Cascade Feed-forward Neural Network Research Articles

Modeling and estimation of water activity for the ionic-liquid-based aqueous ternary systems by smart paradigms

In this study, different artificial intelligence (AI) techniques are employed to determine the water activity of ionic liquid-based ternary systems as a function of pressure, temperature, and the type and concentration of the components. The proposed model has been selected among five artificial neural networks (ANNs), bagged trees (BT), decision trees (DT), gaussian process regression (GPR), adaptive neuro-fuzzy inference systems (ANFIS), and least square support vector machines (LS-SVM) by trial-and-error process and ranking analysis of statistical and computational indicators. The most accurate model is the cascade feed-forward neural network with fifteen hidden neurons which are equipped with logarithm and hyperbolic tangent sigmoid transfer functions. This model estimates 1829 empirical water activity for ionic liquid-based ternary systems with AARD=0.25 %, MSE=1.73 × 10−5, RMSE=4.16 × 10−3, and R2=0.9919.

New achievements on daily reference evapotranspiration forecasting: Potential assessment of multivariate signal decomposition schemes

Reference evapotranspiration (ETo) is a vital climate parameter affecting plants' water use. ETo can generate large deficits in soil moisture and runoff in different regions and seasons, leading to uncertainties in drought warning systems. A novel multivariate variational mode decomposition integrated with a boosted regression tree (i.e., MVMD-BRT) is constructed to forecast daily ETo. Firstly, the correlation matrix based on cross-correlation was computed to investigate the significant input predictor lags of daily ETo. Secondly, the MVMD technique decomposes the significant input lags into signals called intrinsic mode functions (IMFs). Thirdly, the IMFs were then employed in the BRT to build the MVMD-BRT model for daily ETo forecasting. A comparative assessment of MVMD against multivariate empirical mode decomposition (MEMD) was also performed on the same lines to develop the MEMD-BRT model. The MVMD-BRT model is compared against the random forest (RF) and hybrid MVMD-RF, MEMD-RF, extreme learning machine (ELM), and hybrid MVMD-ELM, MEMD-ELM, and cascaded feedforward neural network (CFNN) along with its hybrid MVMD-CFNN models for two stations in Queensland, Australia using a set of goodness-of-fit metrics. The results prove that the MVMD-BRT provide accurate daily ETo forecasting against the benchmark models. The MVMD-BRT model yielded the highest accuracy in terms of (WIE = 0.9070, NSE = 0.8421, LME = 0.6529, KGE = 0.8792) and (WIE = 0.8966, NSE = 0.8396, LME = 0.6521, KGE = 0.8803) for Brisbane and Gympie stations against the comparing models.

Applying feature selection and machine learning techniques to estimate the biomass higher heating value

The biomass higher heating value (HHV) is an important thermal property that determines the amount of recoverable energy from agriculture byproducts. Precise laboratory measurement or accurate prediction of the HHV is essential for designing biomass conversion equipment. The current study combines feature selection scenarios and machine learning tools to establish a general model for estimating biomass HHV. Multiple linear regression and Pearson’s correlation coefficients justified that volatile matter, nitrogen, and oxygen content of biomass samples have a slight effect on the HHV and it is better to ignore them during the HHV modeling. Then, the prediction performance of random forest, multilayer and cascade feedforward neural networks, group method of data handling, and least-squares support vector regressor are compared to determine the intelligent estimator with the highest accuracy toward biomass HHV prediction. The ranking test shows that the multilayer perceptron neural network better predicts the HHV of 532 biomass samples than the other intelligent models. This model presents the outstanding absolute average relative error of 2.75% and 3.12% and regression coefficients of 0.9500 and 0.9418 in the learning and testing stages. The model performance is also superior to a recurrent neural network which was recently developed in the literature using the same databank.

The hydrogen sensing capability of zinc oxide-containing sensors: Modeling by the general regression artificial neural network

This research applies general regression (GR) neural networks to model the hydrogen sensing capability of pure zinc oxide (ZnO) and ZnO nanocomposites containing different additives. The sensor's topology/composition, temperature, and hydrogen concentration in the atmosphere are the independent features of the GR models. Sensitivity analysis approves that the Gaussian spread value must adjust to 23.5 × 10−6 to achieve the most accurate predictions. This model forecasts 297 actual responses of ten ZnO-containing sensors with mean absolute error, relative absolute error, and regression coefficient of 0.29, 1.56%, and 0.9977, respectively. These prediction accuracies are better than those obtained by cascade feedforward, radial basis, Elman recurrent, and multilayer perceptron neural networks. The relevancy test shows that the sensor's hydrogen detection ability improves with the sensor geometry and ZnO additive dose. ZnO–Co3O4 is the best sensor composition to detect even low hydrogen concentrations. Also, the nanowire and nanofiber are the best sensor topologies for the considered task.

Application of artificial neural network models and random forest algorithm for estimation of fracture intensity from petrophysical data

Natural fractures play an essential role in the characterization and modeling of hydrocarbon reservoirs. Modeling fractured reservoirs requires an understanding of fracture characteristics. Fractured zones can be detected by using seismic data, petrophysical logs, well tests, drilling mud loss history and core description. In this study, the feed-forward neural networks (FFNN), cascade feed forward neural networks (CFFN) and random forests (RF) were used to determine fracture density from petrophysical logs. The model performance was assessed using statistical measures including the root mean squared error (RMSE), coefficient of determination (R2), mean absolute error (MAE), Kling Gupta efficiency (KGE) and Willmott’s index (WI). Conventional good logs and full-bore micro-resistivity imaging data were available from three drilled wells of the Mozduran reservoir, Khangiran gas field. According to the findings of this research, the FFNN model showed a higher KGE and WI, and a higher correlation coefficient (R2) compared to the CFNN model. The CFNN model outperformed the FFNN model with lower neurons. The models' performance was also improved by increasing the number of neurons in the hidden layers from 8 to 35. The findings of this study demonstrate that the measured and FFNN calculated fracture intensity is in excellent agreement with image log results showing a correlation coefficient of 92%. The RF algorithm showed higher stability and robustness in predicting fracture intensity with a correlation coefficient of 93%. The results of this study can successfully be used as an aid in a more successful reservoir dynamic modeling and production data analysis.

Experimental Investigation and Modeling of Fluid and Carbonated Rock Interactions with EDTA Chelating Agent during EOR Process

The injection of chemical fluids into oil reservoirs is gaining widespread attention in light of the declining conventional oil resources by recovering more hydrocarbons. This study is focused on using a chemical called ethylenediaminetetraacetic acid (EDTA) chelating agent in a carbonate reservoir to shed light on contact angle differences of 625 aged thin sections and rock dissolution under the influence of different pHs, temperatures, chelating times, and various chelating agent concentrations in seawater. According to a rock dissolution test, at least 5 wt % of EDTA chemical is needed to obtain oil recovery. A ζ potential test and scanning electron microscopy (SEM) images revealed that the mechanism of adsorption at low pH values and the expansion of the electrical double layer (EDL) at high pH values were responsible for wettability alteration, and an increase in EDTA concentration intensified each mechanism. Interfacial tension (IFT) measurements also showed that adding 1 and 10 wt % of the EDTA to the seawater solution reduced the IFT by 67.75% and 76.08%, respectively. The contact angle experiments demonstrated an increase in the mechanism that leads rock to behave more hydrophilically as pH, solution temperature, and chelating agent concentration in saltwater increased. Artificial neural network (ANN) methods also led to the introduction of a model to predict the contact angle employing multilayer perceptron neural networks (MPNN) and cascade feedforward neural networks (CFFNN). The CFFNN with two hidden neurons and trained by the Levenberg–Marquardt backpropagation algorithm is the most accurate model when comparing the accuracy of models for predicting contact angle values. The CFFNN model indicated that the weight percentage of the chelating chemical, which has a share of about 90%, had the greatest influence on the contact angle, and chelating time, with a share of less than 10%, had the least.

Cascade feedforward neural network and deep neural network controller on photovoltaic system with cascaded multilevel inverters: Comparison on standalone and grid integrated system

The introduction of a micro-grid-based power generation network will help to meet the demands of consumers while reducing environmental impact. Several industrialized and emerging countries allocate considerable resources to renewable energy-based power generation and invest significant sums of money in this area. This study examines the challenges involved with electricity generation through photovoltaic (PV) systems and the integration of the same with the grid to mitigate power quality issues and improve the power factor for various loading conditions. An innovative multilayer inverter for grid-connected PV systems has been developed to enhance the voltage profile and resulted in a drop in total harmonic distortion (THD). A cascade multilevel inverter (associated with a grid-integrated PV system and managed using multiple innovative artificial intelligence controllers) was developed in this research project. Various advanced intelligent controllers, such as cascade feedforward neural networks (CFFNN) and deep neural networks (DNN), have been analyzed under various operating situations and observed that the THD of voltage, current at the grid, and the load is less than 3 % as per the IEEE 519 standards along with this power factor is maintained nearly unity for the grid-connected system. The quality of power in terms of voltage, frequency, total harmonics distortion, and power factor are improved by using a novel deep neural network algorithm in a cascaded multilevel inverter and verified according to IEEE 1547 and IEEE 519 standards to determine the efficacy of the proposed system.

ECG Analysis-Based Cardiac Disease Prediction Using Signal Feature Selection with Extraction Based on AI Techniques

ECG (Electrocardiogram) performs classification using a machine learning model for processing different features in the ECG signal. The electrical activity of the heart is computed with the ECG signal with machine learning library. The key issue in the handling of ECG signals is an estimation of irregularities to evaluate the health status of patients. The ECG signal evaluate the impulse waveform for the specialized tissues in the cardiac heart diseases. However, the ECG signal comprises of the different difficulties associated with waveform estimation to derive certain features. Through machine learning (ML) model the input features are computed with input ECG signals. In this paper, proposed a Noise QRS Feature to evaluate the features in the ECG signals for the effective classification. The Noise QRS Feature model computes the ECG signal features of the waveform sequences. Initially, the signal is pre-processed with the Finite Impulse response (FIR) filter for the analysis of ECG signal. The features in the ECG signal are processed and computed with the QRS signal responses in the ECG signal. The Noise QRS Feature evaluate the ECG signal with the kNN for the estimation and classification of features in the ECG signals. The performance of the proposed Noise QRS Feature features are comparatively examined with the Discrete Wavelet Transform (DWT), Dual-Tree Complex Wavelet Transforms (DTCWT) and Discrete Orthonormal Stockwell Transform (DOST) and the machine learning model Cascade Feed Forward Neural Network (CFNN), Feed Forward Neural Network (FFNN). Simulation analysis expressed that the proposed Noise QRS Feature exhibits a higher classification accuracy of 99% which is ~6 – 7% higher than the conventional classifier model.

Cascaded Feedforward Neural Network Based Simultaneously Fast and Precise Multi-Characteristics Extraction and BFS Error Estimation

We present and investigate a scheme to realize fast and precise multi-characteristics extraction and error estimation of Brillouin frequency shift (BFS) simultaneously based on a cascaded feedforward neural network for a Brillouin optical time-domain analyzer (BOTDA) system. The cascaded feedforward neural network is composed of two neural networks in series, where the pre-network is used to extract linewidth (FWHM), signal-to-noise ratio (SNR) and BFS from noised gain spectrum and the post-network utilizes these characteristics and known frequency step to estimate the statistical error of the extracted BFS. The extracted BFS can be used to obtain the environment information along the fiber, while the extracted SNR, FWHM and the estimated BFS error can be utilized to verify the validity of data and monitor the performance of the BOTDA system. The creation process and the performance of the pre-network, post-network and cascaded network are described and analyzed in detail. Both simulation and experimental results show that the cascaded feedforward neural network has higher accuracies on extractions of characteristics (BFS, SNR, FWHM) and error estimation than conventional Lorentzian curve fitting method and error estimation formula when the scanning frequency step of system is large or SNR is low. More importantly, the network is more than 300 times faster than the LCF for simultaneously multi-characteristics extraction and error estimation, which make it possible to real-time monitor the performance of BOTDA system.

Modeling and estimation of fouling factor on the hot wire probe by smart paradigms

The fouling factor is a complex non-linear function of several feature variables. Accurate estimation of this measurable factor is applicable for controlling the heat transfer efficiency of heat exchangers. Artificial neural network (ANN) and adaptive neural-based fuzzy inference system (ANFIS) techniques are accurate approaches for the understanding treatment of the most complicated systems. The cascade-feedforward (CFF), multi-layer perceptron (MLP), radial basis function (RBF), and generalized regression (GR) neural networks and ANFIS were applied for predicting the fouling factor of the hot wire probe from some measured variables using a huge databank, including 1870 empirical dataset. Pearson's and Spearman's techniques confirmed the highest relation between considered features and the first order of fouling factor. The results demonstrate that the cascade feed-forward neural network containing eight hidden neurons was the best artificial intelligence (AI) model with excellent overall AARE = 3.44 %, MSD = 0.0000315, RMSD = 0.0056, and R2 = 0.9982. This work's significance lies in presenting an applicable and accurate tool for estimating fouling factors on hot wire probes to the research community and industry. This model can be considered a reliable replacement for empirical analyses that are often expensive and time-consuming.

Modeling the ammonia absorption capacity of choline chloride-based deep eutectic solvents with artificial neural networks

The ammonia (NH3) absorption capacity of choline chloride-based deep eutectic solvents (DESs) was deeply investigated in laboratory-scale, but little attention was devoted to its modeling. Therefore, the current work aims to cover this gap by developing an artificial neural network (ANN) to estimate the NH3 removal capacity of DESs. Indeed, the prediction accuracy of six ANN variants (multilayer perceptron, wavelet, general regression, radial basis function, cascade feedforward, and recurrent) has been compared to find the model with the lowest errors. Results justified that the cascade feedforward neural network (CFFNN) with only ten hidden nodes is the highest accurate paradigm to predict NH3 solubility in DES-based media. This model simulates 793 NH3 solubility data in eleven DESs with the average absolute relative deviation (AARD = 4.98%), mean absolute error (MAE = 0.163), relative absolute error (RAE = 7.44%), root mean squared error (MSE = 0.318), and regression coefficient (R2 = 0.995). Additionally, the influence of temperature, pressure, and DES composition (water content, choline chloride molar dosage, and type and molar dosage of hydrogen bond donor) on NH3 solubility is comprehensively analyzed by the CFFNN. Increasing the pressure and HBD molar dosage and decreasing the temperature and choline chloride molar dosage enhance NH3 solubility in DES. Pure simulation findings clarified that the NH3 capture capacity of DESs increases by increasing the water content to a critical value, and after that, DES's water content reduces the ammonia solubility. Since glycerol has the maximum number of hydroxyl groups among the investigated HBDs, the choline chloride + glycerol DES showed the highest overall capacity for absorbing NH3 molecules. The simulation results approved that the unit mass of this DES can absorb 14.6 mol of NH3 at 298.2 K and 573.2 kPa.

Short Term Load Forecasting of Residential and Commercial Consumers of Karnataka Electricity Board using CFNN

Electricity use and its access are correlated in the economic development of any country. Economically, electricity cannot be stored, and for stability of an electrical network a balance between generation and consumption is necessary. Electricity demand depends on various factors like temperature, everyday activities, time of day, days of the week days/Holidays. These parameters have led to price volatility and huge spikes in electricity prices. The research work proposes a short term Load prediction Model for LT2 (residential consumers), LT3 (Commercial Consumers) of Karnataka State Electricity Board using Cascaded Feed Forward Neural Network (CFNN). MATLAB software is utilized to design and test the forecasting model for predicting the power consumption. Furthermore, a shallow feed forward neural network-based prediction model is constructed and evaluated for performance comparison. The Performance metrics include Mean Absolute Percentage Error (MAPE) and Mean Squared Error (MSE). The suggested STLF CFNN prediction model outperformed shallow feed forward networks on both performance metrics with prediction errors of less than 1%.

On the evaluation of the interfacial tension of immiscible binary systems of methane, carbon dioxide, and nitrogen-alkanes using robust data-driven approaches

Gas injection has emerged over the recent decades as a promising technology to enhance oil recovery in various fields worldwide. The efficiency and success of a gas injection operation can be assessed through a number of vital experimental studies. Interfacial Tension (IFT) between the injected gas and the displacing fluid is a key parameter playing an eminent role in the foregoing studies. The main scope of this work is making a progress in modeling the IFTs between diverse n-alkanes and Methane (CH4), Carbon Dioxide (CO2), and Nitrogen (N2) natural gases. For this purpose, two smart AI-based approaches of Cascaded Feedforward Neural Network (CFNN) and Decision Tree Learning (DT) were used to simultaneously model the IFTs between foregoing immiscible binary systems as a function of pressure, temperature, the gases properties, and the properties of the liquid. Several statistical measures and graphical descriptions were employed to aid the accuracy analysis of the proposed models. Both developed CFNN and DT networks represented desirable close-to-reality predictions in all binary systems. Besides, CFNN established itself as the most robust model for all studies binary systems with RMSE values of 0.5924, 0.5649, and 0.5870 mN/m, and R2 values of 0.9902, 0.9910, and 0.9904 for the train, test, and overall data, respectively.

Predicting the hydrogen uptake ability of a wide range of zeolites utilizing supervised machine learning methods

This study predicts the hydrogen uptake ability of 28 zeolites using artificial neural networks. The topology-related features of four artificial neural networks and their hyper-parameters determine utilizing 349 experimental data reported for zeolites with different surface areas at wide ranges of pressure and −196 °C. Ranking analysis over various statistical criteria confirms that the most accurate model is the cascade feed-forward neural networks with twelve hidden nodes and logarithm and tangent sigmoid activation functions. This model predicts the experimental data with the absolute average relative deviation of 7.24%, mean absolute error of 0.041, root mean squared error of 0.058, and regression factor of 0.99429. The leverage method approves that 94% of the databank is reliable. Furthermore, the relevancy analysis shows a strong direct relationship between the hydrogen uptake ability of zeolite (HUAZ) and pressure and surface area. This study is the first to develop a model for predicting the HUAZ.

Estimating the Relative Crystallinity of Biodegradable Polylactic Acid and Polyglycolide Polymer Composites by Machine Learning Methodologies.

Biodegradable polymers have recently found significant applications in pharmaceutics processing and drug release/delivery. Composites based on poly (L-lactic acid) (PLLA) have been suggested to enhance the crystallization rate and relative crystallinity of pure PLLA polymers. Despite the large amount of experimental research that has taken place to date, the theoretical aspects of relative crystallinity have not been comprehensively investigated. Therefore, this research uses machine learning methods to estimate the relative crystallinity of biodegradable PLLA/PGA (polyglycolide) composites. Six different artificial intelligent classes were employed to estimate the relative crystallinity of PLLA/PGA polymer composites as a function of crystallization time, temperature, and PGA content. Cumulatively, 1510 machine learning topologies, including 200 multilayer perceptron neural networks, 200 cascade feedforward neural networks (CFFNN), 160 recurrent neural networks, 800 adaptive neuro-fuzzy inference systems, and 150 least-squares support vector regressions, were developed, and their prediction accuracy compared. The modeling results show that a single hidden layer CFFNN with 9 neurons is the most accurate method for estimating 431 experimentally measured datasets. This model predicts an experimental database with an average absolute percentage difference of 8.84%, root mean squared errors of 4.67%, and correlation coefficient (R2) of 0.999008. The modeling results and relevancy studies show that relative crystallinity increases based on the PGA content and crystallization time. Furthermore, the effect of temperature on relative crystallinity is too complex to be easily explained.

Prediction of Oxidation Stability of Biodiesel Derived from Waste and Refined Vegetable Oils by Statistical Approaches

The oxidation stability (OX) of the biodiesel is an essential parameter mainly during storage, which reduces the quality of the biodiesel, thus affecting the engine performance. Moreover, many factors affect oxidation stability. Therefore, determining the most significant parameter is essential for achieving accurate predictions. In this paper, an empirical equation (Poisson Regression Model (PRM)), machine learning models (Multilayer Feed-Forward Neural Network (MFFNN), Cascade Feed-forward Neural Network (CFNN), Radial Basis Neural Network (RBFNN), and Elman neural network (ENN)) with various combinations of input parameters are utilized and employed to identify the most relevant parameters for prediction of the oxidation stability of biodiesel. This study measured the physicochemical properties of 39 samples of waste frying methyl ester and their blends with various percentages of palm biodiesel and refined canola biodiesel. To this aim, 14 parameters including concentration amount of WFME (X1), PME (X2), and RCME (X3) in the mixture, kinematic viscosity (KV) at 40 °C, density at 15 °C (D), cloud point (CP), pour point (PP), the estimation value of the sum of the saturated (∑SFAMs), monounsaturated (∑MUFAMs), polyunsaturated (∑PUFAMs), degree of unsaturation (DU), long-chain saturated factor (LCSF), very-long-chain fatty acid (VLCFA), and ratio (∑MUFAMs+∑PUFAMs∑SFAMs) fatty acid composition were considered. The results demonstrated that the RBFNN model with the combination of X1, X2, X3, ∑SFAMs, ∑MUFAMs, ∑PUFAMs. VLCFA, DU, LCSF, ∑MUFAMs+∑PUFAMs∑SFAMs, KV, and D has the lowest value of root mean squared error and mean absolute error. In the end, the results demonstrated that the RBFNN model performed well and presented high accuracy in estimating the value of OX for the biodiesel samples compared to PRM, MFFNN, CFNN, and ENN.

Predicting viscosity of CO2–N2 gaseous mixtures using advanced intelligent schemes

Acquiring accurate knowledge about the viscosity of carbon dioxide, nitrogen, and their mixtures as an extremely fundamental thermo-physical property for a broad range of temperatures and pressures is crucial not only for carbon capture and utilization (CCU) or carbon capture and storage (CCS) operations but also in chemical and petroleum industries and engineering design process. The proposed study aims at developing a model to predict the viscosity of carbon dioxide and nitrogen mixtures utilizing the Boosted Regression Tree (BRT) model optimized with the Artificial Bee Colony (ABC) and Particle Swarm Optimization (PSO) algorithms, the Cascade Feed-Forward Neural Networks (CFNN) and Multilayer Perception (MLP), General Regression Neural Network (GRNN), and the Genetic Programming (GP) techniques. To this end, an extensive dataset consisted of 3036 data points was gathered from the open-source literature in a broad range of pressures (0.001–453.2 MPa) and temperatures (66.5–973.15 K). The consistency of the employed paradigms was assessed based on graphical and statistical error analyses. The results indicated that the developed models provide a high degree of consistency with experimental values compared to the literature correlations. Among the established intelligent models, BRT-ABC model with a correlation coefficient (R 2 ) of 0.9993 and root mean square error (RMSE) of 1.80 μPa s achieved the most accurate and reliable predictions of the gaseous mixture viscosity. Meanwhile, the GP technique was used to develop two easy-to-use correlations with regard to gas composition, temperature, and pressure with R 2 values of 0.9883 and 0.9900 at temperatures lower and higher than 300 K, respectively. • Advanced machine learning models were evolved for estimating the viscosity of CO 2 –N 2 gaseous mixtures. • An extensive databank covering 3036 data points was considered for implementing the paradigms. • BRT-ABC model achieved the most accurate results.

Low-Complexity Multi-Task Learning Aided Neural Networks for Equalization in Short-Reach Optical Interconnects

With the rapid development of machine learning technologies in recent years, different types of neural network (NN)-based equalizers have been proposed and proved to be efficient digital signal processing tools to deal with the nonlinear impairments in short-reach direct detection optical interconnects. However, one major concern for these NN-based equalizers is their computational complexity (CC), since only a few tens of multiplications per symbol can be practically handled considering real-time implementation. In this paper, we propose an NN-based multi-symbol equalization scheme inspired by multi-task learning. Compared with traditional single-output NN-based equalizers, the CC can be significantly reduced with the help of the proposed scheme. A 50-Gb/s 25-km pulse amplitude modulation (PAM)-4 direct detection optical link is experimentally carried out and 3 different types of NNs, i.e., feedforward NN (FNN), cascade FNN (C-FNN), and recurrent NN (RNN), are employed for the proposed multi-symbol equalization scheme. Experimental results show that a maximum CC reduction of 43.2%/41.1%/44.0% can be achieved with FNN/C-FNN/RNN-based multi-symbol equalization to achieve their best performance, compared with the corresponding single-output NNs. The best performance of the proposed and the traditional FNN/C-FNN-based schemes are the same, while the RNN-based scheme sacrifices a slight system performance. Aided by multi-symbol equalization, the CC needed to recover 1 symbol for various NNs can all be reduced to about a few tens of multiplications, indicating the feasibility of real-time implementation of these NN-based equalizers.

Experimental investigation and artificial intelligent estimation of thermal conductivity of nanofluids with different nanoparticles shapes

Determining and modelling the effective thermal conductivity of nanofluids is always a concern of investigators. The effect of nanoparticles' shape on the effective thermal conductivity of nanofluids has not been accurately correlated with other influencing factors such as temperature and nanoparticle concentration. The main purpose of this study is to reveal the effects of influencing factors on the effective thermal conductivity of nanofluids by experimental investigation and artificial intelligence (AI). We prepared the TiO2/water nanofluids with four shapes of TiO2 nanoparticles (spherical, ellipsoidal, clubbed, and sheet) and measured the effective thermal conductivities of the samples with nanoparticles concentrations from 0.5 to 4.0 vol% as a function of temperature from 20 to 60 °C. Besides, 389 experimental datasets collected from literature and 80 datasets from our experiment were used to determine the optimum structure of AI-based models. Temperature, concentrations, shape factor, and thermal conductivity of nanoparticles were selected as independent variables, and the relative thermal conductivity of nanofluids was selected as the dependent variable. Six AI-based models were examined, including four artificial neural networks of multi-layer perceptron, cascade feedforward, radial basis function, and generalized regression neural network, adaptive neuro-fuzzy inference systems, and least-squares support vector machines, and the estimating performances were compared. The experimental results showed that increasing temperature and nanoparticles concentrations led to a remarkable improvement in the relative thermal conductivity of nanofluids, attributed to the intensified Brownian motion of nanoparticles with the high temperature and the effective collision of nanoparticles with the high concentrations. Besides, nanoparticles with a large aspect ratio provided the fast pathway with seldom crossing an interparticle boundary or junction point for effective heat transfer in the nanofluids, resulting in higher relative thermal conductivity of nanofluids. Statistical analyses confirmed that the cascade feed-forward neural network with ten hidden neurons and the Levenberg–Marquardt training algorithms was the optimized AI-based model for estimating the relative thermal conductivity of nanofluids. This model estimated allover experimental data by MSE = 0.0039, RMSE = 0.0622, AARD% = 2.66%, and R2 = 0.9908.

Intelligent modeling for considering the effect of bio-source type and appearance shape on the biomass heat capacity

Biomass has attracted significant interest as a renewable energy source recently. This study employs different artificial intelligence (AI) scenarios to determine biomass heat capacity (Cp). Cumulatively, 1025 experimental measurements for 25 different biomass types (block and powder forms) were utilized to design/identify the most accurate AI model. Validity checking approves that more than 98% of the data are valid. The Cp is estimated as a function of biomass source, appearance shape, and temperature. The cascade feedforward (CFF) neural network appears as the most precise tool for the concerned matter. This CFF predicts the biomass Cp by absolute average relative deviation (AARD%) and regression coefficient (R2) of 0.42% and 0.99347, respectively. Despite the empirical correlation that only considers the temperature effect on the biomass heat capacity, the CFF paradigm also incorporates the effect of biomass source and appearance shape. The powder form of biomasses has higher Cp than their block form.