A CAD tool based on a group of efficient algorithms to verify,design,and optimize power/ground networks for standard cell model is presented.Nonlinear programming techniques,branch and bound algorithms and incomplete ...A CAD tool based on a group of efficient algorithms to verify,design,and optimize power/ground networks for standard cell model is presented.Nonlinear programming techniques,branch and bound algorithms and incomplete Cholesky decomposition conjugate gradient method (ICCG) are the three main parts of our work.Users can choose nonlinear programming method or branch and bound algorithm to satisfy their different requirements of precision and speed.The experimental results prove that the algorithms can run very fast with lower wiring resources consumption.As a result,the CAD tool based on these algorithms is able to cope with large-scale circuits.展开更多
Nowadays artificial neural networkS （ANNs） with strong ability have been applied widely for prediction of non- linear phenomenon. In this work an optimized ANN with 7 inputs that consist of temperature, pressure, cr...Nowadays artificial neural networkS （ANNs） with strong ability have been applied widely for prediction of non- linear phenomenon. In this work an optimized ANN with 7 inputs that consist of temperature, pressure, critical temperature, critical pressure, density, molecular weight and acentric factor has been used for solubility predic- tion of three disperse dyes in supercritical carbon dioxide （SC-C02） and ethanol as co-solvent. It was shown how a multi-layer perceptron network can be trained to represent the solubility of disperse dyes in SC-C02. Numeric Sensitivity Analysis and Garson equation were utilized to find out the degree of effectiveness of different input variables on the efficiency of the proposed model. Results showed that our proposed ANN model has correlation coefficient, Nash-Sutcliffe model efficiency coefficient and discrepancy ratio about 0.998, 0.992, and 1.053 respectively.展开更多
In order to solve the difficulty of detailed recognition of subdivisions of structural coal types,a differentiation model that combines BP neural network with an ultrasonic reflection method is proposed.Structural coa...In order to solve the difficulty of detailed recognition of subdivisions of structural coal types,a differentiation model that combines BP neural network with an ultrasonic reflection method is proposed.Structural coal types are recognized based on a suitable consideration of ultrasonic speed,an ultrasonic attenuation coefficient,characteristics of ultrasonic transmission and other parameters relating to structural coal types.We have focused on a computational model of ultrasonic speed,attenuation coefficient in coal and differentiation algorithm of structural coal types based on a BP neural network.Experiments demonstrate that the model can distinguish structural coal types effectively.It is important for the improved ultrasonic differentiation model to predict coal and gas outbursts.展开更多
Single image super-resolution(SISR)is a fundamentally challenging problem because a low-resolution(LR)image can correspond to a set of high-resolution(HR)images,while most are not expected.Recently,SISR can be achieve...Single image super-resolution(SISR)is a fundamentally challenging problem because a low-resolution(LR)image can correspond to a set of high-resolution(HR)images,while most are not expected.Recently,SISR can be achieved by a deep learning-based method.By constructing a very deep super-resolution convolutional neural network(VDSRCNN),the LR images can be improved to HR images.This study mainly achieves two objectives:image super-resolution(ISR)and deblurring the image from VDSRCNN.Firstly,by analyzing ISR,we modify different training parameters to test the performance of VDSRCNN.Secondly,we add the motion blurred images to the training set to optimize the performance of VDSRCNN.Finally,we use image quality indexes to evaluate the difference between the images from classical methods and VDSRCNN.The results indicate that the VDSRCNN performs better in generating HR images from LR images using the optimized VDSRCNN in a proper method.展开更多
Transformers utilizing HTS （high temperature superconductors） are considered as a timely invention. The number of power transformers age more than 30 years old and nearing retirement is increasing. If this window of...Transformers utilizing HTS （high temperature superconductors） are considered as a timely invention. The number of power transformers age more than 30 years old and nearing retirement is increasing. If this window of opportunity is not grabbed, there would be great reluctance to replace recently installed highly priced capital asset. Major projects of developing HTS transformers are well making progress in the United States, Europe, Japan, Korea and China which indicate the interest. The efforts must have been appropriately verified through the economic interest of the discounted losses. Consequently, it is very important to develop an understanding of the fundamental HTS transformer design issues that can provide guidance for developing practical devices of interest to the electric utility industry. The parameters of HTS transformer need to be validated before any effort is to carry out to model the behaviour of a distribution network under a range of conditions. The predicted performance and reliability of HTS transformers can then be verified through the network modelling and analysis calculation. The ultimate purpose is to furnish electric utilities precise information as to which HTS transformers work under various applications with greater technical efficiency and proven reliability.展开更多
In this review a series of organic-based open porous networks are discussed, in which hydrogen bonds play an important role in network formation. Using these open networks as molecular templates: 1) a wealth of functi...In this review a series of organic-based open porous networks are discussed, in which hydrogen bonds play an important role in network formation. Using these open networks as molecular templates: 1) a wealth of functional guest species can be immo- bilized; 2) fullerene molecules can be separated and recognized; 3) photoisomerization reactions can be observed by STM; 4) 1D molecular arrays can be constructed; and 5) heterogeneous bilayer structures can be formed. It is envisioned that these su- pramolecular networks might be developed into a new family of useful soft frameworks for studies toward shape-selective ca- talysis, molecular recognition and host-guest supramolecular chemistry.展开更多
文摘A CAD tool based on a group of efficient algorithms to verify,design,and optimize power/ground networks for standard cell model is presented.Nonlinear programming techniques,branch and bound algorithms and incomplete Cholesky decomposition conjugate gradient method (ICCG) are the three main parts of our work.Users can choose nonlinear programming method or branch and bound algorithm to satisfy their different requirements of precision and speed.The experimental results prove that the algorithms can run very fast with lower wiring resources consumption.As a result,the CAD tool based on these algorithms is able to cope with large-scale circuits.
文摘Nowadays artificial neural networkS （ANNs） with strong ability have been applied widely for prediction of non- linear phenomenon. In this work an optimized ANN with 7 inputs that consist of temperature, pressure, critical temperature, critical pressure, density, molecular weight and acentric factor has been used for solubility predic- tion of three disperse dyes in supercritical carbon dioxide （SC-C02） and ethanol as co-solvent. It was shown how a multi-layer perceptron network can be trained to represent the solubility of disperse dyes in SC-C02. Numeric Sensitivity Analysis and Garson equation were utilized to find out the degree of effectiveness of different input variables on the efficiency of the proposed model. Results showed that our proposed ANN model has correlation coefficient, Nash-Sutcliffe model efficiency coefficient and discrepancy ratio about 0.998, 0.992, and 1.053 respectively.
基金Projects 50674093 supported by the National Natural Science Foundation of China20050290010 by the Doctoral Foundation of the Chinese Education Ministry
文摘In order to solve the difficulty of detailed recognition of subdivisions of structural coal types,a differentiation model that combines BP neural network with an ultrasonic reflection method is proposed.Structural coal types are recognized based on a suitable consideration of ultrasonic speed,an ultrasonic attenuation coefficient,characteristics of ultrasonic transmission and other parameters relating to structural coal types.We have focused on a computational model of ultrasonic speed,attenuation coefficient in coal and differentiation algorithm of structural coal types based on a BP neural network.Experiments demonstrate that the model can distinguish structural coal types effectively.It is important for the improved ultrasonic differentiation model to predict coal and gas outbursts.
文摘Single image super-resolution(SISR)is a fundamentally challenging problem because a low-resolution(LR)image can correspond to a set of high-resolution(HR)images,while most are not expected.Recently,SISR can be achieved by a deep learning-based method.By constructing a very deep super-resolution convolutional neural network(VDSRCNN),the LR images can be improved to HR images.This study mainly achieves two objectives:image super-resolution(ISR)and deblurring the image from VDSRCNN.Firstly,by analyzing ISR,we modify different training parameters to test the performance of VDSRCNN.Secondly,we add the motion blurred images to the training set to optimize the performance of VDSRCNN.Finally,we use image quality indexes to evaluate the difference between the images from classical methods and VDSRCNN.The results indicate that the VDSRCNN performs better in generating HR images from LR images using the optimized VDSRCNN in a proper method.
文摘Transformers utilizing HTS （high temperature superconductors） are considered as a timely invention. The number of power transformers age more than 30 years old and nearing retirement is increasing. If this window of opportunity is not grabbed, there would be great reluctance to replace recently installed highly priced capital asset. Major projects of developing HTS transformers are well making progress in the United States, Europe, Japan, Korea and China which indicate the interest. The efforts must have been appropriately verified through the economic interest of the discounted losses. Consequently, it is very important to develop an understanding of the fundamental HTS transformer design issues that can provide guidance for developing practical devices of interest to the electric utility industry. The parameters of HTS transformer need to be validated before any effort is to carry out to model the behaviour of a distribution network under a range of conditions. The predicted performance and reliability of HTS transformers can then be verified through the network modelling and analysis calculation. The ultimate purpose is to furnish electric utilities precise information as to which HTS transformers work under various applications with greater technical efficiency and proven reliability.
文摘In this review a series of organic-based open porous networks are discussed, in which hydrogen bonds play an important role in network formation. Using these open networks as molecular templates: 1) a wealth of functional guest species can be immo- bilized; 2) fullerene molecules can be separated and recognized; 3) photoisomerization reactions can be observed by STM; 4) 1D molecular arrays can be constructed; and 5) heterogeneous bilayer structures can be formed. It is envisioned that these su- pramolecular networks might be developed into a new family of useful soft frameworks for studies toward shape-selective ca- talysis, molecular recognition and host-guest supramolecular chemistry.
