Estimation of sky luminance in the tropics using artificial neural networks: Modeling and performance comparison with the CIE model

An artificial neural network (ANN) model for estimating sky luminance was developed. A 3-year period (2007–2009) of sky luminance data obtained from measurements at Nakhon Pathom (13.82°N, 100.04°E) and a 1-year period (2008) of the same type of data at Songkhla (7.20°N, 100.60°E), Thailand were used in this study. The ANN model was trained using a back propagation algorithm, based on 2 years data (2007–2008) at Nakhon Pathom for clear, partly cloudy and overcast skies. The trained ANN model was used to predict sky luminance at Nakhon Pathom for the year 2009 for the case of clear, partly cloudy and overcast skies. The results were compared with those of the CIE model. It was found that the ANN model performed better than CIE models for most cases. The ANN model trained with Nakhon Pathom data were also used to predict sky luminance at Songkhla and satisfactory results were obtained.     

Dynamic neural network modeling for hydrochloric acid recovery process

This paper describes neural network models for the prediction of the concentration profile of a hydrochloric acid recovery process consisting of double fixed-bed ion exchange columns. The process is used to remove the Fe²⁺ and Fe³⁺ ion from the pickling liquor, resulting in increasing the acid concentration for reusing in the pickling process. Due to the complexity and highly nonlinearity of the process, the modeling of the process based on the first principle is difficult and involve too many unknown parameters. Therefore, an attractive alternative technique, neural network modeling, has been applied to model this system because of its ability to model a complex nonlinear process, even when process understanding is limited. The process data sets are gathered from a real hydrochloric acid recovery pilot plant and used for neural network training and validation. Backpropagation and Lenvenberg-Marquardt techniques are used to train various neural network architectures, and the accuracy of the obtained models have been examined by using test data set. The optimal neural network architectures of this process can be determined by MSE minimization technique. The simulation results have shown that multilayer feedforward neural network models with two hidden layers provide sufficiently accurate prediction of the concentration profile of the process.     

Irradiation effect of carbon negative-ion implantation on polytetrafluoroethylene for controlling cell-adhesion property

We have investigated the irradiation effect of negative-ion implantation on the changes of physical surface property of polytetrafluoroethylene (PTFE) for controlling the adhesion property of stem cells. Carbon negative ions were implanted into PTFE sheets at fluences of 1 × 10¹⁴-1 × 10¹⁶ ions/cm² and energies of 5-20 keV. Wettability and atomic bonding state including the ion-induced functional groups on the modified surfaces were investigated by water contact angle measurement and XPS analysis, respectively. An initial value of water contact angles on PTFE decreased from 104° to 88° with an increase in ion influence to 1 × 10¹⁶ ions/cm², corresponding to the peak shifting of XPS C1s spectra from 292.5 eV to 285 eV with long tail on the left peak-side. The change of peak position was due to decrease of C-F₂ bonds and increase of C-C bonds with the formation of hydrophilic oxygen functional groups of OH and CO bonds after the ion implantation. After culturing rat mesenchymal stem cells (MSC) for 4 days, the cell-adhesion properties on the C⁻-patterned PTFE were observed by fluorescent microscopy with staining the cell nuclei and their actin filament (F-actin). The clear adhesion patterning of MSCs on the PTFE was obtained at energies of 5-10 keV and a fluence of 1 × 10¹⁵ ions/cm². While the sparse patterns and the uncontrollable patterns were found at a low fluence of 3 × 10¹⁴ ions/cm² and a high fluence of 3 × 10¹⁵ ions/cm², respectively. As a result, we could improve the surface wettability of PTFE to control the cell-adhesion property by carbon negative-ion implantation.     

Neural network based model predictive control for a steel pickling process

A multi-layer feedforward neural network model based predictive control scheme is developed for a multivariable nonlinear steel pickling process in this paper. In the acid baths three variables under controlled are the hydrochloric acid concentrations. The baths exhibit the normal features of an industrial system such as nonlinear dynamics and multi-effects among variables. In the modeling, multiple input, single-output recurrent neural network subsystem models are developed using input–output data sets obtaining from mathematical model simulation. The Levenberg–Marquardt algorithm is used to train the process models. In the control (MPC) algorithm, the feedforward neural network models are used to predict the state variables over a prediction horizon within the model predictive control algorithm for searching the optimal control actions via sequential quadratic programming. The proposed algorithm is tested for control of a steel pickling process in several cases in simulation such as for set point tracking, disturbance, model mismatch and presence of noise. The results for the neural network model predictive control (NNMPC) overall show better performance in the control of the system over the conventional PI controller in all cases.     

Physicochemical properties of thermal alkaline treated pigeonpea (Cajanus cajan L.) flour

Thermal alkaline treatment, normally used for corn, was applied to pigeonpea grains. Starch granules were isolated using wet milling and alkaline treatments. Effects of the calcium hydroxide [Ca(OH)₂] concentration in the range of 0–1% (w/v) on granule structure, crystalline structure, chemical composition, and physicochemical, thermal, and pasting properties of isolated starch granules were determined. Compared to native samples, thermal alkaline treated samples had higher protein, lipid, calcium, and phosphorus contents, but lower starch and amylose contents. Thermal alkaline treatment increased starch granular size and gelatinization temperatures, but decreased relative crystallinity, gelatinization enthalpy, swelling power, solubility, amylose leaching, and the pasting viscosity. Amylose-lipid complexes were not found in thermal alkaline treated flours. As the Ca(OH)₂ concentration increased, the amylose content, relative crystallinity, gelatinization temperature, and enthalpy also increased, but the swelling power, solubility, amylose leaching, and paste viscosity decreased. A higher Ca(OH)₂ concentration produced more stable starch granules that resisted re-gelatinization.     

Creating Flavin Reductase Variants with Thermostable and Solvent-Tolerant Properties by Rational-Design Engineering

We have employed computational approaches—FireProt and FRESCO—to predict thermostable variants of the reductase component (C₁) of (4-hydroxyphenyl)acetate 3-hydroxylase. With the additional aid of experimental results, two C₁ variants, A166L and A58P, were identified as thermotolerant enzymes, with thermostability improvements of 2.6–5.6 °C and increased catalytic efficiency of 2- to 3.5-fold. After heat treatment at 45 °C, both of the thermostable C₁ variants remain active and generate reduced flavin mononucleotide (FMNH⁻) for reactions catalyzed by bacterial luciferase and by the monooxygenase C₂ more efficiently than the wild type (WT). In addition to thermotolerance, the A166L and A58P variants also exhibited solvent tolerance. Molecular dynamics (MD) simulations (6 ns) at 300–500 K indicated that mutation of A166 to L and of A58 to P resulted in structural changes with increased stabilization of hydrophobic interactions, and thus in improved thermostability. Our findings demonstrated that improvements in the thermostability of C₁ enzyme can lead to broad-spectrum uses of C₁ as a redox biocatalyst for future industrial applications.     

Dynamic composition estimation for a ternary batch distillation

A novel Kalman estimator has been proposed to provide the estimates of dynamic composition in a ternary batch distillation process operated in an optimal-reflux policy. The estimator is formulated based on a sequence of reduced-order process models representing a whole batch behavior. Therefore, the full-order models are first developed around different pseudo-steady-state operating conditions along batch optimal profiles. Then they reduce their orders to achieve all state observability and controllability by a balanced truncation method. In the estimator scheme, the reduced models as well as relevant covariance matrices of process noise are pre-scheduled and switched according to any desired periods. Four important issues have been studied including selection of a sensor frequency, effects of an integrating step size, a state initialization and a measurement noise. The performances of the reduced estimator have been investigated and compared with those of a conventional nonlinear estimator. Simulation results have demonstrated that the performances of the novel linear estimator are reasonably good and almost identical to the nonlinear estimator in all cases, though the linear estimator performs rather sensitively to the effect of high measurement noise. Nevertheless, it has been found to be applicable to implement in real plants with much lower computation effort, easier state initialization and unrequired a priori knowledge of thermodynamics.     

Roselle (Hibiscus sabdariffa L.) oil as an alternative feedstock for biodiesel production in Thailand

The production of biodiesel fuel from crude roselle oil was evaluated in this study. The process of alkali-catalyzed transesterification with methanol was carried out to examine the effects of reaction variables on the formation of methyl ester: variables which included methanol-to-oil molar ratios of 4:1-10:1, catalyst concentrations of 0.25-2.0% w/w of oil, reaction temperatures of 32-60 °C, and reaction times of 5-80 min. The methyl ester content from each reaction condition was analyzed by gas chromatography (GC), the optimum condition having been achieved at a methanol-to-oil molar ratio of 8:1, a catalyst concentration of 1.5% w/w of oil, a reaction temperature of 60 °C, and a reaction time of 60 min. The resultant methyl ester content of 99.4% w/w, plus all of the other measured properties of the roselle biodiesel, met the Thai biodiesel (B100) specifications and international standards EN 14214:2008 (E) and ASTM D 6751-07b, with the exception of a higher carbon residue and lower oxidation stability.     

High free fatty acid coconut oil as a potential feedstock for biodiesel production in Thailand

Coconut oil having 12.8% free fatty acid (FFA) was used as a feedstock to produce biodiesel by a two-step process. In the first step, FFA level of the coconut oil was reduced to 0.6% by acid-catalyzed esterification. In the second step, triglycerides in product from the first step were transesterified with methanol by using an alkaline catalyst to produce methyl esters and glycerol. Effect of parameters related to these processes was studied and optimized, including methanol-to-oil ratio, catalyst concentration, reaction temperature, and reaction time. Methyl ester content of the coconut biodiesel was determined by GC to be 98.4% under the optimum condition. The viscosity of coconut biodiesel product was very close to that of Thai petroleum diesel and other measured properties met the Thai biodiesel (B100) specification.