Optimized operation scheme of flash-memory-based neural network online training with ultra-high endurance

With the rapid development of machine learning,the demand for high-efficient computing becomes more and more urgent.To break the bottleneck of the traditional Von Neumann architecture,computing-in-memory(CIM)has attracted increasing attention in recent years.In this work,to provide a feasible CIM solution for the large-scale neural networks(NN)requiring continuous weight updating in online training,a flash-based computing-in-memory with high endurance(10^(9) cycles)and ultrafast programming speed is investigated.On the one hand,the proposed programming scheme of channel hot electron injection(CHEI)and hot hole injection(HHI)demonstrate high linearity,symmetric potentiation,and a depression process,which help to improve the training speed and accuracy.On the other hand,the low-damage programming scheme and memory window(MW)optimizations can suppress cell degradation effectively with improved computing accuracy.Even after 109 cycles,the leakage current(I_(off))of cells remains sub-10pA,ensuring the large-scale computing ability of memory.Further characterizations are done on read disturb to demonstrate its robust reliabilities.By processing CIFAR-10 tasks,it is evident that~90%accuracy can be achieved after 109 cycles in both ResNet50 and VGG16 NN.Our results suggest that flash-based CIM has great potential to overcome the limitations of traditional Von Neumann architectures and enable high-performance NN online training,which pave the way for further development of artificial intelligence(AI)accelerators.

Evaluation of soil fabric using elastic waves during load-unload

It is essential to assess the evolution of soil fabric as it has an important role in the mechanical responses of soils during complex loading conditions.This contribution carries out the physical experiments using three granular materials in the laboratory.The variations of compression and shear wave velocities(Vp and Vs)are investigated during load-unload cycles under dry and drained conditions.Supplementary discrete element method(DEM)simulations are performed to understand the evolution of soil fabric during the equivalent load-unload cycles using spherical particles.Vp and Vs are not always reversible even though the stress state regains its isotropic condition after unload,indicating that Vp and Vs are governed by not only the stress state but also the fabric change.The variations of Vp/Vs are density-and stress-dependent;a higher level of stress ratio(s01/s03)threshold is observed for denser packings to trigger a significant change in wave velocity ratio(Vp/Vs)for experimental results using spherical glass beads and simulation data using spherical particles.Considering the particle shape,a higher s01/s03 threshold is found for more angular particles than rounded particles.The DEM result reveals that Vp/Vs of spherical particles can be correlated linearly with the evolution of fabric ratio(Fver/Fhor)during loadunload in a pre-peak range under dry and drained conditions.

Use of various MgO resources for high-purity Mg metal production through molten salt electrolysis and vacuum distillation

A green and effective electrolytic process was developed to produce high-purity Mg metal using primary and secondary resources containing Mg O as a feedstock. The electrolysis of various Mg O resources was conducted using a Cu cathode in MgF2– LiF – KCl molten salt at 1043 K by applying an average current of 1.44 A for 12.5 h. The electrolysis of calcined North Korean magnesite and seawater Mg O clinker yielded Mg alloys of MgCu2and(Cu) phases with current efficiencies of 89.6–92.4%. The electrolysis of oxidized Mg O-C refractory brick, aged ferronickel slag, and ferronickel slag yielded Mg alloys of MgCu2and(Cu) phases with current efficiencies of 59.3–92.3%. The vacuum distillation of Mg alloys obtained was conducted at 1300 K for 10 h to produce high-purity Mg metal. After vacuum distillation, Mg metal with a purity of above 99.994% was obtained. Therefore, this study demonstrates the feasibility of the production of high-purity Mg metal from various Mg O resources using a novel electrolytic process with a Cu cathode, followed by vacuum distillation.

Prediction of fuel cell performance degradation using a combined approach of machine learning and impedance spectroscopy

Accurate prediction of performance degradation in complex systems such as solid oxide fuel cells is crucial for expediting technological advancements.However,significant challenges still persist due to limited comprehension of degradation mechanisms and difficulties in acquiring in-situ features.In this study,we propose an effective approach that integrates long short-term memory(LSTM) neural network and dynamic electrochemical impedance spectroscopy(DEIS).This integrated approach enables precise prediction of future evolutions in both current-voltage and EIS features using historical testing data,without prior knowledge of degradation mechanisms.For short-term predictions spanning hundreds of hours,our approach achieves a prediction accuracy exceeding 0.99,showcasing promising prospects for diagnostic applications.Additionally,for long-term predictions spanning thousands of hours,we quantitatively determine the significance of each degradation mechanism,which is crucial for enhancing cell durability.Moreover,our proposed approach demonstrates satisfactory predictive ability in both time and frequency domains,offering the potential to reduce EIS testing time by more than half.

Three dimensional microstructures of carbon deposition on Ni-YSZ anodes under polarization

In the present study,two Ni/YSZ anodes with different volume ratios of Ni and YSZ,30:70 and 45:55 vol%,are operated in dry methane under open circuit and polarized conditions.Three-dimensional(3D)Ni/YSZ microstructures after carbon deposition are reconstructed by the focused ion beam-scanning electron microscopy(FIB-SEM)with the help of machine learning segmentation.From the reconstructed mircostructures,volume fraction,connectivity,three phase boundary(TPB)density,and tortuosity are quantified.In addition,local carbon microstructures are quantitatively reconstructed,and the effect of polarization on carbon morphology is investigated.It is demonstrated that Ni surface in the vicinity of active TPB near the electrolyte is free from carbon formation,while remaining Ni surface at some distances from TPB exhibits severe carbon deposition.In average,total amount of carbon deposition is larger near the electrolyte.These observations imply complex interplay between the electrochemical steam generation and methane cracking on Ni surface which take place very locally near the active TPB.

Degradation of differently processed Mg-based implants leads to distinct foreign body reactions(FBRs)through dissimilar signaling pathways

Mg alloys have mechanical properties compatible with human bones.However,their rapid degradation and associated foreign body reactions in vivo significantly limit their application for human implants.In this study,three differently processed Mg alloys,pure Mg(PM),cold extruded Mg alloy AZ31(CE AZ31),and fully annealed AZ31 Mg alloy(FA AZ31)were comparatively investigated for their potential as implants using a rat model.All three implanted Mg alloys do not show any impact on hepato-and renal function,nor any signs of observable changes to vital organs.Proteomics analysis of tissues directly contacting the implants 2.5 months post implantation revealed that FA AZ31 activates very few inflammation and immune associated signaling pathways;while the CE AZ31 and PM produce more significant inflammatory responses as confirmed by cytokine array analyses.Further,FA AZ31 activated pathways for cell organization and development that may improve the recovery of injured tissues.Structurally,EBSD analysis reveals that the FA AZ31 alloy has a higher ratio of first-order pyramidal orientated(10–11){10–1–2}grain texture with a value of 0.25,while PM and CE AZ31 alloys have lower ratios of first-order pyramidal orientated texture with the values of 0.16 and 0.17,respectively.This is associated with recovery and recrystallisation during annealing which promotes grain texture which exhibits enhanced degradation behaviours and induces a more limited immune response in vivo.In conclusion,the FA AZ31 demonstrated better biocompatibility and corrosion resistance and is a promising candidate for metal-based degradable implants which warrants further investigation.

Two-stage dynamic recrystallization and texture evolution in Al-7Mg alloy during hot torsion

Hot torsion tests were performed on the Al-7Mg alloy at the temperature ranging from 300 to 500℃ and strain rates between 0.05 and 5 s^(-1) to explore the progressive dynamic recrystallization(DRX)and texture behaviors.The DRX behavior of the alloy manifested two distinct stages:Stage 1 at strain of≤2 and Stage 2 at strains of≥2.In Stage 1,there was a slight increase in the DRXed grain fraction(X_(DRX))with predominance of discontinuous DRX(DDRX),followed by a modest change in X_(DRX) until the transition to Stage 2.Stage 2 was marked by an accelerated rate of DRX,culminating in a substantial final X_(DRX) of~0.9.Electron backscattered diffraction(EBSD)analysis on a sample in Stage 2 revealed that continuous DRX(CDRX)predominantly occurred within the(121)[001]grains,whereas the(111)[110]grains underwent a geometric DRX(GDRX)evolution without a noticeable sub-grain structure.Furthermore,a modified Avrami’s DRX kinetics model was utilized to predict the microstructural refinement in the Al-7Mg alloy during the DRX evolution.Although this kinetics model did not accurately capture the DDRX behavior in Stage 1,it effectively simulated the DRX rate in Stage 2.The texture index was employed to assess the evolution of the texture isotropy during hot-torsion test,demonstrating significant improvement(>75%)in texture randomness before the commencement of Stage 2.This initial texture evolution is attributed to the rotation of parent grains and the substructure evolution,rather than to an increase in X_(DRX).

Metal-to-insulator transitions in 3d-band correlated oxides containing Fe compositions

Metal-to-insulator transitions (MITs),which are achieved in 3d-band correlated transitional metal oxides,trigger abrupt variations in electrical,optical,and/or magnetic properties beyond those of conventional semiconductors.Among such material families,iron(Fe:3d^(6)4s^(2))-containing oxides pique interest owing to their widely tunable MIT properties,which are associated with the various valence states of Fe.Their potential electronic applications also show promise,given the large abundance of Fe on Earth.Representative MIT properties triggered by critical temperature (TMIT) were reported for ReFe_(2)O_(4)(Fe^(2.5+)),ReBaFe_(2)O_(5)(Fe^(2.5+)),Fe_(3)O_(4)(Fe^(2.67+)),Re_(1/3)Sr_(2/3)FeO_(3)(Fe^(3.67+)),Re Cu_(3)Fe_(4)O_(12)(Fe^(3.75+)),and Ca_(1-x)Sr_(x)FeO_(3)(Fe^(4+))(where Re represents rare-earth elements).The common feature of MITs of these Fe-containing oxides is that they are usually accompanied by charge ordering transitions or disproportionation associated with the valence states of Fe.Herein,we review the material family of Fe-containing MIT oxides,their MIT functionalities,and their respective mechanisms.From the perspective of potentially correlated electronic applications,the tunability of the TMITand its resultant resistive change in Fe-containing oxides are summarized and further compared with those of other materials exhibiting MIT functionality.In particular,we highlight the abrupt MIT and wide tunability of TMITof Fe-containing quadruple perovskites,such as Re Cu3Fe4O12.However,their effective material synthesis still needs to be further explored to cater to potential applications.

Multidisciplinary Design Optimization of Vehicle Instrument Panel Based on Multi-objective Genetic Algorithm

Typical multidisciplinary design optimization（MDO） has gradually been proposed to balance performances of lightweight, noise, vibration and harshness（NVH） and safety for instrument panel（IP） structure in the automotive development. Nevertheless, plastic constitutive relation of Polypropylene（PP） under different strain rates, has not been taken into consideration in current reliability-based and collaborative IP MDO design. In this paper, based on tensile test under different strain rates, the constitutive relation of Polypropylene material is studied. Impact simulation tests for head and knee bolster are carried out to meet the regulation of FMVSS 201 and FMVSS 208, respectively. NVH analysis is performed to obtain mainly the natural frequencies and corresponding mode shapes, while the crashworthiness analysis is employed to examine the crash behavior of IP structure. With the consideration of lightweight, NVH, head and knee bolster impact performance, design of experiment（DOE）, response surface model（RSM）, and collaborative optimization（CO） are applied to realize the determined and reliability-based optimizations, respectively. Furthermore, based on multi-objective genetic algorithm（MOGA）, the optimal Pareto sets are completed to solve the multi-objective optimization（MOO） problem. The proposed research ensures the smoothness of Pareto set, enhances the ability of engineers to make a comprehensive decision about multi-objectives and choose the optimal design, and improves the quality and efficiency of MDO.

Thermodynamic evaluation of new metallurgical refining processes for SOG-silicon production

With the aim of developing a new silicon refining process for production of solar grade silicon, a low-temperature refining technique referred to as ＂solidification refining of silicon with a Si-Al solvent at low temperature＂ was studied. The refinability of silicon by the partial solidification from a Si-Al solvent was discussed with thermodynamic evaluation for the impurity segregation between solid silicon and a Si-Al solvent. Impurity segregation ratios were measured by using temperature gradient zone melting method for phosphorus and boron and were estimated by the thermodynamic calculation for metallic impurities. The excellent refinability was clarified from the extremely small segregation ratios of impurities at lower temperature and was also confirmed by the test refining with the partial solidification under the induction heating. Furthermore, silicon crystal growth was studied by directional solidification experiments of a Si-Al alloy, and was estimated to be diffusion controlled.

Application of A* Algorithm for Real-time Path Re-planning of an Unmanned Surface Vehicle Avoiding Underwater Obstacles

This paper describes path re-planning techniques and underwater obstacle avoidance for unmanned surface vehicle(USV) based on multi-beam forward looking sonar(FLS). Near-optimal paths in static and dynamic environments with underwater obstacles are computed using a numerical solution procedure based on an A* algorithm. The USV is modeled with a circular shape in 2 degrees of freedom(surge and yaw). In this paper, two-dimensional(2-D) underwater obstacle avoidance and the robust real-time path re-planning technique for actual USV using multi-beam FLS are developed. Our real-time path re-planning algorithm has been tested to regenerate the optimal path for several updated frames in the field of view of the sonar with a proper update frequency of the FLS. The performance of the proposed method was verified through simulations, and sea experiments. For simulations, the USV model can avoid both a single stationary obstacle, multiple stationary obstacles and moving obstacles with the near-optimal trajectory that are performed both in the vehicle and the world reference frame. For sea experiments, the proposed method for an underwater obstacle avoidance system is implemented with a USV test platform. The actual USV is automatically controlled and succeeded in its real-time avoidance against the stationary undersea obstacle in the field of view of the FLS together with the Global Positioning System(GPS) of the USV.

Decomposition of Cellulose by Continuous Near-Critical Water Reactions

A pilot-scale apparatus for continuous supercritical and near-critical water reaction was set up. A high-pressure slurry supplying system was developed to feed the solid material-water slurries. The apparatus features temperature up to 600℃, pressure up to 40MPa, residence time from 24s to 15min, maximum amount of slurry supply of 2.4 L·h-1, maximum solid content of slurry up to 10%(by mass) for cellulose from Merck, and resistance to corrosion. Long-time runs of decomposition of cellulose were carried out and steady runs were confirmed. Kinetics of cellulose decomposition was studied. The apparent activation energy evaluated was 147kJ·mol-1. In addition, a new three-step pathway for cellulose hydrolysis was proposed. The derived kinetic equation is in good agreement with the experimental data.

Performance Analysis and Improvement of Flat Torque Converters Using DOE Method

Automotive torque converters have recently been designed with an increasingly narrower profile for the purpose of achieving a smaller axial size and reducing weight. Design of experiment(DOE) and computational fluid dynamics(CFD) techniques are applied to improve the performance of a flat torque converter. Four torque converters with different flatness ratios(0.204, 0.186, 0.172, and 0.158) are designed and simulated first to investigate the effects of flatness ratio on their overall performance, including efficiency, torque ratio, and impeller torque factor. The simulation results show that the overall performance tends to deteriorate as the flatness ratio decreases. Then a parametric study covering six geometric parameters, namely, inlet and outlet angles of impeller, turbine, and stator is carried out. The results demonstrate that the inlet and outlet angles play an important role in determining the performance characteristics of a torque converter. Furthermore, the relative importance of the six design parameters is investigated using DOE method for each response(stall torque ratio and peak efficiency). The turbine outlet angle is found to exert the greatest influence on both responses. After DOE analysis, an optimized design for the flat torque converter geometry is obtained. Compared to the conventional product, the width of the optimized flat torque converter torus is reduced by about 20% while the values of stall torque ratio and peak efficiency are only decreased by 0.4% and 1.7%, respectively.The proposed new optimization strategy based on DOE method together with desirability function approach can be used for performance enhancement in the design process of flat torque converters.

Effects of heat treatment on the microstructures and mechanical properties of a new type of nitrogen-containing die steel

Nitrogen can increase the strength of steels without weakening the toughness and improve the corrosion resistance at the same time. Compared with conventional nitrogen-free die steels, a new type of nitrogen-containing die steel was developed with many superior properties, such as high strength, high hardness, and good toughness. This paper focused on the effects of heat treatment on the microstruc- tures and mechanical properties of the new type of nitrogen-containing die steel, which were investigated by the optimized deformation process and heat treatment. Isothermal spheroidal annealing and high-temperature quenching as well as high-temperature tempering were ap- plied in the experiment by means of an orthogonal method after the steel was multiply forged. The mechanical properties of nitro- gen-containing die steel forgings are better than the standard of NADCA #207-2003.

Continuous observation of twinning and dynamic recrystallization in ZM21 magnesium alloy tubes during locally heated dieless drawing

Compared to cold drawing,dieless drawing has shown great potential for manufacturing biodegradable Mg alloy microtubes due to the large reduction in area acquired in a single pass.However,owing to the local heating and local deformation,the deformation mechanism during dieless drawing is not clear,and thus causing difficulties in controlling the microstructure of dieless drawn tubes.For the purpose of acquiring a desired microstructure.in this study the deformation mechanism of ZM21 Mg alloy tube was clarified by conducting continuous observation of the microstructural evolution during dieless drawing.The results show that both SRX and DRX occurred during dieless drawing.SRX occurred before the plastic deformation to soften dieless drawn tubes.With increase of feeding speed,the deformation mechanism changed accordingly:(1) At the low-speed of 0.02 mm/s,the deformation mechanism was dominated by twin-slip sliding,during which {10-12} tension twins were generated inside grains to accommodate the plastic deformation by changing the crystal orientation.(2) At the intermediate-speed of 2 mm/s,a twin-DRX process related to {10-12} tension twin was observed,which was characterized by the generation of abundant {10-12} tension twins and the evolution of misorientation angle of {10-12} tension twins.Moreover,the transformation from twin-DRX to CDRX can be observed at the late stage of plastic deformation,which was attributed to the inhomogeneous conditions of dieless drawing.(3) At the high-speed of 5 mm/s,a CDRX process was observed,during which grain boundary sliding and grain tilting were observed,in addition to the gradual rotation of subgrains.These results show that during dieless drawing,DRX is not only a temperature-dependent phenomenon,but also influenced by the variation of feeding speed.

Shifting Rule Modification Strategy of Automatic Transmission Based on Driver-vehicle-road Environment

Accidental or frequent shift often occurs when the shifting rule is built based on traditional two parameters （i.e., velocity and throttle）, because the speed of engine varies slower than change of throttle opening. Currently, modifying shift point velocity value or throttle by throttle change rate is one of common methods, but the results are not so satisfactory in some working condition such as uphill. The reason is that these methods merely consider throttle change rate which is not enough for a car driving in driver-vehicle-road environment system. So a novel fuzzy control modification strategy is proposed to avoid or reduce those abnormal shift actions. It can adjust shifting rule by the change rate of throttle, current gear position and road environment information, while different gear position and driving environment get corresponding modification value. In order to compare the results of shifting actions, fuel consumption and braking distance, emergent braking in level road and extra-urban driving cycle（EUDC） working conditions with fuzzy shifting schedule modification strategy are simulated digitally. Furthermore, a hardware-in-the-loop simulation platform is introduced to verify its effect in slope road condition according to the ON/OFF numbers of solenoid valve in hydraulic system. The simulation results show that the problem of unexpected shift in those working conditions may be resolved by fuzzy modification strategy. At last, it is concluded that although there is some slight decline in power performance in uphill situation, this fuzzy modification strategy could correctly identify slope of road, decrease braking distance, improve vehicle comfort and fuel economy effectively and prolong the life of clutch system. So, this fuzzy logic shifting strategy provides important references for vehicle intelligent shifting schedule.

Performance Optimization of Torque Converters Based on Modified 1D Flow Model

A methodology for performance optimization of torque converters is put forward based on the one-dimensional (1D) flow model. It is found that the inaccuracy of 1D flow model for predicting hydraulic performance at the low speed ratio is mainly caused by the separation phenomenon at the stator cascade which is induced by large flow impinging at the pressure side of the stator blades. A semi-empirical separation model is presented and incorporated to the original 1D flow model. It is illustrated that the improved model is able to predict the circumferential velocity components accurately, which can be applied to performance optimization. Then, the Pareto front is obtained by using the genetic algorithm (GA) in order to inspect the coupled relationship among stalling impeller torque capacity, stalling torque ratio and efficiency. The efficiency is maximized on the premise that a target stalling impeller torque capacity and torque ratio are achieved. Finally, the optimized result is verified by the computational fluid dynamics(CFD) simulation, which indicates that the maximal efficiency is increased by 0.96%.

Multi-scale Decomposition of Co-seismic Deformation from High Resolution DEMs:a Case Study of the 2004 Mid-Niigata Earthquake

Decomposing co-seismic deformation is an immediate need for researchers who are interested in earthquake inversion analysis and geo-hazard mapping. However, conventional InSAR or digital elevation models （DEMs） imagery analyses only provide the displacement in the Line-of-Sight （LOS） direction or elevation changes. The 2004 Mid-Niigata earthquake in Japan provides lessons on how to decompose co-seismic deformation from two sets of DEMs. If three adjacent points undergo a rigid-body-translation movement, their co-seismic deformation can be decomposed by solving simultaneous equations. Although this method has been successfully used to discuss tectonic deformations, the algorithm needed improvement and a more rigorous algorithm, including a new definition of nominal plane, DEMs comparability improvement and matrix condition check is provided. Even with these procedures, the obtained decomposed displacement often showed remarkable scatter prompting the use of the moving average method, which was used to determine both tectonic and localized displacement characteristics. A cut-off window and a pair of band-pass windows were selected according to the regional geology and construction activities to ease the tectonic and localized displacement calculations, respectively. The displacement field of the tectonic scale shows two major clusters of large lateral components, and coincidently major visible landslides were found mostly within them. The localized displacement helps to reveal hidden landslides in the target area. As far as the Kizawa hamlet is concerned, the obtained vectors show down-slope movements, which are consistent with the observed traces of dislocations that were found in the Kizawa tunnel and irrigation wells. The method proposed has great potential to be applied to understanding post-earthquake rehabilitation in other areas.

The Effect of Bioturbation Activity of the Ark Clam Scapharca subcrenata on the Fluxes of Nutrient Exchange at the Sediment–Water Interface

Filter-feeding shellfish are common benthos and significantly affect the biogeochemical cycle in the shallow coastal ecosystems.Ark clam Scapharca subcrenata is one of the widely cultured bivalve species in many coastal areas owing to its tremendous economic value.However,there is little information regarding the effects of the bioturbation of S.subcrenata on the fluxes of nutrient exchange in the sediment-water interface(SWI).In this regard,S.subcrenata was sampled during October 2016 to determine the effects of its bioturbation activity on the nutrient exchange flux of the SWI.The results showed that the biological activity of S.subcrenata could increase the diffusion depth and the rate of the nutrients exchange in the sediments.The bioturbation of S.subcrenata could allow the nutrients to permeate into the surface sediments at 6-10cm and increase the release rate of nutrients at the SWI.The releasing fluxes of DIN and PO43−-P in the culture area were found to be around three times higher than that in the non-cultured region.The culture of S.subcrenata has been proved to be an important contributor to nutrient exchange across the SWI in the farming area of Haizhou Bay.Nutrients exchange in the SWI contributes a part of 86%DIN,71%PO43−-P and 18%SiO32−-Si for the aquaculture farm.

Improvement in classification accuracy of stainless steel alloys by laser-induced breakdown spectroscopy based on elemental intensity ratio analysis

Laser-induced breakdown spectroscopy(LIBS) is a useful technique for accurate sorting of metal scrap by chemical composition analysis.In this work,a method for intensity-ratiobased LIBS classification of stainless steel applicable to highly fluctuating LIBS signal conditions is proposed.The spectral line pairs for intensity ratio calculation are selected according to elemental concentration and upper levels of emission lines.It is demonstrated that the classification accuracy can be significantly improved from that of full-spectra principal component analysis or intensity-based analysis.The proposed method is considered to be suited to an industrial scrap sorting system that requires minimal maintenance and low system price.