Modeling and Optimization of Anode‐Supported Solid Oxide Fuel Cells on Cell Parameters via Artificial Neural Network and Genetic Algorithm

An artificial neural network (ANN) and a genetic algorithm (GA) are employed to model and optimize cell parameters to improve the performance of singular, intermediate‐temperature, solid oxide fuel cells (IT‐SOFCs). The ANN model uses a feed‐forward neural network with an error back‐propagation algorithm. The ANN is trained using experimental data as a black‐box without using physical models. The developed model is able to predict the performance of the SOFC. An optimization algorithm is utilized to select the optimal SOFC parameters. The optimal values of four cell parameters (anode support thickness, anode support porosity, electrolyte thickness, and functional layer cathode thickness) are determined by using the GA under different conditions. The results show that these optimum cell parameters deliver the highest maximum power density under different constraints on the anode support thickness, porosity, and electrolyte thickness.     

Modeling and optimization of cross-flow ultrafiltration using hybrid neural network-genetic algorithm approach

Precise modeling flux decline under various operating parameters in cross-flow ultrafiltration (UF) of oily wastewaters and afterward, employing an appropriate optimization algorithm in order to optimize operating parameters involved in the process model result in attaining desired permeate flux, is of fundamental great interest from an economical and technical point of view. Accordingly, this current research proposed a hybrid process modeling and optimization based on computational intelligence paradigms where the combination of artificial neural network (ANN) and genetic algorithm (GA) meets the challenge of specified-objective based on two steps: first the development of bio-inspired approach based on ANN, trained, validated and tested successfully with experimental data collected during the polyacrylonitrile (PAN) UF process to treat the oily wastewater of Tehran refinery in a laboratory scale in which the model received feed temperature (T), feed pH, trans-membrane pressure (TMP), cross-flow velocity (CFV), and filtration time as inputs; and gave permeate flux as an output. Subsequently, the 5-dimensional input space of the ANN model portraying process input variables was optimized by applying GA, with a view to realizing maximum or minimum process output variable. The results obtained validate the estimates of the ANN–GA technique with a good accuracy. Finally, the relative importance of the controllable operation factors on flux decline is determined by applying the various correlation statistic techniques. According to the result of the sensitivity analysis based on the correlation coefficient, the filtration time was the most significant one, followed by T, CFV, feed pH and TMP.     

Experimental investigation,modeling and optimization of membrane separation using artificial neural network and multi-objective optimization using genetic algorithm

In this work, treatment of oily wastewaters with commercial polyacrylonitrile (PAN) ultrafiltration (UF) membranes was investigated. In order to do these experiments, the outlet wastewater of the API (American Petroleum Institute) unit of Tehran refinery, is used as the feed. The purpose of this paper was to predict the permeation flux and fouling resistance, by applying artificial neural networks (ANNs), and then to optimize the operating conditions in separation of oil from industrial oily wastewaters, including trans-membrane pressure (TMP), cross-flow velocity (CFV), feed temperature and pH, so that a maximum permeation flux accompanied by a minimum fouling resistance, was acquired by applying genetic algorithm as a powerful soft computing technique. The experimental input data, including TMP, CFV, feed temperature and pH, permeation flux and fouling resistance as outputs, were used to create ANN models. This fact that there is an excellent agreement between the experimental data and the predicted values was shown by the modeling results. Eventually, by multi-objective optimization, using genetic algorithm (GA), an optimization tool was created to predict the optimum operating parameters for desired permeation flux (i.e. maximum flux) and fouling resistance (i.e. minimum fouling) behavior. The accuracy of the model is confirmed by the comparison between the predicted and experimental data.     

Optimization of a Reverse Osmosis System Using Genetic Algorithm

Abstract

Reverse Osmosis (RO) has found extensive application in industry as a highly efficient separation process. In most cases, it is required to select the optimum set of operating variables such that the performance of the system is maximized. In this work, an attempt has been made to optimize the performance of RO system with a cellulose acetate membrane to separate NaCl‐Water system using Genetic Algorithm (GA). The GAs are faster and more efficient than conventional gradient based optimization techniques. The optimization problem was to maximize the observed rejection of the solute by varying the feed flowrate and overall permeate flux across the membrane for a constant feed concentration. To model the system, a well‐established transport model for RO system, the Spiegler‐Kedem model was used. It was found that the GA converged rapidly to the optimal solution at the 8th generation. The effect of varying GA parameters like size of population, crossover probability, and mutation probability on the result was also studied. The algorithm converged to the optimum solution set at the 8th generation. It was also seen that varying the computational parameters significantly affected the results.     

Effluent organic matter removal from reverse osmosis feed by granular activated carbon and purolite A502PS fluidized beds

Applying pre-treatments to remove dissolved organic matter from reverse osmosis (RO) feed can help to reduce organic fouling of the RO membrane. In this study the performance of granular activated carbon (GAC), a popular adsorbent, and purolite A502PS, an anion exchange resin, in removing effluent organic matter (EfOM) from RO feed collected from a water reclamation plant located at Sydney Olympic Park, Australia were evaluated and compared through adsorption equilibrium, kinetics and fluidized bed experiments. The maximum adsorption capacity (Q_max) of GAC calculated from the Langmuir model with RO feed was 13.4 mg/g GAC. The operational conditions of fluidized bed columns packed with GAC and purolite A502PS strongly affected the removal of EfOM. GAC fluidized bed with a bed height of 10 cm and fluidization velocity of 5.7 m/h removed more than 80% of dissolved organic carbon (DOC) during a 7 h experiment. The average DOC removal was 60% when the bed height was reduced to 7 cm. When comparing GAC with purolite A502PS, more of the later was required to remove the same amount of DOC. The poorer performance of purolite A502PS can be explained by the competition provided by other inorganic anions present in RO feed. A plug flow model can be used to predict the impact of the amount of adsorbent and of the flow rate on removal of organic matter from the fluidized bed column.     

Artificial Neural Network Modeling of the Deposition Rate of Lactose Powder in Spray Dryers

A spray dryer is the ideal equipment for the production of food powders because it can easily impart well-defined end product characteristics such as moisture content, particle size, porosity, and bulk density. Wall deposition of particles in spray dryers is a key processing problem and an understanding of wall deposition can guide the selection of operating conditions to minimize this problem. The stickiness of powders causes the deposition of particles on the wall. Operating parameters such as inlet air temperature and feed flow rate affect the air temperature and humidity inside the dryer, which together with the addition of drying aids can affect the stickiness and moisture content of the product and hence its deposition on the wall. In this article, an artificial neural network (ANN) method was used to model the effects of inlet air temperature, feed flow rate, and maltodextrin ratio on wall deposition flux and moisture content of lactose-rich products. An ANN trained by back-propagation algorithms was developed to predict two performance indices based on the three input variables. The results showed good agreement between predicted results using the ANN and the measured data taken under the same conditions. The optimum condition found by the ANN for minimum moisture content and minimum wall deposition rate for lactose-rich feed was inlet air temperature of 140°C, feed rate of 23 mL/min, and maltodextrin ratio of 45%. The ANN technology has been shown to be an excellent investigative and predictive tool for spray drying of lactose-rich products.     

Optimization of solar-powered reverse osmosis desalination pilot plant using response surface methodology

A solar thermal and photovoltaic-powered reverse osmosis (RO) desalination plant has been constructed and optimized for brackish water desalination. The central composite experimental design of orthogonal type and response surface methodology (RSM) have been used to develop predictive models for simulation and optimization of different responses such as the salt rejection coefficient, the specific permeate flux and the RO specific performance index that takes into consideration the salt rejection coefficient, the permeate flux, the energy consumption and the conversion factor. The considered input variables were the feed temperature, the feed flow-rate and the feed pressure. Analysis of variance (ANOVA) has been employed to test the significance of the RSM polynomial models. The optimum operating conditions have been determined using the step adjusting gradient method. An optimum RO specific performance index has been achieved experimentally under the obtained optimal conditions. The RO optimized plant guarantees a potable water production of 0.2 m³/day with energy consumption lower than 1.3 kWh/m³.     

Crevice corrosion behavior of high-alloy stainless steelsin a SWRO pilot plant

This paper presents the results of a study which has been carried out to investigate the crevice corrosion behavior of high-alloy stainless steel in a SWRO pilot plant. The study evaluated the corrosion performance of some austenitic and duplex steels in a crevice-forming environment created in a RO plant, especially in high-pressure feed and brine lines. The study of the effect of chemical dosing on crevice corrosion in the RO plant and electrochemical testing of crevice corrosion in the laboratory were the main objectives of this test program. High-alloy stainless steels, namely AL 6XN® and 254 SMO (superaustenitic), 2205, 2507 and DP3W (Duplex) were used in the test program. The tests were carried out in natural seawater and RO concentrate (conductivity: 75,000 to 80,000 μS/cm) at ambient temperature by operating the test plant at normal SWRO operating feed pressure of 54 bar. FeCl₃ was added as coagulant to maintain a silt density index of ∼3, and H₂SO₄ was added to feed in order to maintain the feed pH of ∼ 6.5. Chlorination and subsequent dechlorination agents were not added to the feed. For crevice corrosion tests in RO unit, the exposure periods were 6 and 12 months, respectively. The results of the tests showed that alloy DP3W has the best pitting resistance in crevice forming environment of seawater. In RO concentrate, alloys 2507 and 254 SMO showed lowest “maximum pit depth”. The results of potentiodynamic cyclic polarization (PCP) indicate that all the alloys have high pitting potential and small hysteresis loop. The results of critical crevice solution pH (CCSpH) indicated excellent resistance of alloys 254 SMO and DP3W against crevice corrosion attack and 2205 had least resistance in most aggressive sodium chloride solution.     

高温二级膜法处理高含盐量高COD工业废水研究

采用相转化法制得杂萘联苯聚芳醚砜酮超滤基膜,采用界面聚合法制成聚酰胺表层反渗透复合膜,对高含盐、高COD含量的工业废水进行处理实验。考察温度、压力等操作条件对二级反渗透废水处理系统眭能的影响。结果表明,60℃下,经过二级反渗透处理后,对总盐度的脱除率（以电导率计）达到93.9％,对COD的脱除率达到98.3％,在为期10d的处理实验中COD截留率变化幅度在1.2％以下,体现出良好的长期热稳定性能。     

Optimization and modeling of the performance of polydimethylsiloxane for pervaporation of ethanol−water mixture

Response surface methodology was used to optimize the performance of pervaporation of ethanol aqueous solution using polydimethylsiloxane hollow-fiber membrane. The effects of four operating conditions, that is, the feed temperature (30–50°C), the feed flow rate (10–50 L/h), ethanol concentration (5–20 wt%), and the vacuum pressure (10–50 KPa) on the membrane selectivity and the total flux of permeation were investigated with response surface methodology. The results showed that a quadratic model was suggested for both selectivity and total flux showing a high accuracy with R² = 0.9999 and 0.9995, respectively. The developed models indicated a significant effect of the four studied factors on both selectivity and total flux with some significant interactions between these factors. The optimum selectivity was 15.56, achieved for a feed temperature of 30°C, feed flow rate of 10 L/h, ethanol concentration of 15 wt%, and a permeate pressure of 10.74 KPa whereas the optimum total flux was 1833.66 g/m².h was observed for at a feed temperature of 50°C, a feed flow rate of 50 L/h, ethanol concentration of 15 wt%, and a permeate pressure of 49.38 KPa.     

The use of RO to remove emerging micropollutants following CAS/UF or MBR treatment of municipal wastewater

The removal of various organic micropollutants (OMPs), including six antibiotics (ERY, ROX, CLA, SMX, SMZ, and TMP), three pharmaceuticals (ibuprofen, salicylic acid, and diclofenac), one industrial product (BPA), and one hormone (cholesterol), was investigated in two pilot plants treating the same raw sewage of the Tel-Aviv WWTP. The effluent production by CAS-UF was 45 m³/h while that of MBR was 40 L/h. Each system's effluent constituted the feed for its RO, which comprised three RO steps after the CAS/UF and a semi-batch RO system after the MBR. Despite significant molecular differences between the selected OMPs, high removal rates were achieved after the RO stage (> 99% for macrolides, pharmaceuticals, cholesterol, and BPA, 95% for diclofenac, and > 93% removal of sulfonamides). However, low antibiotics concentrations and 28–223 ng/L residuals of ibuprofen, diclofenac, salicylic acid, cholesterol, and BPA in the MBR/RO and CAS-UF/RO permeates showed that although RO is an efficient removal solution, it cannot serve as an absolute barrier to OMPs. Therefore, additional treatment techniques should be considered to be incorporated aside the RO to ensure complete removal of such substances.     

Enrichment of glutaminase production by Bacillus subtilis RSP‐GLU in submerged cultivation based on neural network—genetic algorithm approach

BACKGROUND: Owing to the importance of glutaminase in biotech product production, its production with isolated Bacillus subtilis RSP‐GLU (MTCC 9727) was investigated. Fermentation factors play an important role in product enhancement. Hence, glutaminase production was optimized using an artificial neural network (ANN) coupled genetic algorithm (GA). RESULTS: A ‘6–12–1’ topology ANN was constructed to identify the nonlinear relationship between fermentation factors and enzyme yield. ANN‐predicted values were optimized for glutaminase production using a GA. The overall mean absolute predictive error (MAPE) and the mean square errors (MSE) were observed to be 0.00125 and 1.77 and 0.002 and 3.06 for training and testing, respectively. The goodness of neural network prediction (coefficient of R²) was found to be 0.996. The maximum interactive impact on glutaminase production was noted with rpm versus medium volume. The use of ANN–GA hybrid methodology resulted in a significant improvement (47%) in glutaminase yield. CONCLUSION: Five different optimum fermentation conditions out of 500 revealed maximum enzyme production. Glutaminase enzyme production in this Bacillus subtilis RSP‐GLU is strongly influenced by aeration of the fermentation. A hybrid ANN‐GA effectively identifies the different fermentation conditions for optimum production of enzyme in a given large set of conditions. Copyright © 2009 Society of Chemical Industry     

A pilot study for reclamation of a combined rinse from a nickel-plating operation using a dual-membrane UF/RO process

A pilot study for reclamation of a combined rinse from a nickel-plating operation was conducted using a dual-membrane UF/RO process. The pilot plant has a product capacity of 1.5 m³/h. The OF unit, as a pre-treatment, was operated at 90% water recovery. The RO unit was operated with a 2:1 configuration in a feed-and-bleed mode with recirculation. Trial runs on various operating pressures and water recoveries were conducted and effect of feed pH on RO permeate quality was studied. Finding a critical pH value was explored to design the feed pH for RO process to treat this particular wastewater. A long-term run for the RO unit with an optimized 75% water recovery was later carried out to study the stability of the product and the fouling tendency of membranes. The cleaning-in-place methods were investigated for both OF and RO units. The pilot plant had successfully operated for 6 months at the time of reporting, consistently producing a high-quality product water (<95 μS/cm) at an overall water recovery of 67.5%. The quality of reclaimed water was better than town water used at the factory. The product water from the pilot plant was used as a substitute for town water for in-process rinsing at the factory with no detrimental effects for 3 months. The study has successfully developed a process for recycling a combined rinse water with conductivity up to 1700 μS/cm. The design data for a full-scale plant were obtained. An economic evaluation shows that a payback within 32 months is feasible at a treatment capacity of 20 m³/h.     

Correlating thermal conductivity of pure hydrocarbons and aromatics via perceptron artificial neural network (PANN) method

Accurate estimation of liquid thermal conductivity is highly necessary to appropriately design equipments in different industries. Respect to this necessity, in the current investigation a feed-forward artificial neural network(ANN) model is examined to correlate the liquid thermal conductivity of normal and aromatic hydrocarbons at the temperatures range of 257–338 K and atmospheric pressure. For this purpose, 956 experimental thermal conductivities for normal and aromatic hydrocarbons are collected from different previously published literature.During the modeling stage, to discriminate different substances, critical temperature(Tc), critical pressure(Pc)and acentric factor(ω) are utilized as the network inputs besides the temperature. During the examination, effects of different transfer functions and number of neurons in hidden layer are investigated to find the optimum network architecture. Besides, statistical error analysis considering the results obtained from available correlations and group contribution methods and proposed neural network is performed to reliably check the feasibility and accuracy of the proposed method. Respect to the obtained results, it can be concluded that the proposed neural network consisted of three layers namely, input, hidden and output layers with 22 neurons in hidden layer was the optimum ANN model. Generally, the proposed model enables to correlate the thermal conductivity of normal and aromatic hydrocarbons with absolute average relative deviation percent(AARD), mean square error(MSE), and correlation coefficient(R~2) of lower than 0.2%, 1.05 × 10~(-7) and 0.9994, respectively.     

Near critical carbon dioxide extraction of Anise (Pimpinella Anisum L.) seed: Mathematical and artificial neural network modeling

In the current study, two models for estimating essential oil extraction yield from Anise, at high pressure condition, were used: mathematical modeling and artificial neural network (ANN) modeling. The extractor modeled mathematically using material balance in both fluid and solid phases. The model was solved numerically and validated with experimental data. Since the potential of near critical extraction is of consider able economic significance, a multi-layer feed forward ANN has been presented for accurate prediction of the mass of extract at this region of extraction. According to the network's training, validation and testing results, a three layer neural network with fifteen neurons in the hidden layer is selected as the best architecture for accurate prediction of mass of extract from Anise seed. Finally, the influence of pressure and solvent flow rate on the extraction kinetics was studied using ANN model and the optimum pressure range has been determined.     

Plasma assisted nitrogen oxide production from air: Using pulsed powered gliding arc reactor for a containerized plant

The production of NO_x from air and air + O₂ is investigated in a pulsed powered milli‐scale gliding arc (GA) reactor, aiming at a containerized process for fertilizer production. Influence of feed mixture, flow rate, temperature, and Ar and O₂ content are investigated at varying specific energy input. The findings are correlated with high‐speed imaging of the GA dynamics. An O₂ content of 40–48% was optimum, with an enhancement of 11% in NO_x production. Addition of Ar and preheating of the feed resulted in lower NO_x production. Lower flow rates produced higher NO_x concentrations due to longer residence time in the GA. The volume covered by GA depends strongly on the gas flow rate, emphasizing that the gas flow rate has a major impact on the GA dynamics and the reaction kinetics. For 0.5 L/min, 1.4 vol % of NO_x concentration was realized, which is promising for a containerized process plant to produce fertilizer in remote locations. © 2017 American Institute of Chemical Engineers AIChE J, 64: 526–537, 2018     

The integration of a reverse osmosis installation in the fully automated operation of a rural water supply company

A reverse osmosis plant was added to an existing municipal water treatment plant to provide sulphate ion removal capabilities. The RO plant consists of two trains each using a different type of RO membrane and producing 50 m³/h (316.8 kpgd). The plant is fully automated and operated by computer based on conductivity values of the blended product water and the water requirement. The remote water treatment plant is not staffed around the clock. All operating data are transmitted by radio to a central control room where the plant performance is monitored on CRT.     

基于神经网络和遗传算法的螺旋折流板换热器性能预测

将人工神经网络（ANN）应用于非连续螺旋折流板换热器的壳程换热和流阻分析。中试试验研究了具有3个螺旋角和2种管型的换热器。作为人工神经网络最常用的一种类型,将多层感知器神经网络（MLP）应用于本研究,使用一定的实验数据进行网络训练及预测。应用遗传算法（GA）对MLP的初始权值和阈值进行优化,预测结果精确。通过比较不同网络结构的预测误差来选择最适宜的网络结构为9-7-5-2。和关联结果比较可知MLP-GA网络对于换热器性能预测更加适合。此外,当使用MLP-GA方法在训练数据范围以外对壳程换热系数和压降进行预测时,网络预测结果和实验结果吻合程度也较高。因此,MLP-GA混合算法能够用来预测螺旋折流板管壳式换热器的传热和水力学性能。     

Using SDI,SDI+ and MFI to evaluate fouling in a UF/RO desalination pilot plant

This paper assesses the performance of a UF/RO demonstration plant located in the Oosterschelde estuary in the south-western part of the Netherlands. Spring blooms in the seawater pose a challenge to the plant because of the resulting increased fouling potential of the water. Determinations of the fouling indices SDI, SDI⁺ and MFI0.45 were carried out at the plant with different operational conditions, such as of coagulant addition and pH correction. Eight different membranes were used in the tests. In general, the UF performance was found to be good as the SDI values were around 1, provided standard membranes were used, and the MFI0.45 values lower than 1 s/L². The MFI0.45 showed the same tendency as the SDI in most cases. As expected, whereas the SDI showed marked sensitivity to used membrane type and operational conditions, the SDI⁺ did not display this dependency and hence appear to be a more reliable fouling index than the SDI. Storing the RO feed overnight in the feed tank increased the fouling potential of the RO feed, likely caused by continued coagulation.     

Fault Detection and Diagnosis in a Sour Gas Absorption Column Using Neural Networks

Process fault detection and diagnosis is an important problem in plant control at the supervisory level. It is the central component of abnormal event management which has attracted a lot of attention recently. In this study, the use of artificial neural networks (ANN) for fault detection is explored. An ANN can represent nonlinear and complex relations between its inputs (sensor measurements) and outputs (faults). As a test case, absorption of CO₂ gas in monoethanolamine (MEA) by a pilot plant called “automatic absorption and stripping pilot plant” is studied. For detecting and diagnosis of faults, variations in feed rate, feed composition, liquid absorber rate and composition are imposed onto the plant. The faults in this process influence variables such as the composition of absorbed gas (CO₂) and temperature and pressure drop of the column. The CO₂ concentration in the product should not exceed a certain limit. By selecting a proper architecture for the network (5‐9‐10), it is possible to detect the faults accurately. The network is trained using the back propagation method. The developed fault diagnosis algorithm is tested using data that has not been seen by the network.