Potential application of metal-organic frameworks (MOFs) for hydrogen storage: Simulation by artificial intelligent techniques

Metal-organic frameworks are a new class of materials for hydrogen adsorption/storage applications. The hydrogen storage capacity of this structure is typically related to pressure, temperature, surface area, and adsorption enthalpy. Literature provides no reliable correlation for estimating the hydrogen uptake capacity of MOFs from these easy-measured variables. Therefore, this study introduces several straightforward and accurate artificial intelligence (AI) techniques to fill this gap, initially determining the appropriate topology of AI-based methods, then comparing their performances by statistical criteria, and introducing the most accurate. This study used artificial neural networks, hybrid neuro-fuzzy systems, and support vector machines as estimators. The general regression neural networks (GRNN) with a spread of 7.92 × 10⁻⁴ shows the highest correlation with the literature data and provides a relative absolute deviation of 5.34%, mean squared error of 0.059, and coefficient of determination of 0.9946.     

Hydrogen solubility in aromatic/cyclic compounds: Prediction by different machine learning techniques

A systematic procedure based on adaptive neuro-fuzzy inference systems (ANFIS), artificial neural networks, and least-squares support vector machines develop to estimate hydrogen solubility in aromatic and cyclic compounds. The key features of these models are determined through a massive trial-and-error process. The proposed intelligent models estimate hydrogen solubility as a function of critical properties and acentric factor of aromatic/cyclic compounds, temperature, and pressure. The ranking analysis based on seven statistical criteria indicates the priority of the ANFIS method over other paradigms. The proposed ANFIS model estimates 278 experimental hydrogen solubility in eleven aromatic/cyclic compounds by the absolute average relative deviation of 7.88%, the mean absolute error of 0.0023, the relative absolute error of 5.05%, mean squared error of 2.74 × 10^?5, root mean squared error of 0.0052, and regression coefficient of 0.99664. Moreover, the relevancy analysis justifies that the pressure is the strongest influential variable for the hydrogen solubility in aromatic/cyclic compounds.     

Predicting total solar irradiation values using artificial neural networks

This study explores the possibility of developing an artificial neural networks model that could be used to predict monthly average daily total solar irradiation on a horizontal surface for locations in Uganda based on geographical and meteorological data: latitude, longitude, altitude, sunshine duration, relative humidity and maximum temperature. Results have shown good agreement between the predicted and measured values of total solar irradiation. A correlation coefficient of 0.997 was obtained with mean bias error of 0.018 MJ/m² and root mean square error of 0.131 MJ/m². Overall, the artificial neural networks model predicted with an accuracy of 0.1% of the mean absolute percentage error.     

Machine learning methods help accurate estimation of the hydrogen solubility in biomaterials

This study investigates the hydrogen solubility in five industrially relevant biochemicals, i.e., eugenol, furfural, furan, allyl alcohol, and furfuryl alcohol. Gene expression programming, three different artificial neural networks, and least-squares support vector regression (LS-SVR) are checked in this regard. The ranking analysis employing seven well-known statistical criteria confirmed that the LS-SVR is the most accurate model for the given purpose. The developed LS-SVR is superior to the Peng-Robinson, Soave-Redlich-Kwong, and perturbed-chain statistical associating fluid theory thermodynamic-based correlations proposed in the literature. The LS-SVR model predicts hydrogen solubility in the considered biochemicals with the relative absolute deviation of 1.91%, mean squared error of 6.4 × 10^?7, and regression coefficient of 0.99924. Both experimental and modeling observations approved that furan has the maximum tendency to absorb hydrogen molecules. A pure simulation analysis indicates that the maximum hydrogen solubility of 0.11 (mole fraction) can be absorbed by furan at pressure = 14.93 MPa and temperature = 402 K.     

Application of artificial neural networks (ANN) for modeling of industrial hydrogen plant

A back propagation feed forward artificial neural network (ANN) with three layers is used for modeling of industrial hydrogen plant. The required operating data for training of ANN is obtained by modeling and simulation of an industrial hydrogen plant. The operating data are calculated by changing effective parameters such as feed temperature, reformer pressure, steam to carbon ratio and carbon dioxide to methane ratio in feed stream. Tangent sigmoid transfer function is used in the hidden and output layer and the proposed neural network is trained with a gradient descent algorithm. The optimum number of neurons in hidden layer is determined as optimum value with minimizing of the mean square error (MSE). With changing of effective parameters, the model predicts temperature, pressure and mole fraction of hydrogen and carbon monoxide in the product of the hydrogen plant. The result can be used to gain better knowledge and optimize of the hydrogen production plants.     

Performance prediction of a RPF‐fired boiler using artificial neural networks

In order to provide adequate engineering assistance and to improve the energy efficiency in process industries, it is crucial to evaluate the operational performance of a boiler in terms of its practical requirements, viz. temperature, pressure, and mass flow rate of steam. This study was aimed at assessing and optimizing the performance of a refuse plastic fuel‐fired boiler using artificial neural networks. A feed‐forward back propagation neural network model was developed and trained using existing plant data (5 months), to predict temperature, pressure, and mass flow rate of steam, using the following input parameters: feed water pressure, feed water temperature, conveyor speed, and incinerator exit temperature. The predictive capability of the model was evaluated in terms of mean absolute percentage error between the model fitted and actual plant data, while sensitivity analysis was performed on the input parameters by determining the absolute average sensitivity values. The higher absolute average sensitivity value of the incinerator exit temperature in comparison to that of feed water pressure, feed water temperature and conveyor speed suggested that the change of incineration exit temperature has a significant influence on the selected outputs (steam properties). Overall, the good results observed from this work demonstrate the fact that artificial neural networks can efficiently predict the data on steam properties and could serve as a good tool to monitor boiler behavior under real‐time conditions. Copyright © 2013 John Wiley & Sons, Ltd.     

A hybrid intelligent model to predict the hydrogen concentration in the producer gas from a downdraft gasifier

This research work presents an artificial intelligence approach to predicting the hydrogen concentration in the producer gas from biomass gasification. An experimental gasification plant consisting of an air-blown downdraft fixed-bed gasifier fueled with exhausted olive pomace pellets and a producer gas conditioning unit was used to collect the whole dataset. During an extensive experimental campaign, the producer gas volumetric composition was measured and recorded with a portable syngas analyzer at a constant time step of 10 seconds. The resulting dataset comprises nearly 75 hours of plant operation in total. A hybrid intelligent model was developed with the aim of performing fault detection in measuring the hydrogen concentration in the producer gas and still provide reliable values in the event of malfunction. The best performing hybrid model comprises six local internal submodels that combine artificial neural networks and support vector machines for regression. The results are remarkably satisfactory, with a mean absolute prediction error of only 0.134% by volume. Accordingly, the developed model could be used as a virtual sensor to support or even avoid the need for a real sensor that is specific for measuring the hydrogen concentration in the producer gas.     

Prediction of monthly mean daily diffuse solar radiation using artificial neural networks and comparison with other empirical models

An artificial neural network (ANN) model for estimating monthly mean daily diffuse solar radiation is presented in this paper. Solar radiation data from 9 stations having different climatic conditions all over China during 1995–2004 are used for training and testing the ANN. Solar radiation data from eight typical cities are used for training the neural networks and data from the remaining one location are used for testing the estimated values. Estimated values are compared with measured values in terms of mean percentage error (MPE), mean bias error (MBE) and root mean square error (RMSE). The results of the ANN model have been compared with other empirical regression models. The solar radiation estimations by ANN are in good agreement with the actual values and are superior to those of other available models. In addition, ANN model is tested to predict the same components for Zhengzhou station over the same period. Results indicate that ANN model predicts the actual values for Zhengzhou with a good accuracy of 94.81%. Data for Zhengzhou are not included as a part of ANN training set. Hence, these results demonstrate the generalization capability of this approach and its ability to produce accurate estimates in China.     

On comparing three artificial neural networks for wind speed forecasting

Wind speed forecasting is critical for wind energy conversion systems since it greatly influences the issues such as the scheduling of a power system, and the dynamic control of the wind turbine. In this paper, we present a comprehensive comparison study on the application of different artificial neural networks in 1-h-ahead wind speed forecasting. Three types of typical neural networks, namely, adaptive linear element, back propagation, and radial basis function, are investigated. The wind data used are the hourly mean wind speed collected at two observation sites in North Dakota. The performance is evaluated based on three metrics, namely, mean absolute error, root mean square error, and mean absolute percentage error. The results show that even for the same wind dataset, no single neural network model outperforms others universally in terms of all evaluation metrics. Moreover, the selection of the type of neural networks for best performance is also dependent upon the data sources. Among the optimal models obtained, the relative difference in terms of one particular evaluation metric can be as much as 20%. This indicates the need of generating a single robust and reliable forecast by applying a post-processing method.     

Modeling of a continuous fluidized bed dryer using artificial neural networks

The work involves experimentation on drying of solids in a continuous fluidized bed dryer covering different variables like bed temperature, gas flow rate, solids flow rate and initial moisture content of solids. The data are modeled using artificial neural networks. The results obtained from artificial neural networks are compared with those obtained using Tanks-in-series model. It was found that results obtained from ANN fit the experimental data more accurately compared to the RTD model with less percentage error. This indicates a better fit of artificial neural networks to experimental data compared to various mathematical models.     

Catalytic upgrading of 4-methylaniosle as a representative of lignin-derived pyrolysis bio-oil: Process evaluation and optimization via coupled application of design of experiment and artificial neural networks

Catalytic upgrading of 4-methylanisole as a representative of lignin-derived pyrolysis bio-oil was investigated over Pt/γ-Al₂O₃ catalyst. The catalytic upgrading process was conducted at different operating condition to determine the detailed reactions network. Additionally, artificial neural network and design of experiment were applied by feeding the reaction temperature, operating pressure and space velocity to predict 4-methylanisole conversion, main products selectivity, reactions rate and reactions network. The main products of 4-methylanisole upgrading were toluene, phenol derivatives, cyclohexanone, 4-methylcyclohexanone, and 2-tert-butyl-4-methylphenol. The major classes of reactions during the upgrading process were hydrogenolysis, hydrodeoxygenation, alkylation, and hydrogenation. For optimization of experimental data obtained at suggested conditions by design of experiment, the response surface methodology was applied. Artificial neural network model was used to investigate the kinetics behavior of the system due to the complex nature of system. A combination of the response surface methodology, artificial neural network, and design of experiment has revealed its ability to solve a quadratic polynomial model. The coefficients of determination were close to 1, and the mean square error of the artificial neural network model was close to 0 which showed the high accuracy of model predictions. It was inferred that during the upgrading process of 4-methylanisole, increasing temperature and pressure and setting space velocity at the minimum value are the reasons to come close to the optimum reaction rate. The comparison of experimental results with simulated data from the artificial neural network and the response surface methodology models illustrated that the developed model can create an applicable situation for practical design of large-scale production of valuable fuels from renewable resources.     

Artificial neural networks and neuro-fuzzy inference systems as virtual sensors for hydrogen safety prediction

Hydrogen is increasingly investigated as an alternative fuel to petroleum products in running internal combustion engines and as powering remote area power systems using generators. The safety issues related to hydrogen gas are further exasperated by expensive instrumentation required to measure the percentage of explosive limits, flow rates and production pressure. This paper investigates the use of model based virtual sensors (rather than expensive physical sensors) in connection with hydrogen production with a Hogen^®20 electrolyzer system. The virtual sensors are used to predict relevant hydrogen safety parameters, such as the percentage of lower explosive limit, hydrogen pressure and hydrogen flow rate as a function of different input conditions of power supplied (voltage and current), the feed of de-ionized water and Hogen^®20 electrolyzer system parameters. The virtual sensors are developed by means of the application of various Artificial Intelligent techniques. To train and appraise the neural network models as virtual sensors, the Hogen^®20 electrolyzer is instrumented with necessary sensors to gather experimental data which together with MATLAB neural networks toolbox and tailor made adaptive neuro-fuzzy inference systems (ANFIS) were used as predictive tools to estimate hydrogen safety parameters. It was shown that using the neural networks hydrogen safety parameters were predicted to less than 3% of percentage average root mean square error. The most accurate prediction was achieved by using ANFIS.     

A new approach for the prediction of the heat transfer rate of the wire-on-tube type heat exchanger––use of an artificial neural network model

This study presents an application of artificial neural networks (ANNs) to predict the heat transfer rate of the wire-on-tube type heat exchanger. A back propagation algorithm, the most common learning method for ANNs, is used in the training and testing of the network. To solve this algorithm, a computer program was developed by using C++ programming language. The consistence between experimental and ANNs approach results was achieved by a mean absolute relative error <3%. It is suggested that the ANNs model is an easy modeling tool for heat engineers to obtain a quick preliminary assessment of heat transfer rate in response to the engineering modifications to the exchanger.     

Effects of a Nanofluid and Magnetic Field on the Thermal Efficiency of a Two‐Phase Closed Thermosyphon

Heat transfer of a CuO/water nanofluid in a two‐phase closed thermosyphon (TPCT) that is thermally enhanced by a magnetic field has been predicted by an optimized artificial neural network (ANN). The magnetic field strength, volume fraction of nanoparticles in water, and inlet power were used as input parameters and the thermal efficiency was used as the output parameters. The correlation coefficient (R² = 0.924), mean square error (MSE = 0.000340231), mean absolute error (MAE = 0.012410941), and normalized mean‐squared error (NMSE = 0.112417498) between the measured and predicted thermal efficiency by the ANN model were developed. The results were compared with experimental data and it was found that the thermal efficiency estimated by the multi‐layer perception neural network is accurate. In this study, a new approach for the auto‐design of neural networks, based on a genetic algorithm, has been used to predict collection output of a TPCT. © 2013 Wiley Periodicals, Inc. Heat Trans Asian Res, 42(7): 630–650, 2013; Published online in Wiley Online Library ( wileyonlinelibrary.com/journal/htj ). DOI 10.1002/htj.21043     

A correlation development for predicting the pressure drop of various refrigerants during condensation and evaporation in horizontal smooth and micro-fin tubes

This paper predicts the condensation and evaporation pressure drops of R32, R125, R410A, R134a, R22, R502, R507a, R32/R134a (25/75 by wt%), R407C and R12 flowing inside various horizontal smooth and micro-fin tubes by means of the numerical techniques of artificial neural networks (ANNs) and non-linear least squares (NLS). In its analyses, this paper used experimental data from the National Institute of Standards and Technology (NIST) and Eckels and Pate, as presented in Choi et al.'s study provided by NIST. In their experimental setups, the horizontal test sections have 1.587, 3.78, 3.81 and 3.97 m long countercurrent flow double tube heat exchangers with refrigerant flowing in the inner smooth (8, 8.01 and 11.1 mm i.d.) and micro-fin (4.339, 5.45, 7.43 and 8.443 mm i.d.) copper tubes and cooling water flowing in the annulus. Their test runs cover a wide range saturation temperatures, vapor qualities and mass fluxes. The pressure drops are calculated with 1485 measured data points, together with analyses of artificial neural networks and non-linear least squares numerically. Inputs of the ANNs of the best correlation are the measured values of the test sections, such as mass flux, tube length, inlet and outlet vapor qualities, critical pressure, latent heat of condensation, mass fraction of liquid and vapor phases, dynamic viscosities of liquid and vapor phases, hydraulic diameter, two-phase density and the outputs of the ANNs, which comprise the experimental total pressure drops of the evaporation and condensation data from independent laboratories. The total pressure drops of in-tube condensation and in-tube evaporation tests are modeled using the artificial neural network (ANN) method of multi-layer perceptron (MLP) with 12-40-1 architecture. Its average error rate is 7.085%, which came from the cross validation tests of 1485 evaporation and condensation data points. Dependency of the output of the ANNs from 12 numbers of input values is also shown in detail, and new ANN based empirical pressure drop correlations are developed separately for the conditions of condensation and evaporation in smooth and micro-fin tubes as a result of the analyses. In addition, a single empirical correlation for the determination of both evaporation and condensation pressure drops in smooth and micro-fin tubes is proposed with an error rate of 14.556%.     

Optimization of fermentative hydrogen production process using genetic algorithm based on neural network and response surface methodology

A central composite design was carried out to investigate the effect of temperature, initial pH and glucose concentration on fermentative hydrogen production by mixed cultures in batch test. The modeling abilities of the response surface methodology model and neural network model, as well as the optimizing abilities of response surface methodology and the genetic algorithm based on a neural network model were compared. The results showed that the root mean square error and the standard error of prediction for the neural network model were much smaller than those for the response surface methodology model, indicting that the neural network model had a much higher modeling ability than the response surface methodology model. The maximum hydrogen yield of 289.8 mL/g glucose identified by response surface methodology was a little lower than that of 360.5 mL/g glucose identified by the genetic algorithm based on a neural network model, indicating that the genetic algorithm based on a neural network model had a much higher optimizing ability than the response surface methodology. Thus, the genetic algorithm based on a neural network model is a better optimization method than response surface methodology and is recommended to be used during the optimization of fermentative hydrogen production process.     

Data-driven modelling and optimization of hydrogen adsorption on carbon nanostructures

In the present study, using modelling based on experimental data, models for predicting the hydrogen adsorption isotherm were presented. The three Automatic Learning of Algebraic Models (ALAMO), feed-forward artificial neural networks (ANNs), and group method of data handling-type polynomial neural networks (GMDH-PNN) were constructed. The created models were evaluated to predict the equilibrium data of hydrogen storage on carbon nanostructures, including activated carbons doped with palladium (Pd) nanoparticles, fullerene pillared graphene nanocomposites, and nickel (Ni)-decorated carbon nanotubes. The inputs were nanostructure characteristics such as surface area, pore-volume, and thermodynamic conditions such as pressure. The generalization of the trained models was acceptable, and the models successfully predicted the hydrogen adsorption isotherm for new inputs. The relative error percentage for most data points is less than 4%, which demonstrates their applicability in determining adsorption isotherms for any operating conditions. By performing error analysis calculations, it was shown that the ALAMO model has the highest accuracy. Also, sensitivity analysis calculations show that pressure is the most influential parameter in the adsorption process. Besides, by performing Genetic Algorithm (GA) optimization using the ALAMO model, the amount of pressure and adsorbent properties were determined so that the amount of hydrogen adsorption is maximized. According to the optimization results based on the GA, the higher the pressure, the greater the amount of hydrogen adsorption. The nanotubes with a surface area of 194.15 m²/g, a total volume of 1.8 cm³/g, micropore volume of 0.097 cm³/g, and mesopore volume of 0.963 cm³/g, graphene with a surface area of 2977.13 m²/g, a total volume of 1.5134 cm³/g, density of 617.45 kg/m³, and activated carbon at pressures less than 30 bar with a surface of 2546.36 m²/g, a total volume of 1.237 cm³/g, micropore volume of 0.839 cm³/g, and activated carbon at pressures more than 30 bar with a surface of 3027 m²/g, a total volume of 1.343 cm³/g, a micropore volume of 0.9582 cm³/g, and a mesopore volume of 1.23 cm³/g, have the highest amount of stored hydrogen.     

Estimation of monthly average daily global solar irradiation using artificial neural networks

This study explores the possibility of developing a prediction model using artificial neural networks (ANN), which could be used to estimate monthly average daily global solar irradiation on a horizontal surface for locations in Uganda based on weather station data: sunshine duration, maximum temperature, cloud cover and location parameters: latitude, longitude, altitude. Results have shown good agreement between the estimated and measured values of global solar irradiation. A correlation coefficient of 0.974 was obtained with mean bias error of 0.059 MJ/m² and root mean square error of 0.385 MJ/m². The comparison between the ANN and empirical method emphasized the superiority of the proposed ANN prediction model.     

Prediction of convection heat transfer in converging–diverging tube for laminar air flowing using back-propagation neural network

The ability of an artificial neural network (ANN) model for heat transfer analysis in a converging–diverging tube is studied. Back propagation learning algorithm, the most common method for ANNs, was used in training and testing/validation the network. It is trained with selected values of the Reynolds numbers (Re), Prandtl numbers (Pr), half taper angle (θ), aspect ratio (L_cyc/D_max), and Nusselt number (Nu). The trained network is the used to make predictions of the Nusselt numbers. The accuracy between selected data and ANNs results was achieved with a mean absolute relative error less than 1.5%. This shows that well trained neural network model provided fast, accurate and consistent results.     

Prediction of hydrogen production by magnetic field effect water electrolysis using artificial neural network predictive models

Developing an efficient water electrolysis (WE) configuration is essential for high-efficiency hydrogen evolution reaction (HER) activity. In this regard, it has been proven that adding a magnetic field (MF) to the electrolysis system greatly improves the hydrogen output rate. In this study, we developed a method based on a machine learning approach to further improve the hydrogen production (HP) system with MF effect WE. An artificial neural network (ANN) model was developed to estimate the effect of input parameters such as MF, electrode material (cathode type), electrolyte type, supplied power (onset voltage), surface area, temperature, and time on HP in different electrolyzer systems. The network was built using 104 experimental data sets from various electrolysis studies. In the study, the percentage contributions of the input parameters to the HP rate and the optimum network architecture to minimize computation time and maximize network accuracy are presented. The model architecture of 7–12–1 was obtained using the best-hidden neurons. The Levenberg-Marquardt (LM) algorithm was used to train the multi-layer feed-forward neural network. Moreover, the utilization of a range of categorical variables to improve ANN prediction accuracy is a significant novelty in this work. Results demonstrated that the output of the trained ANN model fitted well with the experimental data. The test's correlation coefficient (R) and mean squared error (MSE) were 0.973 and 0.01125, respectively, confirming its powerful predictive performance. This ANN application is the first novel viable model to perform prediction using a neural network algorithm in the electrolysis process for MF effect HP using both categorical and continuous data inputs.