Toward prediction of kinematic viscosity of biodiesel using a robust approach

Especially by using a renewable source of fuels such as biodiesel, a large number of high-quality researches have been performed on the reduction of pollution released from fossil fuels. Transesterification process is a common way for the production of biodiesel from vegetable oil, animal fat, and algae oil in the presence of alcohol and catalyst. Viscosity is one of the important physical fuel properties used in the selection of biodiesel. Experimental measurement of viscosity is a time-consuming task. Hence, in this contribution, applicability and performance of two artificial neural network-based models named least square support vector machine (LSSVM) and genetic algorithm-radial basis function (GA-RBF) for the prediction of kinematic viscosity of biodiesel were investigated. Root-mean-square error, coefficient of determination (R²), and average absolute relative deviation of each modeling were reported for each LSSVM and GA-RBF models. Modeling results show that the proposed LSSVM model is more accurate and robust than GA-RBF model.     

Thermal conductivity and dynamic viscosity modeling of Fe2O3/water nanofluid by applying various connectionist approaches

Abstract

Thermal conductivity and dynamic viscosity play key role in heat transfer capacity of nanofluids. In the present study, thermal conductivity and dynamic viscosity of Fe₂O₃/water are modeled by applying various artificial neural network algorithms. The applied algorithms are MLP, GA-RBF, LSSVM, and CHPSO ANFIS algorithms. The data for modeling procedure are extracted from several experimental studies. Obtained results by the different algorithms are compared and it was concluded that the highest R-squared values belonged to GA-RBF algorithm which were equal to 0.9962 and 0.9982 for thermal conductivity ratio and dynamic viscosity, respectively.     

Application of LSSVM algorithm for estimating higher heating value of biomass based on ultimate analysis

The higher heating value (HHV) is known as one of the energy evaluation parameters for biomass which has wide application in economic aspects investigation of energy sources. In this investigation the LSSVM algorithm as novel predicting model in the purpose of estimation of higher heating value in terms of ultimate analysis. A total number of 78 experimental data for training and testing of the algorithm were gathered from literature.in the purpose of evaluation of estimating algorithm the results are reported graphically and statistically. The calculated statistical indexes for overall data such as Root mean square error (RMSE), average absolute relative deviation (AARD) and the coefficient of determination (R²) are 9.2881, 0.038005 and 0.99996 respectively also the graphical results confirm the potential of LSSVM algorithm to be a predicting tool and a simple software for estimation of HHV as function of ultimate analysis.     

A neural network (CSA-LSSVM) model for the estimation of surface tension of branched alkanes

The current study highlights the application of a model based on least square support vector machine (LSSVM) for prediction of surface tension of branched alkanes. An optimization algorithm, namely, coupled simulated annealing (CSA) was applied to the model. Surface tensions of alkanes show a specific interaction between adjacent molecules of the branched alkanes which affects the anisotropic dispersion force component of the surface energy. In this paper, surface tension of branched alkanes was studied in temperature range of 283.15 and 448.15 K. To evaluate the performance and accuracy of this model, statistical and graphical error analyses have been used simultaneously. By applying CSA-LSSVM on 600 data points and finding optimum parameters, the estimated values of surface tension of branched alkanes were compared with experimental data which showed a reasonable agreement with the experimental results. Results demonstrate that the model is precise and viable for prediction of solubility data. The model shows an overall R² and AARD% estimations of 0.9921 and 0.89%, respectively.     

Computational modeling of biodiesel production using supercritical methanol

Renewable fuels such as biodiesel are introduced as promising environmental friendly fuels and they can be applied as alternative fuels instead of fossil fuels. In the present study, a modeling study based on statistical learning theory was investigated by the least square support vector machine (LSSVM) approach for non-catalytic biodiesel production in supercritical methanol. This model can estimate the biodiesel yield as a function of temperature, pressure, reaction time, and Methanol/oil ratio. The results indicated that the suggested LSSVM model was a satisfactory model to predict biodiesel yield that was confirmed by a high value of R² (0.9961) and low value of absolute deviation (1.17%). In addition, our model has been compared with another previous Artificial neural network (ANN)-based model and great estimations of both models were proved.     

Prediction of Thermal Conductivity and Viscosity of Ionic Liquid-Based Nanofluids Using Adaptive Neuro Fuzzy Inference System

Nowadays, ionic liquid-based nanofluids are introduced as a new class of heat transfer fluids, which exhibit superior thermal properties compared to their base ionic liquids. Potential applications of these nanofluids make it necessary to know their thermophysical properties such as thermal conductivity and viscosity. Therefore, adaptive neuro fuzzy inference system (ANFIS) has been successfully developed to predict thermal conductivity and viscosity of ionic liquid-based nanofluids. The developed models have investigated the influence of temperature, nanoparticle concentration, and ionic liquid molecular weight on the thermophysical properties of nanofluids. After developing ANFIS structure, the capability and accuracy of the developed neuro fuzzy models have been evaluated by comparison of model predictions with experimental data extracted from the literature and calculation of statistical parameters such as coefficient of determination (R²) and average relative deviation (ARD). The ARD of ANFIS model in prediction of thermal conductivity of nanofluids is 0.72%, with a high R² of 0.9959. The values of ARD and R² for estimation of nanofluids viscosity are 5.1% and 0.9934, respectively, which indicates a satisfactory degree of accuracy for the proposed models.     

Evaluation of Turbulent Models for Natural Convection of Compressible Air in a Tall Cavity

Numerical simulation of turbulent natural convection of compressible air in a tall cavity is carried out. In order to evaluate the accuracy of turbulent models, various turbulent models are applied to solve the natural convection in a tall cavity that has different temperatures imposed on two opposing vertical walls. For the large-eddy simulation (LES) model, Smagorinsky subgrid scale (SGS) and dynamic Smagorinsky SGS are also applied to the same cases in order to investigate the differences in temperature and velocity caused by different turbulent models. It is found that the k–? model has a high accuracy of predicting velocity distribution at various sampled lines by comparing with experimental data at Rayleigh number of 2.03 × 10¹⁰ and 3.37 × 10¹⁰, while the LES model has good performance in predicting temperature distributions.     

基于变量优化和IWOA-LSTM的锅炉系统水冷壁温度预测

为进一步提高锅炉系统水冷壁温度的预测精度,提出一种基于变量优化和改进鲸鱼算法优化长短期记忆神经网络的水冷壁温度预测模型。首先,通过互信息算法(MI)进行变量选择,消除初始数据中的冗余变量;其次,使用经验模态分解算法(EMD)对变量选择后的数据进行特征分解,在提取变量有效特征信息的同时降低噪音干扰;最后,使用由非线性递减因子和自适应权值改进后的鲸鱼优化算法(Improved Whale Optimization Algorithm,IWOA)确定长短期记忆神经网络(LSTM)的超参数,得到一种新型锅炉系统水冷壁温度预测模型(MI EMD IWOA LSTM)。实验结果表明,相比传统的最小二乘支持向量机(LSSVM)预测模型,MI EMD IWOA LSTM模型的均方根误差(RMSE=0.306 8)和平均绝对百分比误差(MAPE=0.054 6)最低,能够实现对锅炉系统水冷壁工质温度的精准预测。     

基于PSO-LSSVM时序预测模型的管网漏失信号识别

为解决城市供水管网的漏失问题,基于在供水管网各测压点收集的压力数据,构建粒子群(PSO)算法优化LSSVM的时序预测模型来预测压力监测点下一时刻压力值,并提出了城市供水管网漏失识别模型,通过监测点压力值与预测值的残差值是否在阈值范围内来判断管网是否处于正常工况。测试分析结果表明,改进的时序预测模型预测精度较高,可确定各压力监测点阈值,识别管网是否发生漏失事故,为相似工程提供借鉴。     

基于PSO-LSSVM时序预测模型的管网漏失信号识别

为解决城市供水管网的漏失问题,基于在供水管网各测压点收集的压力数据,构建粒子群(PSO)算法优化LSSVM的时序预测模型来预测压力监测点下一时刻压力值,并提出了城市供水管网漏失识别模型,通过监测点压力值与预测值的残差值是否在阈值范围内来判断管网是否处于正常工况。测试分析结果表明,改进的时序预测模型预测精度较高,可确定各压力监测点阈值,识别管网是否发生漏失事故,为相似工程提供借鉴。     

Modeling and optimizing parabolic trough solar collector systems using the least squares support vector machine method

Investigating the complicated thermal physics mechanisms of the parabolic trough solar collector systems plays a vital role in efficiently utilizing the solar energy. In this paper, the least squares support vector machine (LSSVM) method is developed to model and optimize the parabolic trough solar collector system. Numerical simulations are implemented to evaluate the feasibility and efficiency of the LSSVM method, where the sample data derived from the experiment and the simulation results of two solar collector systems with 30 m² and 600 m² solar fields, and the complicated relationship between the solar collector efficiency and the solar flux, the flow rate and the inlet temperature of the heat transfer fluid (HTF) is extracted. Some basic rules, such as the solar collector efficiency increases with the increase of the solar flux and the flow rate of the heat transfer fluid, and decreases with the increase of the inlet temperature of the HTF, are obtained, which indicates the LSSVM method is competent to optimize the solar collector systems. As a result, the new approach will provide meaningful data for developing the parabolic trough solar thermal power plant in China.     

Numerical simulation of natural convection heat transfer from a heated square cylinder in a square cavity filled with micropolar fluid

A numerical investigation has been performed to visualize the magnetohydrodynamic natural convective heat transfer from a heated square cylinder situated within a square enclosure subjected to nonuniform temperature distributions on the left wall. The flow inside the enclosure is unsteady, incompressible, and laminar and the working fluid is micropolar fluid with constant Prandtl number (Pr = 7). The governing equations of the flow problem are the conservation of mass, energy, and linear momentum, as well as the angular momentum equations. Governing equations formulated in dimensionless velocity and pressure form has been solved by Marker and Cell method with second-order accuracy finite difference scheme. Comprehensive verification of the utilized numerical method and mathematical model has shown a good agreement with numerical data of other authors. The results are discussed in terms of the distribution of streamlines and isotherms and surface-averaged Nusselt number, for combinations of Rayleigh number, Ra (10³–10⁶), Vortex viscosity parameter, K (0–5), and Ha parameter (0–50). It has been shown that an increase in the vortex viscosity parameter leads to attenuation of the convective flow and heat transfer inside the cavity.     

Numerical prediction of performance of a low-temperature-differential gamma-type Stirling engine

Abstract

In this study, a numerical simulation model is used to analyze thermodynamic performance of a low temperature-differential gamma-type Stirling engine by adjusting some values of the operating and geometrical parameters around a designated baseline case. The influences of these operating and geometrical parameters on engine performance such as working fluid materials, the stroke of piston and displacer, charged pressure, the heating temperature, and so on, are concerned. A numerical simulation model is established based on turbulent flow assumption and the realizable k – ε model is employed to solve the flow and thermal fields in the engine. In regard to flow in regenerator, Darcy–Forchheimer model was used to depict dynamic behavior of working fluid. Besides, thermal equilibrium model was used for solving the energy equation. Finally, working fluid in the engine undergoes a wide range of pressure and temperature so the effects of temperature and pressure on the viscosity and thermal conductivity of the working fluid are required to include. Thermal conductivity of porous medium matrix is affected by wide range of temperature as well.     

Machine learning approach to predict the thermal performance of closed-loop thermosyphon

This paper presents the machine learning (ML) algorithm to predict the thermal performance of closed-loop thermosyphon (CLT). The experimentation is carried out on the acetone-charged CLT at different test conditions such as heat inputs, filling ratios, and adiabatic lengths. The test data is used to calculate the performance parameters such as thermal resistance, heat transfer coefficient, and effectiveness of the system. Based on the experimental dataset, the ML algorithms are developed to predict the performance parameters of the CLT system. The ML algorithms such as linear regression, decision tree (DT), random forest (RF), and lasso regression are used for the development of the prediction model. The hyperparameters are well-tuned and optimized. The prediction measuring parameters (mean absolute error, R²) are analyzed carefully. It is noticed that the DT model outperformed the prediction of the other used models. The R² score of the DT model was 98.504; whereas, the R² scores of the RF model and linear regression model were about 94.76 and 92.17, respectively. This study will become a roadmap to the ML approach in the thermosyphon system.     

基于共享最小二乘支持向量机模型的电站锅炉燃烧系统的优化

电站锅炉燃烧系统是一个复杂的多输入多输出系统,为了在同一个模型中实现高效率、低污染物排放的优化目标,对标准最小二乘支持向量机回归方法进行了扩展．借助某电厂1000MW超超临界锅炉的现场燃烧调整试验数据,建立了以锅炉热效率和NOx排放质量浓度为输出的共享最小二乘支持向量机（LSSVM）模型,采用一种改进的粒子群算法对共享模型中的锅炉运行工况进行了寻优．结果表明：在共享LSSVM模型中,锅炉热效率和NOx排放质量浓度的平均预测误差分别可达到0．028％和2．16％,搜索得到的高效率和低NOx排放的参数组合可为电站锅炉优化运行提供指导．     

最小二乘支持向量机在大坝变形预测中的应用

介绍了基于统计学习理论的一种新的机器学习技术———支持向量机(SVM)和其拓展方法———最小二乘支持向量机(LSSVM),并将LSSVM算法应用于混凝土大坝安全监控中的变形预测。根据实测数据,建立了基于LSSVM算法的大坝变形预测模型,同时与经典SVM预测模型进行分析比较。结果表明,LSSVM和经典SVM算法在大坝变形预测中都具有较好的可行性、有效性及较高的预测精度;LSSVM在算法的学习训练效率上比SVM有较大的优势,更适合于解决大规模的数据建模。     

Prediction of compressive strength of geopolymer composites using an artificial neural network

Geopolymers are highly complex materials which involve many variables and make for which modelling the properties is very difficult. There is no systematic approach in mix design for geopolymers. Since the amounts of silica modulus, Na₂O content, w/b ratios and curing time have a great influence on the compressive strength, an ANN (artificial neural network) method has been established for predicting compressive strength of ground pumice based Geopolymers and the possibilities of adapting ANN and artificial intelligence system for predicting the compressive strength have been studied. Consequently, a multilayer ANN by using back propagation architecture can be developed for geopolymer compressive strength prediction. In this study, the coefficient of determination (R²) has been used for investigating the proposed model accuracy. As a result, proposed ANN model can predict the compressive strength of geopolymer with R²?=?0.958.     

Artificial neural network (ANN) modeling of thermal conductivity of supercritical ethane

In this present contribution, thermal conductivity ofethene (TCE) above the critical temperature has been studied. The present data cover the temperature range from 283.46 to 425.00 K and the pressure range from 0.1 to 100 MPa. In the present investigation, various network-based strategies, named as artificial neural network (ANN) optimized with two evolutionary algorithms, including genetic algorithm (GA) and differential evolution (DE), were developed for assessing tTCE in supercritical region. The most comprehensive source of data, including around 256 experimental points, was utilized for ANN modeling. Data index plot, scatter plot, relative deviation diagram and root mean square error (RMSE), and coefficient of determination (R²) as the statistical parameters were used in this examination to evaluate the comprehensiveness of the developed ANN model. Results indicate that the GA-ANN is more accurate than DE-ANN to predict TCE in supercritical region. Also, among optimization algorithms, GA has the largest ability for optimizing the ANN network modeling with the RMSE of 4.2966 and determination coefficient (R²) of 0.9640.     

Effect of viscosity variation on natural convection flow of water–alumina nanofluid in an annulus with internal heat generation

Heat transfer enhancement in a horizontal annulus using the variable viscosity property of an Al₂O₃–water nanofluid is investigated. Two different viscosity models are used to evaluate heat transfer enhancement in the annulus. The base case uses the Pak and Cho model and the Brinkman model for viscosity which take into account the dependence of this property on temperature and nanoparticle volume fraction. The inner surface of the annulus is heated uniformly by a constant heat flux q_w and the outer boundary is kept at a constant temperature T_c. The nanofluid generates heat internally. The governing equations are solved numerically subject to appropriate boundary conditions by a penalty finite‐element method. It is observed that for a fixed Prandtl number Pr = 6.2, Rayleigh number Ra = 10⁴ and solid volume fraction ? = 10%, the average Nusselt number is enhanced by diminishing the heat generation parameter, mean diameter of nanoparticles, and diameter of the inner circle. The mean temperature for the fluids (nanofluid and base fluid) corresponding to the above mentioned parameters is plotted as well. © 2012 Wiley Periodicals, Inc. Heat Trans Asian Res; Published online in Wiley Online Library ( wileyonlinelibrary.com/journal/htj ). DOI 10.1002/htj.21016     

Power Law or Logarithmic Law？—A data Analysis for Zero Pressure Gradient Turbulent Boundary Layers with Low Reδs

The paper presents an analysis of two-dimensional zero pressure gradient(ZPG) turbulent boundary layers(TBL) with regard to the application of power laws,only TBL with low Reynolds number 300<Reδ2<6200 are taken into account.It is found that a certain region of the mean velocity profile can be described with a power law of the form u^+=Cpow^*y^α,This power law region is not a priori identical with the overlap region.An algorithm for the determination of the wall skin friction using the power law is proposed.The method was applied with good result to ZPG TBL and to adverse pressure gradient(APG) TBL.To brdge the gap between the wall and the power law region an approach for the turbulent viscosity is suggested.