Full Domain Research Articles

Revisiting Realistic Test-Time Training: Sequential Inference and Adaptation by Anchored Clustering Regularized Self-Training.

Deploying models on target domain data subject to distribution shift requires adaptation. Test-time training (TTT) emerges as a solution to this adaptation under a realistic scenario where access to full source domain data is not available, and instant inference on the target domain is required. Despite many efforts into TTT, there is a confusion over the experimental settings, thus leading to unfair comparisons. In this work, we first revisit TTT assumptions and categorize TTT protocols by two key factors, i.e., whether testing data is sequentially streamed and whether source model is allowed to be trained with modified loss function. Among the multiple protocols, we adopt a realistic sequential test-time training (sTTT) protocol, under which we develop a test-time anchored clustering (TTAC) approach to enable stronger test-time feature learning. TTAC discovers clusters in both source and target domains and matches the target clusters to the source ones to improve adaptation. When source domain information is strictly absent (i.e., source-free) we further develop an efficient method to infer source domain distributions for anchored clustering. Finally, self-training (ST) has demonstrated great success in learning from unlabeled data and we empirically figure out that applying ST alone to TTT is prone to confirmation bias. Therefore, a more effective TTT approach is introduced by regularizing self-training with anchored clustering, and the improved model is referred to as TTAC++. We demonstrate that, under all TTT protocols, TTAC++ consistently outperforms the state-of-the-art methods on five TTT datasets, including corrupted target domain, selected hard samples, synthetic-to-real adaptation and adversarially attacked target domain. We hope this work will provide a fair benchmarking of TTT methods, and future research should be compared within respective protocols.

Parametric Influence on the Thermofluidic Characteristics of the Branched Wavy Heat Sink

Abstract The branched wavy heat sink (BWHS) is found as a novel design for heat sink application. The aim of this paper is first to investigate the optimum branch angle by varying the branch angle while keeping the constant parametric value such as pitch, wavelength, and secondary width; the idea was established, and we found that a 45-deg branch angle is the best among other branch angles. After that, we varied the next important parametric value, i.e., secondary width, to know the optimum secondary width. Four secondary pass angle (branch angle) configurations (25 deg, 35 deg, 45 deg, and 55 deg) were proposed and numerically investigated by keeping the constant parametric value. The RNG k–ɛ model has been employed with the full domain model approach in the computational analysis. Further, the optimum branch angle has also been investigated with the varying secondary width. The secondary flowrates were strongly affected in all the cases. The pressure loss increases as the secondary pass angle increases. However, a secondary pass maintains pressure uniformity in the spanwise direction through flow migration. It is clear that raising the secondary pass angle reduces the hot spot region. Finally, the optimum branch angle has also been investigated with varying secondary widths of 0.5 mm, 0.25 mm, and 0.20 mm. According to the thermofluidic study, the BWHS design with 45-deg angle and 0.25-mm secondary width has better heat convection performance among all. With a 2% volumetric concentration of nanofluid, the heat transfer rate improves by about 10% compared to water. As a result, the 45-deg BWHS presents a viable avenue for further study. The heat transfer rate of the microchannel heat sink (MCHS) is greatly aided by secondary flow mixing. The branched wavy heat sinks (BWHSs) under forced convection with water/Al2O3 under the Re = 100–500 have been investigated.

Unusual tilted growth and epitaxial relationship of NaCl B1-structured NiO and CrN on r-plane Al2O3

Epitaxial NiO and CrN thin films were deposited on a single-crystal Al2O3(11¯02) (r-plane sapphire) using magnetron sputtering. The two materials were intentionally deposited into two different deposition chamber designs and under different conditions (temperature, pressure, gases, and energy of sputtered particles). Despite the differences in the deposition condition and material system, both materials had the same feature with uncommon tilted epitaxial growth. Through an in-depth x-ray diffraction analysis of the NaCl (B1)-structured materials on r-plane sapphire, the full twin domain epitaxial relations were determined and can be described as (110)NaCl(B1)∥(44¯03)Al2O3 and [11¯2]NaCl(B1)∥[1¯1¯20]Al2O3. This relationship differs from the previously observed orientation of (100)NaCl(B1)∥(11¯02)Al2O3 and [100]NaCl(B1)∥[101¯0]Al2O3. These results are of general relevance for the growth of the extended NaCl (B1)-structured cubic material family onto a r-plane sapphire substrate where similar epitaxial growth can be expected.

Study on tank damage and response of adjacent tanks in full time domain of detonation

The flammable gas volatilized from the storage medium in the tank will burn and explode under the action of an accidental ignition source. Tank explosions formed by the explosive shock wave, explosive debris will not only cause their own damage, but also on the safe operation of adjacent tanks to form a threat. Therefore, in this paper, the vertical dome roof tank is taken as the research object, the criteria of brittle fracture failure of tank materials are constructed, and the multi-field coupling calculation model of tank detonation based on TNT equivalent method is established, taking into account the fluid properties inside and outside the tank, the structure of implosion tank and adjacent tank. By analyzing the full time domain process of gas detonation in tanks under different influence factors, the response to implosion damage of tanks and adjacent tanks under the action of explosion shock wave overpressure and debris was obtained. With the increase of the liquid level, the overpressure and deformation of the shock wave of the detonation tank and adjacent tank decrease, and the type and number of fragments decrease. With the increase of the volume of the tank, the overpressure and stress of the shock wave become larger, the strain rate of the tank becomes larger, and brittle failure is more likely to occur. The types and number of fragments formed increase, but detonation fragments are not formed when the tank is full. The research results in this paper provide a theoretical basis and calculation method for preventing and controlling tank implosion accidents, as well as tank design and tank spacing layout.

Elasto-Thermodiffusive Microtemperature Model Induced by a Mechanical Ramp-Type of Nanoscale Photoexcited Semiconductor

ABSTRACT In this study, a novel theoretical approach to describe thermo-optical elastic materials is proposed. This formulation offers insights into the relationship between plasma waves and thermomechanical waves when the microtemperature properties of semiconductor materials such as silicon are taken into account. The proposed model involves incorporating the concept of microtemperature effect according to photothermal (PT) excitation processes in nonlocal photo-thermoelasticity. The investigation of the electron-hole interaction concerning the elasto-thermodiffusion (ETD) theory in the context of thermoelastic (TD) and electronic (ED) deformation is the main focus of the work. The advanced model is utilized to analyze how the mechanical loading of ramp-like structures impacts the unrestricted nonlocality semiconductor material on a free outer plane. The Laplace transform approach in one-dimensional (1D) provides an analytical solution for main non-dimensional thermo-photo-elastic fields (the hole charge carrier field, the displacement (acoustic) wave, thermal wave, and plasma wave (carrier density). The primary fields in the Laplace domain are obtained analytically, and boundary conditions are established using a mechanical ramp type. Using a numerical inverse Laplace transform, full time domain solutions for the fundamental fields are obtained numerically according to the Riemann-sum approximation approach. The graphical representations illustrate and discuss the outcomes of the main physical fields.

Dynamic Response of Bridge–Tunnel Overlapping Structures under High-Speed Railway and Subway Train Loads

With the rapid development of high-speed railroads and subways, there has been an increasing number of bridge–tunnel overlapping structures. To study the dynamic response characteristics of bridge–tunnel structures under the synergistic effects of the vibration generated by high-speed railway and subway trains, the dynamic response characteristics of a bridge–tunnel structure under single-point vibration loading was analyzed by conducting numerical simulations and model tests, with the frequency response function and peak acceleration as the evaluation indices. The dynamic response characteristics of the overlapping structure under moving vibration loads of the high-speed railway and subway trains were further analyzed. The results showed that the dynamic response of the bridge–tunnel overlapping structure increased with the increase in the frequency under the full frequency domain single-point sweep vibration load. The dynamic response of the tunnel hance near the pile foundation side was significantly greater than the vault and invert. Compared with the effect of high-speed train loads alone, the dynamic response of the bridge–tunnel overlapping structure under the synergistic effects of high-speed railways and subways increased significantly and varied at different locations. This investigation provides theoretical support for the design and construction of bridge–tunnel overlapping structures under the synergistic effects of high-speed railways and subways, contributing to improving engineering quality and safety.

An experimental and numerical investigation of thermofluidic properties on the novel curved microchannel heat sink with slotted section

The current work proposes the development of a novel curved microchannel heat sink by modifying the existing flat microchannel heat sink for enhancing the heat dissipation in the electronic device without significant pressure drop penalty. The inspiration for the curved channel is taken from a NACA airfoil. The microchannel comprises two sections: a curved section and a slotted straight section. Experimental and numerical investigations are performed on the proposed geometry to study the thermofluidic properties for Reynolds number (Re) ranging from 200 to 500 using DI water as working fluid. Fluid flow and heat transfer are analysed by carrying out a full domain numerical simulation in ANSYS Fluent 19.0. The thermofluidic properties of the curved microchannel heat sink (CMCHS) and straight microchannel heat sink (SMCHS) are compared numerically by keeping the geometrical parameters (hydraulic diameter, foot area and fin area) and inlet parameters (inlet velocity and heat flux) identical. The average heat transfer coefficient (havg) of CMCHS is enhanced over SMCHS due to the formation of vorticities in the slotted region owing to the fluid intermixing and boundary layer disruption. The presence of curved section leads to reducing pressure drop penalty pertaining to gravitational effects. Together, the curved section and slotted section enhance the performance factor of CMCHS over SMCHS. The vortices formed increase as the Reynolds number increases, enabling an increase in the heat transfer coefficient. Therefore, the performance factor also increases as the Reynolds number is increased. An enhancement of ≈ 67.6 % in havg is achieved for Re = 500 in CMCHS compared to SMCHS with DI water as the working fluid.

Seasonal electricity consumption forecasting: an approach with novel weakening buffer operator and fractional order accumulation grey model

PurposeAccurate prediction of seasonal power consumption trends with impact disturbances provides a scientific basis for the flexible balance of the long timescale power system. Consequently, it fosters reasonable scheduling plans, ensuring the safety of the system and improving the economic dispatching efficiency of the power system.Design/methodology/approachFirst, a new seasonal grey buffer operator in the longitudinal and transverse dimensional perspectives is designed. Then, a new seasonal grey modeling approach that integrates the new operator, full real domain fractional order accumulation generation technique, grey prediction modeling tool and fruit fly optimization algorithm is proposed. Moreover, the rationality, scientificity and superiority of the new approach are verified by designing 24 seasonal electricity consumption forecasting approaches, incorporating case study and amalgamating qualitative and quantitative research.FindingsCompared with other comparative models, the new approach has superior mean absolute percentage error and mean absolute error. Furthermore, the research results show that the new method provides a scientific and effective mathematical method for solving the seasonal trend power consumption forecasting modeling with impact disturbance.Originality/valueConsidering the development trend of longitudinal and transverse dimensions of seasonal data with impact disturbance and the differences in each stage, a new grey buffer operator is constructed, and a new seasonal grey modeling approach with multi-method fusion is proposed to solve the seasonal power consumption forecasting problem.HighlightsThe highlights of the paper are as follows:A new seasonal grey buffer operator is constructed.The impact of shock perturbations on seasonal data trends is effectively mitigated.A novel seasonal grey forecasting approach with multi-method fusion is proposed.Seasonal electricity consumption is successfully predicted by the novel approach.The way to adjust China's power system flexibility in the future is analyzed.

Practical Aspects of Physics-Informed Neural Networks Applied to Solve Frequency-Domain Acoustic Wave Forward Problem

Abstract Physics-informed neural networks (PINNs) have been used by researchers to solve partial differential equation (PDE)-constrained problems. We evaluate PINNs to solve for frequency-domain acoustic wavefields. PINNs can solely use PDEs to define the loss function for optimization without the need for labels. Partial derivatives of PDEs are calculated by mesh-free automatic differentiations. Thus, PINNs are free of numerical dispersion artifacts. It has been applied to the scattered acoustic wave equation, which relied on boundary conditions (BCs) provided by the background analytical wavefield. For a more direct implementation, we solve the nonscattered acoustic wave equation, avoiding limitations related to relying on the background homogeneous medium for BCs. Experiments support our following insights. Although solving time-domain wave equations using PINNs does not require absorbing boundary conditions (ABCs), ABCs are required to ensure a unique solution for PINNs that solve frequency-domain wave equations, because the single-frequency wavefield is not localized and contains wavefield information over the full domain. However, it is not trivial to include the ABC in the PINN implementation, so we develop an adaptive amplitude-scaled and phase-shifted sine activation function, which performs better than the previous implementations. Because there are only two outputs for the fully connected neural network (FCNN), we validate a linearly shrinking FCNN that can achieve a comparable and even better accuracy with a cheaper computational cost. However, there is a spectral bias problem, that is, PINNs learn low-frequency wavefields far more easily than higher frequencies, and the accuracy of higher frequency wavefields is often poor. Because the shapes of multifrequency wavefields are similar, we initialize the FCNN for higher frequency wavefields by that of the lower frequencies, partly mitigating the spectral bias problem. We further incorporate multiscale positional encoding to alleviate the spectral bias problem. We share our codes, data, and results via a public repository.

Performance analysis of an InAs/GaSb superlattice barrier photodetector covering the full LWIR spectral domain

Performance analysis of an InAs/GaSb superlattice barrier photodetector covering the full LWIR spectral domain

Learning a microlocal prior for limited-angle tomography

Abstract Limited-angle tomography is a highly ill-posed linear inverse problem. It arises in many applications, such as digital breast tomosynthesis. Reconstructions from limited-angle data typically suffer from severe stretching of features along the central direction of projections, leading to poor separation between slices perpendicular to the central direction. In this paper, a new method is introduced, based on machine learning and geometry, producing an estimate for interfaces between regions of different X-ray attenuation. The estimate can be presented on top of the reconstruction, indicating more reliably the separation between features. The method uses directional edge detection, implemented using complex wavelets and enhanced with morphological operations. By using convolutional neural networks, the visible part of the singular support is first extracted and then extended to the full domain, filling in the parts of the singular support that would otherwise be hidden due to the lack of measurement directions.

The Practical Path of Ruling Colleges and Universities According to Law in the New Era Combined with Big Data Analysis

Abstract With the development of the information age, big data analysis has become the basis for social management and the development of new ideas for intelligent governance. In this paper, the governance process of colleges and universities is presented and stored in the form of data by simulating the content of governance through the full domain data set, and the relationship between different subjects and object features is mapped through the generalized Cartesian product. From there, various data patterns are clustered to effectively manage various types of operations’ linkage. The multi-objective optimization algorithm is used to solve the Pareto optimal solution for different governance subjects, i.e., the mutual coordination between teachers and students. The objective optimization function is established to identify the Pareto frontiers of various governance programs and emergency plans. The results show that in the driving pathway of college big data governance, the correct rate of classification of policing grid governance events in 28 colleges and universities is 0.8279 on average, which effectively improves governance efficiency. Big data governance in colleges and universities enables college governance to continuously move from decentralized to centralized, from part to whole, and to achieve accurate and efficient governance modes.

Three-dimensional effects during thermocapillary-driven melting of PCMs in cuboidal containers in microgravity

The melting of a phase change material (PCM) in a cuboidal domain under microgravity conditions is investigated numerically. The upper surface of the PCM is free (in contact with air, for example) and variations in its surface tension drive thermocapillary convection in the liquid phase, which significantly enhances heat transfer and accelerates melting. Furthermore, the change in liquid fraction during melting is associated with transitions among various modes of thermocapillary dynamics, including an oscillatory instability to hydrothermal waves. While the characteristics of PCM melting and thermocapillary dynamics have previously been investigated in this system using a two-dimensional model, the current work examines the important question of transverse dynamics and their effect on the melting process. Careful quantitative comparisons are made between the three- and two-dimensional models in terms of melting times, solid/liquid interface evolution, thermal fields, and spectrograms. The results show that transverse modes are often, but not always, reflection symmetric about the midplane and that their influence on melting and PCM performance is relatively minor in most cases. Thus, two-dimensional models may be used to reduce computational costs while still providing a reasonable approximation of the melting process for high Prandtl number materials, especially when compared to the midplane of the full cuboidal domain.

Investigation of SARS-CoV-2 individual proteins reveals the in vitro and in vivo immunogenicity of membrane protein

Evidence in SARS-CoV-2 patients have identified that viral infection is accompanied by the expression of inflammatory mediators by both immune and stromal cells within the pulmonary system. However, the immunogenicity of individual SARS-CoV-2 proteins has yet to be evaluated. The SARS-CoV-2 virus consists of 29 proteins, categorized either as nonstructural proteins (NSP’s), structural proteins (SP’s) or accessory proteins. Here we sought to evaluate the immunogenicity of NSP 1, 7, 8, 9, 10, 12, 14, 16 and the SP’s spike protein (full length, S1, S2 and receptor binding domain subunits), nucleocapsid and membrane SARS-CoV-2 proteins against THP-1 and human peripheral blood mononuclear cells (PBMCs). Our results indicate that various SARS-CoV-2 proteins elicit a proinflammatory immune response indicated by increases in cytokines TNF, IL-6 and IL-1β. Our results support that SARS-CoV-2 membrane protein induced a robust increase of TNF, IL-6, IL-1β and IL-10 expression in both THP-1 and human PBMC’s. Further evaluation of intranasal membrane protein challenge in male and female BALB/c mice show increases in BALF (bronchalveolar lavage fluid) proinflammatory cytokine expression, elevated cellularity, predominantly neutrophilic, and concomitant peribronchiolar and perivascular lymphomononuclear and neutrophilic inflammation. Our results suggest that individual membrane associated SARS-CoV-2 proteins induce a robust immune response that may contribute to viral induced cytokine release syndrome (CRS) in the lungs of moderate to severe COVID-19 patients. We posit that SARS-CoV-2 membrane challenges in immune-competent mice can serve as an adequate surrogate for the development of novel treatments for SARS-CoV-2 induced pulmonary inflammation, thereby avoiding expensive live virus studies under BSL-3 conditions.

Material removal mechanisms affected by milling modes for defective KDP surfaces

Surface defects which are caused by diamond fly-cutting can degrade the optical characteristics of KDP crystals by inducing the laser damage and resultantly shortening the service life when working in the condition of extremely high-power laser irradiations. The ball-end milling repair can totally remove the defects and correspondingly enhance the ability of the KDP crystals to resist laser damage. However, because of the existence of pull-milling, push-milling, down-milling, and up-milling in the complex machining trajectory, it is of great challenge for soft-brittle KDP crystals to achieve full plastic domain milling of repair points for obtaining ultra-smooth repaired surface. In this study, theoretical models of the maximum uncut chip thickness (UCT) considering the defect are constructed in the various milling modes. The cutting force and surface morphology affected by milling modes on the defective surface are simulated using a three-dimensional (3D) finite element model. Combined with the simulations, micro-milling tests are employed to confirm the influence of milling modes on the material removal mechanisms for defective KDP crystals by cutting force, surface morphology, surface roughness and specific cutting energy (SCE). It is found that defects could reduce the maximum UCT to distinct degrees in the various milling modes. Defects have a greater influence on the maximum UCT in the push-milling mode than that in the pull-milling mode, while the influence is the same in the up-milling and down-milling modes. Interestingly, the ductile removal mode occurs on the both defect-free and defective surfaces in the pull-milling mode when the feed per tooth (fz) is 0.5 µm/z. In contrast, the brittle-ductile transition caused by surface defects occurs in the other three milling modes. Furthermore, defects would lead to severer size effect in the all milling modes. When the fz is 0.1 µm/z, for the defective surface, the SCE is larger and the ploughing effect is more significant in the push-milling mode compared to the pull-milling mode. More importantly, as a result of the difficulty of chip discharge, the SCE is greater in the up-milling mode than in the down-milling mode on the defective surface. This work reveals the influence mechanism of milling modes on the material removal process for defective KDP crystals and offers a theoretical foundation and technical value for improving the repaired surface quality of functional KDP crystals served as the optical devices.

Dataset Design for Building Models of Chemical Reactivity.

Models can codify our understanding of chemical reactivity and serve a useful purpose in the development of new synthetic processes via, for example, evaluating hypothetical reaction conditions or in silico substrate tolerance. Perhaps the most determining factor is the composition of the training data and whether it is sufficient to train a model that can make accurate predictions over the full domain of interest. Here, we discuss the design of reaction datasets in ways that are conducive to data-driven modeling, emphasizing the idea that training set diversity and model generalizability rely on the choice of molecular or reaction representation. We additionally discuss the experimental constraints associated with generating common types of chemistry datasets and how these considerations should influence dataset design and model building.

Abstract A094: Preclinical characterization of HMPL-415, a second-generation SHP2 inhibitor

Abstract Background: Oncogenic activation of the RAS/MAPK signaling pathway is one of the leading causes for driving a variety of cancers. Src homology 2-containing protein tyrosine phosphatase 2 (SHP2) functions downstream of multiple RTKs, and integrates growth factor signals to promote RAS activation, making it an attractive drug target for cancers harboring oncogenic alterations in RAS/MAPK pathway. Herein, we present the preclinical characterization of HMPL-415, a highly potent, selective and non-competitive SHP2 inhibitor, discovered and being currently developed in phase I clinical trial (NCT05886374) by HUTCHMED. Methods: The inhibition of HMPL-415 on the full length and catalytic domain of SHP2 was detected by a biochemical DIFMUP pseudosubstracte-fluorgenic assay. For selectivity, HMPL-415 was assessed against a panel of 413 kinases and 21 phosphatases by Eurofins. The modulation of HMPL-415 on RAS/MAPK signaling cascades was evaluated by western blot. In vitro anti-proliferation activity was measured by CellTiter-Glo luminescent assay. Multiple human xenograft tumors with RAS/MAPK pathway activation were applied in immune-deficient mice to investigate the anti-tumor activity of HMPL-415. Results: In vitro, HMPL-415 displayed higher potency in enzyme (5 fold) and cellular target inhibition (> 20 fold) compared to first-generation SHP inhibitor TNO155. It strongly inhibited full length of human SHP2 with an IC50 of 0.4 nM, while had no inhibitory activity up to 10 µM against the catalytic domain of SHP2. In NCI-H358 (KrasG12C), NCI-H508 (class Ⅲ BRAFG596R), and NCI-H1838 (NF1N184fs) cells, HMPL-415 inhibited p-ERK with IC50s of 1~3 nM, and significantly suppressed downstream molecules p-RSK and p-S6. Through inhibition of RAS/MAPK signaling pathway, HMPL-415 robustly inhibited cell proliferation in a panel of human cancer cell lines with KRAS G12C/V, BRAF class Ⅲ, NF1 loss of function (LOF), RTK mutations or KRAS amplification (median GI50 =1 nM) compared to cell lines harboring KRAS G12D, G13 or Q61 mutations (median GI50 >1 µM), indicating a dependence on specific KRAS nucleotide-cycling for anti-tumor effect. Furthermore, the selectivity across panels of 413 kinases and 21 phosphatases revealed that HMPL-415 should have a low off-target risk. In vivo, oral administration of HMPL-415 showed prolonged and high tumor exposure and sustained inhibition on RAS/MAPK signaling after repeat dosing. In human xenograft models carrying KRAS, BRAF class Ⅲ, NF1LOF, and EGFR alterations, HMPL-415 dose-dependently induced tumor growth inhibition at continuous daily doses that were well tolerated. Tumor regression was observed in most tumor models tested at the dose of 3 mg/kg/day, which demonstrated much superior anti-tumor activity than that of TNO155 at 30 mg/kg/day. Conclusion: HMPL-415 is a highly potent, selective and non-competitive SHP2 inhibitor that demonstrates strong anti-tumor activity in vitro and in vivo, supporting further clinical evaluation. Citation Format: Jia Hu, Jun Ni, Zhihu Gao, Xiaoqing Liu, Hui Zhang, Pan Wang, Min Cheng, Guanglin Wang, Zeyu Zhong, Jian Wang, Yang Sai, Na Yang, Weiguo Qing, Yongxin Ren, Michael Shi, Weiguo Su. Preclinical characterization of HMPL-415, a second-generation SHP2 inhibitor [abstract]. In: Proceedings of the AACR-NCI-EORTC Virtual International Conference on Molecular Targets and Cancer Therapeutics; 2023 Oct 11-15; Boston, MA. Philadelphia (PA): AACR; Mol Cancer Ther 2023;22(12 Suppl):Abstract nr A094.

Sensorless control of dual three-phase permanent magnet synchronous motor based on speed feedback and frequency-variable tracking.

The dual three-phase Permanent Magnet Synchronous Motor (PMSM) control system is characterized by its high reliability, slight torque fluctuation and low harmonic content. It is very suitable for systems requiring high power output and high reliability, for instance, electric vehicles, aerospace and military equipment. In this paper, a full speed domain sensorless control technology for dual three-phase PMSM is proposed which solves the limitations of other sensorless controls, improves system accuracy and stability, and has high practicality in fields such as new energy vehicles. The mathematical model of this motor in a static coordinate system is established, and the sine and cosine signals along with the velocity and angle information are obtained by using the flux linkage observer. Moreover, the estimated angle error parameter is introduced into the flux linkage observer; as a result, the estimation accuracy is improved by the estimated speed feedback, and the current frequency is tracked by the stator current Frequency-Variable Tracker (FVT) to reduce the current error. Meanwhile, to make the observer's estimation more accurate and to improve its ability to resist disturbance, a rotor disturbance is added to act as a disturbance variable. Through the mechanical motion equation of the motor, a fourth-order Extended State Observer (ESO) is built to calculate the rotor position and speed. Finally, the technology accuracy is verified using simulation and experimental results. The findings prove that the sensorless detection technology, with speed feedback introduce in this paper, has good reliability and high precision for dual three-phase PMSM under dynamic and static conditions.

Structural engineering and truncation of α-amylase from the hyperthermophilic archaeon Methanocaldococcus jannaschii

Alpha amylases catalyse the hydrolysis of α-1, 4-glycosidic bonds in starch, yielding glucose, maltose, dextrin, and short oligosaccharides, vital to various industrial processes. Structural and functional insights on α-amylase from Methanocaldococcus jannaschii were computationally explored to evaluate a catalytic domain and its fusion with a small ubiquitin-like modifier (SUMO). The recombinant proteins' production, characterization, ligand binding studies, and structural analysis of the cloned amylase native full gene (MjAFG), catalytic domain (MjAD) and fusion enzymes (S-MjAD) were thoroughly analysed in this comparative study. The MjAD and S-MjAD showed 2-fold and 2.5-fold higher specific activities (μmol min−1 mg −1) than MjAFG at 95 °C at pH 6.0. Molecular modelling and MD simulation results showed that the removal of the extra loop (178 residues) at the C-terminal of the catalytic domain exposed the binding and catalytic residues near its active site, which was buried in the MjAFG enzyme. The temperature ramping and secondary structure analysis of MjAFG, MjAD and S-MjAD through CD spectrometry showed no notable alterations in the secondary structures but verified the correct folding of MjA variants. The chimeric fusion of amylases with thermostable α-glucosidases makes it a potential candidate for the starch degrading processes.

Contrasting augmented features for domain adaptation with limited target domain data

Domain adaptation aims to alleviate distribution gaps between source and target domains. However, when the available target domain data are scarce for training, learning generalizable representations for domain adaptation is challenging. We propose a novel approach, dubbed Contrasting Augmented Features (CAF), to tackle the challenge of insufficient target domain data for domain adaptation, by generating and contrasting augmented features. We introduce a semantic feature generator to generate augmented features by replacing the instance-level feature statistics of one domain with another domain. With the augmented features, we further design the reweighted instance contrastive loss and category contrastive loss to improve feature discrimination and align feature distributions of source and target domains. CAF can be applied to few-shot domain adaptation and unsupervised domain adaptation with limited unlabeled target domain data. Despite its simplicity, extensive experiments show promising results for both applications. In addition, experiments demonstrate that CAF is more robust to the number of target domain data and also effective in vanilla unsupervised domain adaptation setting with full target domain data.