Influence Of Process Variables Research Articles

Enhanced α-amylase production by a marine protist, Ulkenia sp. using response surface methodology and genetic algorithm

ABSTRACTAmylases are a group of enzymes with a wide variety of industrial applications. Enhancement of α-amylase production from the marine protists, thraustochytrids has been attempted for the first time by applying statistical-based experimental designs using response surface methodology (RSM) and genetic algorithm (GA) for optimization of the most influencing process variables. A full factorial central composite experimental design was used to study the cumulative interactive effect of nutritional components viz., glucose, corn starch, and yeast extract. RSM was performed on two objectives, that is, growth of Ulkenia sp. AH-2 (ATCC® PRA­296) and α-amylase activity. When GA was conducted for maximization of the enzyme activity, the optimal α-amylase activity was found to be 71.20 U/mL which was close to that obtained by RSM (71.93 U/mL), both of which were in agreement with the predicted value of 72.37 U/mL. Optimal growth at the optimized process variables was found to be 1.89A660nm. The optimized medium increased α-amylase production by 1.2-fold.

Screening the fluid bed granulation process variables and moisture content determination of pharmaceutical granules by NIR Spectroscopy

The fluid bed granulation (FBG) is a wet granulation technique for producing granules. It is a complex process because many process variables can influence the granule properties. Therefore, an understanding of the influence of the granulation process variables is necessary for controlling the process. The moisture content of granule also plays a critical role in determining the outcome of the batch. The purpose of this work was to apply Plackett-Burman design for screening of process variables in FBG, study the influence of the process variables on granules properties and the use of NIR spectroscopy and partial least squares (PLS) regression to predict the moisture content of the granules. In order to study the influence of the process variables on the granules properties, Plackett-Burman design with six factors, two levels and three replicates at the center point (15 runs) was used. The results revealed that the atomizing pressure and the airflow rate are the process variables that have strong influence on the granules properties. The NIR spectroscopy in conjunction with PLS was used to determine the moisture content of granule in the FBG. The proposed PLS model was fitted and its predictive performance was evaluated by traditional chemometric criteria. The root mean square error of prediction (RMSEP) was 4.15% with 2 latent variables (LVs). The proposed NIR method was validated and the results obtained were compared with those of the reference LOD method using a paired t-test.

Development of a green procedure of citrus fruits waste processing to recover carotenoids

In this study, an original and green procedure of processing waste of the citrus fruits was developed using the concept of bio-refinery, innovative techniques “ultrasound”and “micro-wave”, and a green sol- vent “limonene”to recover carotenoids. Essential oil extraction was performed by Solvent Free Microwave Extraction (SFME) and compared to steam distillation (SD). The essential oil yields were comparable for both processes: 4.02 ±0.23% for SFME and 4.16 ±0.05% for SD. After that, carotenoid extraction from cit- rus peels was performed by ultrasound-assisted extraction (UAE) and conventional extraction (CE) using d -limonene obtained starting from essential oil, as a solvent, and then compared to n-hexane extract. Response surface methodology (RSM) using central composite designs (CCD) approach was launched to investigate the influence of process variables on the ultrasound-assisted extraction (UAE). The statisti- cal analysis revealed that the optimized conditions of ultrasound power, temperature and time were 208 W cm −2 , 20 °C and 5 min giving carotenoid content of 11.25 mg L −1 . Compared to the conventional ex- traction, (UAE) gave an increase of 40% in carotenoid content. The comparison to n-hexane extract gave no significant changes in carotenoid content. Combination of microwave, ultrasound and d -limonene ob- tained from a bio-refinery of a by-product of citrus fruits industry allow us to develop a very good envi- ronmental green approach giving high added values compounds, with a saving of time, and a complete valorisation of waste.

Development of a green procedure of citrus fruits waste processing to recover carotenoids

Abstract In this study, an original and green procedure of processing waste of the citrus fruits was developed using the concept of bio-refinery, innovative techniques “ultrasound” and “micro-wave”, and a green solvent “limonene” to recover carotenoids. Essential oil extraction was performed by Solvent Free Microwave Extraction (SFME) and compared to steam distillation (SD). The essential oil yields were comparable for both processes: 4.02 ± 0.23% for SFME and 4.16 ± 0.05% for SD. After that, carotenoid extraction from citrus peels was performed by ultrasound-assisted extraction (UAE) and conventional extraction (CE) using d -limonene obtained starting from essential oil, as a solvent, and then compared to n-hexane extract. Response surface methodology (RSM) using central composite designs (CCD) approach was launched to investigate the influence of process variables on the ultrasound-assisted extraction (UAE). The statistical analysis revealed that the optimized conditions of ultrasound power, temperature and time were 208 W cm −2 , 20 °C and 5 min giving carotenoid content of 11.25 mg L −1 . Compared to the conventional extraction, (UAE) gave an increase of 40% in carotenoid content. The comparison to n-hexane extract gave no significant changes in carotenoid content. Combination of microwave, ultrasound and d -limonene obtained from a bio-refinery of a by-product of citrus fruits industry allow us to develop a very good environmental green approach giving high added values compounds, with a saving of time, and a complete valorisation of waste.

Influence of Silk-Like Finishing Process Variables on Fabric Properties

A 100% PES woven fabric was subjected to a silk-like finishing treatment using a two-factor (soda concentration and vaporisation time), three-level experimental design on a pilot plant in order to examine the influence of process variables on fabric drape and hand, and also to relate weight loss to the properties of the finished fabric. Weight loss in the fabric was significantly related to the process variables studied. By contrast, the vaporisation time affected no drape indicator, even though it interacted significantly with the soda concentration, which influenced three of the five indicators examined. Also the vaporisation time affected no FAST property, but its interaction with the soda concentration had a significant effect on all drape-related FAST parameters. Regression equations accurately predicting drape indicators and physico-mechanical properties of the fabric studied from its weight loss are proposed.

Unidirectional freezing as a tool for tailoring air permeability in macroporous poly(ethylene glycol)-based cross-linked networks

Unidirectional freezing followed by photopolymerization at subzero temperatures was used to obtain highly air-permeable monoliths with ordered porous structures. Scaffolds were obtained from aqueous solutions of a poly(ethylene glycol)dimethacrylate (PEGDMA) oligomer, a photosensitizer and a reducing agent. Solutions were vertically frozen in liquid nitrogen at a controlled rate to induce the oriented growth of ice crystals and then cryo-photopolymerized under blue-light irradiation. Ice crystals were finally removed under vacuum producing macroporous hydrophilic networks with aligned pores. Porosities ranged between 80 and 95%, depending on the initial concentration of PEGDMA. The influence of processing variables on the final properties of the materials was addressed, concerning particularly the effect of porosity and freezing directionality on air permeability. Compared to porous PEGDMA-based monoliths with non-aligned macropores, gas permeability was two to three times higher for oriented scaffolds at the same porosity level, a fact explained by the easier transport of gas molecules through the aligned structures. However, the role of pore orientation on gas permeability was shown to be less marked as porosity increased. The results demonstrate that the use of unidirectional freezing strongly increases the permeability of monolithic samples up to values usually required, for instance, in tissue engineering applications (higher than 2D). These findings provide new perspectives on pore design principles toward future scaffolding of polymeric cross-linked matrices.

Bio-oil production from hydrothermal liquefaction of waste Cyanophyta biomass: Influence of process variables and their interactions on the product distributions

Hydrothermal liquefaction (HTL) of waste Cyanophyta biomass at different temperatures (factor A, 260–420 °C), times (factor B, 5–75 min) and algae/water (a/w) ratios (factor C, 0.02–0.3) by single reaction condition and Response Surface Method (RSM) experiments was investigated. By single reaction condition runs, maximum total bio-oil yield (29.24%) was obtained at 350 °C, 60 min and 0.25 a/w ratio. Maximum bio-oil HHV of 40.04 MJ/kg and energy recovery of 51.09% was achieved at 350 °C, 30 min, 0.1 a/w ratio and 350 °C, 60 min, 0.25 a/w ratio, respectively. RSM results indicate that effect of AB interaction was significant on light bio-oil yield. Both AC and AB had more remarkable influence than BC on heavy bio-oil yield and aqueous total organic carbon (TOC) recovery whereas BC was noticeable on ammonia nitrogen (NH3N) recovery in aqueous products. By model-based optimization of highest bio-oil yield, the highest bio-oil yield reached 31.79%, increasing by 8.72% after RSM optimization, and light and heavy bio-oil yield was 17.44% and 14.35%, respectively. Long-chain alkanes, alkenes, ketones, fatty acids, phenols, benzenes, amides, naphthalenes were the main components in light bio-oil. Some alcohols, phenols and aromatics were primarily found in heavy bio-oil. Solid residue after HTL consisted of numerous microparticles (∼5 μm) observed by Scanning Electron Microscopy (SEM). Energy Dispersive Spectrometer (EDS) analysis shows these particles primarily contained C, O, Mg, P and microelements, derived from Cyanophyta cells.

Simultaneous removal of dyes onto nanowires adsorbent use of ultrasound assisted adsorption to clean waste water: Chemometrics for modeling and optimization, multicomponent adsorption and kinetic study

In this study, Ni doped Ferric Oxy-hydroxide FeO(OH) nanowires (Ni:FeO(OH)-NWs) were synthesized and loaded on activated carbon (AC). Synthesized adsorbent was characterized by field emission scanning electron microscopy (FESEM) and X-ray diffraction analysis (XRD). Isotherms and kinetic behaviors of Safranin-O (SO) and Indigo Carmine (IC) adsorption onto Ni:FeO(OH)-NWs-AC were explained by extended Freundlich‎ and pseudo second order kinetic models. The adsorption performance was critically analyzed using response surface methodology (RSM), artificial neural network (ANN) and linear algebra based models and compared. The influence of process variables (initial dyes concentration, adsorbent mass and sonication time) on the removal of both dyes was investigated by central composite rotatable design (CCRD) of RSM, Multi-Layer Perceptron (MLP) neural network and Doolittle Factorization Algorithm (DFA). All the models (RSM, ANN and DDA) were statistically compared by the coefficient of determination (R2), root mean square error (RMSE), mean absolute error (MAE) and absolute average deviation (AAD) based on the validation data set. The coefficient of determination (R2) calculated from the validation data for RSM, ANN and DDA models were 0.98, 0.99 and 0.99 For IC and 0.99, 0.99 and 0.99 for SO dye, respectively. The ANN model was found to be more precise compared to the other models. However, it was demonstrated that DDA can reduce the orders of data and needs a little time for analysis. So it has bright prospects in chemometrics and it is feasible that the Doolittle Algorithm could be applied to model the real systems. The sensitivity analysis confirmed that sonication time was the essential factor affecting the removal of SO and IC with the relative importance of 36.63% and 12.60%, respectively. The monolayer adsorption capacity of the IC and SO was 29.09 and 37.85mgg−1, respectively.

A new indigenous green option for turbidity removal from aqueous system

Abstract The quest to develop effective, green, eco-friendly and sustainable indigenous solutions to a global challenge necessitated this study. The water extract of the fresh seed of Margaritarea discoidea was characterised and used as a primary coagulant in turbidity removal from qua system. The optimum coagulant dosage of the fruit seed extract (FSE) was determined and the time-concentration profiles of the coagulation/flocculation (CF) process was determined. The influence of process variables (i.e. turbid water pH value, anionic and cationic interference) on the CF process and the characteristics of the sludge derived from the process were evaluated. Premised on experimental and instrumental analysis, the active ingredient in the FSE was found to be galactomannans, a water soluble non-ionic hydrocolloidal polysaccharide. The kinetic analysis of the time-concentration profile data showed that the reaction pathway may not be straight forward as expected for a typical homogeneous chemical reaction. The nature of the active coagulating ingredients in the FSE revealed that the turbidity removal occurred via adsorption and bridging mechanism. Variations in the process variable parameters had no visible influence on the coagulation efficiency of the FSE. The values of the sludge volume index (31.1 mg/g) and the specific cake resistance ( α = 5.27 × 10 9 m/kg) and the resistance of the filter medium ( R m = 1.58 × 10 4 l/m) obtained for the sludge derived from the CF operations were below the values reported for sludges derived from many other CF processes.

α-L-ARABINOFURANOSIDASE FROM AN EFFICIENT HEMICELLULOLYTIC FUNGUS Penicillium janthinellum CAPABLE OF HYDROLYZING WHEAT AND RYE ARABINOXYLAN TO ARABINOSE

This study reports Penicillium janthinellum strain, producing high levels of α-L-arabinofuranosidase (AFase) as well as other components of hemicelluloytic enzyme system (endoxylanase, β-xylosidase and acetyl xylan esterase) on rice straw and wheat bran containing solidified culture medium. Optimization of culture conditions was carried out using Box–Behnken design of experiment to study the influence of process variables (ammonium sulphate, pH and moisture level) on AFase production. Analysis of data showed R2 (0.9967) and adjusted R2 (0.9925) indicating model to be good fit and robust to predict culture conditions for AFase production. Under optimal culture conditions P. janthinellum produced high levels of AFase (212 units/g dw substrate) in addition to xylanase (1800 units/g dw substrate), β-xylosidase (31 units/g dw substrate), acetyl xylan esterase (231 units/g dw substrate) and feruloyl esterase (27 units/g dw substrate). AFase from P. janthinellum culture extract was purified to homogeneity and characterized to be a 64 kDa protein with a pI of 3.8. The peptide mass fingerprinting showed the AFase belonged to family GH54. The enzyme was optimally active at 50o C at pH 5.5 and its activity was positively modulated in presence of Fe3+ions. The enzyme preferentially catalyzed the hydrolysis of pNP- arabinofuranoside (pNPA) with Km and Vmax of 0.4mM and 260 unit mg-1 protein-1, respectively. Hydrolysis with purified AFase (3.0 units/g substrate) released arabinose from rye arabinoxylans (29.5 mg/g substrate) and wheat arabinoxylan (13.4 mg/g substrate), as the sole product indicating P. janthinellum as the important source of α-L-arabinofuranosidase for bioconversion of agro-residue to value added products.

Modeling of Continuous Ultrasonication to Improve Total Phenolic Content and Antioxidant Activity in Sorghum Flour: A Comparison between Response Surface Methodology and Artificial Neural Network

Abstract Fermentation followed by continuous ultrasonication was applied to release the bound phenolics in sorghum flour (SF). Total phenolic content (TPC) and antioxidant activity (AA) increased with decrease in fermentation time (FT), flour to water ratio (FWR), flow rate (FR) and ultrasonication intensity (UI). The influence of process variables was investigated by Box–Behnken design and multi-layer perceptron neural network. The optimum conditions for maximum TPC and AA were obtained as 12 h FT, 10 % (w/v) FWR, 20 W/cm2 UI, 4 ml/s FR and 120 s UT. The values observed for TPC and AA at optimum conditions were 90.1 mg GAE/100 g dm and 190.1 µmol TE/100 g dm, respectively, while these values for control SF were observed as 63.9 mg GAE/100 g dm and 133.5 µmol TE/100 g dm. Ultrasonication improved the free phenolic acid content by releasing bound phenolics in SF. The ANN model prediction was more precise compared to the RSM model.

PLANEJAMENTO COMPOSTO CENTRAL APLICADO A SISTEMA DE OBTENÇÃO DE BIOQUEROSENE VIA CATÁLISE HETEROGÊNEA

Neste trabalho foi produzido bioquerosene pela reação de esterificação (utilizando o Amberlyst® 15 como catalisador heterogêneo) dos ácidos graxos livres de C6 a C16 átomos de carbono provenientes do óleo de babaçu (Orbignya speciosa) utilizando um planejamento de experimento fatorial 23 com composto central. A análise termogravimétria foi utilizada para indicar a faixa de trabalho do catalisador e a microscopia eletrônica de varredura, a morfologia. Os parâmetros de processo estudados foram a temperatura de reação entre 70 °C e 100 °C, a concentração em massa de catalisador entre 5 % e 10 % e o tempo de reação entre 240 minutos e 480 minutos. Como parâmetro de resposta do planejamento experimental foi avaliado o teor em massa de metil ésteres. A otimização do processo foi realizada através da metodologia de superfície de resposta (MSR) e o modelo polinomial de segunda ordem. Foi observado que a variável de processo de maior influência foi a temperatura de reação, sendo que a maior conversão de ésteres foi obtida com 7,5 % de catalisador; temperatura de reação de 85 °C e com um tempo de 360 minutos de reação. Com esses parâmetros a conversão alcançada foi de 94,7 % em biocombustível. Utilizando a MRS, com modelo polinomial de segunda ordem, as condições otimizadas obtidas para a reação de esterificação com Amberlyst® 15 foram: temperatura de 90,0 °C, quantidade de catalisador de 7,3 % e tempo reacional de 338,2 minutos. O rendimento teórico da conversão previsto para os ácidos graxos livres de C6 a C16 átomos de carbono em bioquerosene foi de 95,7 %. Esta pesquisa foi apoiada pela Universidade Federal do Ceará (UFC), Laboratório de Imagem Central Analítica da UFC, Fundação Núcleo de Tecnologia Industrial do Ceará (NUTEC) e financiada pelo Conselho Nacional de Desenvolvimento Científico e Tecnológico (CNPq) e pela Fundação Cearense de Apoio ao Desenvolvimento Científico e Tecnológico (FUNCAP).

X-ray inspection and Six-Sigma in analysis of LED thermal pad coverage

PurposeThe purpose of this paper is to evaluate selected methods of reduction voidings in lead-free solder joints underneath thermal pads of light-emitting diodes (LEDs), using X-ray inspection and Six Sigma methodology.Design/methodology/approachOn the basis of cause and effect diagram for solder voiding, the potential causes of voids and influence of process variables on void formation were found. Three process variables were chosen: the type of reflow soldering, vacuum incorporation and the type of solder paste. Samples of LEDs were mounted with convection and vapour phase reflow soldering. Vacuum was incorporated into vapour phase soldering. Two types of solder pastes OM338PT and LFS-216LT were used. Algorithm incorporated into X-ray inspection system enabled to calculate the statistical distribution of LED thermal pad coverage and to find the process capability index (Cpk) of applied soldering techniques.FindingsThe evaluation of selected soldering processes of LEDs in respect of their thermal pad coverage and statistical Cpk indices is presented. Vapour-phase soldering with vacuum is capable (Cpk > 1) for OM338PT and LFS-216LT paste. Convection reflow without vacuum with LFS-216LT paste is also capable (Cpk = 1.1). Other technological soldering processes require improvements. Vacuum improves radically the capability of a reflow soldering for an LED assembly. When vacuum is not accessible, some improvement of capability to a lower extent is possible by an application of void-free solder pastes.Originality/valueSix Sigma statistical methodology combined with X-ray diagnosis was used to check whether applied methods of void reduction underneath LED thermal pads are capable processes.

A Study on the Removal of Direct Blue 71 Dye From Textile Wastewater Produced From State Company of Cotton Industries by Electrocoagulation Using Aluminum Electrodes

The removal of direct blue 71 dye from a prepared wastewater was studied employing batch electrocoagulation (EC) cell. The electrodes of aluminum were used. The influence of process variables which include initial pH (2.0-12.0), wastewater conductivity (0.8 -12.57) mS/cm , initial dye concentration (30 -210) mg/L, electrolysis time (3-12) min, current density (10-50) mA/cm2 were studied in order to maximize the color removal from wastewater. Experimental results showed that the color removal yield increases with increasing pH until pH 6.0 after that it decreased with increasing pH. The color removal increased with increasing current density, wastewater conductivity, electrolysis time, and decreased with increasing the concentration of initial dye. The maximum color removal yield of 96.5% was obtained at pH 6.0, wastewater conductivity 9.28 mS/cm , electrolysis time 6 min ,the concentration of initial dye 6 0 mg/L and current density 30 mA/cm2 .

Green Sonoextraction of Protein from Oleaginous Press Rapeseed Cake.

In this study, extraction of soluble proteins from rapeseed cake using different conventional and innovative extraction processes in order to maximize the extraction yield has been investigated. Firstly, various extraction techniques including ultrasound, microwave, and percolation were tested to increase the protein recovery efficiency. Secondly, response surface methodology (RSM) using a central composite design (CCD) approach was applied to investigate the influence of process variables on ultrasound-assisted extraction (UAE). Statistical analysis revealed that the optimized conditions providing a protein yield of 4.24 g/100 g DM were an ultrasound power of 5.6 W·cm−2 and temperature of 45 °C. Quantitatively UAE followed by two stages of conventional extraction gave the best total protein yield of 9.81 g/100 g DM. Qualitatively, the protein efficiency ratio (PER) used as measure of the nutritive value (12S/2S ratio) which indicates protein quality in terms of S-containing essential amino acids, was similar to that of the conventional extraction method. Small amounts of protein aggregate were observed in the HPLC profile of the extract.

INFLUENCE OF PROCESS VARIABLES ON ADSORPTION OF CONGO RED ONTO MANGO LEAF CHAR USING FACTORIAL DESIGN ANALYSIS

Influence of process variables like pH, adsorbent dosage (D), initial dye concentration (C0), particle size (S) and temperature (T) on adsorption of a direct dye on carbon derived from mango leaf were studied using a 25 factorial design. The main effects and interaction effects were determined using statistical methods. Maximum uptake obtained was 55.7 mg/g (pH = 2, T = 40 C, D = 0.5 g/dm3, C0=100 mg/dm3 and S = 53 um). It was found that the effects of process variables namely pH, adsorbent dosage, initial dye concentration, particle size and temperature on dye uptake were statistically significant. In addition, 2-level, 3-level and 4-level interactions between the variables were also found to influence dye uptake significantly. The model explained 96.79% variation in the response leaving only 3.21% for residuals. Dye uptake capacity up to 55.7 mg/g could be obtained with the proposed carbonaceous adsorbent mango leaf char.

Nutrient release characteristics and coating homogeneity of biopolymer coated urea as a function of fluidized bed process variables

AbstractThe present study investigates the effect of fluid‐bed process parameters on the diffusion coefficient of nitrogen release and coating homogeneity of controlled‐release urea (CRU) produced in a rotary fluidized bed using polyvinyl‐alcohol‐modified starch as a coating material. An existing mathematical model was used to estimate the diffusion coefficient. The coefficient of variance of size distribution and coating mass variation are reported as a measure of coating homogeneity.Statistical analysis suggested that the most influential process variables that govern urea release characteristics and coating homogeneity were (a) fluidizing gas temperature and (b) coating time. A moderate spray rate combined with longer coating time yielded the lowest diffusion coefficient for nutrient release. Elutriation as the result of elevated fluidizing gas temperature allowed a higher diffusion coefficient due to lower coating thickness. Burst release patterns were observed for granules with coating imperfections. The augmented temperature of fluidizing gas had a negative effect on coating mass and size distribution of CRU granules but the influence of longer coating time was positive.

Synergistic effect of chloro and sulphonic acid groups on the hydrolysis of microcrystalline cellulose under benign conditions

The hydrolysis of cellulose catalysed by ionic liquid (IL) immobilized on chloromethyl vinyl benzene (CMVB) with different IL/CMVB ratio was carried out in water as a solvent. The influence of process variables like IL/CMVB ratio in catalyst, temperature and time were investigated in a batch reactor. It was found that 40% IL/CMVB catalyst afforded maximum%TRS and glucose yield 58.5 and 47.9% respectively at 160 °C in 6h reaction time. This can be attributed to the synergistic effect of chloro and sulfonic groups, to dissolve the cellulose by breaking the intra and inter molecular hydrogen bonds and cleave the β-1, 4 glycosidic bonds respectively. The catalysts synthesized in the laboratory were characterized by FTIR, CHNS and TGA. The catalyst can be recovered three times without any significant loss in activity and characterized by FTIR to show the peaks of chloro and sulfonic groups in the recovered catalyst.

Screening of process variables using Plackett–Burman design in the fabrication of gedunin-loaded liposomes

This study is to screening the formulation and process variables that produce significant effect on the gedunin-loaded liposome formulations by using quality-by-design approach. Placket–Burman screening design was used to screen the most influencing formulation and process variables. Mean vesicle size, zeta potential, entrapment efficiency, and loading capacity were found in the range of 112–990 nm, −19.39 to −39.20 mV, 45.25–87.60%, and 3.54–10.47%, respectively. Differential scanning calorimetry (DSC) and powder X-ray diffraction (XRD) result suggested that Gedunin encapsulated within liposome as amorphous state. The analysis of Pareto chart represented that selected independent variables had a most significant effect on dependent variables.

Air Extrusion System for Ionotropic Alginate Microgel Particle Formation: A Review

AbstractIonotropic alginate microgels are commonly used for bioencapsulation applications and the air extrusion system is one of the common methods employed to produce these microgels. Although there have been many individual studies, to date no single report has comprehensively summarized the findings and results of the usage of this system. This review gives insight into the air extrusion methods, covering the setup, type, and characteristics of the system, focusing on their innovative aspects. Characterization techniques for the resulting microgels are presented, along with the influence of process variables on the product size and the development of empirical models for size prediction.