary SOTA fully supervised segmentation frameworks and open-vocabulary segmentation frameworks across almost all evaluation metrics for both known and unknown categories on the ADE20K, Cityscapes, and COCO datasets. This highlights SegLD's capability as a universal segmentation framework, with the potential to replace other segmentation frameworks and adapt to various segmentation domains. The project link for SegLD is https://zht-segld.github.io/. • One-shot training achieves multi-task open-vocabulary segmentation tasks. • Two parallel latent diffusion processes deeply infuse multimodal information. • Learning inter-category and inter-task differences is established by contrastive loss. • Achieved SOTA performance in segmentation for both seen and unseen categories. [ABSTRACT FROM AUTHOR]

Microcellular injection moulding is an important injection moulding technique to create foaming plastic parts. However, there are no consistent conclusions on the impact of processing parameters on the cell morphology of microcellular injection moulded parts. This paper investigates the influence of the main processing parameters, such as melt temperature, mould temperature, injection pressure, flow rate, shot volume and gas dosage amount, on the average cell size and weight reduction of a talc-reinforced polypropylene square part (165 mm × 165 mm × 3.2 mm), using the commercial software Moldex 3D. The effect of each parameter is investigated considering a range of values and the simulation results were compared with published experimental results. The differences between numerical and experimental trends are discussed. [ABSTRACT FROM AUTHOR]

Deep transfer learning (DTL) greatly improved the cross-domain generalization of fault diagnosis and makes it more practical and operable. However, existing work focuses on addressing temporal feature shift, while neglecting the modeling and narrow of spectral feature shift. To solve this issue, this work focus on the study of temporal–spectral domain adaption (TSDA) for bearing fault diagnosis and proposes a temporal–spectral domain adaptive network (TSDAN). Specifically, TSDAN constructs a temporal–spectral representation by extracting temporal features and spectral features through two branching modules: a convolutional network and a novel spectral neural network, respectively. To construct spectral neural networks, we introduce spectral convolution, spectral pooling, spectral normalization, and spectral activation. Moreover, a Sinkhorn divergence-based temporal-spectrum adapter is designed to align the temporal-spectrum representations from the source and target domains. Finally, we provide the implementation details of TSDAN-based fault diagnosis on publicly available and self-built datasets, which validate the effectiveness and superiority of the proposed approach. • A novel temporal–spectral domain adaptation network (TSDAN) is proposed. • TSDAN extracts both temporal and spectral features. • TSDAN narrows both temporal and spectral feature shift. • Experimental case studies verified validity and superiority. [ABSTRACT FROM AUTHOR]

As a classical insulated gate bipolar transistor (IGBT) structure, the trench-gate field-stop (FS) IGBT has been widely used. For more convenient application, physics-based trench-gate FS IGBT models have received great attentions due to its effectiveness in simulating key device properties. However, current IGBT modeling studies for this popular IGBT chip seldom investigate the depletion layer behavior under the trench-gate structure. This article first derives the physics-based miller capacitance formula for the trench-gate structure according to finite element method (FEM) simulation. Another haunting issue for these physics-based IGBT models is that the intrinsic device parameters are difficult to be extracted appropriately. Thus, an FEM FS IGBT model is established in Sentaurus TCAD as the benchmark. Then, an improved parameter extraction procedure for a Fourier-series-based (FSB) IGBT model is proposed based on the improved miller capacitance formula. By the particle swarm optimization algorithm, key device parameters for the FSB IGBT model can be successfully extracted by fitting the simulated waveforms to those by FEM simulation at a randomly selected turn-off operating condition. Parameters extracted through the proposed procedure using the FSB IGBT model with the improved miller capacitance formula are very close to the device parameters of the FEM model. More importantly, the reliability and the robustness of the extracted parameters are further validated by comparing the waveforms of FSB model with those of FEM model at different operating conditions. [ABSTRACT FROM AUTHOR]

Transfer learning (TL) and domain adaptation (DA) have aroused the interest of many researchers and are used in the prognostics and health management (PHM) of bearings. However, traditional DA methods tend to lose fine-grained information when aligning the domains globally. Moreover, most DA methods are not suitable for regression tasks, such as remaining useful life (RUL) prediction. This article proposes a deep subdomain adaptive regression network (DSARN), which can align the relevant subdomains in the source and target domains. This method has good cross-domain generalization ability and thus can be used for RUL prediction of bearings under multiple operating conditions. The proposed method takes the multichannel time–frequency representation (TFR) of bearings as the input and uses DSARN to extract the domain-invariant hidden representation and to establish a mapping to RUL. The effectiveness of the proposed method is verified by case studies on bearings. A comparison with other state-of-the-art deep learning (DL) and TL methods also proves the superiority of the proposed method. [ABSTRACT FROM AUTHOR]

Deep learning (DL) has shown effectiveness in the field of Prognostic and Health Management (PHM) of rolling bearings and successfully solved many problems with its powerful feature learning and function fitting capabilities. However, it is still very challenging to transfer a remaining useful life (RUL) prediction model trained by historical data to bearings running under new operating conditions. Domain adaptation (DA) is an effective solution to obtain cross-domain features of bearings. This article presents an RUL prediction method for rolling bearings based on a deep transfer auto-encoder. Two strategies, parameter transfer and feature representation transfer, are used to transfer the feature extraction model and RUL regression model trained in the labeled source domain to the unlabeled target domain. In this way, the full utilization of the historical model and the rapid adaptation to the new operating conditions are realized. A case study on the IEEE PHM Challenge 2012 bearing data set verifies the effectiveness of the proposed method. The comparison with other nontransfer and transfer machine learning models shows the advantages of this method in precision and accuracy. [ABSTRACT FROM AUTHOR]

The Fourier-series-based insulated-gate bipolar transistor (IGBT) model has been very effective in simulating its switching behaviors and meanwhile, reflecting carrier dynamics. However, in turn-on process, such a model has difficulty in distinguishing the boundary between undepleted area and depletion layer of the drift region due to Gibbs phenomenon and therefore, cannot accurately describe the carrier behavior of carrier storage region (CSR). In this study, the Fourier-series-based model is modified to determine the more precise depletion layer edge for turn on. This is achieved by utilizing the actual characteristics of the carrier concentration gradient at the edge of CSR in turn-on process. Such a modified model is implemented in MATLAB/Simulink and it can better reflect the depletion layer and simulate the distribution of both the carrier concentration and electric field during turn-on process. It corrects the unreasonable voltage tail caused by the error of drift region voltage calculation in previous models. The results of the proposed model under both the resistive load and inductive load are in good agreement with TCAD. Such a modified model can more accurately simulate both carrier dynamics and switching behaviors in the turn-on process. [ABSTRACT FROM AUTHOR]

Non-spherical particles have shown increasing importance in the simulation of discrete element methods. Bonded particle model (BPM) is often used to realize the inclusion of non-spherical particles into the calculation using spherical calculation methods. However, the traditional grid subdivision method shows low efficiency and high computational complexity in the face of particle clusters containing a large number of elemental spheres. In this paper, an improved contact detection algorithm for bonded particles based on multi-level grid and bounding box method is proposed. The algorithm treats unbroken or broken new clusters as a unit in the process of contact detection. The contact detection problem including clusters is split into two steps: a pre-detection between the bounding spheres in the primary grid and an accurate contact detection of the elemental spheres in the secondary grid. The hopper experiment and the comminution simulation show that the algorithm has high accuracy and the calculation results are in line with the actual situation. The numerical simulation results of multiple sets of comparisons show that the improved algorithm has higher computing speed than the traditional grid subdivision algorithm, especially when the particle distribution in the calculation domain is relatively loose, the number of particle clusters is large or the composition of the cluster is more complicated. Unlabelled Image • Contact detection problem is split into a pre-detection and an accurate detection. • The improved algorithm has much lower sensitivity of the complexity of the cluster. • The improved algorithm is notably less sensitive to the proportion of cluster. • The improved algorithm is notably more sensitive to the density of particles. • The improved algorithm has computational speed advantage in most cases. [ABSTRACT FROM AUTHOR]

Steel balls of different sizes may have uneven axial distribution in the rotating drum, which will cause different crushing effects in different areas of the ball mill. As a result, the working efficiency of ball mills cannot be accurately reflected by traditional evaluation indicators which assume that the balls are uniformly distributed along the axis. The existence of axial segregation of the steel balls was verified by including balls of different sizes in the drum through the experiment and simulation (Discrete element method) model. Different size balls in the drum tend to separate from each other while the same kind of balls gather along the axis. The segregation of different types of balls causes uneven distribution of collision and energy in the ball mill. Based on the basic assumption that the crushing energy is related to the energy loss in ball collisions, the concept of regional total energy (or RTE, for short) is put forward to evaluate the axial inconsistent crushing effect in different areas caused by axial segregation in the drum. The regional total energy in the ball mill under axial segregation is sampled and calculated by the Monte Carlo method. The calculation results show that the distribution of the broken energy in the region is related to the distribution of the ball, and the region with more balls has greater crushing energy, which also means better crushing effect. Thus, it is necessary to study the crushing effect of ball mills with the consideration of regional total energy. Unlabelled Image • Different sizes of balls inside the cylinder will segregate along the axis. • The axial segregation of balls causes uneven distribution of collision and energy. • Regional total energy of different kinds of collision are distributed non-uniformly along the axis. • Regional total energy shows a distribution consistent with that of balls in ball mill. • Region with more spheres has greater crushing energy and better crushing effect. [ABSTRACT FROM AUTHOR]

Automatic real-time fire characterization is a crucial requirement of future smart firefighting. This work proposes a novel computer vision method to automatically measure the fire heat release rate, even when the camera is moving in real-time. Firstly, a portable binocular stereo camera is used to capture the real-time fire video stream that is fed into a pre-trained computer-vision model frame-by-frame to detect the fire region. By identifying the fire location inside the image, the real-time distance between the camera and the fire source is determined. This fire distance helps re-scale the images to match the input scale of the AI-image Fire Calorimetry. Then, the deep learning model can automatically output the transient fire power in real time. Results show that the stereo vision system is capable of accurately measuring the distance between the camera and the fire source, flame height, and power, with a relative error of less than 20%. This work provides an automatic and flexible way to measure the distance, power and hazard of fire in real-time, and such a method has broad applications in firefighting operations and decision-making. • Propose a smart fire calorimetry method based on computer vision and AI-driven fire identification. • Use binocular camera to measure the real-time fire distance and determine the fire dimension. • Combine fire image from a moving camera and deep learning to quantify fire heat release rate. • Provide a framework of AI-Fire calorimetry for a fast fire hazard assessment in smart firefighting. [ABSTRACT FROM AUTHOR]

The hypothalamic-pituitary-gonadal (HPG) axis plays a critical role in regulating reproductive function. Gonadotropin-releasing hormone (GnRH), which is secreted by the hypothalamus, acts on pituitary gonadotrophs to stimulate luteinizing hormone (LH) and follicle-stimulating hormone (FSH) synthesis and secretion, ultimately affecting the animal's fertility. MicroRNAs are small, non-coding RNAs that are widely expressed throughout the brain and can fine-tune gene expression post-transcriptionally. Recently, growing evidence has unveiled the central position of miRNAs within a key regulatory process involving GnRH secretion and subsequent activation in the pituitary. Although transcriptional regulation of reproduction has been well studied, the post-transcriptional processes are less well understood. In this review, we elaborate comprehensively on the critical role of miRNAs in the reproductive process, including both temporal and spatial aspects. A better understanding of how miRNAs impact the neuroendocrine system may improve our knowledge of reproduction and provide novel targets for therapeutic development. [ABSTRACT FROM AUTHOR]

Since classical deep learning (DL) techniques are hungry for massive data and suffer from domain shift, domain adaptation (DA) methods are broadly adopted in prognostics and health management (PHM) to align source and target domains. However, DA relies on target datasets collected in advance, which are not always available in practice. In this paper, a domain generalization (DG) approach, which learns from multiple source domains and generalizes well to unseen domains, is introduced for remaining useful life (RUL) prediction of bearings under unseen operating conditions. Specifically, we propose an adversarial out-domain augmentation (AOA) framework to generate pseudo-domains, thereby increasing the diversity of available samples. Hence, a generator is trained in an adversarial manner to generate augmented pseudo-domains by maximizing the domain discrepancy of the latent representations. In addition, we add manifold and semantic regularization to its objective function to ensure the consistency of the pseudo-domains. Trained with these available domains, a task predictor can improve the generalization in inaccessible target domain. Based on this, we provide a specific implementation of AOA-based RUL prediction for DG and validate its effectiveness and superiority using experimental datasets. [ABSTRACT FROM AUTHOR]

The tremendous success of deep learning and transfer learning broadened the scope of fault diagnosis, especially the latter further improved the diagnosis accuracy under multiple working conditions. However, most existing attempts assume that label distribution is domain-invariant despite taking into account the different feature distributions. This does not accommodate the diversity of fault mode distributions under different operating conditions and weakens the generalization to imbalanced domain adaptation (IDA) scenarios. Therefore, this work proposed a novel deep imbalanced domain adaptation (DIDA) framework for fault diagnosis of bearings, aiming at the challenging scenario where feature shift and label shift exist simultaneously under different working conditions. Specifically, DIDA overcomes the class-imbalanced label shift and achieves a fine-grained latent space matching by cost-sensitive learning and categorical alignment. Besides, margin loss regularization is introduced to further optimize classification boundaries and improve cross-domain generalization for IDA fault diagnosis tasks. Finally, we simulated the IDA protocols on experimental datasets and conducted case studies under multiple working conditions, thus validating the effectiveness and superiority of the proposed framework. • A novel deep imbalanced domain adaptation (DIDA) is proposed. • DIDA narrows both feature shift and label shift. • DIDA broadens fault diagnosis to IDA scenarios. • Experimental case studies verified validity and superiority. [ABSTRACT FROM AUTHOR]

The scope of data-driven fault diagnosis models is greatly extended through deep learning (DL). However, the classical convolution and recurrent structure have their defects in computational efficiency and feature representation, while the latest Transformer architecture based on attention mechanism has not yet been applied in this field. To solve these problems, we propose a novel time–frequency Transformer (TFT) model inspired by the massive success of vanilla Transformer in sequence processing. Specially, we design a fresh tokenizer and encoder module to extract effective abstractions from the time–frequency representation (TFR) of vibration signals. On this basis, a new end-to-end fault diagnosis framework based on time–frequency Transformer is presented in this paper. Through the case studies on bearing experimental datasets, we construct the optimal Transformer structure and verify its fault diagnosis performance. The superiority of the proposed method is demonstrated in comparison with the benchmark models and other state-of-the-art methods. • A novel model named time–frequency Transformer (TFT) is proposed. • A fresh tokenizer and encoder module are designed to extract effective abstractions. • A new end-to-end fault diagnosis framework based on TFT is presented. • The proposed method shows superiority comparing to other state-of-the-art methods. [ABSTRACT FROM AUTHOR]

Data-driven approaches for prognostic and health management (PHM) increasingly rely on massive historical data, yet annotations are expensive and time-consuming. Learning approaches that utilize semi-labeled or unlabeled data are becoming increasingly popular. In this paper, a self-supervised pre-training via contrast learning (SSPCL) is introduced to learn discriminative representations from unlabeled bearing datasets. Specifically, the SSPCL employs momentum contrast learning (MCL) to investigate the local representation in terms of instance-level discrimination contrast. Further, we propose a specific architecture for SSPCL deployment on bearing vibration signals by presenting several data augmentations for 1D sequences. On this basis, we put forward an incipient fault detection method based on SSPCL for run-to-failure cycle of rolling bearings. This approach transfers the SSPCL pre-trained model to a specific semi-supervised downstream task, effectively utilizing all unlabeled data and relying on only a little priori knowledge. A case study on FEMTO-ST datasets shows that the fine-tuned model is competent for incipient fault detection, outperforming other state-of-the-art methods. Furthermore, a supplemental case on a self-built fault datasets further demonstrate the great potential and superiority of our proposed SSPCL method in PHM. • A self-supervised pretraining via contrast learning (SSPCL) is introduced. • SSPCL implementation on vibration signals with 1D data augmentation. • Incipient fault detection framework based on SSPCL is detailed. • Experimental case studies verified validity and superiority. [ABSTRACT FROM AUTHOR]

• Causal dilated convolution block is built to learn the temporal dependencies. • A residual attention mechanism is proposed to obtain the contribution degree. • Residual connection is used to solve the problem of gradient disappearance. • A novel framework based on TCN-RSA can realize RUL prediction of bearings. Remaining useful life (RUL) prediction has been a hotspot in the engineering field, which is useful to avoid unexpected breakdowns and reduce maintenance costs of the system. Due to the high nonlinearity and complexity of mechanical systems, traditional methods cannot meet the requirements of medium-term and long-term prediction tasks, and often ignore the influence of temporal information on prediction performance. To solve this problem, this paper proposes a new deep learning framework – Temporal convolutional network with residual self-attention mechanism (TCN-RSA), which can learn both time-frequency and temporal information of signals. First, we input the marginal spectrum of vibration signals to TCN. The causal dilated convolution structure in the TCN has the ability to capture long-term dependencies and extract high-level feature representations from the time-frequency domain at the same time. Then, a residual self-attention mechanism is introduced to obtain the feature contribution degree of different moments in the bearing degradation process. Finally, an end-to-end RUL prediction implementation can be established based on TCN-RSA network. The effectiveness of the proposed method is verified by IEEE PHM 2012 Data challenge datasets and XJTU-SY datasets respectively. The comparative study indicates that the proposed TCN-RSA framework outperforms the other state-of-the-art methods in RUL prediction and system prognosis with respect to better accuracy and computation efficiency. [ABSTRACT FROM AUTHOR]

[Display omitted] • Rotational direction of adjacent segments changes the mixing configuration. • Uneven distribution of media makes particle collision energy inconsistent. • Axial RRTE reflects particle collision energy more clearly than radial RRTE. • More serious axial media segregation, the lower the total collision energy. The particle mixture in rotating drums is not always in a completely mixed state, resulting in a complex relationship between the particle mixing and segregation characteristics and the overall collision energy. Based on existing experimental validation, the rotary drums with axial adjacent segments are numerically simulated via the discrete element method to investigate the axial segregation mechanism of the particle mixtures in rotary drums. The results show that the rotational direction of the adjacent axial segments changes the axial flow configuration, resulting in a transformation of the overall particle mixing and segregation characteristics. Based on the proposed novel relative regional total energy (RRTE) index, this work further evaluates the impact of the axial segregation characteristics on the collision energy of binary and ternary particle mixtures within rotary drums. The axial segregation phenomenon makes the RRTE of each radial surface inconsistent, resulting in difficultly reflecting the particle collision energy through the individual radial surface. Compared with the radial drum surface, the RRTE represented by the axial surface can better reflect the particle collision energy in rotating drums. In summary, the more obvious the presence of axial segregation characteristics in rotary drums, the lower the total collision energy used for overall grinding process of the media and abrasive particles. [ABSTRACT FROM AUTHOR]

Microcellular injection moulding (MuCell®) is a polymer processing technology that uses a supercritical fluid inert gas, CO2 or N2, to produce light-weight products. Due to environmental pressures and the requirement of light-weight parts with good mechanical properties, this technology recently gained significant attention. However, poor surface appearance and limited mechanical properties still prevent the wide applications of this technique. This paper reviews the microcellular injection moulding process, main characteristics of the process, bubble nucleation and growth, and major recent developments in the field. Strategies to improve both the surface quality and mechanical properties are discussed in detail as well as the relationships between processing parameters, morphology, and surface and mechanical properties. Modelling approaches to simulate microcellular injection moulding and the mathematical models behind Moldex 3D and Moldflow, the two most commonly used software tools by industry and academia, are reviewed, and the main limitations are highlighted. Finally, future research perspectives to further develop this technology are also discussed. [ABSTRACT FROM AUTHOR]

• A deep transfer metric learning for kernel regression (DTMLKR) model is proposed. • DTMLKR combines the advantages of deep metric learning and transfer learning. • RUL prediction under multiple operating conditions is validated effectively. • DTMLKR shows superiority comparing to other art-of-the-state methods. Accurate estimation of remaining useful life (RUL) is indispensable for the safe operation of rotating machinery, reducing maintenance costs and unnecessary downtime. Numerous data-driven models have been reported to predict the RUL of bearings using historical data. However, it is still very challenging to predict the RUL of bearings under different operating conditions. It is necessary to propose a model which can extract domain invariant deep features and accurately predict the RUL of bearings under new operating condition. In this paper, a novel method called deep transfer metric learning for kernel regression (DTMLKR) is proposed and applied to the RUL prediction of bearings under multiple operating conditions. This method combines deep metric learning with transfer learning (TL) to solve regression problems. Case studies on the IEEE PHM Challenge 2012 dataset demonstrate the effectiveness of the proposed method. Compared with other state-of-the-art methods, the superiority of the proposed method is verified. [ABSTRACT FROM AUTHOR]

BACKGROUND: As a novel type of extracellular polysaccharide produced by Sphingomonas sp., welan gum has been widely applied in various fields because of its excellent properties. The study has improved the fermentation process. Results: The initial sucrose concentration, temperature and stirring speed were set to 20 g L−1, 33 °C and 400 rpm, respectively, and 13.3 g L−1 sucrose was added at 24, 40 and 56 h. The temperature and stirring speed were then set at 28 °C and 600 rpm from 24 to 48 h and 28 °C and 600 rpm from 48 to 72 h, respectively. As a result, welan gum production, dry cell weight, sucrose conversion rate and viscosity were correspondingly increased to 38.60 g L−1, 5.47 g L−1, 0.64 g g−1 and 3779 mPa·s, respectively. In addition, the mechanism by which fermentation strategy promotes welan gum synthesis was investigated by transcriptome analysis. CONCLUSION: Improving respiration and ATP supply, reducing unnecessary protein synthesis, and alleviating competition between cell growth and welan gum synthesis contribute to promoting the fermentation performance of Sphingomonas sp., thus providing a practical strategy for efficient welan gum production. © 2024 Society of Chemical Industry. [ABSTRACT FROM AUTHOR]

• Propose a CV method and digital twin framework to assess real-time tunnel fire risk and evacuation safety. • Apply YOLOv7 object detection model and DeepSORT algorithm to classify and track vehicles in real time. • Evaluate tunnel fire risk via real-time fire load distribution and potential heat release rate in case of fire. • Estimate the evacuation safety by the number of evacuees from vehicles and total evacuation time. The distribution of vehicles during a tunnel fire is a crucial factor that affects fire development and hazards, as well as the following evacuation and rescue operations. This work proposed a novel method using computer vision for assessing the real-time tunnel fire risk and evacuation safety by considering the classification and entry flow of vehicles. The proposed system utilizes YOLOv7 and DeepSORT for vehicle detection, classification, and tracking to enable a real-time digital twin for tunnel fire safety management. Vehicles are divided into 10 categories, in terms of their size, usage, number of passengers, fuel load, and peak fire HRR. After monitoring the vehicle flow at the tunnel portals, the real-time vehicle and fire load distribution are predicted. Then, the real-time tunnel fire scenarios and the safety of the evacuation process are evaluated based on the distribution of vehicles. The system is demonstrated in real road tunnels with traffic video cameras and exhibits a robust performance. The proposed vision-based real-time tunnel fire risk evaluation enables intelligent daily fire safety management and supports fire emergency response and decision-making. [ABSTRACT FROM AUTHOR]

In this paper, a steady-state and dynamic response model is proposed to predict the working characteristics of the electromechanical transmission system including eddy-current magnetic couplers. Considering the influence of mechanical characteristics of motor and load device on the performance of eddy-current magnetic couplers, the system is divided into input and output for kinematic analysis. The expressions of electromagnetic torque versus speed are obtained through layer analysis of the eddy-current magnetic coupler during steady-state and then, the dynamic response model of the system is established by combining the kinematic analysis with the motor, the eddy-current magnetic coupler and three typical loads (constant torque load, quadratic rate load and constant power load). The dynamic characteristics during starting and speed-regulation of the system under constant torque loads are analyzed using the proposed model. The variations of the torque and speed with time of input and output are calculated by numerical integration. Finally, the speed and torque during starting and speed-regulation computed with the dynamic response model are compared with those obtained from the experimental results. The result shows a good agreement. [ABSTRACT FROM AUTHOR]

Nowadays, machinery degradation trend prognostics have made significant progress with sufficient and related data. Limited data may weaken the generalizations of artificial intelligence-based prognostics, which could worsen under the domain shift phenomenon. To overcome the dual challenges and further abstract the prior degraded knowledge from historical data, a novel prediction architecture entitled as statistical alignment-based metagated recurrent unit, profoundly integrating deep learning, knowledge transfer, and metalearning, is proposed in this study and simultaneously features domain adaptation ability and desired generalizations under limited data. To be specific, a designed high-order measure is first developed to align training and testing data distribution. Then, the well-known encoder–decoder framework is employed as the base model, and subtask & cross-subtask learning architectures are developed to distill sensitive prognostics agents for easily generalizing to new domains using limited data. Finally, FEMTO-ST bearing data sets and industrial data sets collected from petrochemical pumps are conducted to prove their effectiveness. [ABSTRACT FROM AUTHOR]

Prognostics and health management of rotating machinery plays an integral role in providing critical monitoring information for stable equipment operation. Thus, an autoregressive model-based prediction of bearing degradation trends considering health indicators with multiscale attention information is proposed. Specifically, a multiscale attention network is developed to construct a health indicator with the representation of degradation trends. Then, a comprehensive evaluation metric is designed to evaluate the metrics of the constructed health indicators. Also, an improved autoregressive integrated moving average model considering the elimination of heteroskedasticity attribute is proposed to predict the degradation trend of health indicators, and the order of the model are determined using the Bayesian information criterion. Finally, the degradation trend of the test set is predicted using the derived model. The results show that the prediction method constructed by a multiscale attention network considering health indicators can effectively utilize the vibration data to predict the degradation trend of bearings. Compared with other related methods, the proposed method has obvious advantages in the health monitoring of rotating machinery. [ABSTRACT FROM AUTHOR]

Currently, fragmentation research based on discrete element method (DEM) has aroused wide interest. Several particle cluster modeling methods for bonded particle model (BPM) have been proposed and shown to be effective. Due to inevitable overlap among elements, however, existing methods bring extra deviation to the simulation. Based on the analysis of these modeling method, a novel method for particle cluster modeling is proposed in this paper. Firstly, a virtual force named internal force between the particles and the boundary points is defined according to their relative positions. Afterward, a dynamic optimization scheme is introduced, which encourages the randomly initialized elemental-spheres to reach an optimal steady state. To evaluate the modeling performance, an index named overlap factor is defined to quantify relative overlap degree. Through several groups of modeling cases in 2D and 3D, it is proved that proposed method has the characteristics of less overlap between particles and high precision of boundary fitting. By comparing with other traditional modeling approaches, superiority of the proposed method is demonstrated. [Display omitted] • A novel method for particle cluster modeling based on internal force is proposed. • Particle cluster model established by this method meets modeling requirements. • The iterative final overlap factor of boundaries in 3D is larger than that in 2D. • The method has less overlap between particles and higher fitting accuracy. [ABSTRACT FROM AUTHOR]

At present, image recognition methods have been widely used in the research on measuring particle mixing and segregation in rotating drums. However, currently commonly used methods tend to study mixing and segregation from the perspective of the radial surface of the particles, but rarely from the perspective of the axial surface. Whether the data obtained from the particle surface (radial segregation or axial segregation) can reflect the degree of mixing and segregation of the entire particle system (overall segregation) is still a question worth studying. In this work, an image analysis method based on stacked images to measure the degree of surface segregation is proposed. The drum length to diameter ratio (L/D) and particle filling rate are used as operating parameters. And the discrete element method (DEM) verified by experiments has studied the difference between the surface segregation and the overall segregation of the binary particle groups (different particle sizes) in the rotating drum. This paper compares the qualitative characteristics (particle segregation characteristic patterns) and quantitative characteristics (particle segregation degree) of surface and overall mixing and segregation in detail. The results show that for short drums, radial segregation index can be used to quantify the overall degree of segregation. While for long drums, axial segregation index can be used to quantify the overall degree of segregation. Unlabelled Image • An image-stacked method to measure particle segregation degree is proposed. • The overall and the axial segregation index reach the maximum meanwhile. • Radial segregation index can quantify overall segregation process for short drums. • Axial segregation index can quantify overall segregation process for long drums. [ABSTRACT FROM AUTHOR]

The external heating test is widely used to evaluate the hazards of battery thermal runaway, but the efficiency and effect of the heating source are rarely quantified. This work performs thermal runaway propagation tests in a 3-layer cylindrical battery pile with a uniform state of charge (SOC) ranging from 30 % to 75 %. A cylindrical heater is in contact with two cells in the first layer and has a power varying from 50 W to 300 W to trigger thermal runaway. Results indicate that for the current system, the heating efficiency to a single cell is around 15 %, and the effective heating power is insensitive to the SOC. The intensity of thermal runaway increases with the external heating power and the cell SOC. The influence of external heating on the propagation of thermal runaway is reflected in the intensity of thermal runaway in the first-layer cells and the preheating effect of subsequent layers. A simplified heat-transfer model is established to quantify the thermal impact on both thermal runaway intensity and preheating depth. Finally, a new approach for selecting the appropriate heating power is proposed to help optimize battery thermal-runaway tests and improve the safety regulations for battery modules. [ABSTRACT FROM AUTHOR]

The external heating test is widely used to evaluate the hazards of battery thermal runaway, but the efficiency and effect of the heating source are rarely quantified. This work performs thermal runaway propagation tests in a 3-layer cylindrical battery pile with a uniform state of charge (SOC) ranging from 30 % to 75 %. A cylindrical heater is in contact with two cells in the first layer and has a power varying from 50 W to 300 W to trigger thermal runaway. Results indicate that for the current system, the heating efficiency to a single cell is around 15 %, and the effective heating power is insensitive to the SOC. The intensity of thermal runaway increases with the external heating power and the cell SOC. The influence of external heating on the propagation of thermal runaway is reflected in the intensity of thermal runaway in the first-layer cells and the preheating effect of subsequent layers. A simplified heat-transfer model is established to quantify the thermal impact on both thermal runaway intensity and preheating depth. Finally, a new approach for selecting the appropriate heating power is proposed to help optimize battery thermal-runaway tests and improve the safety regulations for battery modules. [ABSTRACT FROM AUTHOR]

In the context of industrial big data, deep neural networks have been widely used in fault classification and remaining useful life (RUL) prediction of mechanical equipment due to their powerful nonlinear feature extraction capabilities. However, traditional deep learning models only stay at the level of real domain for feature mining, regardless of the importance of time–frequency domain analysis for rotating machinery. Furthermore, single-channel information from single source limits the ability of networks to learn mechanical degradation trajectories. To solve the above problems, complex domain extension network with multi-channels information fusion is proposed to realize RUL prediction of rotating machinery under different operating conditions. Specifically, the real domain network architecture is first extended to the complex domain, including complex domain convolution, complex domain batch normalization, complex domain weight initialization, complex domain parameter propagation, and complex domain activation functions. On this basis, complex domain extension network with multi-channels information fusion is proposed to extract degenerate features and finally establish an end-to-end mapping between feature information layers and prediction layer. The effectiveness of proposed RUL prediction framework is verified by two case studies with two sets of run-to-failure datasets. The comparison results with current state-of-the-art methods show that the proposed method is a promising method for remaining useful life prediction as its advantages in terms of prediction accuracy, interpretability, and generalization. [ABSTRACT FROM AUTHOR]

A magnet-rotating-type axial magnetic coupler (MAMC), namely, a novel axial magnetic coupler with rotatable monopole permanent magnets (PMs), is proposed. The speed of the MAMC can be adjusted by rotating the rotatable PMs mounted on the PM disk, or changing the axial air gap length between the PM disk and the conductor sheet (CS), or combining these methods. In order to analyze the torque performance, the magnetic circuit model of MAMC is established when the rotatable PMs are not parallel to the CS. Specifically, the influence of the induced eddy current is introduced into a new magnetic equivalent circuit (MEC) model. Combined with Kirchhoff’s law, the calculation formula of torque is derived. Moreover, according to Ampere’s law, the formula for calculating the induced eddy current is derived when the rotatable PMs after rotating are not parallel to the CS. Finally, the effectiveness of the proposed model is verified by the finite element method (FEM). The results obtained by the new MEC model proposed are consistent with those obtained by the FEM, and the formula derived is effective in the case of small rotational speed differences, small air gap lengths, and small thicknesses of the CS. [ABSTRACT FROM AUTHOR]

• A general variable-length limited data driven prediction method is proposed. • Encoder-decoder with attention mechanism framework is deeply integrated with meta -learning. • It provides strong generalizations even when the collected data is reduced to one-twentieth. • Public and real-measured experiments prove the effectiveness of the established prediction method. High accuracy health prognostics are significant to machinery intelligent operation and maintenance. Current data-driven prognostics achieve great success that benefits from amply learning samples. In fact, data scarcity challenge widely exists in machinery prognostics and health management, especially for high-end equipment. This study aims to solve this dilemma and proposes a novel meta learning algorithm reconstructed by classic variable-length prediction mode and attention mechanism, namely meta attention recurrent neural network (MARNN). Specifically, we first develop the encoder-decoder with attention mechanism (EDA) cell to perform episodic learning for the subtask-level upgrade. Then multiple subtasks with EDA as prediction models are aggregated to accomplish meta -level upgrade, thus mining the general degradation knowledge from historical datasets. Finally, cross-domain prognostics tasks can be easily realized through fine-tuning tricks, and three rotating machinery run-to-failed experiments are conducted to prove the generalizations of MARNN, which can obtain desired results even when the on-site adaptation data is reduced to one-twentieth. [ABSTRACT FROM AUTHOR]

The axial mixing and segregation behavior of particles with different properties is a complicated movement process, whose mechanism is still indistinct. Consequently, it is very important to explore dynamic repose angle and flow characteristics of the particles along the axial direction during the mixing and segregation process. Focusing on the mixing and segregation phenomenon in the rotary drum, this paper studies the particle mixing and segregation behavior by changing the speed direction of adjacent axial segmentations. The results show that alternating axial segregation band characteristics can be formed in adjacent axial segmentations with same velocity direction, while the overall mixing degree is better under the condition of the opposite velocity direction. Although the different rotational speed directions of adjacent axial segmentations will cause opposite axial mixing and segregation phenomenon, the general rules of the dynamic angle of repose and axial flow velocity are similar in the rotary drum. In addition, the axial flow structure of the binary particles varies with rotational speed direction of the adjacent axial segments, but not with the drum length to diameter (L / D) ratio and particle filling rate. [Display omitted] • Higher mixing degree, smaller dynamic repose angle between binary particles • Axial flow velocity moves from high dynamic angle repose area to low area. • The greater reduction of dynamic angle of repose, the greater axial velocity. • Axial flow characteristics change with adjacent segment's velocity direction. [ABSTRACT FROM AUTHOR]

Applying an electric field to waxy oil may significantly improve its cold flowability. Our previous study has demonstrated that the accumulation of asphaltenes and resins on the surface of wax particles, i.e., interfacial polarization of wax particles, is the essential mechanism of this electrorheological phenomenon. The influence of asphaltenes concentration on the performance of the electrical treatment has been studied. In this paper, the influence of the concentration of resins, the most abundant non-hydrocarbon component in crude oil, on electrically-treated waxy oil's viscosity and impedance was investigated for the first time by using model oils containing resins in the concentration range of 0.05 wt% - 5.00 wt%. It is found that the effect of resins concentration on viscosity and impedance is not monotonous, with both the viscosity reduction and impedance increment induced by the electric field decreased first and then increased with increasing resins concentration at a given amount of precipitated wax. Increasing the polarity of resins would alleviate the viscosity reduction and impedance increment. This work further demonstrates that resins accumulation on the wax particles surface does occur. • Effect of resins content on electrically-treated oil is studied for the first time. • Effect of resins concentration on electrical efficacy is not monotonous. • Increasing polarity of resins will alleviate the electrical field effect. • Stronger polarity of resins leads to less impedance modulus of model oil. [ABSTRACT FROM AUTHOR]

The novelty of this research is as follows: (1) A new transportable cross-domain adaptation method is developed. (2) A residual self-attention is proposed to fully consider the contextual degradation information. (3) The contrastive loss is introduced to improve the complex transformation invariance of MK-MMD mapping. Many data-driven models normally assume that the training and test data are independent and identically distributed to predict the remaining useful life (RUL) of industrial machines. However, different failure models caused by variable failure behaviors may lead to a domain shift. Meanwhile, conventional methods lack comprehensive attention to temporal information, resulting in a limitation. To handle the aforementioned challenges, a transferable cross-domain approach for RUL estimation is proposed. The hidden features are extracted adaptively by a temporal convolution network in which residual self-attention is able to fully consider the contextual degradation information. Furthermore, a new cross-domain adaption architecture with the contrastive loss and multi-kernel maximum mean discrepancy is designed to learn the domain invariant features. The effectiveness and superiority of the proposed method are proved by the case study on IEEE PHM challenge 2012 bearing dataset and the comparison with other methods. [ABSTRACT FROM AUTHOR]

Improving the energy harvesting efficiency of piezoelectric energy harvesters (PEHs) in the scientific domain is essential for the application of self-powered microelectromechanical devices. This paper proposes an isogeometric optimization method to optimize the material distribution of piezoelectric functionally graded material (PFGM) energy harvester. Firstly, we establish the isogeometric formulation of the PFGM energy harvester based on the plane strain hypothesis and Hamilton's principle. Meanwhile, the same non-uniform rational B-splines basis functions are utilized to represent the distribution of the PFGM. Secondly, three optimization models with the specified eigenfrequency and maximum energy conversion efficiency as objective functions are established. Then we derive the analytical sensitivities by the adjoint method and use the method of moving asymptotes to update the design variables. Furthermore, several numerical examples are performed to study the convergence and accuracy of the numerical solutions and analyze the effects of the metallic volume ratio constraint and the initial design domain on the optimized material distribution. More importantly, the dynamic analysis is executed to further test the energy harvesting efficiency based on the optimized designs. The results show that the presented approach can not only realize the specified eigenfrequency but also effectively improve the energy harvesting efficiency. [ABSTRACT FROM AUTHOR]

Mechanical properties are essential for polyester materials due to their widespread applications as fibers, films, and engineering accessories. The copolymerization strategy has been developed to enhance polyester properties, but facile-synthesized new materials are still under discovery. In this work, we designed a series of terminal functioned polyesters, for introducing plasticization effects and π-π stacking interactions. The fabrication process can be one-pot achieved via melting polycondensation method. The prepared modified polyesters exhibit significantly improved elongation of break (560%), which is 104% improved compared with pristine poly(ethylene terephthalate). Meanwhile, the maximum toughness reaches 206.1 MJ/m3. The thermal properties, viscoelastic properties, crystallinity properties, and fracture morphologies of the as-prepared polyesters are investigated as well. Results show that great elongation of break, higher crystallinity, and good processability are achieved simultaneously. This work provides a one-pot fabricated terminal functioned polyester with advanced tensile properties, expanding its potential application as engineering materials. [Display omitted] • Polyester has been reinforced after furoic acid modification, resulting in advanced comprehensive tensile properties. • The preparation is one-pot achieved by the green method of bulk polymerization. • The reinforcement principles of plasticization effects and π-π stacking interactions are studied in detail. [ABSTRACT FROM AUTHOR]

The paraffin molecules in crude oil crystalize and precipitate out as the oil cools, resulting in a sharp increase in oil viscosity and further the gelation of the oil. The waxy crude oil gel exhibits complex rheological behaviors, such as viscoelasticity, yield stress, and thixotropy, posing challenges to the flow assurance of pipelines. Previous studies have verified that applying a high-voltage electric field to a flowing liquid oil can significantly reduce its viscosity. In this work, the electrorheological behaviors of waxy crude oil gel are explored by a stress-controlled rheometer equipped with in situ DC electric field manipulation. It is reported for the first time that the gel structure can be significantly weakened upon the application of an electric field for a duration of not less than 10 s, with the yield stress, storage modulus, loss modulus, and apparent viscosity reduced by up to 90%. In addition, the gel structure fracture is observed to transit from brittle-like to ductile-like. A stronger structure weakening effect can be obtained at a higher field strength and lower temperature. The structure weakening effect induced by the electric field can be maintained for a certain period of time after the removal of the electric field, with a recovery of yield stress and moduli less than 10% in 48 h. We hold the view that the electrorheological behaviors of waxy oil should be attributed to the weakened attraction between wax particles induced by the electric field. [ABSTRACT FROM AUTHOR]

[Display omitted] • Electric field induces charged particles to adsorb on the surface of wax particles. • The impedance of a waxy oil below the WAT increases after exposure to electric field. • The impedance increase is positively correlated with viscosity reduction. Previous studies have shown that the application of a high-voltage electric field to a liquid or gelled waxy crude oil can significantly improve its cold flowability, such as reducing viscosity and yield strength. However, the mechanism of the electrorheological behaviors of waxy oils remains largely unknown. In this study, the impedance spectroscopy was firstly used to investigate the mechanism of the electrorheological behaviors of waxy oils. The impedance of the waxy oil increases and the conductivity decreases after an oil is exposed to electric fields. Based on this finding and with the aid of further experiments using model oils of controlled compositions, i.e. to control the absence of wax, resins and asphaltenes in the oil, interfacial polarization of wax particles was demonstrated as the primary and essential mechanism of the electrorheological behaviors. The interfacial polarization of wax particles means that charged colloidal particles in the oil, i.e., resins and asphaltenes, would accumulate on the surface of wax particles upon the application of the external electric field, and the resulted electrostatic repulsion will diminish the van der Waals force between wax particles, thus reducing viscosity and the strength of wax particle structure. [ABSTRACT FROM AUTHOR]

• A feature evaluation index based on DTW and Wasserstein distance is constructed. • A new state index - energy entropy moving average cross-correlation is proposed. • Feature distribution discrepancy can be solved by domain adaptation. • Transfer learning based on BiGRU can realize RUL prediction of bearings. Remaining useful life (RUL) prediction, has been a hotspot topic in the engineering field, which can ensure the security, availability, and continuous efficiency of the system. Different degradation trajectories of bearings under various working conditions may lead to the problem of inconsistent feature distribution and difficult acquisition of corresponding training labels, which affects the validity and accuracy of the prediction model. In this paper, a new transfer learning method based on bidirectional Gated Recurrent Unit (TBiGRU) is proposed to accurately predict the RUL of bearings under different working conditions. Firstly, based on dynamic time wraping (DTW) and Wasserstein distance to construct a comprehensive evaluation index of feature, the selection of transferable feature is carried out. Then a new index of energy entropy moving average cross-correlation based on maximal overlap discrete wavelet transform (MODWT) is proposed to realize adaptive recognition of bearings running states and the acquisition of corresponding training labels, which can also get rid of the constraint of setting threshold. Finally, transfer learning is carried out on the BiGRU model to solve the problem of distribution discrepancy, and timing information is also taken into account. The method is applied to the analysis of experimental data, and the results show that the framework can adaptively recognize different running states of bearings and obtain corresponding training labels, and at the same time realize better RUL prediction performance under different working conditions. [ABSTRACT FROM AUTHOR]

• Electric field perpendicular to oil's flow direction can also reduce oil viscosity. • The viscosity reduction is insensitive to the field direction. • Electric field weakens the interaction between wax crystals. • The weakened interaction should be the primary reason for the viscosity reduction. It has been reported that the application of a high-voltage electric field parallel to the oil flow direction or to a quiescent waxy crude oil can reduce the viscosity of the oil. In this work, an electric field perpendicular to the flow direction is applied to a flowing waxy crude oil in a rheometer equipped with in-situ electric field manipulation, and it is observed that viscosity reduction similar to that obtained with electric field parallel to the flow direction can also be achieved. This, therefore, confirms that the treatment efficacy is insensitive to the direction of the electric field compared to the oil flow direction. Moreover, the non-Newtonian characteristics of the treated crude oil become less significant, and more viscosity reduction is achieved at higher electric field strength and lower oil temperature. After the removal of the electric field, the reduced viscosity gradually recovers and the viscosity reduction disappears after about two days. Though it is found that the electric treatment causes an increase in wax crystal size and aggregation of particles to some extent, we think that weakened interaction between wax particles due to electric field exposure should be the primary reason of the viscosity reduction according to the knowledge of suspension rheology and the mechanism of waxy crude oil rheology. [ABSTRACT FROM AUTHOR]

• Combination of electric field and EVA can improve waxy oils' cold flowability. • The combined treatment outperforms the individual treatments under low shearing. • The combined treatment leads to larger size and broader size distribution of wax. Flow assurance issues related to wax precipitation, are one of the most challenging problems in the pipeline transportation of waxy crude oils. Flowability of waxy oils can be improved by suitable pour point depressants (PPDs). Besides chemical treatments, exposing waxy oils to high-voltage electric fields can also significantly improve their cold flowability. However, the combined effect of electric fields and PPD on the flowability of waxy oils is not yet clear. In this investigation, the effect of the combined electrical and EVA treatment on the flow properties of a waxy crude oil was studied. It is found that the combined treatment can lead to more significant reductions in the viscosity, the yield stress and the thixotropy of waxy oils compared to oils treated solely by electric field or EVA. However, the overall improvement in the flowability achieved by the joint treatment is lower than the sum of improvements achieved by the electric field and EVA. The combined treatment outperforms the individual treatments for hours under a low shear of 10 s−1. However, the performance cannot be preserved even for minutes under a high shear of 1200 s−1. Reheating has a negative impact on the electrical treatment, but an intricate impact on the EVA treatment and the combined treatment. The electrical- and chemical performance both undermine upon reheating, but chemical performance can be partially recovered when reheating temperature is above the wax disappearance temperature. Microscopic observation showed that the combined treatment leads to larger size and broader size distribution of wax particles. [ABSTRACT FROM AUTHOR]

Background: Ovarian cancer is a leading cause of the death from gynecologic malignancies. Hypoxia is closely related to the malignant growth of cells. However, the molecular mechanism of hypoxia-regulated ovarian cancer cells remains unclear. Thus, this study was conducted to identify the key genes and pathways implicated in the regulation of hypoxia by bioinformatics analysis. Methods: Using the datasets of GSE53012 downloaded from the Gene Expression Omnibus (GEO), the differentially expressed genes (DEGs) were screened by comparing the RNA expression from cycling hypoxia group, chronic hypoxia group, and control group. Subsequently, cluster analysis was performed followed by the construction of the protein-protein interaction (PPI) network of the overlapping DEGs between the cycling hypoxia and chronic hypoxia using ClusterONE. In addition, gene ontology (GO) functional and pathway enrichment analyses of the DEGs in the most remarkable module were performed using Database for Annotation, Visualization and Integrated Discovery (DAVID) software. Ultimately, the signaling pathways associated with hypoxia were verified by RT-PCR, WB, and MTT assays. Results: A total of 931 overlapping DEGs were identified. Nine hub genes and seven node genes were screened by analyzing the PPI and pathway integration networks, including ESR1, MMP2, ErbB2, MYC, VIM, CYBB, EDN1, SERPINE1, and PDK. Additionally, 11 key pathways closely associated with hypoxia were identified, including focal adhesion, ErbB signaling, and proteoglycans in cancer, among which the ErbB signaling pathway was verified by RT-PCR, WB, and MTT assays. Furthermore, functional enrichment analysis revealed that these genes were mainly involved in the proliferation of ovarian cancer cells, such as regulation of cell proliferation, cell adhesion, positive regulation of cell migration, focal adhesion, and extracellular matrix binding. Conclusion: The results show that hypoxia can promote the proliferation of ovarian cancer cells by affecting the invasion and adhesion functions through the dysregulation of ErbB signaling, which may be governed by the HIF-1α-TGFA-EGFR-ErbB2-MYC axis. These findings will contribute to the identification of new targets for the diagnosis and treatment of ovarian cancer. [ABSTRACT FROM AUTHOR]