Computing the pull-in voltage of fixed–fixed micro-actuators by artificial neural network

Variants of feedforward backpropagation (BP) artificial neural network (ANNs) such as pseudo Mac Laurin power series order 1 (McL1BP), order 2 (McL2BP), order 3 (McL3BP) and order 4 (McL4BP) models are used to fill in mean annual streamflows. The baseline of modeling process is the standard feedforward backpropagation ANN (StandBP). The sigmoid function is used as activation function. Performance comparisons of data infilling models (ANNs) are conducted using the Root Mean Square Error of Predictions (RMSEp) and graphical plots. To test the performance of the data infilling techniques, selected streamflow gauges (i.e. the Diepkloof (control) gauge on the Wonderboomspruit River and the Molteno (target) gauge on Stormbergspruit River) of the Orange drainage river systems of South Africa have been used. The results demonstrated that relatively higher order ANN models; i.e. McL3BP, McL4BP can still outperform the rest of techniques for 7 % (except StandBP), 20 % and 30 % missing data proportions at Molteno gauge. At the same time they have been outperformed by the rest of ANN models at 13 %. Generally higher and low order Pseudo MacLaurin ANNs are acceptable to fill in missing mean annual flows at Molteno gauge. The accuracy of estimated mean annual values at Molteno is substantially negatively affected beyond 20 % gap size for all ANNs. Hence gap size beyond 20 % yields to relatively higher value of RMSEp for both high and low MacLaurin feedforward ANN techniques. A linear relationship could describe accuracy of estimated values and gap size at Molteno. Further work should include other data regimes as well as other South African catchment areas. Other activation functions should also be tested.

An annular ring compact microstrip antenna (ARCMA) constructed by loading a circular slot in the center of the circular patch antenna is a popular microstrip antenna due to its favorable properties. In this study, a method based artificial neural networks (ANNs) has been firstly applied for the computing the resonant frequency of ARCMAs. Multilayered perceptron model based on feed forward back propagation ANN has been utilized, and the constructed model have been separately trained with 8 different learning algorithms to achieve the best results regarding the resonant frequency of ARCMAs at dominant mode. To this end, the resonant frequencies of 80 ARCMAs with varied dimensions and electrical parameters in accordance with UHF band covering GSM, LTE, WLAN and WiMAX applications were simulated with a robust numerical electromagnetic computational tool, $$\hbox {IE3D}^\mathrm{TM}$$IE3DTM, which is based on method of moment. Then, ANN model was constructed with the simulation data, by using 70 ARCMAs for training and the remaining 10 for test. As the performances of the 8 learning algorithms are compared with each other, the best result is obtained with Levenberg---Marquardt algorithm. The proposed ANN model were confirmed by comparing with the suggestions reported elsewhere via measurement data published earlier in the literature, and they have further validated on an ARCMA fabricated in this study. The results achieved in this study show that ANN model learning with LM algorithm can be successfully used to compute the resonant frequency of ARCMAs without involving any sophisticated methods.

Experimental and theoretical evaluation of thermophysical properties for moist air within solar still by using different algorithms of artificial neural network

This article presents a theoretical background for an artificial neural network (ANN) model of the alpha type Stirling engine where thermodynamic dependencies, connected with equations of motion for the piston-crankshaft system with three degrees of freedom were taken into account. Because of the highly nonlinear description of Stirling engine dynamics, the ANN was employed, that modelled output power of Stirling engine as a function of the input power, molar mass, load current, pressure obtained by gas combustion and working parameters of the engine. The ANN model was tested on experimental data, gathered at the laboratory stand, in different working conditions. The proposed ANN model provides good results for both training and testing data-sets.

In this study, three artificial neural network (ANN) models are developed with back propagation (BP) optimization algorithms to predict various lightning damage modes in carbon/epoxy laminates. The proposed ANN models use three input variables associated with lightning waveform parameters (i.e., the peak current amplitude, rising time, and decaying time) to predict fiber damage, matrix damage, and through-thickness damage in the composites. The data used for training and testing the networks was actual lightning damage data collected from peer-reviewed published literature. Various BP training algorithms and network architecture configurations (i.e., data splitting, the number of neurons in a hidden layer, and the number of hidden layers) have been tested to improve the performance of the neural networks. Among the various BP algorithms considered, the Bayesian regularization back propagation (BRBP) showed the overall best performance in lightning damage prediction. When using the BRBP algorithm, as expected, the greater the fraction of the collected data that is allocated to the training dataset, the better the network is trained. In addition, the optimal ANN architecture was found to have a single hidden layer with 20 neurons. The ANN models proposed in this work may prove useful in preliminary assessments of lightning damage and reduce the number of expensive experimental lightning tests.

Artificial neural networks for infectious diarrhea prediction using meteorological factors in Shanghai (China)

Hydraulic jump has numerous applications in the field of hydraulic engineering, such as energy dissipation over spillways, chlorinating of wastewater and many others. The sequent depth ratio is one of the important characteristics of hydraulic jump useful in designing the stilling basin. Despite its importance, the exact value of sequent depth ratio is still undetermined. In the present study an attempt has been made to find out the effects of roughness heights and slopes by conducting an experimental study and artificial neural network (ANN) model. Three different roughness heights of crushed and rounded aggregates and two positive bed slopes were used. The experimental results show that the reductions in sequent depth ratios are more in the case of crushed aggregate (4%–35%) than rounded on the same slope. By increasing bed slope, the sequent depth ratios show increasing trend in the range 3%–45%. The proposed ANN model has the capability to predict the sequent depth ratio with least MAPE (mean absolute percentage error) value 3.15%. Therefore, based on the results obtained from the empirical model and ANN model, it has been concluded that the present study can be better utilized for the estimation of the sequent depth ratio of hydraulic jump.

A prediction on structural stress and deformation of fish cage in waves using machine-learning method

Experimental and numerical study of water sprayed turbulent combustion: Proposal of a neural network modeling for five-dimensional flamelet approach

Estimation of hydraulic jump on corrugated bed using artificial neural networks and genetic programming

Back pressure prediction of the direct air cooled power generating unit using the artificial neural network model

Micellar enhanced ultrafiltration (MEUF) of mercury-contaminated wastewater: Experimental and artificial neural network modeling

This paper deals with the extrusion of gear-like profiles and uses of finite element method (FEM) and artificial neural network (ANN) to predict the extrusion load. In the study, gear-like components has been manufactured by forward extrusion for the AA1070 aluminum alloy and the process was simulated by using a DEFORM-3D software package to establish a database in order to provide the data for ANN modeling. Serious experiments were performed for only one die set and four teeth gear profile to obtain data for comparing with DEFORM-3D results. After verifying a highly appropriate FEM simulation with the experiment at the same conditions, Results were enhanced for different die lengths, extrusion ratios, and two extra teeth number as three and six using FEM simulations. Subsequently, the data from the performed FEM simulations were submitted for the best obtained ANN model. Finally, a good agreement between FE-simulated and ANN-predicted results was obtained. The proposed ANN model is found to be useful in predicting the forming load of the different die set variations based on the reliable test data.

The generalization ability of artificial neural networks (ANNs) and M5 model tree (M5Tree) in modeling reference evapotranspiration (ET 0 ) is investigated in this study. Daily climatic data, average temperature, solar radiation, wind speed, and relative humidity from six different stations operated by California Irrigation Management Information System (CIMIS) located in two different regions of the USA were used in the applications. King-City Oasis Rd., Arroyo Seco, and Salinas North stations are located in San Joaquin region, and San Luis Obispo, Santa Monica, and Santa Barbara stations are located in the Southern region. In the first part of the study, the ANN and M5Tree models were used for estimating ET 0 of six stations and results were compared with the empirical methods. The ANN and M5Tree models were found to be better than the empirical models. In the second part of the study, the ANN and M5Tree models obtained from one station were tested using the data from the other two stations for each region. ANN models performed better than the CIMIS Penman, Hargreaves, Ritchie, and Turc models in two stations while the M5Tree models generally showed better accuracy than the corresponding empirical models in all stations. In the third part of the study, the ANN and M5Tree models were calibrated using three stations located in San Joaquin region and tested using the data from the other three stations located in the Southern region. Four-input ANN and M5Tree models performed better than the CIMIS Penman in only one station while the two-input ANN models were found to be better than the Hargreaves, Ritchie, and Turc models in two stations.

This study aimed to develop an artificial neural network (ANN) model to estimate the nitrate content in the surface water of the Gökırmak River. Samplings were carried out during 12 months from six stations between 2020 and 2021. Nitrate content varied between 0.20 and 2.70 mg l-1 while the mean value was 1.18 mg l-1 during the study period. The developed model consists of two input layers (month and station) and one output layer (nitrate content). Feed-forward backprop was used as the network type. Levenberg-Marquardt (TRAINLM) was used as a training function, LEARNGDM was used as an adaption learning function and mean squared error (MSE) was used as a performance function. The number of neurons was 10 and TANSIG was selected as transfer function. Epoch number adjusted 1000 iterations. ANN model predicted the nitrate content between 0.24 and 2.61 with a mean value of 1.16 mg l-1. The results showed that the best validation performance is 0.61264 at epoch 30. R values are 0.96257 and 0.84231 for training and testing, respectively. R-value was found 0.85352 for all data. In conclusion, this study presents the conception of an artificial neural network (ANN) model designed to predict nitrate concentrations in river water. The developed ANN model provides reasonable results for predicting the nitrate content using only given time and location inputs. More inputs can be included in future studies to ensure higher accuracy in the development of ANN models.