As widely used for secondary energy storage, lithium-ion batteries have become the core component of the power supply system and accurate remaining useful life prediction is the key to ensure its ...reliability. Because of the complex working characteristics of lithium-ion batteries as well as the model parameter changing along with the aging process, the accuracy of the online remaining useful life prediction is difficult but urgent to be improved for the reliable power supply application. The deep learning algorithm improves the accuracy of the remaining useful life prediction, which also reduces the characteristic testing time requirement, providing the possibility to improve the power profitability of predictive energy management. This article analyzes, reviews, classifies, and compares different adaptive mathematical models on deep learning algorithms for the remaining useful life prediction. The features are identified for the modeling ability, according to which the adaptive prediction methods are classified. The specific criteria are defined to evaluate different modeling accuracy in the deep learning calculation procedure. The key features of effective life prediction are used to draw relevant conclusions and suggestions are provided, in which the high-accuracy deep convolutional neural network — extreme learning machine algorithm is chosen to be utilized for the stable remaining useful life prediction of lithium-ion batteries.
•Review, discussion, and classification for remaining useful life prediction methods.•Mathematical modeling of deep learning and DCNN-based calculation flowcharts.•Suitable DCNN-ETL-based deep learning model for the RUL prediction.•Comparative analysis of different methods by RMSE, MaxE, Speed, and Accuracy.
With age, our blood vessels are prone to aging, which induces cardiovascular disease. As an important basis for diagnosing heart disease and evaluating heart function, the electrocardiogram (ECG) ...records cardiac physiological electrical activity. Abnormalities in cardiac physiological activity are directly reflected in the ECG. Thus, ECG research is conducive to heart disease diagnosis. Considering the complexity of arrhythmia detection, we present an improved convolutional neural network (CNN) model for accurate classification. Compared with the traditional machine learning methods, CNN requires no additional feature extraction steps due to the automatic feature processing layers. In this paper, an improved CNN is proposed to automatically classify the heartbeat of arrhythmia. Firstly, all the heartbeats are divided from the original signals. After segmentation, the ECG heartbeats can be inputted into the first convolutional layers. In the proposed structure, kernels with different sizes are used in each convolution layer, which takes full advantage of the features in different scales. Then a max-pooling layer followed. The outputs of the last pooling layer are merged and as the input to fully-connected layers. Our experiment is in accordance with the AAMI inter-patient standard, which included normal beats (N), supraventricular ectopic beats (S), ventricular ectopic beats (V), fusion beats (F), and unknown beats (Q). For verification, the MIT arrhythmia database is introduced to confirm the accuracy of the proposed method, then, comparative experiments are conducted. The experiment demonstrates that our proposed method has high performance for arrhythmia detection, the accuracy is 99.06%. When properly trained, the proposed improved CNN model can be employed as a tool to automatically detect different kinds of arrhythmia from ECG.
Glioblastoma multiforme (GBM) is the most common and the most aggressive type of primary brain malignancy. Glioblastoma stem-like cells (GSCs) can migrate in vascular niches within or away from the ...tumour mass, increasing tumour resistance to treatments and contributing to relapses. To study individual GSC migration and their interactions with the perivasculature of the tumour microenvironment, there is a need to develop a human organotypic in vitro model. Herein, we demonstrated a perivascular niche-on-a-chip, in a serum-free condition with gravity-driven flow, that supported the stemness of patient-derived GSCs and foetal neural stem cells grown in a three-dimensional environment (3D). Endothelial cells from three organ origins, (i) human brain microvascular endothelial cells (hCMEC/D3), (ii) human umbilical vein endothelial cells (HUVECs) and, (iii) human lung microvascular endothelial cells (HMVEC-L) formed rounded microvessels within the extracellular-matrix integrated microfluidic chip. By optimising cell extraction protocols, systematic studies were performed to evaluate the effects of serum-free media, 3D cell cultures, and the application of gravity-driven flow on the characteristics of endothelial cells and their co-culture with GSCs. Our results showed the maintenance of adherent and tight junction markers of hCMEC/D3 in the serum-free culture and that gravity-driven flow was essential to support adequate viability of both the microvessel and the GSCs in co-culture (>80% viability at day 3). Endpoint biological assays showed upregulation of neovascularization-related genes (e.g., angiopoietins, vascular endothelial growth factor receptors) in endothelial cells co-cultured with GSCs in contrast to the neural stem cell reference that showed insignificant changes. The on-chip platform further permitted live-cell imaging of GSC - microvessel interaction, enabling quantitative analysis of GSC polarization and migration. Overall, our comparative genotypic (i.e. qPCR) and phenotypic (i.e. vessel permeability and GSC migration) studies showed that organotypic (brain cancer cells-brain endothelial microvessel) interactions differed from those within non-tissue specific vascular niches of human origin. The development and optimization of this on-chip perivascular niche, in a serum-free flowable culture, could provide the next level of complexity of an in vitro system to study the influence of glioma stem cells on brain endothelium.
Our understanding of the dielectric response of interfacial water, which underlies the solvation properties and reaction rates at aqueous interfaces, relies on the linear response approximation: an ...external electric field induces a linearly proportional polarization. This implies antisymmetry with respect to the sign of the field. Atomistic simulations have suggested, however, that the polarization of interfacial water may deviate considerably from the linear response. Here we present an experimental study addressing this issue. We measured vibrational sum-frequency generation spectra of heavy water (D
O) near a monolayer graphene electrode, to study its response to an external electric field under controlled electrochemical conditions. The spectra of the OD stretch show a pronounced asymmetry for positive versus negative electrode charge. At negative charge below 5 × 10
electrons per square centimetre, a peak of the non-hydrogen-bonded OD groups pointing towards the graphene surface is observed at a frequency of 2,700 per centimetre. At neutral or positive electrode potentials, this 'free-OD' peak disappears abruptly, and the spectra display broad peaks of hydrogen-bonded OD species (at 2,300-2,650 per centimetre). Miller's rule
connects the vibrational sum-frequency generation response to the dielectric constant. The observed deviation from the linear response for electric fields of about ±3 × 10
volts per metre calls into question the validity of treating interfacial water as a simple dielectric medium.
The new thermal insulating shotcrete is of great significance for the management of thermal damage in deep mines, and its own strength has a greater impact on the roadway insulation and safe ...production, so it is very necessary to study the shear strength of the new thermal insulating shotcrete under the influence of the deep hot and humid environment and the stress of mining. For the heat-insulating shotcrete, firstly, we carried out the concrete variable angle shear test under different loading rates, which concluded that the shear rate and peak shear stress showed a trend of increasing and then decreasing; as the angle increases, the different rates have a greater impact on the peak shear stress of the specimen. Secondly, the concrete variable angle shear test was carried out under the temperature and humidity cycle, which revealed that the shear strength of thermal insulated shotcrete increased firstly and then decreased with the increase of temperature at the same number of cycles. Finally, the empirical equations between the cohesive force c, the angle of internal friction ϕ and the number of warm and wet cycles n and the temperature of warm and wet cycles T are fitted with the MATLAB software respectively, and the research results provide technical references for the management of geothermal temperature in deep well projects.
Celotno besedilo
Dostopno za:
DOBA, IZUM, KILJ, NUK, PILJ, PNG, SAZU, SIK, UILJ, UKNU, UL, UM, UPUK
Graphene microribbons (rGO-MRs) are highly desired for their high electrical conductivities and specific surface areas, which contribute to multiple applications in thin, flexible, textile ...supercapacitors, sensors, and actuators. Herein, we demonstrate a facile method for creating reduced graphene oxide microribbons with microscale architecture utilizing a simple blue-violet diode laser under ambient conditions. This method takes advantage of the photochemical reduction mechanism of self-assembled graphene oxide liquid crystals (GO-LC), allowing rGO-MR patterns to be directly printed on the solution surface. The rGO-MR films demonstrated tunable diameters and can be tailored into any geometries. A maximum intrinsic electrical conductivity for rGO-MR reaching 325.8 S/m was observed. The rGO-MR textile electrodes can be assembled into microsupercapacitors with a high areal specific capacitance of 14.4 mF/cm2, a low charge-transfer impedance, and an exceptional cycling performance with a retained 96.8% capacitance after 10 000 cycles. The rGO-MR films also experience changes in resistance in response to the moisture adsorption from human breaths and therefore can also be employed as a breathing sensor for health monitoring. The presented facile method for creating multilayered rGO-MR films directly on liquid surfaces can further expand the potential for three-dimensional printing graphitic materials for various multifunctional applications.
A novel polyaniline nanorod (PAniNR) three-dimensional structure was successfully grown on flexible polyacrylonitrile (PAN) nanofiber substrate as the electrode material for electrochemical ...capacitors (ECs), constructed via self-stabilized dispersion polymerization process. The electrode offered desired mechanical properties such as flexibility and bendability, whereas it maintained optimal electrochemical characteristics. The electrode and the assembled EC cell also achieved intrinsic piezoresistive sensing properties, leading to real-time monitoring of excess mechanical pressure and bending during cell operations. The PAniNR@PAN electrodes show an average diameter of 173.6 nm, with the PAniNR growth of 50.7 nm in length. Compared to the electrodes made from pristine PAni, the gravimetric capacitance increased by 39.8% to 629.6 F/g with aqueous acidic electrolyte. The electrode and the assembled EC cell with gel electrolyte were responsive to tensile, compressive, and bending stresses with a sensitivity of 0.95 MPa–1.
Induced by flexibility of the industrial robot, cutting tool or the workpiece, chatter in robotic machining process has detrimental effects on the surface quality, tool life and machining ...productivity. Consequently, accurate detection and timely suppression for such undesirable vibration is desperately needed to achieve high performance robotic machining. This paper presents a novel approach combining the notch filter and local maximum synchrosqueezing transform for the timely chatter identification in robotic drilling. The proposed approach is accomplished through the following steps. In the first step, the optimal matrix notch filter is designed to eliminate the interference of the spindle frequency and corresponding harmonic components to the measured acceleration signal. Subsequently, the high-resolution time–frequency information of the non-stationary filtered acceleration signal is acquired by employing local maximum synchrosqueezing transform (LMSST). On this basis, the filtered acceleration signal is divided into a finite number of equal-width frequency bands, and the corresponding sub-signal for each frequency band is obtained by summing the corresponding coefficient of the LMSST. Finally, to accurately depict the non-uniformity of energy distribution during the chatter incubation process, the statistical energy entropy is calculated and utilized as the indicator to detect chatter online. The effectiveness of the proposed approach is validated by a large number of robot drilling experiments with different cutting tools, workpiece materials and machining parameters. The results show that the presented local maximum synchrosqueezing-based approach can effectively recognize the chatter at an early stage during its incubation and development process.
Motor fault diagnosis based on deep learning frameworks has gained much attention from academic research and industry to guarantee motor reliability. Those methods are commonly under two default ...assumptions: 1) massive labeled training samples and 2) the training and test data share a similar distribution under unvarying working conditions. Unfortunately, these assumptions are nearly invalid in a real-world scenario, where the signals are unlabeled and the working condition changes constantly, resulting in the diagnosis models of the previous studies that always fail in classifying the unlabeled data in real applications. To deal with those issues, in this paper, we propose a novel feature adaptive motor fault diagnosis using deep transfer learning to improve the performance by transferring the knowledge learned from labeled data under invariant working conditions to the unlabeled data under constantly changing working conditions. A convolutional neural network (CNN) is adopted as the base framework to extract multi-level features from raw vibration signals. Then, the regularization term of maximum mean discrepancy (MMD) is incorporated in the training process to impose constraints on the CNN parameters to reduce the distribution mismatch between the features in the source and target domains. To verify the effectiveness of our proposal, data from the motor tests of European driving cycle (NEDC) for simulating the real working scenario and the motor tests under invariant working conditions are, respectively, conducted as the target domain and the source domain. The results show that the proposal presents higher diagnosis accuracy for the unlabeled target data than other methods, and it is of applicability to bridge the discrepancy between different domains.