Pub Date : 2015-07-28DOI: 10.33899/rengj.2015.101541
M. Mohammad, A. Saadallah
The different types of mathematical models which are physically based models or the models based on laboratory experimental and/or field measurements are considered to estimate the runoff depth and runoff coefficient relating rainfall to runoff especially for the non-measured watersheds. Soil and Water Assessment Tool (SWAT) was applied to estimate the surface runoff depth of Al-Khoser River watershed, 696km 2 . The model was calibrated based on measured single storms data at the watershed outlet, while data measured at the outlet of sub watershed (36km 2) are considered for model verification. For model calibration, the resultant values are 0.99, 0.95, 0.73 and 0.19 for determination coefficient, index of agreement, model efficiency, and t-test respectively, while verification of the model revealed values of 0.97, 0.89, 0.72 and 0.32 for the same parameters considered respectively. The model was applied as a daily continuous simulation for the period 1991-2011. Relationships between the accumulated daily runoff depths and accumulated daily rainfall depths have been obtained, additionally the yearly runoff depth related to the yearly rainfall depth was also derived. Those relations can be used to estimate the runoff depth and runoff coefficient for different studies of the watershed
{"title":"Estimate the Runoff Depth by Using of SWAT Model at a Watershed in Northern Iraq","authors":"M. Mohammad, A. Saadallah","doi":"10.33899/rengj.2015.101541","DOIUrl":"https://doi.org/10.33899/rengj.2015.101541","url":null,"abstract":"The different types of mathematical models which are physically based models or the models based on laboratory experimental and/or field measurements are considered to estimate the runoff depth and runoff coefficient relating rainfall to runoff especially for the non-measured watersheds. Soil and Water Assessment Tool (SWAT) was applied to estimate the surface runoff depth of Al-Khoser River watershed, 696km 2 . The model was calibrated based on measured single storms data at the watershed outlet, while data measured at the outlet of sub watershed (36km 2) are considered for model verification. For model calibration, the resultant values are 0.99, 0.95, 0.73 and 0.19 for determination coefficient, index of agreement, model efficiency, and t-test respectively, while verification of the model revealed values of 0.97, 0.89, 0.72 and 0.32 for the same parameters considered respectively. The model was applied as a daily continuous simulation for the period 1991-2011. Relationships between the accumulated daily runoff depths and accumulated daily rainfall depths have been obtained, additionally the yearly runoff depth related to the yearly rainfall depth was also derived. Those relations can be used to estimate the runoff depth and runoff coefficient for different studies of the watershed","PeriodicalId":339890,"journal":{"name":"AL Rafdain Engineering Journal","volume":"59 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2015-07-28","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"133347819","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2015-07-28DOI: 10.33899/rengj.2015.101546
Abbas S. Hussein, A. Khedher
This research investigates windcatcher usage in Mosul for ventilation . The choice of a square shape windcatcher diagonally partitioned along its height attached to a room of dimensions 4x6x4 m. Indoor air motion speed and movement were studied to assess a windcatcher’s performance .A computational model is used with Autodesk Simulation CFD software for the simulation case. Part of this study was dedicated to find the geometrical characteristics of the windcatcher parameters: height (4 cases), crosssection (3 cases) and location with respect to the room (4 cases). Another part is to analyze a yearly weather data of Mosul to find the prevailing speed and direction for all cases (a range of 20-30C is chosen as a thermal comfort temperature range) and to use them as an input for the simulation to find the best case of the windcatcher. The results shows that the best heights of 4m , the best cross-section of 1x1 m and the best location of the wind catcher is found to be at the middle of long side of the room. The weather data shows that the prevailing wind speed between 1.62 to 4.83 m/s. The final result of the study confirm that the direction of the windcatcher and the room which can give the best performance using all the above selected parameters in Mosul is 337 NNW. :ملتسأ 20 – 5 2014 :لبق 28 – 4 2015 Al-Rafidain Engineering Vol. 23 No. 3 June 2015
{"title":"Simulation For the Buildings Ventilation using Windcatchers in Mosul","authors":"Abbas S. Hussein, A. Khedher","doi":"10.33899/rengj.2015.101546","DOIUrl":"https://doi.org/10.33899/rengj.2015.101546","url":null,"abstract":"This research investigates windcatcher usage in Mosul for ventilation . The choice of a square shape windcatcher diagonally partitioned along its height attached to a room of dimensions 4x6x4 m. Indoor air motion speed and movement were studied to assess a windcatcher’s performance .A computational model is used with Autodesk Simulation CFD software for the simulation case. Part of this study was dedicated to find the geometrical characteristics of the windcatcher parameters: height (4 cases), crosssection (3 cases) and location with respect to the room (4 cases). Another part is to analyze a yearly weather data of Mosul to find the prevailing speed and direction for all cases (a range of 20-30C is chosen as a thermal comfort temperature range) and to use them as an input for the simulation to find the best case of the windcatcher. The results shows that the best heights of 4m , the best cross-section of 1x1 m and the best location of the wind catcher is found to be at the middle of long side of the room. The weather data shows that the prevailing wind speed between 1.62 to 4.83 m/s. The final result of the study confirm that the direction of the windcatcher and the room which can give the best performance using all the above selected parameters in Mosul is 337 NNW. :ملتسأ 20 – 5 2014 :لبق 28 – 4 2015 Al-Rafidain Engineering Vol. 23 No. 3 June 2015","PeriodicalId":339890,"journal":{"name":"AL Rafdain Engineering Journal","volume":"25 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2015-07-28","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"134644998","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2015-07-28DOI: 10.33899/rengj.2015.101564
Basil M. Saied, A.H. Ahmed
This paper presents an air gap eccentricity fault Diagnosis of a salient Pole three phase synchronous generator. The detection and diagnosis of this type of fault is based on analyzing armature and field current signatures, using FFT . Therefore, the synchronous generator, under this type of fault, is modeled by using finite element method (FEM). Also practical results are performed in the lab with deliberately air gap eccentricity faults, by using samples of bearings malfunctioning or the removal of single or pair pole. The modeling and experimental results show that FEM is a reliable approach that can be used to model synchronous generator under electrical and mechanical faults.
{"title":"Mixed Air Gap Eccentricity Fault Diagnosis of a Salient Pole Synchronous Generator Based on Current Signature","authors":"Basil M. Saied, A.H. Ahmed","doi":"10.33899/rengj.2015.101564","DOIUrl":"https://doi.org/10.33899/rengj.2015.101564","url":null,"abstract":"This paper presents an air gap eccentricity fault Diagnosis of a salient Pole three phase synchronous generator. The detection and diagnosis of this type of fault is based on analyzing armature and field current signatures, using FFT . Therefore, the synchronous generator, under this type of fault, is modeled by using finite element method (FEM). Also practical results are performed in the lab with deliberately air gap eccentricity faults, by using samples of bearings malfunctioning or the removal of single or pair pole. The modeling and experimental results show that FEM is a reliable approach that can be used to model synchronous generator under electrical and mechanical faults.","PeriodicalId":339890,"journal":{"name":"AL Rafdain Engineering Journal","volume":"55 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2015-07-28","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"121031276","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2015-07-28DOI: 10.33899/rengj.2015.101570
A. A. Ramadhan
An experimental study of Kevlar-29/epoxy composite thin wall cylinders under internal pressure loading has been investigated in this work. Kevlar/epoxy composite specimens were fabricated and tested. Specimens were tested using the hydraulic pump and subjected to a static internal pressure. The results obtained from this work to find the pressure increase ratio and the failure load were found and a comparison has been achieved. According to the comparison results the tough material was specified. The efficient and effective composite cylinders have been tested experimentally. And this provided a considerable advantage over using a woven roving fiber in pressure vessels applications.
{"title":"An experimental Study of Kevlar/epoxy Composite cylinders under internal pressure loading","authors":"A. A. Ramadhan","doi":"10.33899/rengj.2015.101570","DOIUrl":"https://doi.org/10.33899/rengj.2015.101570","url":null,"abstract":"An experimental study of Kevlar-29/epoxy composite thin wall cylinders under internal pressure loading has been investigated in this work. Kevlar/epoxy composite specimens were fabricated and tested. Specimens were tested using the hydraulic pump and subjected to a static internal pressure. The results obtained from this work to find the pressure increase ratio and the failure load were found and a comparison has been achieved. According to the comparison results the tough material was specified. The efficient and effective composite cylinders have been tested experimentally. And this provided a considerable advantage over using a woven roving fiber in pressure vessels applications.","PeriodicalId":339890,"journal":{"name":"AL Rafdain Engineering Journal","volume":"253 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2015-07-28","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"121894047","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2015-07-28DOI: 10.33899/rengj.2015.101565
K. Al-Ahmady, Hussain Muhsin Obeed
In this research, measurements of total (TSP) and fine suspended particulate PM1, PM2.5, PM7 and PM10 at Badoosh cement factory were carried out. Seventeen locations inside the factory are selected. The effect on soil around the cement factory was also studied by collecting samples in the four directions North, South, East and West to a distance of five kilometers from the factory with one kilometer spacing. The result of the study revealed that; the concentration of very fine suspended particulate PM2.5 was over the international standard limits at different locations of the factory and the overall average concentration for the factory was (47 ). For PM10 suspended particulates, the results at all locations were over the international limits and the overall average concentration reached 869 m, which was much more than the standards of 150 . On the other hand, the total suspended particulate concentration exceed the standard at several locations in the factory. Significant correlations were found among the suspended particulate concentrations according to their size. The results of soil test around the cement factory showed significant differences of oxides percentages among the directions. Higher percentages were recorded in the South and East directions than North-West direction. The results also revealed that the percentages of SiO2 and Al2O3 were 40.94% and 10.53% in the east direction whereas higher percentage for Fe2O3 (4.86%), CaO (36.81%), MgO (4.49%), and SO3 (0.76%) were recorded in the south direction of factory.
{"title":"Assessment of Air Particulate Pollution in New Badoosh Cement Factory/ Iraq","authors":"K. Al-Ahmady, Hussain Muhsin Obeed","doi":"10.33899/rengj.2015.101565","DOIUrl":"https://doi.org/10.33899/rengj.2015.101565","url":null,"abstract":"In this research, measurements of total (TSP) and fine suspended particulate PM1, PM2.5, PM7 and PM10 at Badoosh cement factory were carried out. Seventeen locations inside the factory are selected. The effect on soil around the cement factory was also studied by collecting samples in the four directions North, South, East and West to a distance of five kilometers from the factory with one kilometer spacing. The result of the study revealed that; the concentration of very fine suspended particulate PM2.5 was over the international standard limits at different locations of the factory and the overall average concentration for the factory was (47 ). For PM10 suspended particulates, the results at all locations were over the international limits and the overall average concentration reached 869 m, which was much more than the standards of 150 . On the other hand, the total suspended particulate concentration exceed the standard at several locations in the factory. Significant correlations were found among the suspended particulate concentrations according to their size. The results of soil test around the cement factory showed significant differences of oxides percentages among the directions. Higher percentages were recorded in the South and East directions than North-West direction. The results also revealed that the percentages of SiO2 and Al2O3 were 40.94% and 10.53% in the east direction whereas higher percentage for Fe2O3 (4.86%), CaO (36.81%), MgO (4.49%), and SO3 (0.76%) were recorded in the south direction of factory.","PeriodicalId":339890,"journal":{"name":"AL Rafdain Engineering Journal","volume":"77 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2015-07-28","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"130331010","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2015-07-28DOI: 10.33899/rengj.2015.101537
M. Mwafaq, K. Bilal
Scour process at the abutment of bridge is influenced by the contraction of the stream flow specially when the abutment immure an effective width of the stream flow, also the shape of the abutment has an important effects on the scour depth at the abutment. In this paper a series of laboratory experiments were carried out to study the effect of contraction ratio and abutment shape on the scour and its maximum depth at the abutment. Three shapes of abutments with three-contraction ratio with different flow rates were used. The result of the experiments showed that the circular shape and (30) angle abutments reduce scour compared with that, which was, occur at the rectangular abutment. The reduction of scour depth with circular shape is about (22%) where that with (30) angle is about (27%), also a dimensionless general empirical formula was created for the calculation the scour depth by using the (SPSS 11.5) Statistical program a Coefficient (R=0.933) of determination.
{"title":"Effect of the Contraction ratio and the Shape of the Abutment on the Local Scour at a Bridge Abutments","authors":"M. Mwafaq, K. Bilal","doi":"10.33899/rengj.2015.101537","DOIUrl":"https://doi.org/10.33899/rengj.2015.101537","url":null,"abstract":"Scour process at the abutment of bridge is influenced by the contraction of the stream flow specially when the abutment immure an effective width of the stream flow, also the shape of the abutment has an important effects on the scour depth at the abutment. In this paper a series of laboratory experiments were carried out to study the effect of contraction ratio and abutment shape on the scour and its maximum depth at the abutment. Three shapes of abutments with three-contraction ratio with different flow rates were used. The result of the experiments showed that the circular shape and (30) angle abutments reduce scour compared with that, which was, occur at the rectangular abutment. The reduction of scour depth with circular shape is about (22%) where that with (30) angle is about (27%), also a dimensionless general empirical formula was created for the calculation the scour depth by using the (SPSS 11.5) Statistical program a Coefficient (R=0.933) of determination.","PeriodicalId":339890,"journal":{"name":"AL Rafdain Engineering Journal","volume":"60 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2015-07-28","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"114179542","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2015-07-28DOI: 10.33899/rengj.2015.101567
Zena N. Abdulkader
In this paper, the design of an efficient bio 9/7 filter bank for the discrete wavelet transform (DWT) is proposed. The design starts with the standard design constraints of the lifting scheme architecture. A modification in the resulting 9/7 filter bank coefficients after lifting is made leading to new sets of simple and less-number coefficients for both low-pass and high-pass branches of the direct structure of the DWT filter bank. In order to speed up the operations of the designed filter bank, the need for multipliers is eliminated by realizing the designed filter coefficients in a multiplier-less manner. That’s each of the existing non-zero coefficient is represented as sum-of-powers-of-two ( ). The complexity of resulting bio 9/7 filter bank structure is the lowest among other recent issued structures. The performance of resulting structure is then examined objectively by applying three different test images and computing the peak signal-to-noise-ratio ( ) values for the reconstructed image. Subjective evaluation of these images is also achieved by calculating the average mean opinion score ( ). Objective and subjective evaluation indicate the eligibility of the designed bio 9/7 DWT filter bank for image processing.
{"title":"Design and Realization of An Efficient Bio 9/7 Discrete Wavelet Transform","authors":"Zena N. Abdulkader","doi":"10.33899/rengj.2015.101567","DOIUrl":"https://doi.org/10.33899/rengj.2015.101567","url":null,"abstract":"In this paper, the design of an efficient bio 9/7 filter bank for the discrete wavelet transform (DWT) is proposed. The design starts with the standard design constraints of the lifting scheme architecture. A modification in the resulting 9/7 filter bank coefficients after lifting is made leading to new sets of simple and less-number coefficients for both low-pass and high-pass branches of the direct structure of the DWT filter bank. In order to speed up the operations of the designed filter bank, the need for multipliers is eliminated by realizing the designed filter coefficients in a multiplier-less manner. That’s each of the existing non-zero coefficient is represented as sum-of-powers-of-two ( ). The complexity of resulting bio 9/7 filter bank structure is the lowest among other recent issued structures. The performance of resulting structure is then examined objectively by applying three different test images and computing the peak signal-to-noise-ratio ( ) values for the reconstructed image. Subjective evaluation of these images is also achieved by calculating the average mean opinion score ( ). Objective and subjective evaluation indicate the eligibility of the designed bio 9/7 DWT filter bank for image processing.","PeriodicalId":339890,"journal":{"name":"AL Rafdain Engineering Journal","volume":"3 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2015-07-28","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"130527325","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2015-07-28DOI: 10.33899/rengj.2015.101566
Shefa A. Dawwd, Noor M. AL Layla
Shefa A. Dawwd / Assist. Prof. Noor M. AL Layla shefa.dawwd.2014@ieee.org aneng.noor@gmail.com Computer Engineering Department College of Engineering, University of Mosul, Mosul-Iraq Abstract Training of Artificial Neural Networks (ANNs) for large data sets is a time consuming mission. In this paper, accelerating the training of artificial neural network is achieved by a parallel training using either Multicore Central Processing Unit (CPU) or General Purpose Graphics Processing Unit (GPGPU). The training is implemented using five datasets with diverse amounts of patterns and with different neural network parameters in Multilayer Perceptron (MLP). The results show a significant increase in computation speed, which is increased nearly linear with the number of cores in multicore processor for problems with medium and large training datasets. Also, a considerable speed up is achieved when the GPU is used to train the MLP with the large training datasets. While a single core processor is a better choice when the data set size is small. The optimal number of cores or the type of the parallel platform should be employed according to the load of computation.
Shefa . Dawwd / Assist。Noor M. AL Layla教授shefa.dawwd.2014@ieee.org aneng.noor@gmail.com计算机工程系摩苏尔大学工程学院大型数据集人工神经网络(ann)的训练是一项耗时的任务。本文通过使用多核中央处理器(CPU)或通用图形处理器(GPGPU)进行并行训练来加快人工神经网络的训练速度。在多层感知器(Multilayer Perceptron, MLP)中，使用五个具有不同模式数量和不同神经网络参数的数据集来实现训练。结果表明，对于大中型训练数据集的问题，计算速度与多核处理器的核数呈近似线性增长。此外，当使用GPU使用大型训练数据集训练MLP时，可以实现相当大的速度提升。而单核处理器在数据集较小时是更好的选择。应根据计算负荷选择最优核数或并行平台类型。
{"title":"Training Acceleration of Multi-Layer Perceptron using Multicore CPU and GPU under MATLAB Environment","authors":"Shefa A. Dawwd, Noor M. AL Layla","doi":"10.33899/rengj.2015.101566","DOIUrl":"https://doi.org/10.33899/rengj.2015.101566","url":null,"abstract":"Shefa A. Dawwd / Assist. Prof. Noor M. AL Layla shefa.dawwd.2014@ieee.org aneng.noor@gmail.com Computer Engineering Department College of Engineering, University of Mosul, Mosul-Iraq Abstract Training of Artificial Neural Networks (ANNs) for large data sets is a time consuming mission. In this paper, accelerating the training of artificial neural network is achieved by a parallel training using either Multicore Central Processing Unit (CPU) or General Purpose Graphics Processing Unit (GPGPU). The training is implemented using five datasets with diverse amounts of patterns and with different neural network parameters in Multilayer Perceptron (MLP). The results show a significant increase in computation speed, which is increased nearly linear with the number of cores in multicore processor for problems with medium and large training datasets. Also, a considerable speed up is achieved when the GPU is used to train the MLP with the large training datasets. While a single core processor is a better choice when the data set size is small. The optimal number of cores or the type of the parallel platform should be employed according to the load of computation.","PeriodicalId":339890,"journal":{"name":"AL Rafdain Engineering Journal","volume":"60 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2015-07-28","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"129429460","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2015-07-28DOI: 10.33899/rengj.2015.101563
Ekbal Alsoofee, Nahedh Taha Al-Qemaqchi
{"title":"The effect of design tasks on changing the conceptual structure for architectural engineering students","authors":"Ekbal Alsoofee, Nahedh Taha Al-Qemaqchi","doi":"10.33899/rengj.2015.101563","DOIUrl":"https://doi.org/10.33899/rengj.2015.101563","url":null,"abstract":"","PeriodicalId":339890,"journal":{"name":"AL Rafdain Engineering Journal","volume":"51 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2015-07-28","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"124017459","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2015-04-28DOI: 10.33899/rengj.2015.101082
Shefa A. Dawwd, A. Al-Saegh
Artificial neural networks (ANNs) are widely used in different areas of nowadays applications. Many challenges are imposed on the practical implementation of ANNs. Some of them are: the number of samples required to train the network; the number of adders, multipliers, nonlinear transfer functions, storage elements; and the speed of calculations in either training phase or recall phase. In this paper, the RAM-based neural network is investigated. No weights, adders, multipliers, transfer functions are required to implement it neither in hardware nor in software, but at a cost of large RAM utilization. In addition, a small number of samples are required for training. However, in hardware implementation, a large size of memory is required to train it. The network is implemented on the FPGA platform. The Stratix IV GX FPGA development board, which is provided on large on board RAM, is used. A considerable speedup of 237 is achieved in either training or recalling phases. A comparable error rate of 7.6 is achieved when MNIST (Mixed National Institute of Standards and Technology) database are used to train the network on handwritten digit recognition.
人工神经网络在当今应用的各个领域都有广泛的应用。人工神经网络的实际应用面临许多挑战。其中包括:训练网络所需的样本数量;加法器、乘法器、非线性传递函数、存储元件的数量;以及训练阶段和回忆阶段的计算速度。本文研究了基于ram的神经网络。在硬件和软件中都不需要权重、加法器、乘数、传递函数来实现它，但代价是大量的RAM利用率。另外，需要少量的样本进行训练。然而，在硬件实现中，需要大量的内存来训练它。该网络在FPGA平台上实现。使用Stratix IV GX FPGA开发板，该开发板提供了大板上RAM。无论是在训练阶段还是在回顾阶段，都实现了237的相当大的加速。当使用MNIST (Mixed National Institute of Standards and Technology)数据库训练网络进行手写数字识别时，错误率为7.6。
{"title":"RAM-Based Neural Network Parallel Implementation on a Reconfigurable Platform and Its Application for Handwritten Digits Recognition","authors":"Shefa A. Dawwd, A. Al-Saegh","doi":"10.33899/rengj.2015.101082","DOIUrl":"https://doi.org/10.33899/rengj.2015.101082","url":null,"abstract":"Artificial neural networks (ANNs) are widely used in different areas of nowadays applications. Many challenges are imposed on the practical implementation of ANNs. Some of them are: the number of samples required to train the network; the number of adders, multipliers, nonlinear transfer functions, storage elements; and the speed of calculations in either training phase or recall phase. In this paper, the RAM-based neural network is investigated. No weights, adders, multipliers, transfer functions are required to implement it neither in hardware nor in software, but at a cost of large RAM utilization. In addition, a small number of samples are required for training. However, in hardware implementation, a large size of memory is required to train it. The network is implemented on the FPGA platform. The Stratix IV GX FPGA development board, which is provided on large on board RAM, is used. A considerable speedup of 237 is achieved in either training or recalling phases. A comparable error rate of 7.6 is achieved when MNIST (Mixed National Institute of Standards and Technology) database are used to train the network on handwritten digit recognition.","PeriodicalId":339890,"journal":{"name":"AL Rafdain Engineering Journal","volume":"167 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2015-04-28","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"116035108","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
京公网安备 11010802042870号
京ICP备2023020795号-1
扫码关注我们
求助内容：
应助结果提醒方式：