Publications Scientifiques

Permanent URI for this communityhttps://dspace.univ-boumerdes.dz/handle/123456789/10

Browse

Subject: search.filters.subject.MLP

Search Results

Now showing 1 - 10 of 15
  • No Thumbnail Available
    Item
    On the capability of multilayer perceptron to predict total organic carbon and elemental capture spectroscopy data in unconventional hydrocarbon reservoirs: the case of the Barnett Shale and Bakken oil field
    (Obstetrics & Gynecology Science, 2025) Aliouane , L.; Ouadfeul, S.-A.
    This study explores the capability of a Multilayer Perceptron (MLP) neural network (NN) machine to predict missing or expensive core rocks and well-log data measurements such as the Total Organic Carbon (TOC) and Elemental Capture Spectroscopy (ECS) measurements. Data of boreholes drilled in the Lower Barnett Shale and Bakken oil and gas fields, located in the USA, are used. TOC estimation is first addressed using the Schmoker method in the Barnett Shale gas and Bakken oil reservoirs, followed by the implementation of MLP NNs trained with various learning algorithms such as the Hidden Weight Optimisation, the Conjugate Gradient, and the Levenberg-Marquardt. Input data include standard well logs such as sonic, gamma ray, resistivity, and neutron porosity. The MLP models are validated and generalised using both horizontal and vertical well data. Furthermore, ECS data prediction is performed using MLPs trained on elementary analysis-derived log parameters, offering a cost-effective alternative to direct ECS logging. The results demonstrate that the efficiency and reliability of MLP-based approaches in enhancing geochemical and petrophysical characterisation of subsurface formations is conditioned by the choice of the learning algorithm, the reservoir complexity, number of wells, and their distribution
  • No Thumbnail Available
    Item
    On the capability of multilayer perceptron to predict total organic carbon and elemental capture spectroscopy data in unconventional hydrocarbon reservoirs: the case of the Barnett Shale and Bakken oil field
    (Obstetrics & Gynecology Science, 2025) Aliouane, L.; Ouadfeul, S.-A.
    This study explores the capability of a Multilayer Perceptron (MLP) neural network (NN) machine to predict missing or expensive core rocks and well-log data measurements such as the Total Organic Carbon (TOC) and Elemental Capture Spectroscopy (ECS) measurements. Data of boreholes drilled in the Lower Barnett Shale and Bakken oil and gas fields, located in the USA, are used. TOC estimation is first addressed using the Schmoker method in the Barnett Shale gas and Bakken oil reservoirs, followed by the implementation of MLP NNs trained with various learning algorithms such as the Hidden Weight Optimisation, the Conjugate Gradient, and the Levenberg-Marquardt. Input data include standard well logs such as sonic, gamma ray, resistivity, and neutron porosity. The MLP models are validated and generalised using both horizontal and vertical well data. Furthermore, ECS data prediction is performed using MLPs trained on elementary analysis-derived log parameters, offering a cost-effective alternative to direct ECS logging. The results demonstrate that the efficiency and reliability of MLP-based approaches in enhancing geochemical and petrophysical characterisation of subsurface formations is conditioned by the choice of the learning algorithm, the reservoir complexity, number of wells, and their distribution
  • No Thumbnail Available
    Item
    A comparative study between convolutional and multilayer perceptron neural networks classification models
    (2019) Bachiri, Mohamed Elssaleh; Harrar, Khaled
    Image classification plays an important role in image processing, computer vision, and machine learning. This paper deals with image classification using deep learning. For this, a conventional neural network (CNN) and multilayer perceptron neural network (MLP) models were used for the classification. The two models were implemented on the MNIST dataset which was used at 100% and half of capacity, The models were trained with fixed and flexible number of epochs in two runs. CNN provided an accuracy of 98,43% with a loss of 4,44%, where MLP reached 92,80% of classification with a loss of 25,87%. Indeed, for each model, variables as number of filters, size, and activation functions were discussed. The CNN demonstrated a good performance providing high accuracy for image and also proved to be a better candidate for data applications.
  • No Thumbnail Available
    Item
    Gear fault detection, identification and classification using MLP neural network
    (Springer, 2023) Afia, Adel; Ouelmokhtar, Hand; Gougam, Fawzi; Touzout, Walid; Rahmoune, Chemseddine; Benazzouz, Djamel
    Gear fault detection, identification and classification are highly complicated tasks, as the faults which affect gearboxes tend to share similar frequency signatures. Therefore, load and speed changes in a rotating machinery inevitably provide inaccurate results. However, identifying the fault remains critical, as each individual gear fault influences overall mechanism operation in different manners. Therefore, defect identification and classification appear as the hardest challenge for a geared systems. An automatic method to detect, identify and classify different gear failures is presented in this paper. The intelligent approach consists of a combination of MODWPT, entropy and MLPNN. MODWPT was developed to decompose the signals with a uniform frequency bandwidth. Entropy is employed to build the feature matrix in the feature extraction phase. Then, MLP offers a very efficient classification tool for features classification stage. Based on data sets taken from a gearbox bench test with a good and five varied gear states under various loads and speeds, experimental results presented the efficiency of our technique
  • No Thumbnail Available
    Item
    Levenberg-Marquardt algorithm neural network for clay volume estimation from well-log data in an unconventional tight sand gas reservoir of Ahnet basin (Algerian Sahara)
    (Istituto Nazionale di Oceanografia e di Geofisica Sperimentale, 2022) Aliouane, Leila
    The main goal of this paper is to show the contribution of artificial intelligence, namely a neural network, in reservoir characterisation to predict the clay volume in an unconventional tight sand gas reservoir. Clay volume is usually estimated using the natural gamma ray log, which can give bad results if non-clayey radioactive minerals are present in the reservoir. Our purpose is to implement a multilayer perceptron neural network machine to predict the clay volume using the conventional well-log data as an input and the measured mineralogical component, as desired output with a Levenberg-Marquardt algorithm. Application to two Ordovician reservoir intervals of a borehole located in the Ahnet basin in the Algerian Sahara shows the contribution and the efficacy of the implemented neural network machine in unconventional tight sand reservoirs characterisation
  • No Thumbnail Available
    Item
    Prediction of Wax Appearance Temperature Using Artificial Intelligent Techniques
    (Springer, 2020) Benamara, Chahrazed; Gharb, Kheira; Nait Amar, Menad; Hamada, Boudjema
    The paraffin particles can promote and be involved in the formation of deposits which can lead to plugging of oil production facilities. In this work, an experimental prediction of wax appearance temperature (WAT) has been performed on 59 Algerian crude oil samples using a pour point tester. In addition, a modeling investigation was done to create reliable WAT paradigms. To do so, gene expression programming and multilayers perceptron optimized with Levenberg–Marquardt algorithm (MLP-LMA) and Bayesian regularization algorithm were implemented. To generate these models, some parameters, namely density, viscosity, pour point, freezing point and wax content in crude oils, have been used as input parameters. The results reveal that the developed models provide satisfactory results. Furthermore, the comparison between these models in terms of accuracy indicates that MLP-LMA has the best performances with an overall average absolute relative error of 0.23% and a correlation coefficient of 0.9475.
  • No Thumbnail Available
    Item
    Real-Time prediction of plastic viscosity and apparent viscosity for Oil-Based drilling fluids using a committee machine with intelligent systems
    (Springer, 2022) Youcefi, Mohamed Riad; Hadjadj, Ahmed; Bentriou, Abdelak; Boukredera, Farouk Said
    he prediction of drilling mud rheological properties is a crucial topic with significant importance in analyzing frictional pressure loss and modeling the hole cleaning. Based on Marsh viscosity, mud density, and solid percent, this paper implements a committee machine intelligent system (CMIS) to predict apparent viscosity (AV) and plastic viscosity (PV) of oil-based mud. The established CMIS combines radial basis function neural network (RBFNN) and multilayer perceptron (MLP) via a quadratic model. Levenberg–Marquardt algorithm was applied to optimize the MLP, while differential evolution, genetic algorithm, artificial bee colony, and particle swarm optimization were used to optimize the RBFNN. A databank of 440 and 486 data points for AV and PV, respectively, gathered from various Algerian fields was considered to build the proposed models. Statistical and graphical assessment criteria were employed for investigating the performance of the proposed CMIS. The obtained results reveal that the developed CMIS models exhibit high performance in predicting AV and PV, with an overall average absolute relative deviation (AARD %) of 2.5485 and 4.1009 for AV and PV, respectively, and a coefficient of determination (R2) of 0.9806 and 0.9753 for AV and PV, respectively. A comparison of the CMIS-AV with Pitt's and Almahdawi's models demonstrates its higher prediction capability than these previously published correlations
  • No Thumbnail Available
    Item
    Modeling viscosity of CO 2 at high temperature and pressure conditions
    (Elsevier, 2020) Nait Amar, Menad; Ghriga, Mohammed Abdelfetah; Ouaer, Hocine; Ben Seghier, Mohamed El Amine; Thai Pham, Binh
    The present work aims at applying Machine Learning approaches to predict CO2 viscosity at different thermodynamical conditions. Various data-driven techniques including multilayer perceptron (MLP), gene expression programming (GEP) and group method of data handling (GMDH) were implemented using 1124 experimental points covering temperature from 220 to 673 K and pressure from 0.1 to 7960 MPa. Viscosity was modelled as function of temperature and density measured at the stated conditions. Four backpropagation-based techniques were considered in the MLP training phase; Levenberg-Marquardt (LM), bayesian regularization (BR), scaled conjugate gradient (SCG) and resilient backpropagation (RB). MLP-LM was the most fit of the proposed models with an overall root mean square error (RMSE) of 0.0012 mPa s and coefficient of determination (R2) of 0.9999. A comparison showed that our MLP-LM model outperformed the best preexisting Machine Learning CO2 viscosity models, and that our GEP correlation was superior to preexisting explicit correlations.
  • No Thumbnail Available
    Item
    Hybrid soft computational approaches for modeling the maximum ultimate bond strength between the corroded steel reinforcement and surrounding concrete
    (Springer, 2020) Ben Seghier, Mohamed El Amine; Ouaer, Hocine; Ghriga, Mohammed Abdelfetah; Nait Amar, Menad; Duc-Kien, Thai
    The capacity efficiency of load carrying with the accurate serviceability performances of reinforced concrete (RC) structure is an important aspect, which is mainly dependent on the values of the ultimate bond strength between the corroded steel reinforcements and the surrounding concrete. Therefore, the precise determination of the ultimate bond strength degradation is of paramount importance for maintaining the safety levels of RC structures affected by corrosion. In this regard, hybrid intelligence and machine learning techniques are proposed to build a new framework to predict the ultimate bond strength in between the corroded steel reinforcements and the surrounding concrete. The proposed computational techniques include the multilayer perceptron (MLP), the radial basis function neural network and the genetic expression programming methods. In addition to that, the Levenberg–Marquardt (LM) deterministic approach and two meta-heuristic optimization approaches, namely the artificial bee colony algorithm and the particle swarm optimization algorithm, are employed in order to guarantee an optimum selection of the hyper-parameters of the proposed techniques. The latter were implemented based on an experimental published database consisted of 218 experimental tests, which cover various factors related to the ultimate bond strength, such as compressive strength of the concrete, concrete cover, the type steel, steel bar diameter, length of the bond and the level of corrosion. Based on their performance evaluation through several statistical assessment tools, the proposed models were shown to predict the ultimate bond strength accurately; outperforming the existing hybrid artificial intelligence models developed based on the same collected database. More precisely, the MLP-LM model was, by far, the best model with a determination coefficient (R2) as high as 0.97 and 0.96 for the training and the overall data, respectively.
  • No Thumbnail Available
    Item
    An efficient methodology for multi-objective optimization of water alternating CO2 EOR process
    (Elsevier, 2019) Nait Amar, Menad; Zeraibi, Noureddine