Output Variables Of Interest Research Articles

Development of a novel silk sericin-based hydrogel film by mixture design

Sericin has been used in functional and potentially biodegradable materials for cosmetics, biomedical, agricultural, and food applications. It is a natural polymer with applications in absorbent materials, such as hydrogels, because of its hydrophilic character. However, sericin by itself is brittle, and in contact with water has low structural stability, being necessary its blending with other polymers or the application of crosslinking processes. In this work, hydrogel films were prepared from different mixtures containing sericin (SS), carboxymethylcellulose (CMC), and polyvinyl alcohol (PVA), using a simple and environmentally friendly method consisting of a gelling process followed by solvent casting. A mixture design was applied to assess the incidence of each component and its interaction with the output variables of interest. Two response variables were evaluated in each formulation: water absorption capacity (WA) and gel fraction (GF). It was also possible to model the output variables based on the proportions of the sample components. In addition, a set of formulations were used to produce hydrogels with high water absorption rates while maintaining their structural stability. The optimal hydrogel formulation (HF) was structurally and thermally characterized by FTIR and TGA, respectively. Hydrogel morphology was also studied by scanning electron microscopy (SEM). The results of this study constitute an important contribution to the design of novel processing routes to extend the use of silk sericin in the development of new materials.

72 Developing Machine Learning Models When Data is Limiting

Abstract Animal scientists have become more enthusiastic about developing machine learning (ML) models to improve the predictability of variables of interest. However, adequate data is limited either because ways to collect the data are scarce or because the process is expensive, time- or labor-consuming, or it simply takes too long to be collected. In addition to the usual hurdles in developing ML models, e.g., appropriate technique, the number of layers, and their activation, enough high-quality data is an essential requirement for ML that is often neglected. Should the data come from longitudinal or cross-sectional experiments of how many subjects? When enough high-quality, reliable data is limiting, one alternative is to create synthetic datasets that reflect the correlation among input and output variables of interest. The probability distribution for each variable needs to be well defined, and the correlation among variables must be taken into account. The normal distribution may not always be a reasonable assumption for all variables; thus, variable-specific distributions must be used with their appropriate parameters. An adequate range (min and max) must be provided for each variable to represent it. Shortcomings might occur when nonlinear relationships occur between variables. The synthetic dataset will also fail to provide good predictability when the new inputs do not have similar correlations, as did the variables used to build the synthetic dataset. It is common to standardize or normalize each variable during ML development. Each variable is normalized based on its minimum and maximum values. A common mistake during the ML development process is that training and evaluation subsets are normalized independently when both should be normalized using the range of the complete dataset; otherwise, normalization becomes dependent on the subset, and the ML weights will differ from one epoch to another. This might weaken the ML predictability because the correct range for back-transforming the output to its original value is unknown.

Numerical investigation and sensitivity analysis on bioconvective tangent hyperbolic nanofluid flow towards stretching surface by response surface methodology

In a suspension of tangent hyperbolic bionanofluid keeping both nanoparticles and motile microorganisms, the thermobioconvective boundary layer flow was studied through an exponentially stretching surface utilizing response surface methodology (RSM). The constructed model of a tangent hyperbolic nanofluid in boundary layer flow is studied with implications of thermophoresis and Brownian motion. Condition of zero normal flux of nanomaterials is added at the surface to scatter the nanomaterials from the plate surface. The rate of heat transfer is analyzed using convective boundary condition. Numerical shooting strategy with Runge-Kutta scheme is to follow intently behind the similarity transformation to solve the system of governing equations. It is assumed that the output variables of interest are dependent on the governing input parameters. The sensitivity analysis is additionally introduced. It is discovered that the sensitivity of local Nusselt number increments by expanding Lewis and thermophoresis number while the highest non-dimensional Nusselt number appears close to the significant level for the thermophoresis and low level for the Brownian motion variable. Additionally, it is demonstrated that the average maximum mean thickness of motile microorganism appears at the highest level of Brownian motion and thermophoresis number and thermophoresis and Lewis numbers. The results would provide initial guidance for potential manufacture of devices.

Optimization of the formulation of an original hydrogel-based bone cement using a mixture design

Highly swelling polymers, i.e. superabsorbent hydrogels, are hydrophilic, three dimensional networks that can easily absorb a significant amount water, fluid or drug. They are widely used in various applications such as foods, cosmetics, and medical devices.Bone cements are used in orthopaedics as a filling biomaterial or as a grout enhancing the embedding of a prosthesis into bone and fixation is achieved by mechanical interlock with metal or bone surfaces. Recently, hydrophilic bone cements have attracted the attention for bone tissue-engineering applications.Here a bone cement containing an acrylic hydrogel (HEMA) as a liquid phase and a blend of corn starch, cellulose acetate and bioceramic filler as a solid phase is investigated by means of a mixture design which is a special topic within statistical Design of Experiments (DoE). Output variables of interest, complex shear modulus, compressive modulus and swelling rate related to rheological, mechanical and swelling properties respectively, are measured for each cement formulation.Applying the mixture design strategy enables to assess the impact of the three powder components on each variable of interest and to determine the optimal formulation in order to achieve the required properties of this HEMA-based bone cement, especially the rheology adapted to the desired clinical application, but also appropriate mechanical and swelling properties.

Optimizing initial oil production of horizontal Wolfcamp wells utilizing data analytics

Machine learning techniques have fundamentally altered how oil and gas industry practitioners design fracture operations. In this paper, we perform data analytics utilizing response surface methodology (RSM), a group of statistical techniques that develop a functional relationship between an output variable of interest and several associated input variables, to optimize the output. We apply RSM to optimize horizontal well production based on initial production (IP) of horizontal oil wells for 180 days (IP180 Oil), as a function of five input variables: reservoir type, fracturing fluid (gal/ft), proppants (lbm/ft), cluster spacing, and stage length (ft). The RSM model correlates the initial production of each well to the input variables via a single equation, thus allowing for exploration of the fitted response surface in order to maximize production. Although the choice of the five inputs is made based primarily after consultation with industry professionals, we validate our selection by also applying an assortment of data-analytics-based methods that attempt to rank variable importance and thereby identify completion variables that may be predictive of initial production. The findings rank all five variables above the 50th percentile, thus indicating that the chosen variables have merit. This procedure is applied to a dataset of 201 horizontal wells from the Wolfcamp formations. The model fits reasonably well, with R2 = 61%, a very significant F-statistic p value, and a predicted versus observed values scatterplot indicating a good fit. The RSM analysis suggests that, within the feasible space defined by this dataset, maximum values of IP 180 Oil may be obtained by setting the fracturing fluid in gal/ft at approximately 1972, while simultaneously maximizing the remaining input variables (proppant loading, cluster spacing, stage length). The outcome indicates the possible directions to be taken in seeking a global optimum initial production for the setting of completion variables. Iteration of this scheme may lead to a near-optimum global solution. The real utility of this work may be indicating the way different studies may be designed to optimize production, each with its own selection of inputs, and ultimately be combined in a meta-analysis.

Copula based generalized additive models for location, scale and shape with non-random sample selection

Non-random sample selection is a commonplace amongst many empirical studies and it appears when an output variable of interest is available only for a restricted non-random sub-sample of data. An extension of the generalized additive models for location, scale and shape which accounts for non-random sample selection by introducing a selection equation is discussed. The proposed approach allows for potentially any parametric distribution for the outcome variable, any parametric link function for the selection equation, several dependence structures between the (outcome and selection) equations through the use of copulae, and various types of covariate effects. Using a special case of the proposed model, it is shown how the score equations are corrected for the bias deriving from non-random sample selection. Parameter estimation is carried out within a penalized likelihood based framework. The empirical effectiveness of the approach is demonstrated through a simulation study and a case study. The models can be easily employed via the gjrm() function in the R package GJRM.

Multi-scale mathematical modeling of methane-fueled SOFCs: Predicting limiting current density using a modified Fick’s model

In this work, a direct internal reforming methane-fed solid oxide fuel cell based on a multi-physics channel-level mathematical model considering the effect of limiting current density, is studied. A modified Fick’s model is adopted to refine the gas species concentration at the triple phase boundary. The model assumes competitive absorption of reactants followed by surface diffusion to the reactive sites. The percolation theory is employed to model the micro-scale behavior of the cell. Safe operations of the cell in terms of carbon deposition boundaries for different operating conditions by obtaining the extent of hydrogen oxidation based on fuel utilization are also discussed. A quantitative analysis is presented to show the effects of critical system parameters on the output variables of interest. Porosity and particles size from the micro-model as well as fuel utilization, temperature and pre-reforming rate from the macro-model are some of those estimated parameters. The obtained results show that the suggested rate limiting mechanism based on Fick’s model more accurately predict the effect of limiting current density compared to those of dusty gas model. Further analysis illustrate that the limiting current density increases in both cases of higher inlet fuel concentrations and lower fuel utilizations. In addition, increasing pre-reforming rate and current density causes the system efficiency to diminish while power density is improved.

Variance-Based Sensitivity of Localized Sulphation to Microporous Separator Properties Using a Distributed Parameter Model of a Valve-Regulated Lead-Acid Battery

Modern applications such as electric and hybrid-electric vehicles require high-rate partial state-of-charge (HRPSoC) duty from battery systems. A major cause of premature capacity loss in lead-acid batteries under HRPSoC duty is irreversible passivation by progressive sulphation of the negative electrode. Notable advances have been made in the form of desulphating charge algorithms as well as carbon additives to the negative electrode that seem to restrict PbSO4 crystal growth. However, a method that substantially curbs irreversible sulphation has yet to be developed. During a high-rate discharge, the PbSO4 crystals that nucleate and grow on the porous electrodes are small and needle-like, implying a high crystal geometry factor. The available surface area for dissolution during a subsequent charge is significantly larger than if preceded by a low-rate discharge. Apart from the size, shape and distribution of the PbSO4 crystals, their dissolution also depends on the local electrolyte concentration. A high-rate discharge results in a particularly non-uniform concentration profile and, consequently, areas with low crystal solubility. Our focus is on a new valve-regulated lead-acid (VRLA) battery with immobilized electrolyte undergoing a high-rate discharge. We investigate the microporous separator’s design for healthier end-of-discharge (EOD) conditions. More specifically, we consider EOD conditions that would be favorable to the dissolution of PbSO4 crystals in a subsequent charge. We use a comprehensive electrochemical model that accounts for full mass transport over three regions: the positive electrode, the separator and the negative electrode. Partial differential equations describing the local PbSO4 crystals are also included. The model is verified in COMSOL Multiphysics®and used to simulate a high-rate discharge at 728 A until 12 Ah (50% DOD) followed by a short rest period, all at 25 °C. A variance-based sensitivity analysis (VBSA) is performed to determine the sensitivity of several output variables to changes in selected input parameters. The input space, consisting of the separator’s porosity, thickness and tortuosity, is explored in a Monte Carlo experiment using Quasi-random sampling. The output variables of interest are the local PbSO4 volume fraction and solubility, and the crystal geometry factor. First-order sensitivity explains the effect of a single parameter whereas the total sensitivity takes parameter interactions into account. Figure 1 shows the local PbSO4 volume fraction and solubility during a high-rate discharge. As the battery discharges, the electrolyte concentration decreases and the PbSO4 solubility increases. This solubility is much lower at the negative electrode and more crystals form at the separator interfaces than in other electrode regions. A greater volume fraction of PbSO4 has formed at the positive electrode but considering the local solubility, these crystals should dissolve during the following charge. Table 1 summarizes the sensitivity at the EOD for both separator interfaces calculated using 500 model evaluations. The separator tortuosity seems to have a negligible effect on the PbSO4 crystals at the interfaces during a high-rate discharge. Figure 2 shows the output variables at the EOD against the most influential parameter for each interface. For faster dissolution of the PbSO4 crystals at the negative electrode, we would like the highest crystal geometry factor and solubility but the smallest fraction of PbSO4 crystals. Figure 2 indicates the selection of a separator with thickness between 0.05 and 0.1 cm and porosity between 0.7 and 0.75. Figure 3 illustrates how the electrolyte concentration becomes more uniform in the negative electrode during a rest period after discharge, but the concentration in the positive electrode becomes more non-uniform. Equation (1) is used to calculate the concentration non-uniformity. Figure 3 also indicates that homogenization of the concentration in the negative electrode occurs faster when the separator thickness is roughly between 0.08 and 0.1 cm. At separator thicknesses greater than 0.125 cm, it seems that diffusion takes longer within the negative electrode. The VBSA results suggest that the separator thickness has the greatest effect on the PbSO4 crystals at the negative electrode simply because it influences the concentration profile so severely. We conclude from the VBSA results that optimal values for the separator thickness and porosity, to enable healthier EOD conditions, do exist. It is possible to design a microporous separator that improves the dissolution, during recharge, of the PbSO4 film on the negative electrode that initially formed during a high-rate discharge. Irreversible sulphation of the negative electrode in a lead-acid battery under HRPSoC duty might be curbed by simple yet calculated changes to the separator sandwiched between two porous electrodes. The quantitative approach we have followed can be used to quickly evaluate thousands of different separator designs and determine a suitable starting point for experimental work. Figure 1

The interactions between genotype, management and environment in regional crop modelling

Abstract Biophysical models to simulate crop yield are increasingly applied in regional climate impact assessments. When performing large-area simulations, there is often a paucity of data to spatially represent changes in genotype (G) and management (M) across different environments (E). The importance of this uncertainty source in simulation results is currently unclear. In this study, we used a variance-based sensitivity analysis to quantify the relative contribution of maize hybrid (i.e. G) and sowing date (i.e. M) to the variability in biomass yield (Y T , total above-ground biomass) and harvest index (HI, fraction of grain in total yield) of irrigated silage maize, across the extent of arable lands in New Zealand (i.e. E). Using a locally calibrated crop model (APSIM-maize), 25 G x M scenarios were simulated at a 5 arc minute resolution (∼5 km grid cell) using 30 years of historical weather data. Our results indicate that the impact of limited knowledge on G and M parameters depends on E and differs between model outputs. Specifically, the sensitivity of Y T and HI to genotype and sowing date combinations showed different patterns across locations. The absolute impact of G and M factors was consistently greater in the colder southern regions of New Zealand. However, the relative share of total variability explained by each factor, the sensitivity index (S i ), showed distinct spatial patterns for the two output variables. The Y T was more sensitive than HI in the warmer northern regions where absolute variability was the smallest. These patterns were characterised by a systematic response of S i to environmental drivers. For example, the sensitivity of Y T and HI to hybrid maturity consistently increased with temperature. For the irrigated conditions assumed in our study, inter-annual weather conditions explained a higher share of total variability in the southern colder regions. Our results suggest that the development of methods and datasets to more accurately represent spatio-temporal G and M variability can reduce uncertainty in regional modelling assessments at different degrees, depending on prevailing environmental conditions and the output variable of interest.

Comparative evaluation of 1D and quasi-2D hydraulic models based on benchmark and real-world applications for uncertainty assessment in flood mapping

Summary One-dimensional and quasi-two-dimensional hydraulic freeware models (HEC-RAS, LISFLOOD-FP and FLO-2d) are widely used for flood inundation mapping. These models are tested on a benchmark test with a mixed rectangular–triangular channel cross section. Using a Monte-Carlo approach, we employ extended sensitivity analysis by simultaneously varying the input discharge, longitudinal and lateral gradients and roughness coefficients, as well as the grid cell size. Based on statistical analysis of three output variables of interest, i.e. water depths at the inflow and outflow locations and total flood volume, we investigate the uncertainty enclosed in different model configurations and flow conditions, without the influence of errors and other assumptions on topography, channel geometry and boundary conditions. Moreover, we estimate the uncertainty associated to each input variable and we compare it to the overall one. The outcomes of the benchmark analysis are further highlighted by applying the three models to real-world flood propagation problems, in the context of two challenging case studies in Greece.

A conceptual model that simulates the influence of thermal inertia in building structures

The energy use for maintaining comfortable indoor temperatures are to a certain extent dependent on the thermal storage capacity of materials in contact with the indoor air. This article describes a conceptual model for investigating the effects of increasing the thermal storage capacity of building materials. A building is modeled as an exterior wall, an indoor air volume and a thermally heavy inner wall. The input of thermal energy (heating, solar) is also included, as is a varying external temperature. The result is shown in three-dimensional graphs with different output variables of interest as functions of volumetric heat capacity and thermal conductivity of the material in the heavy construction part (with a standard concrete as a reference). Output variables of interest are for example energy consumption, peak power consumption and thermal comfort parameters. Influence of factors such as the thickness of the interior wall, wall surface area, and the influence of free solar radiation can be tested. The aim is to present a minimal and thus fully comprehensible model that can be used as a qualitative tool to investigate the influence of thermal mass on building performance. The model was tested for a cold-climate case and the results show that passive energy storage through high thermal mass can significantly change the power consumption pattern, which can give significant benefits, while the total energy consumption in most cases is not much influenced.

Controllability study of an ethanol steam reforming process for hydrogen production

A system for ethanol steam reforming and purification of carbon monoxide (CO) designed to feed a PEM fuel cell has been modelled. From the model, we study the sensitivity and controllability emphasizing the study of the influence of the temperature on the output variables of interest. The results of the study of controllability are used for the identification of the best control structures.

Model-based process monitoring using robust generalized linear models

Model-based process-monitoring procedures are extremely useful in situations where an output variable of interest is impacted by one or more inputs to the process, and where there are multistage processes with multiple inputs and outputs. To build the model relating input and output variables, the procedure uses historical data, which often contain outliers. To accommodate the presence of these outliers, a robust fitting scheme is introduced for the Generalized Linear Model in process monitoring. Robust deviance residuals are defined and used as the basis of the monitoring procedure. An example and a simulation study for a gamma-distributed response are included. The average run length performance reveals that the procedure is effective for detecting small process shifts when outliers are present.

Toward a New Model for Probabilistic Household Forecasts

Summary Household projections are key components of analyses of several issues of social concern, including the welfare of the elderly, housing, and environmentally significant consumption patterns. Researchers or policy makers that use such projections need appropriate representations of uncertainty in order to inform their analyses. However, the weaknesses of the traditional approach of providing alternative variants to single “best guess” projection are magnified in household projections, which have many output variables of interest, and many input variables beyond fertility, mortality, and migration. We review current methods of household projections and the potential for using them to produce probabilistic projections, which would address many of these weaknesses. We then propose a new framework for a household projection method of intermediate complexity that we believe is a good candidate for providing a basis for further development of probabilistic household projections. An extension of the traditional headship rate approach, this method is based on modelling changes in headship rates decomposed by household size as a function of variables describing demographic events such as parity specific fertility, union formation and dissolution, and leaving home. It has moderate data requirements, manageable complexity, allows for direct specification of demographic events, and produces output that includes the most important household characteristics for many applications. An illustration of how such a model might be constructed, using data on the U.S. and China over the past several decades, demonstrates the viability of the approach.

WATER QUALITY MODEL OUTPUT UNCERTAINTY AS AFFECTED BY SPATIAL RESOLUTION OF INPUT DATA1

ABSTRACT: Resolution of the input GIS data used to parameterize distributed‐parameter hydrologic/water quality models may affect uncertainty in model outputs and impact the subsequent application of model results in watershed management. In this study we evaluated the impact of varying spatial resolutions of DEM, land use, and soil data (30 × 30 m, 100 × 100 m, 150 × 150 m, 200 × 200 m, 300 × 300 m, 500 × 500 m, and 1,000 × 1,000 m) on the uncertainty of SWAT predicted flow, sediment, NO3‐N, and TP transport. Inputs included measured hydrologic, meteorological, and watershed characteristics as well as water quality data from the Moores Creek watershed in Washington County, Arkansas. The SWAT model output was most affected by input DEM data resolution. A coarser DEM data resolution resulted in decreased representation of watershed area and slope and increased slope length. Distribution of pasture, forest, and urban areas within the watershed was significantly affected at coarser resolution of land use and resulted in significant uncertainty in predicted sediment, NO3‐N, and TP output. Soils data resolution had no significant effect on flow and NO3‐N predictions; however, sediment was overpredicted by 26 percent, and TP was underpredicted by 26 percent at 1,000 m resolution. This may be due to change in relative distribution of various hydrologic soils groups (HSGs) in the watershed. Minimum resolution for input GIS data to achieve less than 10 percent model output error depended upon the output variable of interest. For flow, sediment, NO3‐N, and TP predictions, minimum DEM data resolution should range from 30 to 300 m, whereas minimum land use and soils data resolution should range from 300 to 500 m.

Optimization of a bonded leads process using statistically designed experiments

In manufacturing high-technology products using automated manufacturing equipment it is critical to obtain process-operating conditions that simultaneously optimize several output variables of interest. Often these output variables are non-traditional responses such as standard deviations, and there may even be discrete (count or binary) responses. Statistically designed experiments are very useful in optimizing these processes. We illustrate how factorial experiments, statistical models for the responses of interest, and simple optimization techniques can be successfully applied to a bonded leads process.

Uncertainty and sensitivity analyses of OH-initiated dimethyl sulphide (DMS) oxidation kinetics

A numerical experiment has been conducted on the OH-initiated tropospheric oxidation of DMS. This involved the selection of a set of reactions describing the OH-initiated oxidation kinetics and the conversion of the present level of uncertainty of the system into uncertainty ranges and distributions for the relevant system parameters (kinetic constants and initial concentrations). Uncertainties have been propagated through the model onto the output variables of interest. This has allowed (a) the uncertainty in model prediction to be quantified and compared with observations (uncertainty analysis) and (b) the relative importance of each input parameter in determining the output uncertainty to be quantified (sensitivity analysis). Output considered were the ratio of MSA/(SO2 + H2SO4) concentration at a given time, the ratio SO2/H2SO4, the total peroxynitrate species concentrations and the relative fraction of SO2 and H2SO4 formed through the various pathways. Conditional upon the model and data assumptions underlying the experiment, the following main conclusions were drawn: (1) The possibility of direct formation of SO3 without SO2 as intermediate as suggested by Bandyet al. (1992) and Yinet al. (1990), involving direct thermal decomposition of CH3SO3 · does not seem to play a major role in the overall generation of sulphate. This is relevant to the issue of gas to particle conversion over remote areas. (2) Reaction of CH3SOO · intermediate may be the most important pathway to the formation of SO2. (3) The dominating peroxynitrate is CH3S(O)2O2NO2.

Response Surface Development Using RETRAN

The Electric Power Research Institute (EPRI) and GPU Nuclear Corporation have completed a demonstration project that provides justification for relaxing the high-pressure setpoints for the Oyster Creek Nuclear Generating Station. The project was undertaken because an undesirable overlap had been identified in the high-pressure setpoints when accounting for measurement uncertainties experienced during plant operation. The project employed a statistical combination of uncertainties (SCU) process to provide increased margin for measurement uncertainties. This approach was used because previous experience indicated that there was insufficient margin to justify the desired setpoints using conventional deterministic inputs to the safety analysis and plant performance analysis processes. Through the use of SCU methodology and other deterministic analyses, it is possible to provide comprehensive bases for the desired technical specification changes to the high-pressure setpoints.The SCU process is based on the EPRI setpoint analysis guidelines, and it requires the development of response surfaces to simulate RETRAN peak pressure calculations for the limiting transient events. The use of response surfaces adds an intermediate step to the SCU process, but reduces the number of RETRAN cases required to make appropriate statistical statements about the result probabilities. Basically, each response surface is an approximation of the RETRAN code for one particular event and one output variable of interest, which is valid over a limited region. The response surfaces can be sampled very inexpensively using simple Monte Carlo methods.The basic input to the development of a response surface is a set of results obtained from specific RETRAN cases. Each case includes a particular set of parameters consistent with an experimental design selected to ensure that all of the parameter dependencies are carefully considered and that the response surface fit has a reasonably small fitting error. The parameters selected for incorporation into the response surface are identified through a screening process that uses RETRAN analyses to establish the sensitivity of the event results to the parameter uncertainty. The parameter screening process, the selection of the experimental design, and the development of the response surfaces are described, and the analysis results are provided.