Prediction model for corrosion rate of low-alloy steels under atmospheric conditions using machine learning algorithms

This work constructed a machine learning(ML)model to predict the atmospheric corrosion rate of low-alloy steels(LAS).The material properties of LAS,environmental factors,and exposure time were used as the input,while the corrosion rate as the output.6 dif-ferent ML algorithms were used to construct the proposed model.Through optimization and filtering,the eXtreme gradient boosting(XG-Boost)model exhibited good corrosion rate prediction accuracy.The features of material properties were then transformed into atomic and physical features using the proposed property transformation approach,and the dominant descriptors that affected the corrosion rate were filtered using the recursive feature elimination(RFE)as well as XGBoost methods.The established ML models exhibited better predic-tion performance and generalization ability via property transformation descriptors.In addition,the SHapley additive exPlanations(SHAP)method was applied to analyze the relationship between the descriptors and corrosion rate.The results showed that the property transformation model could effectively help with analyzing the corrosion behavior,thereby significantly improving the generalization ability of corrosion rate prediction models.

Jetting-based bioprinting:process,dispense physics,and applications

Jetting-based bioprinting facilitates contactless drop-on-demand deposition of subnanoliter droplets at well-defined positions to control the spatial arrangement of cells,growth factors,drugs,and biomaterials in a highly automated layer-by-layer fabrication approach.Due to its immense versatility,jetting-based bioprinting has been used for various applications,including tissue engineering and regenerative medicine,wound healing,and drug development.A lack of in-depth understanding exists in the processes that occur during jetting-based bioprinting.This review paper will comprehensively discuss the physical considerations for bioinks and printing conditions used in jetting-based bioprinting.We first present an overview of different jetting-based bioprinting techniques such as inkjet bioprinting,laser-induced forward transfer bioprinting,electrohydrodynamic jet bioprinting,acoustic bioprinting and microvalve bioprinting.Next,we provide an in-depth discussion of various considerations for bioink formulation relating to cell deposition,print chamber design,droplet formation and droplet impact.Finally,we highlight recent accomplishments in jetting-based bioprinting.We present the advantages and challenges of each method,discuss considerations relating to cell viability and protein stability,and conclude by providing insights into future directions of jetting-based bioprinting.

Time-lagged Effects of the Spring Atmospheric Heat Source over the Tibetan Plateau on Summer Precipitation in Northeast China during 1961–2020:Role of Soil Moisture

The spring atmospheric heat source(AHS)over the Tibetan Plateau(TP)has been suggested to affect the Asian summer monsoon and summer precipitation over South China.However,its influence on the summer precipitation in Northeast China(NEC)remains unknown.The connection between spring TP AHS and subsequent summer precipitation over NEC from 1961 to 2020 is analyzed in this study.Results illustrate that stronger spring TP AHS can enhance subsequent summer NEC precipitation,and higher soil moisture in the Yellow River Valley-North China region(YRVNC)acts as a bridge.During spring,the strong TP AHS could strengthen the transportation of water vapor to East China and lead to excessive rainfall in the YRVNC.Thus,soil moisture increases,which regulates local thermal conditions by decreasing local surface skin temperature and sensible heat.Owing to the memory of soil moisture,the lower spring sensible heat over the YRVNC can last until mid-summer,decrease the land–sea thermal contrast,and weaken the southerly winds over the East Asia–western Pacific region and convective activities over the South China Sea and tropical western Pacific.This modulates the East Asia–Pacific teleconnection pattern,which leads to a cyclonic anomaly and excessive summer precipitation over NEC.

Relative Impacts of Sea Ice Loss and Atmospheric Internal Variability on the Winter Arctic to East Asian Surface Air Temperature Based on Large-Ensemble Simulations with NorESM2

To quantify the relative contributions of Arctic sea ice and unforced atmospheric internal variability to the “warm Arctic, cold East Asia”(WACE) teleconnection, this study analyses three sets of large-ensemble simulations carried out by the Norwegian Earth System Model with a coupled atmosphere–land surface model, forced by seasonal sea ice conditions from preindustrial, present-day, and future periods. Each ensemble member within the same set uses the same forcing but with small perturbations to the atmospheric initial state. Hence, the difference between the present-day(or future) ensemble mean and the preindustrial ensemble mean provides the ice-loss-induced response, while the difference of the individual members within the present-day(or future) set is the effect of atmospheric internal variability. Results indicate that both present-day and future sea ice loss can force a negative phase of the Arctic Oscillation with a WACE pattern in winter. The magnitude of ice-induced Arctic warming is over four(ten) times larger than the ice-induced East Asian cooling in the present-day(future) experiment;the latter having a magnitude that is about 30% of the observed cooling. Sea ice loss contributes about 60%(80%) to the Arctic winter warming in the present-day(future) experiment. Atmospheric internal variability can also induce a WACE pattern with comparable magnitudes between the Arctic and East Asia. Ice-lossinduced East Asian cooling can easily be masked by atmospheric internal variability effects because random atmospheric internal variability may induce a larger magnitude warming. The observed WACE pattern occurs as a result of both Arctic sea ice loss and atmospheric internal variability, with the former dominating Arctic warming and the latter dominating East Asian cooling.

Probing signals of atmospheric gravity waves excited by the July 29,2021 M_(W)8.2 Alaska earthquake

It is commonly believed that the atmosphere is decoupled from the solid Earth.Thus,it is difficult for the seismic wave energy inside the Earth to propagate into the atmosphere,and atmospheric pressure wave signals excited by earthquakes are unlikely to exist in atmospheric observations.An increasing number of studies have shown that earthquakes,volcanoes,and tsunamis can perturb the Earth's atmosphere due to various coupling effects.However,the observations mainly focus on acoustic waves with periods of less than 10 min and inertial gravity waves with periods of greater than 1 h.There are almost no clear observations of gravity waves that coincide with observations of low-frequency signals of the Earth's free oscillation frequency band within 1 h.This paper investigates atmospheric gravity wave signals within1 h of surface-atmosphere observations using the periodogram method based on seismometer and microbarometer observations from the global seismic network before and after the July 29,2021 M_(w)8.2 Alaska earthquake in the United States.The numerical results show that the atmospheric gravity wave signals with frequencies similar to those of the Earth's free oscillations _(0)S_(2) and _(0)T_(2) can be detected in the microbaro meter observations.The results con firm the existence of atmospheric gravity waves,indicating that the atmosphere and the solid Earth are not decoupled within this frequency band and that seismic wave energy excited by earthquakes can propagate from the interior of the Earth to the atmosphere and enhance the atmospheric gravity wave signals within 1 h.

Different El Niño Flavors and Associated Atmospheric Teleconnections as Simulated in a Hybrid Coupled Model

A previously developed hybrid coupled model(HCM)is composed of an intermediate tropical Pacific Ocean model and a global atmospheric general circulation model(AGCM),denoted as HCMAGCM.In this study,different El Niño flavors,namely the Eastern-Pacific(EP)and Central-Pacific(CP)types,and the associated global atmospheric teleconnections are examined in a 1000-yr control simulation of the HCMAGCM.The HCMAGCM indicates profoundly different characteristics among EP and CP El Niño events in terms of related oceanic and atmospheric variables in the tropical Pacific,including the amplitude and spatial patterns of sea surface temperature(SST),zonal wind stress,and precipitation anomalies.An SST budget analysis indicates that the thermocline feedback and zonal advective feedback dominantly contribute to the growth of EP and CP El Niño events,respectively.Corresponding to the shifts in the tropical rainfall and deep convection during EP and CP El Niño events,the model also reproduces the differences in the extratropical atmospheric responses during the boreal winter.In particular,the EP El Niño tends to be dominant in exciting a poleward wave train pattern to the Northern Hemisphere,while the CP El Niño tends to preferably produce a wave train similar to the Pacific North American(PNA)pattern.As a result,different climatic impacts exist in North American regions,with a warm-north and cold-south pattern during an EP El Niño and a warm-northeast and cold-southwest pattern during a CP El Niño,respectively.This modeling result highlights the importance of internal natural processes within the tropical Pacific as they relate to the genesis of ENSO diversity because the active ocean–atmosphere coupling is allowed only in the tropical Pacific within the framework of the HCMAGCM.

Characteristics Variation of Atmospheric CO_(2)and CH_(4)in Yongxing Island,South China Sea

Using the observation data in Yongxing Island,South China Sea(SCS)from December 2013 to November 2018,the multiple time scales variation of atmospheric CO_(2)and CH_(4)were analyzed to understand their temporal variation characteristics and controlling factors.The regional-averaged background mole fractions of CO_(2)and CH_(4)both show a single-period sinusoidal variation with a lower value at noon and a higher value in the wee hours.In the seasonal scale,they exhibited a significant seasonal difference with higher values in winter and lower values in summer.In the annual scale,CO_(2)and CH_(4)both show an increasing trend,with an annual growth rate of approximately 3.2 ppm and 12 ppb,respectively.The annual growth rate at this site was higher than the global average.The change in atmospheric CO_(2)and CH_(4)in Yongxing Island was probably caused by the higher emission of the surrounding areas and the airflows driven by monsoon.Hopefully,the long-term and high-resolution greenhouse gases(GHGs)dataset will aid relevent researchers and decision-makers in performing more in-depth studies for GHG sources in order to derive effective strategies.

Developing an atmospheric aging evaluation model of acrylic coatings:A semi-supervised machine learning algorithm

To study the atmospheric aging of acrylic coatings,a two-year aging exposure experiment was conducted in 13 representative climatic environments in China.An atmospheric aging evaluation model of acrylic coatings was developed based on aging data including11 environmental factors from 567 cities.A hybrid method of random forest and Spearman correlation analysis was used to reduce the redundancy and multicollinearity of the data set by dimensionality reduction.A semi-supervised collaborative trained regression model was developed with the environmental factors as input and the low-frequency impedance modulus values of the electrochemical impedance spectra of acrylic coatings in 3.5wt%NaCl solution as output.The model improves accuracy compared to supervised learning algorithms model(support vector machines model).The model provides a new method for the rapid evaluation of the aging performance of acrylic coatings,and may also serve as a reference to evaluate the aging performance of other organic coatings.

A Neural-network-based Alternative Scheme to Include Nonhydrostatic Processes in an Atmospheric Dynamical Core

Here,a nonhydrostatic alternative scheme(NAS)is proposed for the grey zone where the nonhydrostatic impact on the atmosphere is evident but not large enough to justify the necessity to include an implicit nonhydrostatic solver in an atmospheric dynamical core.The NAS is designed to replace this solver,which can be incorporated into any hydrostatic models so that existing well-developed hydrostatic models can effectively serve for a longer time.Recent advances in machine learning(ML)provide a potential tool for capturing the main complicated nonlinear-nonhydrostatic relationship.In this study,an ML approach called a neural network(NN)was adopted to select leading input features and develop the NAS.The NNs were trained and evaluated with 12-day simulation results of dry baroclinic-wave tests by the Weather Research and Forecasting(WRF)model.The forward time difference of the nonhydrostatic tendency was used as the target variable,and the five selected features were the nonhydrostatic tendency at the last time step,and four hydrostatic variables at the current step including geopotential height,pressure in two different forms,and potential temperature,respectively.Finally,a practical NAS was developed with these features and trained layer by layer at a 20-km horizontal resolution,which can accurately reproduce the temporal variation and vertical distribution of the nonhydrostatic tendency.Corrected by the NN-based NAS,the improved hydrostatic solver at different horizontal resolutions can run stably for at least one month and effectively reduce most of the nonhydrostatic errors in terms of system bias,anomaly root-mean-square error,and the error of the wave spatial pattern,which proves the feasibility and superiority of this scheme.

Distribution of Atmospheric Pollution in Southern Benin

This study focuses on air quality in southern Benin in order to show public authorities what the Beninese populations are exposed to for efficient decision-making. Two sampling campaigns were carried out, one in the wet period and the other in the dry season. The measurements were taken using a monitor called an “Air Quality Monitor”. For data processing, the multiple comparison methods of Dun (1961) and the Wilcoxon test were used. To maintain legitimacy, all spatial data were included in the official cartographic repository of Benin: WGS 1984, Transverse Mercator Universe Projection (UTM), Zone 31 North. The Moran statistic was used to measure the levels of spatial autocorrelation of the variables studied and to test the significance. In order to locate the spatial subsets, the local spatial association indices of Anselin Local Moran and Getis-Ord, Gi* were used. In terms of results, on the 13 monitoring sites and the 8 parameters chosen to determine air quality, we do not note any significant inter-seasonal difference. Of the eight parameters, only three parameters present spatial autocorrelation leading to predictions of ambient air quality over the entire study area based on the distance separating the points, namely, PM2.5, PM10 and ambient air quality index (AQI). The localities affected by atmospheric pollution in South Benin are located in the south-western part of Benin, headed by Cotonou, which is heavily polluted by CO2, TCOV, PM10 and PM2.5.

The corrosion characteristic and mechanism of Mg-5Y-1.5Nd-xZn-0.5Zr(x=0,2,4,6 wt.%)alloys in marine atmospheric environment

The microstructure and precipitated phases of as-cast Mg-5Y-1.5Nd-x Zn-0.5Zr(x=0,2,4,6 wt.%)alloys were investigated by optical microscopy,scanning electron microscopy,energy-dispersive spectrometry and X-ray Diffraction.The exposure corrosion experiment of these magnesium alloys was tested in South China Sea and KEXUE vessel atmospheric environment.The corrosion characteristic and mechanism of magnesium alloys of Mg-5Y-1.5Nd-x Zn-0.5Zr(x=0,2,4,6 wt.%)alloys were analyzed by weight loss rate,corrosion depth,corrosion products and corrosion morphologies.The electrochemical corrosion tests were also measured in the natural seawater.The comprehensive results showed that Mg-5Y-1.5Nd-4Zn-0.5Zr magnesium alloy existed the best corrosion resistance whether in the marine atmospheric environment and natural seawater environment.That depended on the microstructure,type and distribution of precipitated phases in Mg-5Y-1.5Nd-4Zn-0.5Zr magnesium alloy.Sufficient quantity anodic precipitated phases in the microstructure of Mg-5Y-1.5Nd-4Zn-0.5Zr alloy played the key role in the corrosion resistance.

A novel encoding mechanism for particle physics

This study proposes a novel particle encoding mechanism that seamlessly incorporates the quantum properties of particles,with a specific emphasis on constituent quarks.The primary objective of this mechanism is to facilitate the digital registration and identification of a wide range of particle information.Its design ensures easy integration with different event generators and digital simulations commonly used in high-energy experiments.Moreover,this innovative framework can be easily expanded to encode complex multi-quark states comprising up to nine valence quarks and accommodating an angular momentum of up to 99/2.This versatility and scalability make it a valuable tool.

Implications of 1.5 K climate warming on warm-season ozone exposure and atmospheric oxidation capacity in China

Surface ozone(O_(3))poses significant threats to public health,agricultural crops,and plants in natural ecosystems.Global warming is likely to increase future O_(3)mainly by altering atmospheric photochemical reactions and enhancing biogenic volatile organic compound(BVOC)emissions.To assess the impacts of the future 1.5 K climate target on O_(3)concentrations and ecological O_(3)exposure in China,numerical simulations were conducted using the CMAQ(Community Multiscale Air Quality)model during April-October 2018.Ecological O_(3)exposure was estimated using six indices(i.e.,M7,M24,N100,SUM60,W126,and AOT40f).The results show that the temperature rise increases the MDA8 O_(3)(maximum daily eight-hour average O_(3))concentrations by∼3 ppb and the number of O_(3)exceedance days by 10-20 days in the North China Plain(NCP),Yangtze River Delta(YRD),and Sichuan Basin(SCB)regions.All O_(3)exposure indices show substantial increases.M24 and M7 in eastern and southern China will rise by 1-3 ppb and 2-4 ppb,respectively.N100 increases by more than 120 h in the surrounding regions of Beijing.SUM60 increases by greater than 9 ppm h^(−1),W126 increases by greater than 15 ppm h^(−1)in Shaanxi and SCB,and AOT40f increases by 6 ppm h^(−1)in NCP and SCB.The temperature increase also promotes atmospheric oxidation capacity(AOC)levels,with the higher AOC contributed by OH radicals in southern China but by NO_(3)radicals in northern China.The change in the reaction rate caused by the temperature increase has a greater influence on O_(3)exposure and AOC than the change in BVOC emissions.

Assessing the corrosion protection property of coatings loaded with corrosion inhibitors using the real-time atmospheric corrosion monitoring technique

The atmospheric corrosion monitoring(ACM)technique has been widely employed to track the real-time corrosion behavior of metal materials.However,limited studies have applied ACM to the corrosion protection properties of organic coatings.This study compared a bare epoxy coating with one containing zinc phosphate corrosion inhibitors,both applied on ACM sensors,to observe their corrosion protection properties over time.Coatings with artificial damage via scratches were exposed to immersion and alternating dry and wet environments,which allowed for monitoring galvanic corrosion currents in real-time.Throughout the corrosion tests,the ACM currents of the zinc phosphate/epoxy coating were considerably lower than those of the blank epoxy coating.The trend in ACM current variations closely matched the results obtained from regular electrochemical tests and surface analysis.This alignment highlights the potential of the ACM technique in evaluating the corrosion protection capabilities of organic coatings.Compared with the blank epoxy coating,the zinc phosphate/epoxy coating showed much-decreased ACM current values that confirmed the effective inhibition of zinc phosphate against steel corrosion beneath the damaged coating.

Discharge and mass transfer characteristics of atmospheric pressure gas-solid two-phase gliding arc

In this work,a gas-solid two-phase gliding arc discharge(GS-GAD)reactor was built.Gliding arc was formed in the gap between the blade electrodes,and solid powder was deposited on the sieve plate positioned beneath the blade electrodes.A range of experimental parameters,including the inter-electrode spacing,gas flow rate,applied voltage,and the type of the powder,were systematically varied to elucidate the influence of solid powder matter on the dynamics of gliding arc discharge(GAD).The discharge images were captured by ICCD and digital camera to investigate the mass transfer characteristics of GS-GAD,and the electrical parameters,such as the effective values of voltage,current,and discharge power were record to reveal the discharge characteristics of GS-GAD.The results demonstrate that powder undergoes spontaneous movement towards the upper region of the gliding arc due to the influence of electric field force.Increasing the discharge voltage,decreasing relative dielectric constant of the powder and reducing the electrode-to-sieve-plate distance all contribute to a greater involvement of powder in the GAD process,subsequently resulting in an enhanced powder concentration within the GAD region.Additionally,powder located beneath the gliding arc experiences downward resistance caused by the opposing gas flow and arc.Excessive gas flow rate notably hampers the powder concentration within the discharge region,and the velocity of powder motion in the upper part of the GAD region is reduced.Under the condition of electrode-to-sieve-plate distance of 30 mm,gas flow rate of 1.5 L/min,and peak-to-peak voltage of 31 kV,the best combination of arc gliding and powder spark discharge phenomena can be achieved with the addition of Al_(2)O_(3) powder.

Characteristics of Lightning Activity in Southeast China and its Relation to the Atmospheric Background

Based on the lightning observation data from the Fengyun-4A(FY-4A)Lightning Mapping Imager(FY-4A/LMI)and the Lightning Imaging Sensor(LIS)on the International Space Station(ISS),we extract the“event”type data as the lightning detection results.These observations are then compared with the cloud-to-ground(CG)lightning observation data from the China Meteorological Administration.This study focuses on the characteristics of lightning activity in Southeast China,primarily in Jiangxi Province and its adjacent areas,from April to September,2017–2022.In addition,with the fifth-generation European Centre for Medium-Range Weather Forecasts reanalysis data,we further delved into the potential factors influencing the distribution and variations in lightning activity and their primary related factors.Our findings indicate that the lightning frequency and density of the FY-4A/LMI,ISS-LIS and CG data are higher in southern and central Jiangxi,central Fujian Province,and western and central Guangdong Province,while they tend to be lower in eastern Hunan Province.In general,the high-value areas of lightning density for the FY-4A/LMI are located in inland mountainous areas.The lower the latitude is,the higher the CG lightning density is.High-value areas of the CG lightning density are more likely to be located in eastern Fujian and southeastern Zhejiang Province.However,the high-value areas of lightning density for the ISS-LIS are more dispersed,with a scattered distribution in inland mountainous areas and along the coast of eastern Fujian.Thus,the mountainous terrain is closely related to the high-value areas of the lightning density.The locations of the high-value areas of the lightning density for the FY-4A/LMI correspond well with those for the CG observations,and the seasonal variations are also consistent.In contrast,the distribution of the high-value areas of the lightning density for the ISS-LIS is more dispersed.The positions of the peak frequency of the FY-4A/LMI lightning and CG lightning contrast with local altitudes,primarily located at lower altitudes or near mountainsides.K-index and convective available potential energy(CAPE)can better reflect the local boundary layer conditions,where the lightning density is higher and lightning seasonal variations are apparent.There are strong correlations in the annual variations between the dew-point temperature(Td)and CG lightning frequency,and the monthly variations of the dew-point temperature and CAPE are also strongly correlated with monthly variations of CG lightning,while they are weakly correlated with the lightning frequency for the FY-4A/LMI and ISS-LIS.This result reflects that the CAPE shows a remarkable effect on the CG lightning frequency during seasonal transitions.

The state-of-the-art of atmospheric pressure plasma for transdermal drug delivery

Plasma-enhanced transdermal drug delivery(TDD) presents advantages over traditional methods,including painless application, minimal skin damage, and rapid recovery of permeability. To harness its clinical potential, factors related to plasma’s unique properties, such as reactive species and electric fields, must be carefully considered.This review provides a concise summary of conventional TDD methods and subsequently offers a comprehensive examination of the current state-of-the-art in plasma-enhanced TDD. This includes an analysis of the impact of plasma on HaCaT human keratinocyte cells, ex vivo/in vivo studies, and clinical research on plasma-assisted TDD. Moreover, the review explores the effects of plasma on skin physical characteristics such as microhole formation, transepidermal water loss(TEWL), molecular structure of the stratum corneum(SC), and skin resistance. Additionally, it discusses the involvement of various reactive agents in plasma-enhanced TDD, encompassing electric fields,charged particles, UV/VUV radiation, heat, and reactive species. Lastly, the review briefly addresses the temporal behavior of the skin after plasma treatment, safety considerations, and potential risks associated with plasma-enhanced TDD.

Influence of bias voltage and oxygen addition on the discharge aspects of a diffuse argon plume in an atmospheric pressure plasma jet

A remote plasma,also referred to as a plasma plume(diffuse or filamentary),is normally formed downstream of an atmospheric pressure plasma jet.In this study,a diffuse plume is formed by increasing the bias voltage(U_(b))applied to the downstream electrode of an argon plasma jet excited by a negatively pulsed voltage.The results indicate that the plume is filamentary when U_(b)is low,which transits to the diffuse plume with increasing U_(b).The discharge initiated at the rising edge of the pulsed voltage is attributed to the diffuse plume,while that at the falling edge contributes to the filament in the plume.For the diffuse plume,the discharge intensity decreases with the increasing oxygen content(C_o).Fast photography reveals that the diffuse plume results from a negative streamer,which has a dark region near the nozzle with C_o=0%.However,the dark region is absent with C_o=0.5%.From the optical emission spectrum,the electron density,electron excitation temperature,gas temperature,and oxygen atom concentration are investigated.

CAS-ESM2.0 Successfully Reproduces Historical Atmospheric CO_(2) in a Coupled Carbon−Climate Simulation

The atmospheric carbon dioxide(CO_(2))concentration has been increasing rapidly since the Industrial Revolution,which has led to unequivocal global warming and crucial environmental change.It is extremely important to investigate the interactions among atmospheric CO_(2),the physical climate system,and the carbon cycle of the underlying surface for a better understanding of the Earth system.Earth system models are widely used to investigate these interactions via coupled carbon-climate simulations.The Chinese Academy of Sciences Earth System Model version 2(CAS-ESM2.0)has successfully fixed a two-way coupling of atmospheric CO_(2)with the climate and carbon cycle on land and in the ocean.Using CAS-ESM2.0,we conducted a coupled carbon-climate simulation by following the CMIP6 proposal of a historical emissions-driven experiment.This paper examines the modeled CO_(2)by comparison with observed CO_(2)at the sites of Mauna Loa and Barrow,and the Greenhouse Gases Observing Satellite(GOSAT)CO_(2)product.The results showed that CAS-ESM2.0 agrees very well with observations in reproducing the increasing trend of annual CO_(2)during the period 1850-2014,and in capturing the seasonal cycle of CO_(2)at the two baseline sites,as well as over northern high latitudes.These agreements illustrate a good ability of CAS-ESM2.0 in simulating carbon-climate interactions,even though uncertainties remain in the processes involved.This paper reports an important stage of the development of CAS-ESM with the coupling of carbon and climate,which will provide significant scientific support for climate research and China’s goal of carbon neutrality.

Exploring device physics of perovskite solar cell via machine learning with limited samples

Perovskite solar cells(PsCs)have developed tremendously over the past decade.However,the key factors influencing the power conversion efficiency(PCE)of PSCs remain incompletely understood,due to the complexity and coupling of these structural and compositional parameters.In this research,we demon-strate an effective approach to optimize PSCs performance via machine learning(ML).To address chal-lenges posed by limited samples,we propose a feature mask(FM)method,which augments training samples through feature transformation rather than synthetic data.Using this approach,squeeze-and-excitation residual network(SEResNet)model achieves an accuracy with a root-mean-square-error(RMSE)of 0.833%and a Pearson's correlation coefficient(r)of 0.980.Furthermore,we employ the permu-tation importance(PI)algorithm to investigate key features for PCE.Subsequently,we predict PCE through high-throughput screenings,in which we study the relationship between PCE and chemical com-positions.After that,we conduct experiments to validate the consistency between predicted results by ML and experimental results.In this work,ML demonstrates the capability to predict device performance,extract key parameters from complex systems,and accelerate the transition from laboratory findings to commercialapplications.