Number Of Experimental Data Points Research Articles

Experiment‐Oriented Machine Learning of Polymer:Non‐Fullerene Organic Solar Cells

AbstractDespite the capacity of conjugated materials for enhanced power conversion efficiency (PCE) of organic photovoltaics (OPV), a comprehensive survey of unexplored materials is beyond the reach of most researchers’ resources. In such instances, a data‐driven approach using machine learning (ML) is an efficient alternative; however, bridging the gap between experimental observations and data science requires a number of refinements. In this investigation, using a random forest model based on an experimental dataset, a high correlation coefficient of 0.85 is achieved for the ML of polymer and non‐fullerene small molecule acceptor OPVs and performed virtual screening of 200,932 conjugated polymers generated by the combinatorial coupling of donor and acceptor units. Further, to evaluate the effectiveness of the ML model, a series of conjugated polymers (based on benzodithiophene and thiazolothiazole) were designed, synthesized, and characterized with different alkyl chains. Among these, PBDTTzEH:IT‐4F showed a PCE of 10.10%, which is in good correspondence with ML predictions with respect to the choice of alkyl chains. Thus, the current study demonstrates how ML can be utilized for developing OPVs using a relatively small number of experimental data points (566) and screening numerous molecular structures.

Predicting the solubility, thermodynamic properties and preferential solvation of sulphamethazine in {acetonitrile + water} mixtures using a minimum number of experimental data points

ABSTRACT The all experimental solubility data of sulphamethazine in {acetonitrile (1) + water (2)} binary solvent mixtures at temperature range from (278.15 to 323.15) K have been carefully reanalysed. Then, a minimum number of experimental solubility data has been chosen to predict the solubility data at all possible solvent compositions and temperatures using interpolation technique. The predicted data was compared with the experimental data employing the average absolute percentage deviation (AAPD) as an accuracy criterion. The preferential solvation analyses based on the inverse Kirkwood-Buff integrals and also apparent thermodynamic properties were conducted employing the simulated data and the obtained results were compared with those obtained employing all data points. Gibbs energy and preferential solvation follow similar trends but enthalpy and entropy of dissolution exhibit significant differences.

Review of the cosolvency models for predicting solubility in solvent mixtures: An update.

The cosolvency models frequently used in solubility data modeling of drugs in mixed solvents were reviewed and their accuracies for calculating the solubility of solutes were briefly discussed. The models could be used either for correlation of the generated solubility data with temperature, solvent composition etc or for prediction of unmeasured solubility data using interpolation/extrapolation technique. Concerning the correlation results employing a given number of independent variables, the accuracies of the investigated models were comparable, since they could be converted to a single mathematical form, however, the accuracies were decreased when models emplyed more independent variables. The accurate correlative models could be employed for prediction purpose and/or screening the experimental solubility data to detect possible outliers. With regard to prediction results, the best predictions were made using the cosolvency models trained by a minimum number of experimental data points and an ab initio accurate prediction is not possible so far and further mathematical efforts are needed to provide such a tool. To connect this gap between available accurate correlative models with the ab initio predictive model, the generally trained models for calculating the solubility of various drugs in different binary mixtures, various drugs in a given binary solvent and also a given drug in various binary solvents at isothermal condition and/or different temperatures were reported. Available accuracy criteria used in the recent publications were reviewed including mean percentage deviation (MPD). The MPD for correlative models is 1-10% whereas the corresponding range for predictive models is 10-80% depend on the model capability and the number of independent variables employed by the model. This is an update for a review article published in this journal in 2008.

Experimental study on the removal of FCCS catalyst particles by electrostatic separation

ABSTRACT Electrostatic separation is a technology for removing catalyst particles from fluid catalytic cracking slurry (FCCS). In this study, the particle trajectory was simulated by COMSOL Multiphysics based on theories and deductions. To verify whether these theories are reasonable for an electrostatic separation simulation, a mixture of alkylbenzene and catalysts was used to characterize the FCCS, and a cold model experiment was carried out, Moreover, the effects of the raw material properties, operating conditions and material properties on the efficiency of the separation were investigated. The results showed that the separation efficiency increased with decreasing glass sphere diameter because the porosity decreased. With the increase in the size of the catalyst particles, the dielectrophoretic (DEP) force increased rapidly and played a more important role, and the separation efficiency increased. Increasing the separation time helped adsorb the catalyst particles that were far from the contact point of the glass ball, and some particles in these areas were unable to be adsorbed. The electrostatic separator has the optimal voltage, and this value decreased with decreasing glass sphere diameter. The experimental results were consistent with the theoretical analysis. Through the nonlinear fitting of a large number of experimental data points, the prediction model of the device was obtained.

Solubility Prediction of Drugs in Binary Solvent Mixtures at Various Temperatures Using a Minimum Number of Experimental Data Points.

This study aimed to provide a rational experimental design to collect a minimum number of experimental data points for a drug dissolved in a given binary solvent mixture at various temperatures, and to describe a computational procedure to predict the solubility of the drugs in any solvent composition and temperature of interest. We gathered available solubility data sets from papers published from 2012 to 2016 (56 data sets, 3488 data points totally). The mean percentage deviations (MPD) used to check the accuracy of predictions was calculated by Eq. 10. Fifty-six datasets were analyzed using 8 training data points which the overall MPD was calculated to be 15.5% ± 15.1%, and for 52 datasets after excluding 5 outlier sets was 12.1% ± 8.9%. The paired t test was conducted to compare the MPD values obtained from the models trained by 7 and 8 training data points and the reduction in prediction overall MPD (from 17.7% to 15.5%) was statistically significant (p < 0.04). To further reduction in MPD values, the computations were also conducted using 9 training data points, which did not reveal any significant difference comparing to the predictions using 8 training data points (p > 0.88). This observation revealed that the model adequately trained using 8 data points and could be used as a practical strategy for predicting the solubility of drugs in binary solvent mixtures at various temperatures with acceptable prediction error and using minimum experimental efforts. These sorts of predictions are highly in demand in the pharmaceutical industry.

Synchrotron X-ray diffraction based method for stress intensity factor evaluation in the bulk of materials

A generalised approach for determining the stress intensity factor in the bulk of non-transparent materials is presented. The new approach combines experimental elastic strain data, measured with powerful synchrotron X-ray diffraction, with an elastic model based on Williams’ series development. The stress intensity factor is calculated using a multi-point over-deterministic method where the number of experimental data points is higher than the number of unknowns describing the elastic field surrounding the crack-tip. The tool is tested on X-ray strain measurements collected on a bainitic steel. In contrast to surface techniques the approach provides insights into the crack tip mechanics deep within the sample and makes full use of the plastic zone data. Satisfactory results are obtained with 1 term in the series development and a smaller region of interest. By increasing the number of terms, it is possible to improve the accuracy in the SIF predictions but this requires a larger region of interest.

The ENBIS‐17 Quality and Reliability Engineering International Special Issue

The ENBIS‐17 Quality and Reliability Engineering International Special Issue

Examining potential reductions in wind tunnel testing requirements

AbstractWind tunnel test campaigns involve a large number of experimental data points, can take a long time to accomplish, and can consume tremendous resources. Design of experiments is a systematic approach to experimental design and analysis that has the potential to improve the efficiency and effectiveness of wind tunnel testing. We use a legacy wind tunnel test campaign and compare the resulting data from that campaign to data generated using smaller experimental design strategies. The comparison is accomplished using a Monte Carlo sampling methodology coupled with a statistical comparison of the estimated surfaces. Initial results from a 2‐phase experiment suggest a tremendous opportunity to reduce wind tunnel test efforts without losing test information and provide suggestions on the types of designs to use to best realize the test matrix reductions.

Statistical investigation of the full concentration range of fasted and fed simulated intestinal fluid on the equilibrium solubility of oral drugs

Upon oral administration the solubility of a drug in intestinal fluid is a key property influencing bioavailability. It is also recognised that simple aqueous solubility does not reflect intestinal solubility and to optimise in vitro investigations simulated intestinal media systems have been developed. Simulated intestinal media which can mimic either the fasted or fed state consists of multiple components each of which either singly or in combination may influence drug solubility, a property that can be investigated by a statistical design of experiment technique. In this study a design of experiment covering the full range from the lower limit of fasted to the upper limit of fed parameters and using a small number of experiments has been performed. The measured equilibrium solubility values are comparable with literature values for simulated fasted and fed intestinal fluids as well as human fasted and fed intestinal fluids. The equilibrium solubility data range is statistically equivalent to a combination of published fasted and fed design of experiment data in six (indomethacin, phenytoin, zafirlukast, carvedilol, fenofibrate and probucol) drugs with three (aprepitant, tadalafil and felodipine) drugs not equivalent. In addition the measured equilibrium solubility data sets were not normally distributed. Further studies will be required to determine the reasons for these results however it implies that a single solubility measurement without knowledge of the solubility distribution will be of limited value. The statistically significant media factors which promote equilibrium solubility (pH, sodium oleate and bile salt) were in agreement with published results but the number of determined significant factors and factor interactions was fewer in this study, lecithin for example did not influence solubility. This may be due to the reduction in statistical sensitivity from the lower number of experimental data points or the fact that using the full range will examine media parameters ratios that are not biorelevant. Overall the approach will provide an estimate of the solubility range and the most important media factors but will not be equivalent to larger scale focussed studies. Further investigations will be required to determine why some drugs do not produce equivalent DoE solubility distributions, for example combined fasted and fed DoE, but this simply may be due to the complexity and individuality of the interactions between a drug and the media components.

Improving the estimation of complete field soil water characteristic curves through field monitoring data

In this work, Soil Water Characteristic Curves (SWCCs) were reconstructed through simultaneous field measurements of soil pore water pressure and water content. The objective was to evaluate whether field-based monitoring can allow for the improvement of the accuracy in SWCCs estimation with respect to the use of laboratory techniques. Moreover, field assessment of SWCCs allowed to: a) quantify the hydrological hysteresis affecting SWCCs through field data; b) analyze the effect of different temporal resolution of field measures; c) highlight the differences in SWCCs reconstructed for a particular soil during different hydrological years; d) evaluate the reliability of field reconstructed SWCCs, by the comparison between assessed and measured trends of a component of the soil water balance. These aspects were fundamental for assessing the reliability of the field reconstructed SWCCs. Field data at two Italian test-sites were measured. These test-sites were used to evaluate the goodness of field reconstructed SWCCs for soils characterized by different geomorphological, geological, physical and pedological features. Field measured or laboratory measured SWCCs data of 5 soil horizons (3 in a predominantly silty soil, 2 in a predominantly clayey one) were fitted by Van Genuchten model. Different field drying and wetting periods were identified, based on monthly meteorological conditions, in terms of rainfall and evapotranspiration amounts, of different cycles. This method allowed for a correct discrimination of the main drying and the main wetting paths from field data related and for a more reliable quantification of soil hydrological properties with respect to laboratory methodologies. Particular patterns of changes in SWCCs forms along depth could be also identified. Field SWCCs estimation is not affected by the temporal resolution of the acquisition (hours or days), as testified by similar values of Van Genuchten equation fitting parameters. Instead, hourly data may offer a clearer vision of the drying and wetting paths, due to the highest number of experimental data points. Moreover, in temperate climate situations as those of the test-sites, main drying curves and main wetting curves of a particular soil were substantially similar also for different hydrological cycles with peculiar meteorological conditions. SWCCs parameters were implemented in a numerical code (HYDRUS-1D) to simulate soil water storage for different soil horizons. Field reconstructed SWCCs allowed for simulating with a higher precision these trends, confirming the reliability of the reconstructed field curves by a quantitative point of view. Moreover, best results were obtained considering hysteresis in the modeling.

Modeling of Asphaltene Onset Precipitation Conditions with Cubic Plus Association (CPA) and Perturbed Chain Statistical Associating Fluid Theory (PC-SAFT) Equations of State

Asphaltene precipitation has been one of the major problems in the oil industry, and its modeling is still believed to be a quite complex issue due to the different characteristics of thousands of heavy components in crude oil. There have been several attempts to model asphaltene precipitation using various equations of state and empirical models. In the past few years, association models based on CPA and SAFT equations of state have been found to be promising models for studies of asphaltene precipitation. In this work, we compare asphaltene precipitation results obtained from different modeling approaches based on CPA, PC-SAFT with association (PC-SAFT (WA)), and PC-SAFT without association (PC-SAFT (WOA)) models. While the modeling approaches for the CPA and PC-SAFT (WOA) have been described before in various literature, the modeling approach for PC-SAFT (WA) is proposed in this work. All three models require the same number of experimental data points (at least three upper onset pressures and one bubb...

Shock tube and modeling study of 2,7-dimethyloctane pyrolysis and oxidation

High molecular weight iso-paraffinic molecules are found in conventional petroleum, Fischer–Tropsch (FT), and other alternative hydrocarbon fuels, yet fundamental combustion studies on this class of compounds are lacking. In the present work, ignition delay time measurements in 2,7-dimethyloctane/air were carried out behind reflected shock waves using conventional and constrained reaction volume (CRV) methods. The ignition delay time measurements covered the temperature range 666–1216K, pressure range 12–27atm, and equivalence ratio of 0.5 and 1. The ignition delay time temperatures span the low-, intermediate- and high-temperature regimes for 2,7-dimethyloctane (2,7-DMO) oxidation. Clear evidence of negative temperature coefficient behavior was observed near 800K. Fuel time-history measurements were also carried out in pyrolysis experiments in mixtures of 2000ppm 2,7-DMO/argon at pressures near 16 and 35atm, and in the temperature range of 1126–1455K. Based on the fuel removal rates, the overall 2,7-DMO decomposition rate constant can be represented with k=4.47×105 exp(−23.4[kcal/mol]/RT) [1/s]. Ethylene time-history measurements in pyrolysis experiments at 16atm are also provided. The current shock tube dataset was simulated using a novel chemical kinetic model for 2,7-DMO. The reaction mechanism includes comprehensive low- and high-temperature reaction classes with rate constants assigned using established rules. Comparisons between the simulated and experimental data show simulations reproduce the qualitative trends across the entire range of conditions tested. However, the present kinetic modeling simulations cannot quantitatively reproduce a number of experimental data points, and these are analyzed herein.

New modified UNIFAC parameters using critically evaluated phase equilibrium data

New modified UNIFAC property prediction model parameters are reported for 89 main groups and 984 group–group interactions using critically evaluated phase equilibrium data including vapor–liquid equilibrium (VLE), liquid–liquid equilibrium (LLE), solid–liquid equilibrium (SLE), excess enthalpy (HE), infinite dilution activity coefficient (AINF) and excess heat capacity (CPE) data. The new algorithmic framework for quality assessment of phase equilibrium data is applied for qualifying the consistency of data and screening out possible erroneous data. Substantial improvement over previous versions of UNIFAC is observed due to inclusion of experimental data from recent publications and proper weighting based on a quality assessment procedure. The systems requiring further verification of phase equilibrium data were identified where insufficient number of experimental data points is available or where existing data are conflicting.

Fitting approach to liquid–liquid displacement porosimetry based on the log-normal pore size distribution

Liquid–liquid displacement porosimetry (LLDP) has been used to characterize several UF membranes in a wide range of molecular weight cut-offs (MWCO). A new method to convert porosimetric data into pore size distributions and related information has been developed based on assuming log-normal pore size distributions. The results of this are in good agreement with those from the customary data conversion algorithm (as derived by Grabar and Nikitine). The proposed method can also be used when a reduced number of experimental data points is available, leading to a significant reduction of data acquisition time needed to complete a reliable analysis.

Adequate Treatment of Correlated Experimental Data in Nuclear Data Evaluations Avoiding Peelle's Pertinent Puzzle

Peelle's Pertinent Puzzle denotes the occurrence of unexpected mean values and variances when experimental data affected by statistical and systematic uncertainties are combined by weighing with an experimental covariance matrix. It was shown for an arbitrary number of experimental data points for the same physical quantity that this puzzle is primarily caused by an improper estimate of the experimental covariance matrix. In this contribution, the solution is tested for more than one physical quantity and is applied to a nuclear data problem.

Keynote Lecture: Residual Stress Gradient Analysis by Multiple Diffraction Line Methods

The paper deals with methods for X-ray stress analysis (XSA), which allow for the evaluation of near surface in-plane residual stress gradients σ||(τ) and σ||(z) in the LAPLACE- and the real space, respectively. Since the ‘robustness’ of residual stress gradient analysis strongly depends on both, the quality of the measured strain data and the number of experimental data points, the discussion aims at those approaches which are based on processing various diffraction lines or even complete diffraction patterns. It is shown that these techniques, which were originally developed for angle-dispersive (AD) diffraction, can be adapted and enhanced for energy-dispersive (ED) diffraction employing high-energy synchrotron radiation. With the example of a shot-peened ferritic steel it is demonstrated, that sin²ψ-data measured in the Ψ-mode of XSA employing the ED diffraction technique can be analyzed on different levels of approximation.

The extended TILAR approach: a novel tool for dynamic modeling of the transcription factor network regulating the adaption to in vitro cultivation of murine hepatocytes

BackgroundNetwork inference is an important tool to reveal the underlying interactions of biological systems. In the liver, a complex system of transcription factors is active to distribute signals and induce the cellular response following extracellular stimuli. Plenty of information is available about single transcription factors important for the different functions of the liver, but little is known about their causal relations to each other.ResultsGiven a DNA microarray time series dataset of collagen monolayers cultured murine hepatocytes, we identified 22 differentially expressed genes for which the corresponding protein is known to exhibit transcription factor activity. We developed the Extended TILAR (ExTILAR) network inference algorithm based on the modeling concept of the previously published TILAR algorithm. Using ExTILAR, we inferred a transcription factor network based on gene expression data which puts these important genes into a functional context. This way, we identified a previously unknown relationship between Tgif1 and Atf3 which we validated experimentally. Beside its known role in metabolic processes, this extends the knowledge about Tgif1 in hepatocytes towards a possible influence of processes such as proliferation and cell cycle. Moreover, two positive (i.e. double negative) regulatory loops were predicted that could give rise to bistable behavior. We further evaluated the performance of ExTILAR by systematic inference of an in silico network.ConclusionsWe present the ExTILAR algorithm, which combines the advantages of the regression based inference algorithm TILAR, like large network sizes processable and low computational costs, with the advantages of dynamic network models based on ordinary differential equation (i.e. in silico knock-down simulations). Like TILAR, ExTILAR makes use of various prior-knowledge types such as transcription factor binding site information and gene interaction knowledge to infer biologically meaningful gene regulatory networks. Therefore, ExTILAR is especially useful when a large number of genes is modeled using a small number of experimental data points.

Modeling Methodology Based on Stacked Neural Networks for Inferential Prediction of Polypropylene Melt Index

A modeling methodology based on stacked neural networks by combining several individual networks in parallel is proposed. Stacked neural network as an effective method for modeling of inherently complex and nonlinear systems especially a system with a limited number of experimental data points is chosen for yield prediction. Proper determination of the stacking weights is essential for good stacked neural networks model performance, so determination of appropriate weights for combining individual networks using robust least squares estimation is proposed. Inferential prediction of melt index as the most important characteristic process of polypropylene polymerization has been carried out. The application of the proposed modeling method based on stacked neural networks to the development of melt index soft sensor in an industrial propylene polymerization plant demonstrates its effectiveness.

Examining improved experimental designs for wind tunnel testing using Monte Carlo sampling methods

Wind tunnels are used in the design and testing of a wide variety of systems and products. Wind tunnel test campaigns involve a large number of experimental data points, can take a long time to accomplish, and can consume tremendous resources. Design of Experiments is a systematic, statistically based approach to experimental design and analysis that has the potential to improve the efficiency and effectiveness of wind tunnel testing. In Defense Acquisition, wind tunnel testing of aircraft systems may require years of effort to fully characterize the system of interest. We employ data from a fairly large legacy wind tunnel test campaign and compare that data's corresponding response surface to the response surfaces derived from data generated using smaller, statistically motivated experimental design strategies. The comparison is accomplished using a Monte Carlo sampling methodology coupled with a statistical comparison of the system's estimated response surfaces. Initial results suggest a tremendous opportunity to reduce wind tunnel test efforts without losing test information. Published in 2010 by John Wiley &amp; Sons, Ltd.

Finite Element Model Tuning with Spatially-Dense 3D Modes

An investigation into varying the spatial density of three-dimensional (3D) modes in order to accurately tune a finite element (FE) model to larger numbers of modes is conducted. This FE model tuning approach is evaluated on a 60,000+ FE degree of freedom (DOF) model of the United States Air Force Academy’s fifth small satellite, FalconSAT-5 (FS-5), structural engineering model (SEM). The evaluation of this FE model tuning approach starts with the collection of natural frequencies and 3D modes extracted from scanning laser Doppler velocimeter (LDV) frequency response function measurements at 2,165 closely-spaced points on the surface of the test article. The measured modes and their associated natural frequencies serve as target values in a gradient-based tuning approach. The FE model is tuned to have differences with the measured natural frequencies less than 2% and in many cases cross-orthogonality values greater than 0.90. Using both a QR-decomposition and the Triaxial Effective Independence (EfI3 + ) sensor selection strategies, the effect of adding more tuning points on tuning accuracy is studied. This study demonstrates that increasing the number of experimental data points using the EfI3 + sensor selection strategy for FE model optimization results in increased tuning accuracy.