Levenberg-Marquardt Back-propagation Learning Algorithm Research Articles

Deploying artificial neural network to predict hybrid biodiesel fuel properties from their fatty acid compositions

Measurement-related problems have spurred fuel properties prediction using machine learning techniques. Improved fuel properties offered by hybrid biodiesel (HB) via mixed oils were predicted from their fatty acid compositions (FACs) using artificial neural network (ANN). FACs and fuel properties of HB sourced from the literature were used to develop ANN models. FAC data were used as the input parameters to predict the fuel properties data (kinematic viscosity (KV), density, calorific value (CV), and flash point (FP)) considered as the output parameters of the models. Using the multilayer perception ANN, the models were trained using Levenberg-Marquardt back propagation learning algorithm coupled with different numbers of neurons and activation functions for the prediction of the fuel properties. The models were observed to accurately predict these fuel properties with high prediction accuracy (R 2 = 1). The evaluated model performance errors were 0.1014 and 0.0504, 0.2905 and 0.4225, 0.1848, and 0.1038, and 0.4726 and 0.7833 for KV, density, CV, and FP using root mean square error and average absolute deviation respectively. Prediction performance and error estimates were slightly better than those for single feedstock biodiesel. Hence, this study shows the ability of ANN to accurately predict the fuel properties of HB from the FAs.

Comparison of Linear Regression and Artificial Neural Network Models for the Dimensional Control of the Welded Stamped Steel Arms

The production of parts by pressing and subsequent welding is commonly used in the automotive industry. The disadvantage of this method of production is that inaccuracies arising during pressing significantly affect the final dimension of the part. However, this can be corrected by the choice of the technological parameters of the following operation—welding. Suitably designed parameters make it possible to partially eliminate inaccuracies arising during pressing and thus increase the overall applicability of this technology. The paper is focused on the upper arm geometry of a car produced in this manner. There have been two neural networks proposed in which the optimal welding parameters are determined based on the stamped dimensions and the desired final dimensions. The Levenberg–Marquardt back-propagation algorithm and the Bayesian regularised back-propagation algorithm were used as the learning algorithm for ANNs in multi-layer feed-forward networks. The outputs obtained from the neural networks were compared with a linear prediction model based on a on the design of experiment methodology. The mean absolute percentage error of the linear regression model on the entire dataset was 3 × 10−3%. A neural network with Levenberg–Marquardt back-propagation learning algorithm had a mean absolute percentage error of 4 × 10−3. Similarly, a neural network with a Bayesian regularised back-propagation learning algorithm had a mean absolute percentage error of 3 × 10−3%.

A novel prediction and control method for solar energy dispatch based on the battery energy storage system using an experimental dataset

The high power generation growth by photovoltaic systems needs to forecast the power generation profile during a day. It is also required to evolve the high-efficient and optimal on-grid/off-grid photovoltaic power generation units. Furthermore, some advantages can be achieved by integrating photovoltaic systems with storage devices such as battery energy storage systems. Thus, optimizing the hybrid systems comprising photovoltaic and battery energy storage systems is needed to evaluate the best capacity. In the present work, a novel control and sizing scheme is proposed for the battery energy storage system in a photovoltaic power generation plant in one-hour ahead and one-day ahead during the dispatching phase. Then, the proposed prediction strategy is recommended for solar irradiation and power utilization. The control approach comprises a predictive control method concerning a Radial Basis Function network optimized by Levenberg-Marquardt back-propagation learning algorithm. Using the RBF network for simulation leads to a WAPE% =1.68 %.

Artificial Neural Network to Predict the Thermal Drawdown of Enhanced Geothermal System

Abstract This work presents the prediction of thermal drawdown of an enhanced geothermal system (EGS) using artificial neural network (ANN). A three-dimensional numerical model of EGS was developed to generate the training and testing data sets for ANN. We have performed a quantitative study of geothermal energy production for various injection operating conditions and reservoir fracture aperture. Input parameters for ANN include temperature, mass flux, pressure, and fracture transmissivity, while the production well temperature is the output parameter. The Levenberg–Marquardt back-propagation learning algorithm, the tan-sigmoid, and the linear transfer function were used for the ANN optimization. The best results were obtained with an ANN architecture composed of eight hidden layers and 20 neurons in the hidden layer, which made it possible to predict the production temperature with a satisfactory range (R2 > 0.99). An appropriate accuracy of the ANN model was obtained with a percentage error less than (± 4.5). The results from the numerical simulations suggest that fracture transmissivity has less effect on thermal drawdown than the injection mass flux and temperature. From our results, we confirm that ANN modeling may predict the thermal drawdown of an EGS system with high accuracy.

Energy Storage Control for Dispatching Photovoltaic Power

The strong growth of the solar power generation industry requires an increasing need to predict the profile of solar power production over a day and develop highly efficient and optimized stand-alone and grid-connected photovoltaic systems. Moreover, the opportunities offered by battery energy storage systems (BESSs) coupled with photovoltaic (PV) systems require an ability to forecast the load power to optimize the size of the entire system composed of PV panels and storage devices. This paper presents a sizing and control strategy of BESSs for dispatching a photovoltaic generation farm in the 1-h ahead and day-ahead markets. The forecasting of the solar irradiation and load power consumption is performed by developing a predictive model based on a feed-forward neural network trained with the Levenberg–Marquardt back-propagation learning algorithm.

Modeling and Experimental Prediction of Wastewater Treatment Efficiency in Oil Refineries Using Activated Sludge Process

In this study, activated sludge process for wastewater treatment in a refinery was investigated. For such purpose, a laboratory scale rig was built. The effect of several parameters such as temperature, residence time, effect of Leca (filling-in percentage of the reactor by Leca) and UV radiation on COD removal efficiency were experimentally examined. Maximum COD removal efficiency was obtained to be 94% after final testing. An artificial neural network (ANN) was applied to evaluate the effect of operational parameters on the efficiency as the next step. A two-layered ANN provided the best results, using Levenberg–Marquardt back propagation learning algorithm (trainLM) in which tansig and purelin used as transfer functions in the hidden and output layers. Furthermore, the application of three neurons in the hidden layer caused to gratify network training while overfitting was hindered. ANN model, provided a good estimation for correlation coefficient and the mean square error (MSE) which calculated 0.997 and 0.5 × 10-3 respectively.

Operation simulation of a recycled electrochemical ozone generator using artificial neural network

The present work has focused on the modeling and simulation of a recycled ozone generator system via electrochemical oxidation of water. To produce ozone, a Pyrex glass electrochemical reactor, comprised of two separate half-cell by Nafion 117 membrane was applied. The applied anode and cathode electrodes were Ti/Sn-Sb-Ni and platinized titanium, respectively. The modeling and simulation of the reactor operation were done via artificial neural network (ANN) technique. In this regard, four important operational parameters (i.e. electrolyte concentration, applied voltage, flow rate and electrolysis time) and the generated ozone concentration were considered as the independent inputs and the network output, respectively. To find out the best model, six numbers of three-layered ANNs with different functions were constructed and optimized. Best simulation was related to a model, consist of Levenberg–Marquardt Back propagation learning algorithm (trainlm) and tangent sigmoid (tansig) as transfer function in the both hidden and output layers. Also, application of 10 hidden neurons and 80 iterations for the network calibration caused to satisfy the network training while overfitting was prevented. The K-fold cross-validation method, employed for the model evaluation, showed high correlation coefficient (0.9936) and low mean square error (3.58×10−4) for the testing data. Sensitivity analysis indicated order of relative importance the operational parameters on the ozone production as: time>[electrolyte]>voltage>flow rate.

Predicting the Heating Value of Municipal Solid Waste-based Materials: An Artificial Neural Network Model

Municipal solid waste has been recently considered as one of the important renewable energy resources. For municipal solid waste-based energy systems, municipal solid waste heating value is a critical variable. In this study, an artificial neural network model based on Levenberg-Marquardt backpropagation learning algorithm has been developed to predict higher heating value of municipal solid waste using their contents of water, carbon, hydrogen, nitrogen, oxygen, sulfur, and ash. Artificial neural network model has been able to predict the higher heating values of municipal solid waste with a tangent sigmoid transfer function at hidden layer with five neurons and a linear transfer function at output layer. Mean-squared error of prediction results in the validation process has been found to be 0.0137. The regression between network outputs and corresponding targets has been showed to be acceptable with a high correlation coefficient of 0.991. It can be deduced that the developed model is a robust estimation tool for higher heating value of municipal solid waste-based materials.