Artificial Neural Network Modeling Approach Research Articles

Development of a decision support system tool to predict the pulmonary function using artificial neural network approach

SummaryThe spirometry is considered a preclinical tool for the evaluation of the respiratory system. The formal lung volumes measurement and health status lung system are made using spirometry. Artificial neural network (ANN) has been introduced in solving complex problems in a large number of different settings, including medical diagnosis support system as predictive power. An objective of this research was intended to investigate the development of a new decision support system (DSS) using ANN modeling approaches and algorithms to predict pulmonary function in people. The spirometry data and general characteristics, anthropometric data, and body composition parameters (N = 130) were obtained from subjects. The classification of pulmonary function was performed by the multi‐layer perceptron (MLP) model. Findings show that the MLP model is capable of classifying respiratory abnormalities in different people. The ANN model was totally 93.6%, 92.3%, 84.6%, and 91.5% successful in correctly classified in training, validation, test, and all data, respectively. Also, a DSS tool was created that allows the evaluation and classification of the results of spirometry data. It appears that ANNs are useful in classification pulmonary function.

Artificial neural network (ANN) for prediction indoor airborne particle concentration

Due to experimental data insufficiency for results validation realized by Computation Fluid Dynamics method (CFD), we are proposed new numerical simulations to determined concentration distribution of fine particles in indoor air for transient regime. The ANN model approach of multi-layer perceptron type with three layers is applied successfully. This model requires learning through a database which deduced from the bibliographic literature, composed by 2271 measurement points of which 80% assigned to ANN model training, 10% to test model and so the remaining (10%) assigned to validation part. The ANN model developed in this paper is beneficial and easy to predict fine particles distribution in air indoor when compared to the CFD method. The results average error found by this model does not reach 5%, when compared to the CFD method with an error of 16%. This model is used to treat the effect of the velocity and air exhaust section positions on the stability and flow regime establishment time.

Artificial neural network (ANN) approach to predict unconfined compressive strength (UCS) of oil and gas well cement reinforced with nanoparticles

The prediction of unconfined compressive strength (UCS) of oil well cement class “H” based on the artificial neural network (ANN) modeling approach is presented in this study. 195 cement samples were embedded with varying dosages of strength enhancing pre-dispersed nanoparticles consisting of nanosilica (nano-SiO2), nanoalumina (nano-Al2O3), and nanotitanium dioxide (nano-TiO2) at various simulated wellbore temperatures. The efficacy of the pre-dispersed nanoparticle solutions was analyzed by transmission electron microscope (TEM) images. Nano-SiO2 and nano-Al2O3 displayed excellent dispersibility throughout the solution. However, nano-TiO2 readily agglomerates which, at high concentrations, is detrimental to the UCS of cement. 70% of the data set was used to train the ANN model, 15% was used for validation, and 15% was used to test the model. The model consisted of one input layer with five nodes, one hidden layer with 12 nodes, and one output layer with one node. 12 nodes in the hidden layer resulted in the lowest mean squared error (MSE). The model parameters were saved and used after seven epochs during training, at which point the validation error began to increase leading to overfitting. The statistical performance measures consisting of MSE, the square root of the coefficient of determination (R), and the mean absolute percentage error (MAPE) showed values close to zero, one, and less than five percent, respectively. The statistical performance measures of the ANN model displayed superior results when compared to the measures obtained by the multi-linear (MLR) and random forest (RF) regression algorithms. The developed ANN model displays high predictive accuracy and can replace, or be used in combination with, destructive UCS tests which can save the petroleum industry time, resources, and capital.

Environmental modeling of landscape aesthetic value in natural urban parks using artificial neural network technique

An accurate assessment of the landscapes in a region requires having sufficient information of the influential factors as well as the type, manner, and the impact rate of each of them on the user’s perception of the landscape quality. The purpose of this study is modeling landscape aesthetic quality of urban parks using an artificial neural network to predict the value of landscape aesthetic and prioritizing the influential variables of the model. To evaluate the landscape aesthetic quality in the urban park, a combination of user perspective and artificial neural network modeling approach has been used. The aesthetic quality of 100 urban park landscapes was quantified based on citizens' perception. Totally, 15 landscape attributes were recorded as influential variables on visual quality of landscape. According to the results, the multi-layer perceptron model with structure of 15-8-1 (15 input variables, 8 neurons in the hidden layer, and one output variable) and the maximum value of coefficient of determination (R2) in three data sets, namely training, validation, and test which are 0.97, 0.88 and 0.90, present the best performance of structure optimization. Accordingly, land slope, flowers and bushes, buildings, and hard surfaces ratio with a model sensitivity coefficient of 0.56, 0.24, 0.07, and 0.07, respectively, show the maximum effect on the landscape aesthetic quality in urban parks. The developed multi-layer perceptron model in MATLAB software is known as a decision support system in designing the structure of urban parks and also provides possibility of predicting the landscape aesthetic value in new parks.

Artificial neural network for prediction of rock properties using acoustic frequencies recorded during rock drilling operations

Determining properties of rocks in rock mechanics/engineering applications such as underground tunnelling, slope stability, foundations, dam design and rock blasting is often difficult due to the requirements of high quality of core rock samples and accurate test apparatus. Prediction of the geomechanical properties of rock material has been an area of interest for rock mechanics for several years now. Nowadays, soft computing methods are used as a powerful tool to estimate the rock properties, cost and duration of the project. This has led to a lack of necessity to develop a model to predict rock properties in the field of rock mechanics. ANN (artificial neural network) models were developed to predict geomechanical properties of the sedimentary rock types using dominant frequencies of an acoustic signal during rock drilling operations. A set of experimental drilling operations test conditions around 875 were used as input parameters including drill bit spindle speeds (rpm), drill bit penetration rates (mm/min), drill bit diameters (mm) and dominant frequencies of the acoustic signal (Hz). The response (output) was uniaxial compressive strength (UCS), Brazilian tensile strength (BTS), density ( $$\rho $$ ) and abrasivity (%). The developed models were checked using various performance indices. The results from the analysis show that the suggested ANN model approach is efficient, fits the data and accurately reflects the experimental results. The ANN models predicted physico-mechanical rock properties through the dominant frequency of acoustic signals during rock drilling operations.

Black-, gray-, and white-box modeling of biogas production rate from a real-scale anaerobic sludge digestion system in a biological and advanced biological treatment plant

Artificial intelligence-based methodology [artificial neural network (ANN) and fuzzy logic] and a multiple regression-based analysis were conducted for modeling of the biogas production rate from a real full-scale sludge digestion process. In the computational analysis, five process-related parameters such as influent sludge flow rate, total solids content, total volatile solids content, alkalinity, and volatile fatty acids concentration were considered as the input variables for the proposed models. In the ANN-based modeling, a benchmark comparison of 11 backpropagation (BP) algorithms was employed in the first step, and the scaled conjugate gradient algorithm was chosen as the best BP algorithm in terms of their respective mean squared errors. According to the selected BP algorithm (scaled conjugate gradient BP), the number of neurons at the hidden layer was optimized as 14, and the coefficient of determination (R2) was obtained as 0.65 for the optimal three-layer ANN structure (5:14:1). In the second part of the study, a MISO (multiple-input single-output)-type fuzzy-logic model was developed, and five input variables were fuzzified in a knowledge-based manner. For the fuzzy subsets, trapezoidal membership functions were implemented with 10 and 20 levels, and a Mamdani-type fuzzy inference system (FIS) was used to employ 394 rules in the if–then form. The product, summation, and centroid methods were employed, respectively, for implication, aggregation, and defuzzification processes conducted in the FIS. Fuzzy logic-produced forecasts were compared with the estimations obtained from both ANN-based model and a polynomial multiple regression-based model (employed as the third part of this study) derived in this study. Findings of this study clearly indicated that compared to ANN model and conventional multiple regression approach, the proposed MISO fuzzy logic-based model produced smaller deviations and exhibited a superior predictive performance on forecasting biogas production rate from a full-scale treatment plant with a satisfactory R2 value of 0.88.

Modeling and optimization of chlorophenol rejection for spiral wound reverse osmosis membrane modules

This study shows an artificial neural network (ANN) model of chlorophenol rejection from aqueous solutions and predicting the performance of spiral wound reverse osmosis (SWRO) modules. This type of rejection shows complex non-linear dependencies on feed pressure, feed temperature, concentration, and feed flow rate. It provides a demanding test of the application of ANN model analysis to SWRO modules. The predictions are compared with experimental data obtained with SWRO modules. The overall agreement between the experimental and ANN model predicted was almost 99.9% accuracy for the chlorophenol rejection. The ANN model approach has the advantage of understanding the complex chlorophenol rejection phenomena as a function of SWRO process parameters.

Photocatalytic-persulfate- oxidation for diclofenac removal from aqueous solutions: Modeling, optimization and biotoxicity test assessment

In this paper, the influence of several aquatic factors (the nature of catalyst, the initial pH and the initial concentration of the pollutant) on the photocatalytic degradation of diclofenac (DFC), one of the most widely prescribed anti-inflammatory non-steroidal drug, was studied. Also, in order to examine the intensification process, the variation of the photocatalytic DFC degradation in the presence of sodium persulfate (PPS) was analyzed. It was found that, compared to titanium dioxide (TiO2), the zinc oxide (ZnO) photocatalyst performed exceptionally well, with a 96.13% DFC degradation efficiency after 150 min. The photodegradation of DFC by ZnO catalyst fitted well the Langmuir–Hinshelwood kinetic model. The maximum efficiency is 97.27% for simulated solar-UVA/ZnO/PPS and 77% for simulated solar-UVA/ZnO. In order to determine the optimal conditions leading to the maximization of DFC removal, an artificial neural network (ANN) modeling approach combined with genetic algorithm (GA) was applied. The best ANN determined had a correlation of 0.999 and it was further used in the process optimization where a 99.7% degradation efficiency was identified as the optimum under the following conditions: DFC initial concentration 37,9 mg L−1, pH 5,88 and PPS initial concentration 500 mg L−1. The effectiveness of the process and the toxicity of the pharmaceutical pollutants and their by-products were also evaluated and confirmed by the biological tests using liver and kidney of Mus musculus mice.

A Novel Assisted Artificial Neural Network Modeling Approach for Improved Accuracy Using Small Datasets: Application in Residual Strength Evaluation of Panels with Multiple Site Damage Cracks

An artificial neural network (ANN) extracts knowledge from a training dataset and uses this acquired knowledge to forecast outputs for any new set of inputs. When the input/output relations are complex and highly non-linear, the ANN needs a relatively large training dataset (hundreds of data points) to capture these relations adequately. This paper introduces a novel assisted-ANN modeling approach that enables the development of ANNs using small datasets, while maintaining high prediction accuracy. This approach uses parameters that are obtained using the known input/output relations (partial or full relations). These so called assistance parameters are included as ANN inputs in addition to the traditional direct independent inputs. The proposed assisted approach is applied for predicting the residual strength of panels with multiple site damage (MSD) cracks. Different assistance levels (four levels) and different training dataset sizes (from 75 down to 22 data points) are investigated, and the results are compared to the traditional approach. The results show that the assisted approach helps in achieving high predictions’ accuracy (<3% average error). The relative accuracy improvement is higher (up to 46%) for ANN learning algorithms that give lower prediction accuracy. Also, the relative accuracy improvement becomes more significant (up to 38%) for smaller dataset sizes.

Artificial neural network modelling approach of drying kinetics evolution for hot air oven, microwave, microwave convective and freeze dried pineapple

This study presents the dehydration characteristics of the drying of the pineapple slice. Hot air oven (60 °C), microwave (100 W), microwave convective (600 W, 40 °C) and freeze drying (0.1 mbar, − 40 °C) of pineapple slice through artificial neural network (ANN) modeling approach (Backpropagation algorithm for feed-forward network). Considering the sample thickness of 0.15 cm, effective moisture diffusivity is calculated. To study the reconstitution property of the product, the rehydration ratio is considered at 20 °C, 40 °C, 60 °C and 80 °C. The microwave convective drying (96 min) is the fastest while the freeze drying is slowest (20 h). The coefficient of correlation through the ANN model for hot air oven, microwave, microwave convective and freeze dried sample are 0.99936, 0.9994, 0.99948, and 0.99926 respectively. In terms of rehydration ratio, freeze drying (5.49) is found best followed by microwave (4.62), hot air (4.50), and microwave convective drying (2.00). At 80 °C the coefficient of rehydration value attains the maximum of 0.969 for freeze dried sample, whereas the minimum value of 0.369 is observed at 20 °C for the microwave-convective dried sample. From the observations obtain it may be stated that the ANN model predicts the dehydration kinetics adequately for the entire drying process. Freeze dried product appears to be the best in terms of reconstitution properties.

Prediction of uniaxial tensile flow using finite element-based indentation and optimized artificial neural networks

This study derives a uniaxial tensile flow from spherical indentation data using an artificial neural network (ANN) combined with finite element (FE) analysis. The feasibility of the FE-based simulations is confirmed through experimental indentation for various steels. Parametric studies of the FE simulation are performed to generate an ANN training database. An encoding for feature extraction and a hyperparameter optimization is implemented to design the ANN with high predictive performance. The indentation load–depth curves are converted into hardening parameters through the trained ANN. The predictive performance of the FE–ANN model using real-life indentation data is investigated in-depth with thorough error evaluation, and verified by uniaxial tensile tests. The emphasis is made that the mean absolute percentage error between the experimental and simulated indentation data is required to be meticulously controlled below 1% to accurately predict the tensile properties. The validations demonstrate that the applied FE–ANN modeling approach is very robust and captures the tensile properties well. Furthermore, the Taguchi orthogonal array (OA) method that can achieve high efficiency and fidelity with less training data is discussed. The FE–ANN model is concisely designed using the Taguchi OA method and can predict elasticity as well as plasticity.

Artificial neural network modeling approach for the prediction of five-day biological oxygen demand and wastewater treatment plant performance

Abstract The measurement of the wastewater BOD5 level requires five days, and the use of a prediction model to estimate BOD­5 saves time and enables the adoption of an online control system. This study investigates the application of artificial neural networks (ANNs) in predicting the influent BOD5 concentration and the performance of WWTPs. The WWTP performance was defined in terms of the COD, BOD, and TSS concentrations in the effluent. Sensitivity analysis was performed to identify the best-performing ANN network structure and configuration. The results showed that the ANN model developed to predict the BOD concentration performed the best among the three outputs. The top-performing ANN models yielded R2 values of 0.752, 0.612, and 0.631 for the prediction of the BOD, COD, and TSS concentrations, respectively. The optimal performing models were obtained (three inputs – one output), which indicated that the influent temperature and conductivity greatly affect the WWTP performance as inputs in all models. The developed prediction model for the influent BOD5 concentration attained a high accuracy, i.e., R2 = 0.754, which implies that the model is viable as a soft sensor for online control and management systems for WWTPs. Overall, the ANN model provides a simple approach for the prediction of the complex processes of WWTPs.

Application of Artificial Neural Network for Prediction of Key Indexes of Corn Industrial Drying by Considering the Ambient Conditions

Uncontrollable ambient conditions are the main factors limiting the self-adaption control of an industrial drying system. To achieve the goal of accurate control of the drying process, the influence of the ambient conditions on the drying behavior should be taken into consideration when modeling the drying process. Present work introduced an industrial drying system with a loading capacity of 50 t, two artificial neural network prediction models with (IANN) and without (OANN) considering the ambient conditions were established using artificial neural network modeling approach. The ambient conditions on the moisture content (MC), exergy efficiency of the heat exchanger (ηex,h) and specific recovered radiant energy (Er) of the drying process were also investigated. The results showed that the ηex,h and Er increase with the increase of ambient temperature while the drying time decrease with the increase of the ambient temperature. The IANN model has a better prediction performance that that of OANN model. An optimal architecture of 9-2-12-3 artificial neuron network model was developed and the best prediction performance of the artificial neural network (ANN) model were found at a training epoch number of 30, and a momentum coefficient of 0.4, where the coefficient of determination of moisture content, exergy efficiency of heat exchanger, and the specific recovered radiant energy, respectively are 0.998, 0.992, and 0.980, indicating that the model has an excellent prediction performance and can be used in engineering practice.

An artificial neural network modeling approach for short and long fatigue crack propagation

Fatigue crack growth-based damage modeling approaches have received great interest due to its critical importance in the industry. However, a substantial deficiency of an explicit fatigue damage model to quantify the accurate fatigue crack growth behavior remains due to the complex fatigue crack growth behavior in different length scales. This complexity arises from the fact that fatigue crack growth (FCG) in different length scales depends on many damage controlling parameters. Machine learning-based fatigue damage modeling approaches have received noticeable attention for fatigue crack growth analysis due to their abilities to account for numerous damage parameters simultaneously. In the presented paper, a radial basis function artificial neural network (RBF-ANN) model has been developed to predict the FCG behavior, including the short and long crack regimes. The presented RBF-ANN model has been trained and verified by experimental data sets of Ti-6Al-4V titanium alloy, 2024-T3 and 7075-T6 aluminium alloys. The predictions showed that the RBF-ANN model has a good interpolation capability to predict the nonlinearity of both short and long crack growth behavior. However, the model shows poor extrapolation capability for accurate short crack growth predictions for cases that there are limited data sets in hand. The model effectiveness greatly depends on sufficient available input data.

Prediction of Surface Roughness Based on Cutting Parameters and Machining Vibration in End Milling Using Regression Method and Artificial Neural Network

This study presents surface roughness modeling for machined parts based on cutting parameters (spindle speed, cutting depth, and feed rate) and machining vibration in the end milling process. Prediction models were developed using multiple regression analysis and an artificial neural network (ANN) modeling approach. To reduce the effect of chatter, machining tests were conducted under varying cutting parameters as defined in the stable regions of the milling tool. The surface roughness and machining vibration level are modeled with nonlinear quadratic forms based on the cutting parameters and their interactions through multiple regression analysis methods, respectively. Analysis of variance was employed to determine the significance of cutting parameters on surface roughness. The results show that the combined effects of spindle speed and cutting depth significantly influence surface roughness. The comparison between the prediction performance of the multiple regression and neural network-based models reveal that the ANN models achieve higher prediction accuracy for all training data with R = 0.96 and root mean square error (RMSE) = 3.0% compared with regression models with R = 0.82 and RMSE = 7.57%. Independent machining tests were conducted to validate the predictive models; the results conclude that the ANN model based on cutting parameters with machining vibration has a higher average prediction accuracy (93.14%) than those of models with three cutting parameters. Finally, the feasibility of the predictive model as the base to develop an online surface roughness recognition system has been successfully demonstrated based on contour surface milling test. This study reveals that the predictive models derived on the cutting conditions with consideration of machining stability can ensure the prediction accuracy for application in milling process.

Artificial neural networking to estimate the leaf area for invasive plant Wedelia trilobata

Leaf area are very important parameter for the understanding of growth and physiological responses of invasive plant species under different environmental factors. This study was conducted to build non‐destructive leaf area model of Wedelia trilobata that were grown in greenhouse. Regression analysis and artificial neural network (ANN) approaches were used for the development of leaf area model with the help of leaf length and width of 262 plants samples. In selection of best method under both techniques, the lower value of mean absolute error (MAE), root mean square error (RMSE), mean absolute percentage error (MAPE) and higher value of R2 were considered. According to the results it was found that ANN have higher value of (R2 = 0.96) and lower value of error (MAE = 0.023, RMSE = 0.379, MAPE = 0.001) than regression analysis (R2 = 0.94, MAE = 0.111, RMSE = 1.798, MAPE = 0.0005). It was concluded that error between predicted and actual value was less under ANN. Therefore, ANN model approach can be used as an alternating method for the estimation of leaf area. Through estimation of leaf area, invasive plant growth can predict under different environment conditions.

Data-driven modeling to optimize the injection well placement for waterflooding in heterogeneous reservoirs applying artificial neural networks and reducing observation cost

Secondary recovery methods such as waterflooding are often applied to depleted reservoirs for enhancing oil and gas production. Given that a large number of discretized elements are required in the numerical simulations of heterogeneous reservoirs, it is not feasible to run multiple full-physics simulations. In this regard, we propose a data-driven modeling approach to efficiently predict the hydrocarbon production and greatly reduce the computational and observation cost in such problems. We predict the fluid productions as a function of heterogeneity and injection well placement by applying artificial neural network with small number of training dataset, which are obtained with full-physics simulation models. To improve the accuracy of predictions, we utilize well data at producer and injector to achieve economic and efficient prediction without requiring any geological information on reservoir. The suggested artificial neural network modeling approach only utilizing well data enables the efficient decision making with reduced computational and observation cost.

MLR and ANN Approaches for Prediction of Synthetic/Natural Nanofibers Diameter in the Environmental and Medical Applications

Fiber diameter plays an important role in the properties of electrospinning of nanofibers. However, one major problem is the lack of a comprehensive method that can link processing parameters to nanofibers’ diameter. The objective of this study is to develope an artificial neural network (ANN) modeling and multiple regression (MLR) analysis approaches to predict the diameter of nanofibers. Processing parameters, including weight ratio, voltage, injection rate, and distance, were considered as independent variables and the nanofiber diameter as the dependent variable of the ANN model. The results of ANN modeling, especially its high accuracy (R2 = 0.959) in comparison with MLR results (R2 = 0.564), introduced the prediction the diameter of nanofibers model (PDNFM) as a comparative model for predicting the diameter of poly (3-caprolactone) (PCL)/gelatin (Gt) nanofibers. According to the result of sensitivity analysis of the model, the values of weight ratio, distance, injection rate, and voltage, respectively, were identified as the most significant parameters which influence PDNFM.

Modeling of UCS value of stabilized pond ashes using adaptive neuro-fuzzy inference system and artificial neural network

This paper investigates the capability of adaptive neuro-fuzzy inference system (ANFIS) and artificial neural network (ANN) modeling approach to predict the unconfined compressive strength (UCS) of stabilized pond ashes with lime alone and in combination with lime sludge. Out of 170 data set, a total of 119 data were randomly selected for training, whereas remaining 51 were used for testing the model. Four membership’s functions (MFs) such as Gaussian, generalized bell-shaped, triangular, and trapezoidal were used with ANFIS model. Statistical parameters were used to compare the performance of four MF-based ANFIS and ANN models. A comparison of results suggests that Triangular MF-based ANFIS model exhibit better predictive performance with higher CC = 0.980 and lower MSE = 3028.515 and RMSE = 55.032 than other MF-based ANFIS and ANN model. The results of single-factor analysis of variance indicate that there is an insignificant difference between measured and predicted values of UCS using different models. Further, results of sensitivity analysis depict that the curing period, lime sludge, and lime are the most important parameters which affect the performance of Triangular MF-based ANFIS in predicting the UCS of stabilized pond ashes. Thus, the Triangular MF-based ANFIS model could be a useful tool in predicting the UCS value of stabilized pond ashes because of its adequacy in handling uncertainties in the test results with accuracy.

A spatial-and-temporal-based method for rapid particle concentration estimations in an urban environment

The increasing construction of buildings and infrastructure in cities heavily influences pollutant dispersion and causes a spread of increased particle concentrations. Real-time data and information on local pollution levels are highly desired by residents, urban planners and policy-makers. Such information is scarce due to the high cost of real-time measurement. To fill the gap, the aim of this research is to develop a model that can rapidly estimate particulate pollution based on a data-driven artificial neural network modelling approach. The key influential factors such as background pollution level, weather conditions and urban morphology are embedded in the model in association with local emission sources of pollution relating to construction activities and traffic flows. The data for urban spatial-variables (building and road) and traffic information is processed with the aid of the Geographic Information System using self-developed Python scripts. The geographic dataset containing the required information for each grid is integrated with the artificial neural network model to perform forecasting of particle concentrations. The model has been verified with measurements from a case study with 20 sample locations in Chongqing, China, showing that the average relative error of particle concentration estimation compared to measurement is 17.56% for PM10 and 16.04% for PM2.5. A map of a time-specific spatial interpolation of particle concentrations which visualises real-time pollution is consequently produced based on the method. The newly proposed method is novel and holistic which integrates spatial and time information covering aspects of urban form, weather file, traffic, construction sites. The rigorously validated model has been transformed to a robust tool for a fast estimation of real-time particle concentration in an urban environment.