ANN Design Research Articles

Investigation on Sliding Wear Properties of Nano Metallic Particle Reinforced Hybrid Composites Through Design of Experiments and ANN

Investigation on Sliding Wear Properties of Nano Metallic Particle Reinforced Hybrid Composites Through Design of Experiments and ANN

Optimization and design of ANN with Levenberg-Marquardt algorithm to increase the accuracy in predicting the viscosity of SAE40 oil-based hybrid nano-lubricant

In recent years, the use of artificial neural networks (ANNs) has increased due to their good performance and high accuracy in predicting laboratory data. In this study, experimental data of viscosity ( μ nf ) of MWCNT-Al 2 O 3 (10:90)/SAE40 nano-lubricant in volume fraction φ =0.0625%–1% and temperature range T = 25 to 50 °C were designed and checked by ANN. For modeling by ANN, a multilayer perceptron (MLP) ANN with the Levenberg-Marquardt (ML) algorithm, which is one of the most accurate ANN modeling methods, is used. To design the ANN, the parameters of φ , temperature and shear rate are considered input variables and μ nf is considered as an output variable. Finally, the values of the R regression coefficient and the mean square error (MSE) for the optimal structure were 0.9999 and 0.000615042, respectively. The range of the margin of deviation (MOD) of the predicted values is −2% < MOD <2%. A comparison of three data sets, namely laboratory data, correlation output and ANN output, shows that The ANN method is much more powerful and more accurate than mathematical relationship data and laboratory data. • Design of an ANN and experimental evaluation of the viscosity of a nano-lubricant are investigated. • A multilayer perceptron network with Levenberg-Marquardt algorithm is used. • φ , temperature and shear rate are considered as input variables and μ nf as an output variable. • The range of the margin of deviation (MOD) of the predicted values is −2% < MOD <2%. • the ANN method is much more capable than the new relation and estimates laboratory data more accurately.

Optimal artificial neural network architecture design for modeling an industrial ethylene oxide plant

Optimum selection of input variables, number of hidden neurons and connections among the network elements deliver the best configuration of an ANN, usually resulting in reduced over-fitting and improved test performance. This study focuses on the development of a superstructure-oriented feedforward ANN design and training algorithm whose impacts are demonstrated on an industrial Ethylene Oxide (EO) plant for the prediction of product related variables. Proposed method brings about a mixed integer nonlinear programming problem (MINLP) to be solved, which takes the existence of inputs, neurons, and connections among the network elements into account by binary variables in addition to continuous weights of existing connections. Our investigations show that almost 85% of the ANN connections are removed compared to the fully connected ANN (FC-ANN) with 50% decrease in the number of inputs of the ANN. The modified ANN delivers a better prediction performance over FC-ANN, since FC-ANN suffers from over-fitting.

Prototyping Feed-Forward Artificial Neural Network on Spartan 3S1000 FPGA for Blood Type Classification

In this research, a Feed-Forward Artificial Neural Network design was implemented on Xilinx Spartan 3S1000 Field Programable Gate Array using XSA-3S Board and prototyped blood type classification device. This research uses blood sample images as a system input. The system was built using VHSIC Hardware Description Language to describe the feed-forward propagation with a backpropagation neural network algorithm. We use three layers for the feed-forward ANN design with two hidden layers. The hidden layer designed has two neurons. In this study, the accuracy of detection obtained for four-type blood image resolutions results from 86%-92%, respectively.

Investigation of ANN architecture for predicting the compressive strength of concrete containing GGBFS.

An extensive simulation program is used in this study to discover the best ANN model for predicting the compressive strength of concrete containing Ground Granulated Blast Furnace Slag (GGBFS). To accomplish this purpose, an experimental database of 595 samples is compiled from the literature and utilized to find the best ANN architecture. The cement content, water content, coarse aggregate content, fine aggregate content, GGBFS content, carboxylic type hyper plasticizing content, superplasticizer content, and testing age are the eight inputs in this database. As a result, the optimal selection of the ANN design is carried out and evaluated using conventional statistical metrics. The results demonstrate that utilizing the best architecture [8–14–4–1] among the 240 investigated architectures, and the best ANN model, is a very efficient predictor of the compressive strength of concrete using GGBFS, with a maximum R2 value of 0.968 on the training part and 0.965 on the testing part. Furthermore, a sensitivity analysis is performed over 500 Monte Carlo simulations using the best ANN model to determine the reliability of ANN model in predicting the compressive strength of concrete. The findings of this research may make it easier and more efficient to apply the ANN model to many civil engineering challenges.

Design of ANN Based Machine Learning Method for Crop Prediction

In agriculture, crop yield estimation is critical. Remote sensing is being used in farming systems to increase yield efficiency and lower operating costs. Remote sensing-based strategies, on the other hand, necessitate extensive processing, necessitating the use of machine learning models for crop yield prediction. Descriptive analytics is a form of analytics that is used to accurately estimate crop yields. This paper discusses the most recent research on machine learning-based strategies for efficient crop yield prediction. In general, the training model's accuracy should be higher, and the error rate should be low. As a result, significant effort is being put forward to propose a machine learning technique that will provide high precision in crop yield prediction.

Application of Artificial Neural Network for Internal Combustion Engines: A State of the Art Review

The automotive industry is facing a crucial time. The transformation from internal combustion engines to new electrical technologies requires enormous investment, and hence the IC engines are likely to serve as a means of transportation for the coming decades. The search for sustainable green alternative fuel and operating parameter optimization is a current feasible solution and is a critical issue among the scientific community. Engine experiments are complicated, costly, and time-consuming, especially when the global economy is drastically down due to the COVID-19 pandemic and putting the limitation of social distancing. Industries are looking for proven computational solutions to address these issues. Recently, artificial neural network has been proven beneficial in several areas of engineering to reduce the time and experimentation cost. The IC engine is one of them. ANN has been used to predict and analyze different characteristics such as performance, combustion, and emissions of the IC engine to save time and energy. The complex nature of ANN may lead to computation time, energy, and space. Recent studies are centered on changing the network topology, deep learning, and design of ANN to get the highest performance. The present study summarizes the application of ANN to predict and optimize the complicated characteristics of various types of engines with different fuels. The study aims to investigate the network topologies adopted to design the model and thereafter statistical evaluation of the developed ANN models. A comparison of the ANN model with other prediction models is also presented.

Surface qualitative analysis and ANN modelling for pool boiling heat transfer using Al2O3-water based nanofluids

The present investigation is focussed on the surface associated qualitative analysis for nucleate pool boiling heat transfer characteristics of alumina-water based nanofluids. An optimal ANN design has been developed with various training algorithms and hidden layer based on experimental results. The nanofluids were prepared using distilled water and alumina nanoparticles of 40 nm size. The particle concentrations varied form 0.01 wt.% to 1 wt.%. The characterisation of nanoparticles and nanofluids was done using FESEM, EDAX, zeta potential stability test and thermophysical analysis. The pool boiling setup validation was done by comparing experimental results with Rohsenow’s correlation outputs. The boiling heat transfer was performed by varying input heat flux and calculating heat transfer coefficients with surface temperatures. The post boiling analysis was also performed on the nanoparticle deposited surfaces. The effect of nanoparticle concentrations and microlayer particle deposition is studied using surface roughness analysis, contact angle measurement and microscopic images. The pool boiling heat transfer with nanofluids showed an enhancement of up to 70 % as compared to distilled water. The improved thermophysical properties of nanofluids and the enhanced surface wettability provided an overall increase in heat transfer coefficients. The particle deposited surfaces showed heat transfer deterioration of up to 40 % with water. The ANN model analysis showed that the LM (Levenberg Marquardt) algorithm provided the most optimised structure. The combination of number of epochs and MSEs (Mean square errors) were taken for selecting the best model.

Neural network-based design and evaluation of performance metrics using adaptive line enhancer with adaptive algorithms for auscultation analysis

Auscultation is an important key for the physical (respiratory and circulatory) examination and is helpful in diagnosing various disorders. Auscultation is performed for the purposes of examining the circulatory and respiratory sounds and gastrointestinal system (bowel sounds). Besides inconsistencies in the propagation of the normal sounds, there are also several types of specific irregularities that can be heard in respiratory sounds: commonly known abnormal sounds in lung sound (wheezes, crackles, stridor, squawks, rhonchi and crackles) and heart sounds (heart murmurs). However, detection of abnormal sounds during auscultation needs extensive training and experience. Real-time separation of these heart sound signals from the lung sound signals is of great research interest and difficult to achieve. In this work, the authors proposed a novel adaptive line enhancer using nonlinear ANN design used for auscultation analysis. Our proposed designs are trained using different networks training algorithms which resulted in better mean square error reduction within compact time.

Heterogeneous Sensor Data Analysis Using Efficient Adaptive Artificial Neural Network on FPGA Based Edge Gateway

We propose a FPGA based design that performs real-time power-efficient analysis of heterogeneous sensor data using adaptive ANN on edge gateway of smart military wearables. In this work, four independent ANN classifiers are developed with optimum topologies. Out of which human activity, BP and toxic gas classifier are multiclass and ECG classifier is binary. These classifiers are later integrated into a single adaptive ANN hardware with a select line( s) that switches the hardware architecture as per the sensor type. Five versions of adaptive ANN with different precisions have been synthesized into IP cores. These IP cores are implemented and tested on Xilinx Artix-7 FPGA using Microblaze test system and LabVIEW based sensor simulators. The hardware analysis shows that the adaptive ANN even with 8-bit precision is the most efficient IP core in terms of hardware resource utilization and power consumption without compromising much on classification accuracy. This IP core requires only 31 microseconds for classification by consuming only 12 milliwatts of power. The proposed adaptive ANN design saves 61% to 97% of different FPGA resources and 44% of power as compared with the independent implementations. In addition, 96.87% to 98.75% of data throughput reduction is achieved by this edge gateway.

Performance Analysis for Smooth Power Generation of WECS System Using ANN/PI Based Control Techniques

In this paper, the physical control techniques of tip speed ratio and blade pitch angle of wind energy conversion system for smooth power generation has been demonstrated. To extract smooth power from WECS, it is required to control the system in such a way that it must produce a constant torque. ANN and PI are the most effective control techniques for the optimal control of TSR and BPA to maintain the constant torque. The proposed system contains the design of ANN and PI based controller for performance analysis of WECS. MATLAB/Simulink environment is used to implement the system to test on different uncertain conditions of wind. The obtained results are found suitable with the design ANN and PI based controller WECS system.

Thrice Filtered Information Energy Based Particle Swarm Feature Selection (TFIE-PSFS) Method Based Artificial Neural Network Classification for Improving Heart Disease Diagnosis

Diagnosing the existence of heart disease is really tedious process, as it entails deep knowledge and opulent experience. As a whole, the forecast of heart disease lies upon the conventional method of analysing medical report such as ECG (The Electrocardiogram), MRI (Magnetic Resonance Imaging), Blood Pressure, Stress tests by a Medicinal expert. Nowadays, a large volume of medical statistics is obtainable in medical industry and turns as a excessive source of forecasting valuable and concealed facts in almost all medical complications. Thus, these facts would really aid the doctors to create exact predictions. The innovative methods of Artificial Neural Network models have also been contributing themselves in yielding the main prediction accuracy over medical statistics. This paper targets to predict the presence of heart disease utilizing Back Propagation MLP (Multilayer Perceptron) of Artificial Neural Network. The proposed ANN design targeted to generate the three outputs Yes (Patient having heart disease), No (Patient not having heart disease), and Hesitant (Patient those who are in between yes and no category).

Electrical Energy Management Based on a Hybrid Artificial Neural Network-Particle Swarm Optimization-Integrated Two-Stage Non-Intrusive Load Monitoring Process in Smart Homes

Concerning electrical energy used in today’s modern society, electrical energy demands requested from downstream sectors in a smart grid are continuously increasing. One way to meet the electrical demands requested is to monitor and manage industrial, commercial, as well as residential electrical appliances efficiently in response to Demand Response (DR) programs for Demand-Side Management (DSM). Monitoring and managing electrical appliances that consume electrical energy in fields of interest can be realized through use of Energy Management Systems (EMS) with Non-Intrusive Load Monitoring (NILM). This paper presents an Internet of Things (IoT)-oriented Home EMS (HEMS). Also, a novel hybrid Artificial Neural Network-Particle Swarm Optimization (ANN-PSO)-integrated NILM approach is proposed and used to model and identify electrical appliances for DSM in the HEMS. ANN can be applied in NILM as a load identification task. Nevertheless, the performance of ANN used for load identification depends on three principal design factors: The network topology designed, the type of activation functions chosen, and the training algorithm adopted. As a result, PSO is conducted and used to incorporate meta-heuristics with ANN considering the three principal design factors relating to an ANN design. The HEMS with the novel hybrid ANN-PSO-integrated NILM proposed in this paper was deployed and evaluated in a realistic residential house environment. As the experimentation reported in this paper shows, the presented HEMS utilizing the proposed novel hybrid ANN-PSO-integrated NILM to model and identify monitored electrical appliances is feasible and workable, with an overall classification rate of 91.67% in load classification for DSM.

Intelligent tools to model photocatalytic degradation of beta‐naphtol by titanium dioxide nanoparticles

Feasibility of applying intelligent tools in prediction and optimization of photocatalytic degradation of beta‐naphthol using the titanium dioxide (TiO2) nanoparticles were conducted in this study. Biphasic TiO2 nanoparticles were synthesized using the controlled hydrolysis of TiCl4, and their properties were studied using the X‐ray diffraction and transmission electron microscopy methods. Therefore, factors affecting photocatalytic degradation of beta‐naphthol including impurity concentration, catalyst content, acidity, and aeration rate were monitored and controlled. The laboratory data showed that degradation rate of beta‐naphthol is a complicated nonlinear function of monitored variables. Two models including artificial network trained with particle swarm optimization (ANN‐PSO) and adaptive neuro‐fuzzy interference system trained with particle swarm optimization (ANFIS‐PSO) were used for prediction of this system. The results showed presence of a significant relation between the real and predicted data of these 2 models. However, ANFIS‐PSO can be more efficiently applied for prediction and optimization of photocatalytic behavior of TiO2 nanoparticles as for degradation of beta‐naphthol as compared to ANN‐PSO. As an advantage, ANFIS eliminates the problems of fuzzy logic, such as creation of membership functions, and local minima, which should be located in design of ANN, and through PSO algorithm, it could be a very powerful tool for simulating kinds of processes.

Artificial Neural Network model for the determination of GSM Rxlevel from atmospheric parameters

Accurate received signal level (Rxlevel) values are useful for mobile telecommunication network planning. Rxlevel is affected by the dynamics of the atmosphere through which it propagates. Adequate knowledge of the prevailing atmospheric conditions in an environment is essential for proper network planning. However most of the existing GSM received signal determination model are function of distance between point of signal reception and transmitting site thus necessitating the development of a model that involve the use of atmospheric parameters in the determination of received GSM signal level. In this paper, a three stage approach was used in the development of the model using some atmospheric parameters such as atmospheric temperature, relative humidity and dew point. The selected and easily measurable atmospheric parameters were used as input parameters in developing two new models for computing the Rxlevel of GSM signal using a three-step approach. Data acquisition and pre-processing serves as the first stage and formulation of ANN design and the development of parametric model for the Rxlevel using ANN synaptic weights form the second stage of the proposed approach. The third stage involves the use of ANN weight and bias values, and network architecture in the development of the model equation. In evaluating the performance of the proposed models, network parameters were varied and the results obtained using mean squared error (MSE) as performance measure showed the developed model with 33 neurons in the hidden layer and tansig activation, function in both the hidden and output layers as the optimal model with least MSE value of 0.056. Thus showing that the developed model has an acceptable accuracy value as demonstrated from comparison of results with actual measured values.

Effect of Operating Parameters on Ethanol–Water Vacuum Separation in an Ethanol Dehydration Apparatus and Process Modeling with ANN

Abstract Bioethanol has been found to be a suitable substitute for gasoline in internal combustion engines. It could be used either in an undiluted form or blended with gasoline. To blend the ethanol and gasoline, the water content of ethanol should reach 0.5% or less. In the present research work, 3A Zeolite was used as an absorbent with vacuum distillation. The effects of the operating parameters such as temperature, vacuum pressure and vapor flow rate on ethanol–water separation were investigated. Final ethanol concentration was obtained at the end of every run as well as the concentration of outlet ethanol. Both linear regression and ANN design were used to determine the best fit for two final parameters. The optimum condition was obtained at 0.4 bar vacuum pressure and 20 l/min ethanol–water vapor flow rate. ANN model is more qualified to the simulation of outspread data while the linear regression is not. L10L10 mode and L5T10 mode provide the best results for final concentration and total time, respectively. The Trainlm Algorithm like the previous research training algorithm is the best.

Design of ANN for prediction of operating performance Gas Sweetening Process used in Natural Gas Purification

Gas sweetening unit is an essential industrial processthat exhibits significant non-linear behavior. It comprises of absorption and regeneration towers.Conventional linear control schemes based on rigorous mathematical models, implemented in sweetening units, show poor performance and lead to off-specification products. A feed forward back-propagationneural networkmodelwith an efficiency of 95% has been developed for the Gas Sweetening Process used in Natural gas purification process, as per the data extracted from the GSU specifics.

An Integrated RS and ANN Design Method for City’s Industry Development Stage Identification

An Integrated RS and ANN Design Method for City’s Industry Development Stage Identification

Solar radiation: Cloudiness forecasting using a soft computing approach

Solar energy is one of the most important energy sources with increasing penetration into the power supply systems ofmany countries, due to the reduced environmental impact of its operation. One of the important factors of the efficiency ofphotovoltaic systems is the predictability of solar radiation, which depends on the clouds and the meteorologicalconditions, the occurrence of which is a non-linear process. Prediction of clouds’ amount that affects solar radiation usingArtificial Neural Networks (ANNs) is presented in this paper. To our knowledge, this approach is the first computationalintelligence method that refers to cloudiness forecasting. Input parameter selection is very critical in ANN design. In thisstudy, ten meteorological and temporal variables were selected in the implementation of the ANN structure. Theappropriate ANN architecture consists of a hidden layer with hyperbolic tangent sigmoid transfer function and an outputlayer with a saturating linear transfer function. The most important benefits of the proposed method are the cloudinessforecasting in the next half hour, the embedment of temporal parameters for cloudiness prediction and the highest quantityof data comparatively with other similar studies.

A DOE based approach for the design of RBF artificial neural networks applied to prediction of surface roughness in AISI 52100 hardened steel turning

The use of artificial neural networks for prediction in hard turning has received considerable attention in literature. An often quoted drawback of ANNs is the lack of a systematic way for the design of high performance networks. This study presents a DOE based approach for the design of ANNs of Radial Basis Function (RBF) architecture applied to surface roughness prediction in turning of AISI 52100 hardened steel. Experimental factors are the number of radial units on the hidden layer, the algorithm employed to calculate the spread factor of radial units and the algorithm employed to calculate radial function centers. DOE is employed to select levels of factors that benefit network prediction skills. Experiments with data sets of distinct sizes were conducted and network configurations leading to high performance were identified. ANN models obtained proved capable to predict roughness in accurate, precise and affordable way. Results pointed significant factors for network design and revealed that interaction effects between design parameters have significant influence on network performance for the task proposed. The work concludes that the DOE methodology constitutes a better approach to the design of RBF networks for roughness prediction than the most common trial and error approach.