Average Percent Relative Error Research Articles

Recent Developments in the Viscosity Modeling of Concentrated Monodisperse Emulsions.

Emulsions form a large group of food materials. Many foods are either partly or wholly emulsions or are in the form of emulsion at some stage of the production process. A good understanding of the rheological properties of emulsions, especially their shear viscosity, is essential in the design, formulation, and processing of food emulsions. The texture and mouthfeel of food emulsions are also largely influenced by emulsion viscosity. Therefore, it is of practical importance to be able to correlate and predict emulsion viscosity as a function of droplet concentration and other relevant variables. In this article, the recent developments made in the viscosity modeling of concentrated emulsions are reviewed. The viscosity models for concentrated emulsions published in the twenty-first century are discussed, compared, and evaluated using a large body of experimental viscosity data available on emulsions. The effects of droplet size distribution and capillary number on the viscosity of concentrated emulsions are also discussed in detail. A new generalized viscosity model is developed for concentrated emulsions that includes the effect of capillary number and is accurate with small average percent relative error (within 3%).

Petrophysical log-driven kerogen typing: unveiling the potential of hybrid machine learning

The importance of characterizing kerogen type in evaluating source rock and the nature of hydrocarbon yield is emphasized. However, traditional laboratory geochemical assessments can be time-intensive and costly. In this study, an innovative approach was taken to bridge this gap by utilizing machine learning techniques to ascertain key parameters—Organic Oxygen Index (OI), Hydrogen Index (HI), and kerogen type—from petrophysical logs of a well in the Perth Basin, Western Australia. This approach assembled geochemical data from 138 cutting samples of the Kockatea and Woodada formations and petrophysical log data. Subsequently, six machine learning algorithms were applied to predict the OI and HI parameters. The efficacy of these methods was assessed using statistical parameters, including Coefficient of Determination (R2), Average Percentage Relative Error, Average Absolute Percentage Relative Error, Root Mean Square Error, and Standard Deviation. The Support Vector Machines method emerged as the standout performer, with an R2 of 0.993 for the OI and 0.989 for the HI, establishing itself as an optimal tool for predicting these indices. Additionally, six classifiers were employed to determine kerogen types, with accuracy tested using precision, recall, F1-Score, and accuracy parameters.The study's findings highlight the superiority of the Gradient Boosting method in kerogen-type classification, achieving an impressive accuracy rate of 93.54%. It is concluded that when utilized with petrophysical logs, machine learning methodologies offer a powerful, efficient, and cost-effective alternative for determining OI, HI, and kerogen type. The novelty of this approach lies in its ability to accurately predict these crucial parameters using readily available well-log data, potentially revolutionizing traditional geochemical analysis practices.Graphical abstract

Modeling of the Sintered Density in Cu-Al Alloy Using Machine Learning Approaches.

In powder metallurgy materials, sintered density in Cu-Al alloy plays a critical role in detecting mechanical properties. Experimental measurement of this property is costly and time-consuming. In this study, adaptive boosting decision tree, support vector regression, k-nearest neighbors, extreme gradient boosting, and four multilayer perceptron (MLP) models tuned by resilient backpropagation, Levenberg-Marquardt (LM), scaled conjugate gradient, and Bayesian regularization were employed for predicting powder densification through sintering. Yield strength, Young's modulus, volume variation caused by the phase transformation, hardness, liquid volume, liquidus temperature, the solubility ratio among the liquid phase and the solid phase, sintered temperature, solidus temperature, sintered atmosphere, holding time, compaction pressure, particle size, and specific shape factor were regarded as the input parameters of the suggested models. The cross plot, error distribution curve, and cumulative frequency diagram as graphical tools and average percent relative error (APRE), average absolute percent relative error (AAPRE), root mean square error (RMSE), standard deviation (SD), and coefficient of correlation (R) as the statistical evaluations were utilized to estimate the models' accuracy. All of the developed models were compared with preexisting approaches, and the results exhibited that the developed models in the present work are more precise and valid than the existing ones. The designed MLP-LM model was found to be the most precise approach with AAPRE = 1.292%, APRE = -0.032%, SD = 0.020, RMSE = 0.016, and R = 0.989. Lately, outlier detection was applied performing the leverage technique to detect the suspected data points. The outlier detection discovered that few points are located out of the applicability domain of the proposed MLP-LM model.

A gated recurrent unit model to predict Poisson's ratio using deep learning

Static Poisson's ratio (νs) is crucial for determining geomechanical properties in petroleum applications, namely sand production. Some models have been used to predict νs; however, the published models were limited to specific data ranges with an average absolute percentage relative error (AAPRE) of more than 10%. The published gated recurrent unit (GRU) models do not consider trend analysis to show physical behaviors. In this study, we aim to develop a GRU model using trend analysis and three inputs for predicting νs based on a broad range of data, νs (value of 0.1627–0.4492), bulk formation density (RHOB) (0.315–2.994 g/mL), compressional time (DTc) (44.43–186.9 μs/ft), and shear time (DTs) (72.9–341.2 μs/ft). The GRU model was evaluated using different approaches, including statistical error analyses. The GRU model showed the proper trends, and the model data ranges were wider than previous ones. The GRU model has the largest correlation coefficient (R) of 0.967 and the lowest AAPRE, average percent relative error (APRE), root mean square error (RMSE), and standard deviation (SD) of 3.228%, −1.054%, 4.389, and 0.013, respectively, compared to other models. The GRU model has a high accuracy for the different datasets: training, validation, testing, and the whole datasets with R and AAPRE values were 0.981 and 2.601%, 0.966 and 3.274%, 0.967 and 3.228%, and 0.977 and 2.861%, respectively. The group error analyses of all inputs show that the GRU model has less than 5% AAPRE for all input ranges, which is superior to other models that have different AAPRE values of more than 10% at various ranges of inputs.

Deep Learning-Based Emulation of Radiative Transfer Models for Top-of-Atmosphere BRDF Modelling Using Sentinel-3 OLCI

The Bidirectional Reflectance Distribution Function (BRDF) defines the anisotropy of surface reflectance and plays a fundamental role in many remote sensing applications. This study proposes a new machine learning-based model for characterizing the BRDF. The model integrates the capability of Radiative Transfer Models (RTMs) to generate simulated remote sensing data with the power of deep neural networks to emulate, learn and approximate the complex pattern of physical RTMs for BRDF modeling. To implement this idea, we used a one-dimensional convolutional neural network (1D-CNN) trained with a dataset simulated using two widely used RTMs: PROSAIL and 6S. The proposed 1D-CNN consists of convolutional, max poling, and dropout layers that collaborate to establish a more efficient relationship between the input and output variables from the coupled PROSAIL and 6S yielding a robust, fast, and accurate BRDF model. We evaluated the proposed approach performance using a collection of an independent testing dataset. The results indicated that the proposed framework for BRDF modeling performed well at four simulated Sentinel-3 OLCI bands, including Oa04 (blue), Oa06 (green), Oa08 (red), and Oa17 (NIR), with a mean correlation coefficient of around 0.97, and RMSE around 0.003 and an average relative percentage error of under 4%. Furthermore, to assess the performance of the developed network in the real domain, a collection of multi-temporals OLCI real data was used. The results indicated that the proposed framework has a good performance in the real domain with a coefficient correlation (R2), 0.88, 0.76, 0.7527, and 0.7560 respectively for the blue, green, red, and NIR bands.

A reservoir bubble point pressure prediction model using the Adaptive Neuro-Fuzzy Inference System (ANFIS) technique with trend analysis.

The bubble point pressure (Pb) could be obtained from pressure-volume-temperature (PVT) measurements; nonetheless, these measurements have drawbacks such as time, cost, and difficulties associated with conducting experiments at high-pressure-high-temperature conditions. Therefore, numerous attempts have been made using several approaches (such as regressions and machine learning) to accurately develop models for predicting the Pb. However, some previous models did not study the trend analysis to prove the correct relationships between inputs and outputs to show the proper physical behavior. Thus, this study aims to build a robust and more accurate model to predict the Pb using the adaptive neuro-fuzzy inference system (ANFIS) and trend analysis approaches for the first time. More than 700 global datasets have been used to develop and validate the model to robustly and accurately predict the Pb. The proposed ANFIS model is compared with 21 existing models using statistical error analysis such as correlation coefficient (R), standard deviation (SD), average absolute percentage relative error (AAPRE), average percentage relative error (APRE), and root mean square error (RMSE). The ANFIS model shows the proper relationships between independent and dependent parameters that indicate the correct physical behavior. The ANFIS model outperformed all 21 models with the highest R of 0.994 and the lowest AAPRE, APRE, SD, and RMSE of 6.38%, -0.99%, 0.074 psi, and 9.73 psi, respectively, as the first rank model. The second rank model has the R, AAPRE, APRE, SD, and RMSE of 0.9724, 9%, -1.58%, 0.095 psi, and 13.04 psi, respectively. It is concluded that the proposed ANFIS model is validated to follow the correct physical behavior with higher accuracy than all studied models.

Ultrasound estimation of strain time constant and vascular permeability in tumors using a CEEMDAN and linear regression-based method

Ultrasound poroelastography focuses on the estimation of the spatio-temporal mechanical behavior of tissues using data often corrupted with non-stationary noise. The cumulative strain calculated from prolonged temporal acquisition of RF data can face the problem of aggregate noise. This noise can significantly affect the accuracy of curve fitting techniques necessary to estimate the clinically significant strain Time Constant (TC) and related parameters. We present a new technique, which decomposes the non-linear temporal behavior of the differential strain to extract the monotonic decaying trend by using the time-domain and data-driven Complete Ensemble Empirical Mode Decomposition with Adaptive Noise (CEEMDAN) algorithm. A linear regression scheme is then used to obtain the slope of the transformed non-linear trend, which carries information about the strain TC. Assessment of Vascular Permeability (VP), a transport parameter indicative of tumor growth, requires accurate strain TC estimations. Finite Element (FE), ultrasound simulations and in vivo experiments are used to investigate the performance of the proposed technique. Based on the simulation analysis, the average Percentage Relative Error (PRE) values of our method are 4.15% (for TC estimation) and 5.00% (for VP estimation) at 20 dB SNR level for different Percentage of Good Frames (PGF) (i.e., 20%, 50%, 75%, and 100%). These PRE values are substantially lower than those obtained using other conventional elastographic techniques. Our proposed method could become a new data-adaptive tool for analyzing the non-linear time-dependent response of complex tissues such as cancers.

Industrial water consumption forecasting based on combined CEEMD-ARIMA model for Henan province, central chain: A case study.

Industrial water consumption is a major component of the total regional water consumption. Accurate and scientific prediction of industrial water consumption is an essential guide to the rational use of natural resources. In this paper, we proposed a combined model of CEEMD (collective empirical modal decomposition) and ARIMA (autoregressive integrated moving average) for forecasting industrial water consumption to establish an accurate and efficient forecasting model, because of the poor generalization ability of most current industrial water consumption forecasting models. The influencing factors of industrial water consumption are complex, and the data are non-stationary. "Decomposition-prediction-reconstruction" is one of the significant methods for forecasting time series data, and the data decomposition has a suppressive influence on the modal mixing problem in the EMD decomposition procedure. Based on the smoothing ability of CEEMD for non-smooth signals and the better adaptation of the autoregressive moving average prediction model (ARIMA), a combined CEEMD-ARIMA model was established for industrial water consumption forecasting. This study was conducted for industrial water consumption in Henan Province in central China. The results suggest the combined CEEMD-ARIMA model has a favorable forecasting effect, with an average relative percentage error of 1.96%, and mean square error (MSE) of 0.35, a Nash efficiency coefficient (NSE) of 0.95, a prediction pass rate of 100%, and a better prediction accuracy than the ARIMA model and the combined EEMD-ARIMA model. It provides an effective prediction method for the prediction of industrial water consumption and has good application prospects.

An Accurate Reservoir's Bubble Point Pressure Correlation.

Bubble point pressure (Pb) is essential for determining petroleum production, simulation, and reservoir characterization calculations. The Pb can be measured from the pressure–volume–temperature (PVT) experiments. Nonetheless, the PVT measurements have limitations, such as being costly and time-consuming. Therefore, some studies used alternative methods, namely, empirical correlations and machine learning techniques, to obtain the Pb. However, the previously published methods have restrictions like accuracy, and some use specific data to build their models. In addition, most of the previously published models have not shown the proper relationships between the features and targets to indicate the correct physical behavior. Therefore, this study develops an accurate and robust correlation to obtain the Pb applying the Group Method of Data Handling (GMDH). The GMDH combines neural networks and statistical methods that generate relationships among the feature and target parameters. A total of 760 global datasets were used to develop the GMDH model. The GMDH model is verified using trend analysis and indicates that the GMDH model follows all input parameters’ exact physical behavior. In addition, different statistical analyses were conducted to investigate the GMDH and the published models’ robustness. The GMDH model follows the correct trend for four input parameters (gas solubility, gas specific gravity, oil specific gravity, and reservoir temperature). The GMDH correlation has the lowest average percent relative error, root mean square error, and standard deviation of 8.51%, 12.70, and 0.09, respectively, and the highest correlation coefficient of 0.9883 compared to published models. The different statistical analyses indicated that the GMDH is the first rank model to accurately and robustly predict the Pb.

High-Throughput Legume Seed Phenotyping Using a Handheld 3D Laser Scanner

High-throughput phenotyping involves many samples and diverse trait types. For the goal of automatic measurement and batch data processing, a novel method for high-throughput legume seed phenotyping is proposed. A pipeline of automatic data acquisition and processing, including point cloud acquisition, single-seed extraction, pose normalization, three-dimensional (3D) reconstruction, and trait estimation, is proposed. First, a handheld laser scanner is used to obtain the legume seed point clouds in batches. Second, a combined segmentation method using the RANSAC method, the Euclidean segmentation method, and the dimensionality of the features is proposed to conduct single-seed extraction. Third, a coordinate rotation method based on PCA and the table normal is proposed to conduct pose normalization. Fourth, a fast symmetry-based 3D reconstruction method is built to reconstruct a 3D model of the single seed, and the Poisson surface reconstruction method is used for surface reconstruction. Finally, 34 traits, including 11 morphological traits, 11 scale factors, and 12 shape factors, are automatically calculated. A total of 2500 samples of five kinds of legume seeds are measured. Experimental results show that the average accuracies of scanning and segmentation are 99.52% and 100%, respectively. The overall average reconstruction error is 0.014 mm. The average morphological trait measurement accuracy is submillimeter, and the average relative percentage error is within 3%. The proposed method provides a feasible method of batch data acquisition and processing, which will facilitate the automation in high-throughput legume seed phenotyping.

Theoretical model of natural circulation flow and heat transfer within one-ended inclined pipe

The most commonplace natural flow inside one-ended inclined pipes today is water heating systems. In this study, a model was created for the estimation of the pipe outlet temperature of the fluid with the energy balance for the inside of the tank, flow rate calculation of natural circulation, and other thermal calculations in a one-ended inclined pipe. In addition, this model has been compared with the Li model in the literature, and it is easier and more successful. 100 W, 200 W, 400 W, 600 W, and 800 W thermal power was applied to a one-ended inclined pipe, and the temperature values were recorded in five minute periods in the experiments that lasted a total of six hours. As the average of the experiments, the estimation results for the current model and the Li model are: the average per-cent relative error rates are 6.02 and 15.1 and the coefficient of determination, R2, are 0.9865 and 0.9683, respectively.

A design tool for a parabolic trough collector system for industrial process heat based on dynamic simulation

The industrial sector is one of the biggest oil consumers in Cyprus, corresponding to 20% of the fuel consumption. A parabolic trough collector (PTC) system seems the best option to produce clean thermal energy at higher temperatures than those that can be achieved from the already widely used flat plate collectors on the island. This paper presents a simulation model built to investigate the performance of the first industrial PTC system in Cyprus, installed at the Cyprus biggest soft drinks factory. The simulation model is built in TRNSYS and is validated using data from the real PTC system installed at the factory. The results show a very good fitting between the operating parameters and the power output of the Solar Field (SF), Concrete Thermal Storage System (CTES), and the Steam Generator (SG). The average percentage relative error of the system's contribution to the process is less than 6.32% for the daily steam production and during a week did not exceed 6.45%. The novelty of this study is the development of a design tool that can be used by potential interested industries to identify the suitable system that fits their needs. All data are provided in the form of graphs and allow anyone to use as input data the thermal energy demand and required steam temperature of the industry to retrieve information about the size of a suitable system which satisfies these requirements depending on each case. The payback period for all cases examined varies from 2 to 6 years, depending on the size of the system.

Non-parametric probabilistic load flow using Gaussian process learning

The load flow problem is fundamental to characterize the equilibrium behavior of a power system. Uncertain power injections such as those due to demand variations and intermittent renewable resources will change the system’s equilibrium unexpectedly, and thus potentially jeopardizing the system’s reliability and stability. Understanding load flow solutions under uncertainty becomes imperative to ensure the seamless operation of a power system. In this work, we propose a non-parametric probabilistic load flow (NP-PLF) technique based on the Gaussian Process (GP) learning to understand the power system behavior under uncertainty for better operational decisions. The technique can provide “semi-explicit” form of load flow solutions by implementing the learning and testing steps that map control variables to inputs. The proposed NP-PLF leverages upon GP upper confidence bound (GP-UCB) sampling algorithm. The salient features of this NP-PLF method are: i) applicable for power flow problem having power injection uncertainty with an unknown class of distribution; ii) providing probabilistic learning bound (PLB) which further provides control over the error and convergence; iii) capable of handling intermittent distributed generation as well as load uncertainties. The simulation results performed on the IEEE 30-bus and IEEE 118-bus system show that the proposed method can learn the voltage function over the power injection subspace using a small number of training samples. Further, the testing with different input uncertainty distributions indicates that complete statistical information can be obtained for the probabilistic load flow problem with an average percentage relative error of the order of 10−3% on 50,000 test points.

New model for standpipe pressure prediction while drilling using Group Method of Data Handling

The continuous evaluation of the measured Stand Pipe Pressure (SPP) against a modeled SPP value in real-time involves the automatic detection of undesirable drilling events such as drill string washouts and mud pump failures. Numerous theoretical and experimental studies have been established to calculate the friction pressure losses using different rheological models and based on an extension of pipe flow correlations to an annular geometry. However, it would not be feasible to employ these models for real-time applications since they are limited to some conditions and intervals of application and require input parameters that might not be available in real-time on each rig. In this study, The Group Method of Data Handling (GMDH) is applied to develop a trustworthy model that can predict the SPP in real-time as a function of mud flow, well depth, RPM and the Fan VG viscometer reading at 600 and 300 rpm. In order to accomplish the modeling task, 3351 data points were collected from two wells from Algerian fields. Graphical and statistical assessment criteria disclosed that the model predictions are in excellent agreement with the experimental data with a coefficient of determination of 0.9666 and an average percent relative error less than 2.401%. Furthermore, another dataset (1594 data points) from well-3 was employed to validate the developed correlation for SPP. The obtained results confirmed that the proposed GMDH-SPP model can be applied in real-time to estimate the SPP with high accuracy. Besides, it was found that the proposed GMDH correlation follows the physically expected trends with respect to the employed input parameters. Lastly, the findings of this study can help for the early detection of downhole problems such as drill string washout, pump failure, and bit balling.

A RADIAL BASIS FUNCTION APPROACH TO ESTIMATE PRECIPITATIONS IN BRASOV COUNTY, ROMANIA

This research tackles the problems of spatial distribution of precipitation using five deterministic sub methods related to the radial basis function (RBF) group: thin-plate spline (TPS), spline with tension (ST), multiquadric (MQ), inverse multiquadric (IMQ) and completely regularized spline (CRS). The study is used for retrieving annual precipitation over Brasov County, Romania for a period of two decades (2000-2019) using data from the fifth-generation reanalysis dataset (ERA5) provided by the European Centre for Medium-range Weather Forecasts (ECMWF). Each method was tested for 10 and 60 neighbours, and the results were evaluated through cross-validation, Taylor diagram and six statistical indicators: root mean square error (RMSE), mean error (ME), correlation coefficient (R), determination coefficient (R2), average absolute percent relative error (AAPRE) and average percent relative error (APRE). The result of the study shows a similar pattern between all methods, where the predicted precipitation increases from south to north. The southern part of Brasov County recorded in two decades 1547.7 mm precipitation, and the northern part 2149.5 mm. From all the methods analysed, the most accurate method of predicting precipitation in Brasov County is ST60. The study reveals that the number of neighbours influence the accuracy of prediction. As a result, the best prediction of precipitation was generated by ST60. Not for all methods the increase number of neighbours lead to better results. From all outputs, 66.6% have better results when is used 60 neighbours and 34.4% with 10 neighbours.

Modeling interfacial tension of methane-brine systems at high pressure and high salinity conditions

Natural gas which consists mainly of methane (usually more than 90% in volume), is becoming increasingly an important and efficient source of energy because of the lower greenhouse gas emissions and air pollution. Achieving satisfactory recovery factors in gas reservoirs is sensitive to the methane-brine/water interfaces induced by the interfacial tension (IFT) between these systems. Accordingly, accurate determination of IFT of the systems methane-brine/water is extremely important for natural gas production. In this paper, several intelligent models were implemented to accurately estimate interfacial tension (IFT) of the systems brine/pure water-methane under wide temperature, pressure and salinity ranges of (278.1–477.59 K), (0.01–260 MPa) and (0–200,000 ppm), respectively. The established models were based on an extensive databank including 879 experimental measurements. The implemented intelligent models in this study were Extreme Learning Machine (ELM), Radial Basis Function (RBF) neural network, Multilayer Perceptron (MLP), Least Square Vector Machine (LSSVM), and Generalized Regression Neural Network (GRNN). Various optimization algorithms were applied for improving the learning phase of these models. Furthermore, a Committee Machine Intelligent System (CMIS) scheme was proposed by linking the best-found paradigm under a linear single model. The results showed that all the developed intelligent-based paradigms exhibit reliable prediction abilities. In addition, it was found that CMIS and GRNN are the fittest paradigms with overall absolute average percent relative error (AAPRE) values of 1.117% and 1.003%, respectively. Besides, the performance assessment revealed that our best paradigms outperform the existing approaches. Finally, the sensitivity analysis revealed that salinity has a slight impact on IFT.

Multi‐objective hybrid decomposition and local dominance based meter placement for distribution system state estimation

The study proposes a new hybrid multi-objective evolutionary optimisation algorithm based on decomposition and local dominance for meter placement in distribution system state estimation. The evenly distributed qualitative and diverse solutions on the Pareto front are required for a decision-maker for selecting a final optimal solution. Such a Pareto front can be achieved by obtaining the balance between convergence and diversity of multi-objective optimisation algorithm. Therefore, the proposed method combined dominance and decomposition techniques, modelled meter placement as a constrained combinatorial multi-objective optimisation. The meter placement is designed as a trade-off between three objectives that are minimising the cost of the meters, average relative percentage error (ARPE) of voltage magnitude and ARPE of voltage angle. As the meter placement problem is a combinatorial optimisation, the binomial distribution-based Monte Carlo method is utilised to initialise the population, which aims to improve the diversity, as a consequence it improves the convergence, which is a by-product of this method. The results of the proposed method are compared with multi-objective evolutionary algorithm based on decomposition, non-dominated sorting genetic algorithm-II and with multi-objective hybrid particle swarm optimisation-krill herd algorithm, multi-objective hybrid estimation of distribution algorithm-interior point method and demonstrated on PG&E 69-bus distribution system and Practical Indian 85-bus distribution system.

Prediction of a wellhead separator efficiency and risk assessment in a gas condensate reservoir

In this paper artificial neural network (ANN) and adaptive neuro-fuzzy inference system (ANFIS) were applied to predict the separation percentage of gas and gas condensate in a wellhead separator in Naar oil field (Boushehr province, IRAN). The operating parameters including valve opening percentage, gas flow, design pressure, and design temperature are considered as the inputs of the models. The accuracy of the proposed models were evaluated using statistical parameters such as correlation coefficient (R2), average percent relative error (APRE), average absolute percent relative error (AAPRE), and root mean square error (RMSE). Based on the achieved data, R2 values were 0.9691 and 0.9807 for ANN and ANFIS models, respectively, while the values of RMSE were 6.117 and 4.57 for the applied models, which denote the higher accuracy of ANFIS model. Moreover, risk analyzing and consequence assessment of probable explosion of separator using PHAST (Process Hazard Analysis Software) software showed that inspection of separators is very important. Considering the calculated results, it can be concluded that ANFIS was better than ANN in prediction of gas and gas condensate separation percentages, since its output showed higher affinity to the real data. Generally, the findings obtained from the current work suggest that it is possible to predict the separation efficiency of a wellhead separator using intelligent systems.

Intelligent Road Inspection with Advanced Machine Learning; Hybrid Prediction Models for Smart Mobility and Transportation Maintenance Systems

Prediction models in mobility and transportation maintenance systems have been dramatically improved by using machine learning methods. This paper proposes novel machine learning models for an intelligent road inspection. The traditional road inspection systems based on the pavement condition index (PCI) are often associated with the critical safety, energy and cost issues. Alternatively, the proposed models utilize surface deflection data from falling weight deflectometer (FWD) tests to predict the PCI. Machine learning methods are the single multi-layer perceptron (MLP) and radial basis function (RBF) neural networks as well as their hybrids, i.e., Levenberg–Marquardt (MLP-LM), scaled conjugate gradient (MLP-SCG), imperialist competitive (RBF-ICA), and genetic algorithms (RBF-GA). Furthermore, the committee machine intelligent systems (CMIS) method was adopted to combine the results and improve the accuracy of the modeling. The results of the analysis have been verified through using four criteria of average percent relative error (APRE), average absolute percent relative error (AAPRE), root mean square error (RMSE) and standard error (SE). The CMIS model outperforms other models with the promising results of APRE = 2.3303, AAPRE = 11.6768, RMSE = 12.0056 and SD = 0.0210.

Numerical studies on the electromagnetic and thermal performances of radio frequency disinfestation treatments for dried apricots

Abstract In recent years, radio frequency (RF) dielectric heating has been increasingly studied as a promising novel disinfestation method for agricultural commodities, especially dried fruit. Three-dimensional multiphysics-based models have been established with a commercial finite element (FE) simulation package (COMSOL Multiphysics) to analyze the RF disinfestation process of dried apricots by defining six sub-domains: the flesh, a thin air layer, pit, air layer, kernel, and Indian meal moth larva located on the apricot surface. Thermo-physical and dielectric properties of the apricot flesh, pit and kernel were measured, respectively, by considering the heterogeneous structures of each individual dried fruit. The models were then experimentally validated by RF heating (27.12 MHz, 6 kW) 0.2 kg dried apricots filled in a rectangular container. The simulated point temperatures matched well with the experimental results both for the dried apricots and larvae, with a smallest average relative percentage error (RPE) of 1.45 %. Spatial distributions of simulated temperatures also indicated good agreement with the available experimental data, indicating that the maximum temperature of insect larvae was 11.3 °C higher than that of the flesh. Model prediction results demonstrated that compared with other physical locations (insect in perfect/partial contact with the flesh), the focusing effect of electric fields caused relatively fast heating rate (10.2 °C min−1) and high average temperature (68 °C) when the insects were located on the fruit surface (in point contact with flesh). Thus, the focusing effect of electric fields within insect body plays an important role in efficacy of RF disinfestation treatment for dried fruit.