Fully connected neural networks(FCNNs)have been developed for the closure of subgrid-scale(SGS)stress and SGS heat flux in large-eddy simulations of compressible turbulent channel flow.The FCNNbased SGS model trained ...Fully connected neural networks(FCNNs)have been developed for the closure of subgrid-scale(SGS)stress and SGS heat flux in large-eddy simulations of compressible turbulent channel flow.The FCNNbased SGS model trained using data with Mach number Ma=3.0 and Reynolds number Re=3000 was applied to situations with different Mach numbers and Reynolds numbers.The input variables of the neural network model were the filtered velocity gradients and temperature gradients at a single spatial grid point.The a priori test showed that the FCNN model had a correlation coefficient larger than 0.91 and a relative error smaller than 0.43,with much better reconstructions of SGS unclosed terms than the dynamic Smagorinsky model(DSM).In a posteriori test,the behavior of the FCNN model was marginally better than that of the DSM in predicting the mean velocity profiles,mean temperature profiles,turbulent intensities,total Reynolds stress,total Reynolds heat flux,and mean SGS flux of kinetic energy,and outperformed the Smagorinsky model.展开更多
The Heihe River drainage basin is one of the endangered ecological regions of China. The shortage of water resources is the bottleneck, which constrains the sustainable development of the region. Many scholars in Chin...The Heihe River drainage basin is one of the endangered ecological regions of China. The shortage of water resources is the bottleneck, which constrains the sustainable development of the region. Many scholars in China have done researches concerning this problem. Based on previous researches, this paper analyzed characteristics, tendencies, and causes of annual runoff variations in the Yingluo Gorge (1944-2005) and the Zhengyi Gorge (1954-2005), which are the boundaries of the upper reaches, the middle reaches, and the lower reaches of the Heihe River drainage basin, by wavelet analysis, wavelet neural network model, and GIS spatial analysis. The results show that: (1) annual runoff variations of the Yingluo Gorge have principal periods of 7 years and 25 years, and its increasing rate is 1.04 m^3/s.10y; (2) annual runoff variations of the Zhengyi Gorge have principal periods of 6 years and 27 years, and its decreasing rate is 2.25 m^3/s.10y; (3) prediction results show that: during 2006-2015, annual runoff variations of the Yingluo and Zhengyi gorges have ascending tendencies, and the increasing rates are respectively 2.04 m^3/s.10y and 1.61 m^3/s.10y; (4) the increase of annual runoff in the Yingluo Gorge has causal relationship with increased temperature and precipitation in the upper reaches, and the decrease of annual runoff in the Zhengyi Gorge in the past decades was mainly caused by the increased human consumption of water resources in the middle researches. The study results will provide scientific basis for making rational use and allocation schemes of water resources in the Heihe River drainage basin.展开更多
Bayesian regularized BP neural network(BRBPNN) technique was applied in the chlorophyll-α prediction of Nanzui water area in Dongting Lake. Through BP network interpolation method, the input and output samples of t...Bayesian regularized BP neural network(BRBPNN) technique was applied in the chlorophyll-α prediction of Nanzui water area in Dongting Lake. Through BP network interpolation method, the input and output samples of the network were obtained. After the selection of input variables using stepwise/multiple linear regression method in SPSS i1.0 software, the BRBPNN model was established between chlorophyll-α and environmental parameters, biological parameters. The achieved optimal network structure was 3-11-1 with the correlation coefficients and the mean square errors for the training set and the test set as 0.999 and 0.000?8426, 0.981 and 0.0216 respectively. The sum of square weights between each input neuron and the hidden layer of optimal BRBPNN models of different structures indicated that the effect of individual input parameter on chlorophyll- α declined in the order of alga amount 〉 secchi disc depth(SD) 〉 electrical conductivity (EC). Additionally, it also demonstrated that the contributions of these three factors were the maximal for the change of chlorophyll-α concentration, total phosphorus(TP) and total nitrogen(TN) were the minimal. All the results showed that BRBPNN model was capable of automated regularization parameter selection and thus it may ensure the excellent generation ability and robustness. Thus, this study laid the foundation for the application of BRBPNN model in the analysis of aquatic ecological data(chlorophyll-α prediction) and the explanation about the effective eutrophication treatment measures for Nanzui water area in Dongting Lake.展开更多
Background: Leaf Area Index(LAI) is an important parameter used in monitoring and modeling of forest ecosystems. The aim of this study was to evaluate performance of the artificial neural network(ANN) models to predic...Background: Leaf Area Index(LAI) is an important parameter used in monitoring and modeling of forest ecosystems. The aim of this study was to evaluate performance of the artificial neural network(ANN) models to predict the LAI by comparing the regression analysis models as the classical method in these pure and even-aged Crimean pine forest stands.Methods: One hundred eight temporary sample plots were collected from Crimean pine forest stands to estimate stand parameters. Each sample plot was imaged with hemispherical photographs to detect the LAI. The partial correlation analysis was used to assess the relationships between the stand LAI values and stand parameters, and the multivariate linear regression analysis was used to predict the LAI from stand parameters. Different artificial neural network models comprising different number of neuron and transfer functions were trained and used to predict the LAI of forest stands.Results: The correlation coefficients between LAI and stand parameters(stand number of trees, basal area, the quadratic mean diameter, stand density and stand age) were significant at the level of 0.01. The stand age, number of trees, site index, and basal area were independent parameters in the most successful regression model predicted LAI values using stand parameters(R_(adj)~2=0.5431). As corresponding method to predict the interactions between the stand LAI values and stand parameters, the neural network architecture based on the RBF 4-19-1 with Gaussian activation function in hidden layer and the identity activation function in output layer performed better in predicting LAI(SSE(12.1040), MSE(0.1223), RMSE(0.3497), AIC(0.1040), BIC(-77.7310) and R^2(0.6392)) compared to the other studied techniques.Conclusion: The ANN outperformed the multivariate regression techniques in predicting LAI from stand parameters. The ANN models, developed in this study, may aid in making forest management planning in study forest stands.展开更多
In the context of global warming,precipitation forms are likely to transform from snowfall to rainfall with a more pronounced trend.The change in precipitation forms will inevitably affect the processes of regional ru...In the context of global warming,precipitation forms are likely to transform from snowfall to rainfall with a more pronounced trend.The change in precipitation forms will inevitably affect the processes of regional runoff generation and confluence as well as the annual distribution of runoff.Most researchers used precipitation data from the CMIP5 model directly to study future precipitation trends without distinguishing between snowfall and rainfall.CMIP5 models have been proven to have better performance in simulating temperature but poorer performance in simulating precipitation.To overcome the above limitations,this paper used a Back Propagation Neural Network(BNN)to predict the rainfall-to-precipitation ratio(RPR)in months experiencing freezing-thawing transitions(FTTs).We utilized the meteorological(air pressure,air temperature,evaporation,relative humidity,wind speed,sunshine hours,surface temperature),topographic(altitude,slope,aspect)and geographic(longitude,latitude)data from 28 meteorological stations in the Chinese Tianshan Mountains region(CTMR)from 1961 to 2018 to calculate the RPR and constructed an index system of impact factors.Based on the BNN,decision-making trial and evaluation laboratory method(BP-DEMATEL),the key factors driving the transformation of the RPR in the CTMR were identified.We found that temperature was the only key factor affecting the transformation of the RPR in the BP-DEMATEL model.Considering the relationship between temperature and the RPR,the future temperature under different representative concentration pathways(RCPs)(RCP2.6/RCP4.5/RCP8.5)provided by 21 CMIP5 models and the meteorological factors from meteorological stations were input into the BNN model to acquire the future RPR from 2011 to 2100.The results showed that under the three scenarios,the RPR in the number of months experiencing FTTs during 2011-2100 will be higher than that in the historical period(1981-2010)in the CTMR.Furthermore,in terms of spatial variation,the RPR values on the south slope will be larger than those on the north slope under the three emission scenarios.Moreover,the RPR values exhibited different variation characteristics under different emission scenarios.Under the low-emission scenario(RCP2.6),as time passed,the RPR values changed slightly at more stations.Under the mediumemission scenario(RCP4.5),the RPR increased in the whole CTMR and stabilized on the north slope by the end of this century.Under the high-emission scenario(RCP8.5),the RPR values increased significantly through the 21 st century in the whole CTMR.This study may help to provide a scientific management basis for agricultural production and hydrology.展开更多
This paper develops a joint model utilizing the principal component analysis(PCA)and the back propagation(BP)neural network model optimized by the Levenberg Marquardt(LM)algorithm,and as an application of the joint mo...This paper develops a joint model utilizing the principal component analysis(PCA)and the back propagation(BP)neural network model optimized by the Levenberg Marquardt(LM)algorithm,and as an application of the joint model to investigate the damages caused by typhoons for a coastal province,Fujian Province,China in 2005-2015(latest).First,the PCA is applied to analyze comprehensively the relationship between hazard factors,hazard bearing factors and disaster factors.Then five integrated indices,overall disaster level,typhoon intensity,damaged condition of houses,medical rescue and self-rescue capability,are extracted through the PCA;Finally,the BP neural network model,which takes the principal component scores as input and is optimized by the LM algorithm,is implemented to forecast the comprehensive loss of typhoons.It is estimated that an average annual loss of 138.514 billion RMB occurred for 2005-2015,with a maximum loss of 215.582 in 2006 and a decreasing trend since 2010 though the typhoon intensity increases.The model was validated using three typhoon events and it is found that the error is less than 1%.These results provide information for the government to increase medical institutions and medical workers and for the communities to promote residents’self-rescue capability.展开更多
The recurrent neural network (RNN) model based on projective operator was studied. Different from the former study, the value region of projective operator in the neural network in this paper is a general closed con...The recurrent neural network (RNN) model based on projective operator was studied. Different from the former study, the value region of projective operator in the neural network in this paper is a general closed convex subset of n-dimensional Euclidean space and it is not a compact convex set in general, that is, the value region of projective operator is probably unbounded. It was proved that the network has a global solution and its solution trajectory converges to some equilibrium set whenever objective function satisfies some conditions. After that, the model was applied to continuously differentiable optimization and nonlinear or implicit complementarity problems. In addition, simulation experiments confirm the efficiency of the RNN.展开更多
Accurate estimation of the solubility of a chemical compound is an important issue for many industrial proce sses.To overcome the defects of some thermodynamic models and simple correlations,a parallel neural network(...Accurate estimation of the solubility of a chemical compound is an important issue for many industrial proce sses.To overcome the defects of some thermodynamic models and simple correlations,a parallel neural network(PNN) model was conceived and optimized to predict the solubility of diosgenin in seven n-alkanols(C_(1)-C_(7)).The linear regression analysis of the parity plots indicates that the PNN model can give more accurate descriptions of the solubility of diosgenin than the ordinary neural network(ONN) model.The comparison of the average root mean square deviation(RMSD) shows that the suggested model has a slight advantage over the thermodynamic NRTL model in terms of the calculating precision.Moreover,the PNN model can reflect the effects of the temperature and the chain length of the alcohol solvent on the solution behavior of diosgenin correctly and can estimate its solubility in the n-alkanols with more carbon atoms.展开更多
The factors of geomorphology, geological setting, effect of ground water and environment dynamic factors (e.g. rainfall and artificial water recharge) should be integrated in the discrimination of the stability of the...The factors of geomorphology, geological setting, effect of ground water and environment dynamic factors (e.g. rainfall and artificial water recharge) should be integrated in the discrimination of the stability of the ancient landslide. As the criterion of landslide stability has been studied, the artificial neural network model was then applied to discriminate the stability of the ancient landslide in the impounding area of the Three Gorges project on the Yangtze River, China. The model has the property of self adaptive identifying and integrating complex qualitative factors and quantitative factors. The results of the artificial neural network model are coincided well with what were gained by classical limit equilibrium analysis (the Bishop method and Janbu method) and by other comprehensive discrimination methods.展开更多
We propose a new approach for analyzing the global asymptotic stability of the extended discrete-time bidirectional associative memory (BAM) neural networks. By using the Euler rule, we discretize the continuous-tim...We propose a new approach for analyzing the global asymptotic stability of the extended discrete-time bidirectional associative memory (BAM) neural networks. By using the Euler rule, we discretize the continuous-time BAM neural networks as the extended discrete-time BAM neural networks with non-threshold activation functions. Here we present some conditions under which the neural networks have unique equilibrium points. To judge the global asymptotic stability of the equilibrium points, we introduce a new neural network model - standard neural network model (SNNM). For the SNNMs, we derive the sufficient conditions for the global asymptotic stability of the equilibrium points, which are formulated as some linear matrix inequalities (LMIs). We transform the discrete-time BAM into the SNNM and apply the general result about the SNNM to the determination of global asymptotic stability of the discrete-time BAM. The approach proposed extends the known stability results, has lower conservativeness, can be verified easily, and can also be applied to other forms of recurrent neural networks.展开更多
Intelligent systems and methods such as the neural network (NN) are usually used in electric power systems for short-term electrical load forecasting. However, a vast amount of electrical load data is often redundan...Intelligent systems and methods such as the neural network (NN) are usually used in electric power systems for short-term electrical load forecasting. However, a vast amount of electrical load data is often redundant, and linearly or nonlinearly correlated with each other. Highly correlated input data can result in erroneous prediction results given out by an NN model. Besides this, the determination of the topological structure of an NN model has always been a problem for designers. This paper presents a new artificial intelligence hybrid procedure for next day electric load forecasting based on partial least squares (PLS) and NN. PLS is used for the compression of data input space, and helps to determine the structure of the NN model. The hybrid PLS-NN model can be used to predict hourly electric load on weekdays and weekends. The advantage of this methodology is that the hybrid model can provide faster convergence and more precise prediction results in comparison with abductive networks algorithm. Extensive testing on the electrical load data of the Puget power utility in the USA confirms the validity of the proposed approach.展开更多
Data from the deformation on Split Hopkinson Bar were used for constructing an artificial neural network model. When putting the thermodynamic parameters of the metals into the trained network model, the corresponding...Data from the deformation on Split Hopkinson Bar were used for constructing an artificial neural network model. When putting the thermodynamic parameters of the metals into the trained network model, the corresponding yielding stress can be predicted. The results show that the systematic error is small when the objective function is 0.5 , the number of the nodes in the hidden layer is 6 and the learning rate is about 0.1 , and the accuracy of the rate error is less than 3%. [展开更多
One synthetical control method of AGC/LPC system based on intelligence control theory-neural networks internal model control method is presented. Genetic algorithm (GA) is applied to optimize the parameters of the neu...One synthetical control method of AGC/LPC system based on intelligence control theory-neural networks internal model control method is presented. Genetic algorithm (GA) is applied to optimize the parameters of the neural networks. Simulation results prove that this method is effective.展开更多
Training neural network to recognize targets needs a lot of samples.People usually get these samples in a non-systematic way,which can miss or overemphasize some target information.To improve this situation,a new meth...Training neural network to recognize targets needs a lot of samples.People usually get these samples in a non-systematic way,which can miss or overemphasize some target information.To improve this situation,a new method based on virtual model and invariant moments was proposed to generate training samples.The method was composed of the following steps:use computer and simulation software to build target object's virtual model and then simulate the environment,light condition,camera parameter,etc.;rotate the model by spin and nutation of inclination to get the image sequence by virtual camera;preprocess each image and transfer them into binary image;calculate the invariant moments for each image and get a vectors' sequence.The vectors' sequence which was proved to be complete became the training samples together with the target outputs.The simulated results showed that the proposed method could be used to recognize the real targets and improve the accuracy of target recognition effectively when the sampling interval was short enough and the circumstance simulation was close enough.展开更多
The polymer electrolyte membrane(PEM) fuel cell has been regarded as a potential alternative power source,and a model is necessary for its design,control and power management.A hybrid dynamic model of PEM fuel cell,...The polymer electrolyte membrane(PEM) fuel cell has been regarded as a potential alternative power source,and a model is necessary for its design,control and power management.A hybrid dynamic model of PEM fuel cell,which combines the advantages of mechanism model and black-box model,is proposed in this paper.To improve the performance,the static neural network and variable neural network are used to build the black-box model.The static neural network can significantly improve the static performance of the hybrid model,and the variable neural network makes the hybrid dynamic model predict the real PEM fuel cell behavior with required accuracy.Finally,the hybrid dynamic model is validated with a 500 W PEM fuel cell.The static and transient experiment results show that the hybrid dynamic model can predict the behavior of the fuel cell stack accurately and therefore can be effectively utilized in practical application.展开更多
On the basis of Artificial Neural Network theory, a back propagation neural network with one middle layer is building in this paper, and its algorithms is also given, Using this BP network model, study the case of Mal...On the basis of Artificial Neural Network theory, a back propagation neural network with one middle layer is building in this paper, and its algorithms is also given, Using this BP network model, study the case of Malian-River basin. The results by calculating show that the solution based on BP algorithms are consis- tent with those based multiple - variables linear regression model. They also indicate that BP model in this paper is reasonable and BP algorithms are feasible.展开更多
This paper develops a feedforward neural network based input output model for a general unknown nonlinear dynamic system identification when only the inputs and outputs are accessible observations. In the developed m...This paper develops a feedforward neural network based input output model for a general unknown nonlinear dynamic system identification when only the inputs and outputs are accessible observations. In the developed model, the size of the input space is directly related to the system order. By monitoring the identification error characteristic curve, we are able to determine the system order and subsequently an appropriate network structure for systems identification. Simulation results are promising and show that generic nonlinear systems can be identified, different cases of the same system can also be discriminated by our model.展开更多
This study aims to reveal the spatial structural characteristics of 1,652 Ethnic-Minority Villages(EMV)in China and to analyze the mechanisms driving their spatial heterogeneity.EMV are a special type of settlement sp...This study aims to reveal the spatial structural characteristics of 1,652 Ethnic-Minority Villages(EMV)in China and to analyze the mechanisms driving their spatial heterogeneity.EMV are a special type of settlement space that preserve a large number of historical traces of the ethnic culture of ancient China.They are important carriers of China’s excellent traditional culture and are key to the implementation of rural revitalization strategies.In this study,1652 EMV in China were selected as the research subjects.The Nearest Neighbor Index,kernel density,and spatial autocorrelation index were employed to reveal the spatial structural characteristics of minority villages.Neural network models,spatial lag models,and geographical detectors were used to analyze the formation mechanism of spatial heterogeneity in EMV.The results indicate that:(1)EMV exhibit significant spatial differentiation characterized by“single-core with multiple surrounding sub-centers,”“polarization between east and west,”“decreasing quantity from southwest to east coast to northeast to northwest,”and“large dispersion with small agglomeration.”(2)EMV are mainly distributed in areas rich in intangible cultural heritage,with high vegetation coverage and low altitude,far from central cities,and having limited arable land and an underdeveloped economy and transportation,particularly in shaded or riverbank areas.(3)Distance from the nearest river(X3),distance from central cities(X8),national intangible cultural heritage(X9),and NDVI(X10)were the main driving factors affecting the spatial distribution of EMV,whereas elevation(X1)and GDP(X5)had the weakest influence.As EMV are a relatively unique territorial spatial unit,the identification of their spatial heterogeneity characteristics not only deepens the research content of settlement geography,but also involves the assessment,protection,and development of Minority Villages,which is of great significance for the inheritance and utilization of excellent ethnic cultures in the era.展开更多
Based on data from a petrochemical company’s MIP unit over the past three years,19 input variables and 2 output variables were selected for modeling using the maximum information coefficient and Pearson correlation c...Based on data from a petrochemical company’s MIP unit over the past three years,19 input variables and 2 output variables were selected for modeling using the maximum information coefficient and Pearson correlation coefficient among 155 variables,which included properties of feedstock oil and spent catalyst,operational variables,and material flows.The distillation range variables were reduced using factor analysis,and the feedstock oils were clustered into three types using the K-means++algorithm.Each feedstock oil type was then used as an input variable for modeling.An XGBoost model and a back propagation(BP)neural network model with a structure of 20-15-15-2 were developed to predict the combined yield of gasoline and propylene,as well as the coke yield.In the test set,the BP neural network model demonstrated better fitting and generalization abilities with a mean absolute percentage error and determination coefficient of 1.48%and 0.738,respectively,compared to the XGBoost model.It was therefore chosen for further optimization work.The genetic algorithm was utilized to optimize operational variables in order to increase the combined yield of gasoline and propylene while controlling the growth of coke yield.Seven commercial test results in the MIP unit showed an average increase of 1.39 percentage points for the combined yield of gasoline and propylene and an average decrease of 0.11 percentage points for coke yield.These results indicate that the model effectively improves the combined yield of gasoline and propylene while controlling the increase in coke yield.展开更多
The quantization algorithm compresses the original network by reducing the numerical bit width of the model,which improves the computation speed. Because different layers have different redundancy and sensitivity to d...The quantization algorithm compresses the original network by reducing the numerical bit width of the model,which improves the computation speed. Because different layers have different redundancy and sensitivity to databit width. Reducing the data bit width will result in a loss of accuracy. Therefore, it is difficult to determinethe optimal bit width for different parts of the network with guaranteed accuracy. Mixed precision quantizationcan effectively reduce the amount of computation while keeping the model accuracy basically unchanged. In thispaper, a hardware-aware mixed precision quantization strategy optimal assignment algorithm adapted to low bitwidth is proposed, and reinforcement learning is used to automatically predict the mixed precision that meets theconstraints of hardware resources. In the state-space design, the standard deviation of weights is used to measurethe distribution difference of data, the execution speed feedback of simulated neural network accelerator inferenceis used as the environment to limit the action space of the agent, and the accuracy of the quantization model afterretraining is used as the reward function to guide the agent to carry out deep reinforcement learning training. Theexperimental results show that the proposed method obtains a suitable model layer-by-layer quantization strategyunder the condition that the computational resources are satisfied, and themodel accuracy is effectively improved.The proposed method has strong intelligence and certain universality and has strong application potential in thefield of mixed precision quantization and embedded neural network model deployment.展开更多
基金Financial support provided by the National Natural Science Foundation of China(Grant Nos.11702042 and 91952104)。
文摘Fully connected neural networks(FCNNs)have been developed for the closure of subgrid-scale(SGS)stress and SGS heat flux in large-eddy simulations of compressible turbulent channel flow.The FCNNbased SGS model trained using data with Mach number Ma=3.0 and Reynolds number Re=3000 was applied to situations with different Mach numbers and Reynolds numbers.The input variables of the neural network model were the filtered velocity gradients and temperature gradients at a single spatial grid point.The a priori test showed that the FCNN model had a correlation coefficient larger than 0.91 and a relative error smaller than 0.43,with much better reconstructions of SGS unclosed terms than the dynamic Smagorinsky model(DSM).In a posteriori test,the behavior of the FCNN model was marginally better than that of the DSM in predicting the mean velocity profiles,mean temperature profiles,turbulent intensities,total Reynolds stress,total Reynolds heat flux,and mean SGS flux of kinetic energy,and outperformed the Smagorinsky model.
基金National Natural Science Foundation of China, No.40335046
文摘The Heihe River drainage basin is one of the endangered ecological regions of China. The shortage of water resources is the bottleneck, which constrains the sustainable development of the region. Many scholars in China have done researches concerning this problem. Based on previous researches, this paper analyzed characteristics, tendencies, and causes of annual runoff variations in the Yingluo Gorge (1944-2005) and the Zhengyi Gorge (1954-2005), which are the boundaries of the upper reaches, the middle reaches, and the lower reaches of the Heihe River drainage basin, by wavelet analysis, wavelet neural network model, and GIS spatial analysis. The results show that: (1) annual runoff variations of the Yingluo Gorge have principal periods of 7 years and 25 years, and its increasing rate is 1.04 m^3/s.10y; (2) annual runoff variations of the Zhengyi Gorge have principal periods of 6 years and 27 years, and its decreasing rate is 2.25 m^3/s.10y; (3) prediction results show that: during 2006-2015, annual runoff variations of the Yingluo and Zhengyi gorges have ascending tendencies, and the increasing rates are respectively 2.04 m^3/s.10y and 1.61 m^3/s.10y; (4) the increase of annual runoff in the Yingluo Gorge has causal relationship with increased temperature and precipitation in the upper reaches, and the decrease of annual runoff in the Zhengyi Gorge in the past decades was mainly caused by the increased human consumption of water resources in the middle researches. The study results will provide scientific basis for making rational use and allocation schemes of water resources in the Heihe River drainage basin.
文摘Bayesian regularized BP neural network(BRBPNN) technique was applied in the chlorophyll-α prediction of Nanzui water area in Dongting Lake. Through BP network interpolation method, the input and output samples of the network were obtained. After the selection of input variables using stepwise/multiple linear regression method in SPSS i1.0 software, the BRBPNN model was established between chlorophyll-α and environmental parameters, biological parameters. The achieved optimal network structure was 3-11-1 with the correlation coefficients and the mean square errors for the training set and the test set as 0.999 and 0.000?8426, 0.981 and 0.0216 respectively. The sum of square weights between each input neuron and the hidden layer of optimal BRBPNN models of different structures indicated that the effect of individual input parameter on chlorophyll- α declined in the order of alga amount 〉 secchi disc depth(SD) 〉 electrical conductivity (EC). Additionally, it also demonstrated that the contributions of these three factors were the maximal for the change of chlorophyll-α concentration, total phosphorus(TP) and total nitrogen(TN) were the minimal. All the results showed that BRBPNN model was capable of automated regularization parameter selection and thus it may ensure the excellent generation ability and robustness. Thus, this study laid the foundation for the application of BRBPNN model in the analysis of aquatic ecological data(chlorophyll-α prediction) and the explanation about the effective eutrophication treatment measures for Nanzui water area in Dongting Lake.
基金Funding from The Scientific and Technological Research Council of Turkey(Project No:2130026)is gratefully acknowledged
文摘Background: Leaf Area Index(LAI) is an important parameter used in monitoring and modeling of forest ecosystems. The aim of this study was to evaluate performance of the artificial neural network(ANN) models to predict the LAI by comparing the regression analysis models as the classical method in these pure and even-aged Crimean pine forest stands.Methods: One hundred eight temporary sample plots were collected from Crimean pine forest stands to estimate stand parameters. Each sample plot was imaged with hemispherical photographs to detect the LAI. The partial correlation analysis was used to assess the relationships between the stand LAI values and stand parameters, and the multivariate linear regression analysis was used to predict the LAI from stand parameters. Different artificial neural network models comprising different number of neuron and transfer functions were trained and used to predict the LAI of forest stands.Results: The correlation coefficients between LAI and stand parameters(stand number of trees, basal area, the quadratic mean diameter, stand density and stand age) were significant at the level of 0.01. The stand age, number of trees, site index, and basal area were independent parameters in the most successful regression model predicted LAI values using stand parameters(R_(adj)~2=0.5431). As corresponding method to predict the interactions between the stand LAI values and stand parameters, the neural network architecture based on the RBF 4-19-1 with Gaussian activation function in hidden layer and the identity activation function in output layer performed better in predicting LAI(SSE(12.1040), MSE(0.1223), RMSE(0.3497), AIC(0.1040), BIC(-77.7310) and R^2(0.6392)) compared to the other studied techniques.Conclusion: The ANN outperformed the multivariate regression techniques in predicting LAI from stand parameters. The ANN models, developed in this study, may aid in making forest management planning in study forest stands.
基金financially supported by the National Natural Science Foundation of China(41761014,42161025,42101096)the Strategic Priority Research Program of Chinese Academy of Sciences(Grant No.XDA20020201)the Foundation of A Hundred Youth Talents Training Program of Lanzhou Jiaotong University,and the Excellent Platform of Lanzhou Jiaotong University。
文摘In the context of global warming,precipitation forms are likely to transform from snowfall to rainfall with a more pronounced trend.The change in precipitation forms will inevitably affect the processes of regional runoff generation and confluence as well as the annual distribution of runoff.Most researchers used precipitation data from the CMIP5 model directly to study future precipitation trends without distinguishing between snowfall and rainfall.CMIP5 models have been proven to have better performance in simulating temperature but poorer performance in simulating precipitation.To overcome the above limitations,this paper used a Back Propagation Neural Network(BNN)to predict the rainfall-to-precipitation ratio(RPR)in months experiencing freezing-thawing transitions(FTTs).We utilized the meteorological(air pressure,air temperature,evaporation,relative humidity,wind speed,sunshine hours,surface temperature),topographic(altitude,slope,aspect)and geographic(longitude,latitude)data from 28 meteorological stations in the Chinese Tianshan Mountains region(CTMR)from 1961 to 2018 to calculate the RPR and constructed an index system of impact factors.Based on the BNN,decision-making trial and evaluation laboratory method(BP-DEMATEL),the key factors driving the transformation of the RPR in the CTMR were identified.We found that temperature was the only key factor affecting the transformation of the RPR in the BP-DEMATEL model.Considering the relationship between temperature and the RPR,the future temperature under different representative concentration pathways(RCPs)(RCP2.6/RCP4.5/RCP8.5)provided by 21 CMIP5 models and the meteorological factors from meteorological stations were input into the BNN model to acquire the future RPR from 2011 to 2100.The results showed that under the three scenarios,the RPR in the number of months experiencing FTTs during 2011-2100 will be higher than that in the historical period(1981-2010)in the CTMR.Furthermore,in terms of spatial variation,the RPR values on the south slope will be larger than those on the north slope under the three emission scenarios.Moreover,the RPR values exhibited different variation characteristics under different emission scenarios.Under the low-emission scenario(RCP2.6),as time passed,the RPR values changed slightly at more stations.Under the mediumemission scenario(RCP4.5),the RPR increased in the whole CTMR and stabilized on the north slope by the end of this century.Under the high-emission scenario(RCP8.5),the RPR values increased significantly through the 21 st century in the whole CTMR.This study may help to provide a scientific management basis for agricultural production and hydrology.
文摘This paper develops a joint model utilizing the principal component analysis(PCA)and the back propagation(BP)neural network model optimized by the Levenberg Marquardt(LM)algorithm,and as an application of the joint model to investigate the damages caused by typhoons for a coastal province,Fujian Province,China in 2005-2015(latest).First,the PCA is applied to analyze comprehensively the relationship between hazard factors,hazard bearing factors and disaster factors.Then five integrated indices,overall disaster level,typhoon intensity,damaged condition of houses,medical rescue and self-rescue capability,are extracted through the PCA;Finally,the BP neural network model,which takes the principal component scores as input and is optimized by the LM algorithm,is implemented to forecast the comprehensive loss of typhoons.It is estimated that an average annual loss of 138.514 billion RMB occurred for 2005-2015,with a maximum loss of 215.582 in 2006 and a decreasing trend since 2010 though the typhoon intensity increases.The model was validated using three typhoon events and it is found that the error is less than 1%.These results provide information for the government to increase medical institutions and medical workers and for the communities to promote residents’self-rescue capability.
文摘The recurrent neural network (RNN) model based on projective operator was studied. Different from the former study, the value region of projective operator in the neural network in this paper is a general closed convex subset of n-dimensional Euclidean space and it is not a compact convex set in general, that is, the value region of projective operator is probably unbounded. It was proved that the network has a global solution and its solution trajectory converges to some equilibrium set whenever objective function satisfies some conditions. After that, the model was applied to continuously differentiable optimization and nonlinear or implicit complementarity problems. In addition, simulation experiments confirm the efficiency of the RNN.
基金supported by the Science and Technology Plan Project of Henan Province (No. 192102310232)。
文摘Accurate estimation of the solubility of a chemical compound is an important issue for many industrial proce sses.To overcome the defects of some thermodynamic models and simple correlations,a parallel neural network(PNN) model was conceived and optimized to predict the solubility of diosgenin in seven n-alkanols(C_(1)-C_(7)).The linear regression analysis of the parity plots indicates that the PNN model can give more accurate descriptions of the solubility of diosgenin than the ordinary neural network(ONN) model.The comparison of the average root mean square deviation(RMSD) shows that the suggested model has a slight advantage over the thermodynamic NRTL model in terms of the calculating precision.Moreover,the PNN model can reflect the effects of the temperature and the chain length of the alcohol solvent on the solution behavior of diosgenin correctly and can estimate its solubility in the n-alkanols with more carbon atoms.
文摘The factors of geomorphology, geological setting, effect of ground water and environment dynamic factors (e.g. rainfall and artificial water recharge) should be integrated in the discrimination of the stability of the ancient landslide. As the criterion of landslide stability has been studied, the artificial neural network model was then applied to discriminate the stability of the ancient landslide in the impounding area of the Three Gorges project on the Yangtze River, China. The model has the property of self adaptive identifying and integrating complex qualitative factors and quantitative factors. The results of the artificial neural network model are coincided well with what were gained by classical limit equilibrium analysis (the Bishop method and Janbu method) and by other comprehensive discrimination methods.
基金This project was supported by the National Natural Science Foundation of China (60074008) .
文摘We propose a new approach for analyzing the global asymptotic stability of the extended discrete-time bidirectional associative memory (BAM) neural networks. By using the Euler rule, we discretize the continuous-time BAM neural networks as the extended discrete-time BAM neural networks with non-threshold activation functions. Here we present some conditions under which the neural networks have unique equilibrium points. To judge the global asymptotic stability of the equilibrium points, we introduce a new neural network model - standard neural network model (SNNM). For the SNNMs, we derive the sufficient conditions for the global asymptotic stability of the equilibrium points, which are formulated as some linear matrix inequalities (LMIs). We transform the discrete-time BAM into the SNNM and apply the general result about the SNNM to the determination of global asymptotic stability of the discrete-time BAM. The approach proposed extends the known stability results, has lower conservativeness, can be verified easily, and can also be applied to other forms of recurrent neural networks.
文摘Intelligent systems and methods such as the neural network (NN) are usually used in electric power systems for short-term electrical load forecasting. However, a vast amount of electrical load data is often redundant, and linearly or nonlinearly correlated with each other. Highly correlated input data can result in erroneous prediction results given out by an NN model. Besides this, the determination of the topological structure of an NN model has always been a problem for designers. This paper presents a new artificial intelligence hybrid procedure for next day electric load forecasting based on partial least squares (PLS) and NN. PLS is used for the compression of data input space, and helps to determine the structure of the NN model. The hybrid PLS-NN model can be used to predict hourly electric load on weekdays and weekends. The advantage of this methodology is that the hybrid model can provide faster convergence and more precise prediction results in comparison with abductive networks algorithm. Extensive testing on the electrical load data of the Puget power utility in the USA confirms the validity of the proposed approach.
文摘Data from the deformation on Split Hopkinson Bar were used for constructing an artificial neural network model. When putting the thermodynamic parameters of the metals into the trained network model, the corresponding yielding stress can be predicted. The results show that the systematic error is small when the objective function is 0.5 , the number of the nodes in the hidden layer is 6 and the learning rate is about 0.1 , and the accuracy of the rate error is less than 3%. [
文摘One synthetical control method of AGC/LPC system based on intelligence control theory-neural networks internal model control method is presented. Genetic algorithm (GA) is applied to optimize the parameters of the neural networks. Simulation results prove that this method is effective.
基金Supported by the Ministerial Level Research Foundation(404040401)
文摘Training neural network to recognize targets needs a lot of samples.People usually get these samples in a non-systematic way,which can miss or overemphasize some target information.To improve this situation,a new method based on virtual model and invariant moments was proposed to generate training samples.The method was composed of the following steps:use computer and simulation software to build target object's virtual model and then simulate the environment,light condition,camera parameter,etc.;rotate the model by spin and nutation of inclination to get the image sequence by virtual camera;preprocess each image and transfer them into binary image;calculate the invariant moments for each image and get a vectors' sequence.The vectors' sequence which was proved to be complete became the training samples together with the target outputs.The simulated results showed that the proposed method could be used to recognize the real targets and improve the accuracy of target recognition effectively when the sampling interval was short enough and the circumstance simulation was close enough.
基金Supported by the National Science Fund for Distinguished Young Scholars of China (60925011)
文摘The polymer electrolyte membrane(PEM) fuel cell has been regarded as a potential alternative power source,and a model is necessary for its design,control and power management.A hybrid dynamic model of PEM fuel cell,which combines the advantages of mechanism model and black-box model,is proposed in this paper.To improve the performance,the static neural network and variable neural network are used to build the black-box model.The static neural network can significantly improve the static performance of the hybrid model,and the variable neural network makes the hybrid dynamic model predict the real PEM fuel cell behavior with required accuracy.Finally,the hybrid dynamic model is validated with a 500 W PEM fuel cell.The static and transient experiment results show that the hybrid dynamic model can predict the behavior of the fuel cell stack accurately and therefore can be effectively utilized in practical application.
基金Supported by Brilliant Youth Fund in Hebei Province
文摘On the basis of Artificial Neural Network theory, a back propagation neural network with one middle layer is building in this paper, and its algorithms is also given, Using this BP network model, study the case of Malian-River basin. The results by calculating show that the solution based on BP algorithms are consis- tent with those based multiple - variables linear regression model. They also indicate that BP model in this paper is reasonable and BP algorithms are feasible.
文摘This paper develops a feedforward neural network based input output model for a general unknown nonlinear dynamic system identification when only the inputs and outputs are accessible observations. In the developed model, the size of the input space is directly related to the system order. By monitoring the identification error characteristic curve, we are able to determine the system order and subsequently an appropriate network structure for systems identification. Simulation results are promising and show that generic nonlinear systems can be identified, different cases of the same system can also be discriminated by our model.
文摘This study aims to reveal the spatial structural characteristics of 1,652 Ethnic-Minority Villages(EMV)in China and to analyze the mechanisms driving their spatial heterogeneity.EMV are a special type of settlement space that preserve a large number of historical traces of the ethnic culture of ancient China.They are important carriers of China’s excellent traditional culture and are key to the implementation of rural revitalization strategies.In this study,1652 EMV in China were selected as the research subjects.The Nearest Neighbor Index,kernel density,and spatial autocorrelation index were employed to reveal the spatial structural characteristics of minority villages.Neural network models,spatial lag models,and geographical detectors were used to analyze the formation mechanism of spatial heterogeneity in EMV.The results indicate that:(1)EMV exhibit significant spatial differentiation characterized by“single-core with multiple surrounding sub-centers,”“polarization between east and west,”“decreasing quantity from southwest to east coast to northeast to northwest,”and“large dispersion with small agglomeration.”(2)EMV are mainly distributed in areas rich in intangible cultural heritage,with high vegetation coverage and low altitude,far from central cities,and having limited arable land and an underdeveloped economy and transportation,particularly in shaded or riverbank areas.(3)Distance from the nearest river(X3),distance from central cities(X8),national intangible cultural heritage(X9),and NDVI(X10)were the main driving factors affecting the spatial distribution of EMV,whereas elevation(X1)and GDP(X5)had the weakest influence.As EMV are a relatively unique territorial spatial unit,the identification of their spatial heterogeneity characteristics not only deepens the research content of settlement geography,but also involves the assessment,protection,and development of Minority Villages,which is of great significance for the inheritance and utilization of excellent ethnic cultures in the era.
基金the National Natural Science Foundation of China(No.U22B20141)the SINOPEC funded project(No.31900000-21-ZC0607-0009).
文摘Based on data from a petrochemical company’s MIP unit over the past three years,19 input variables and 2 output variables were selected for modeling using the maximum information coefficient and Pearson correlation coefficient among 155 variables,which included properties of feedstock oil and spent catalyst,operational variables,and material flows.The distillation range variables were reduced using factor analysis,and the feedstock oils were clustered into three types using the K-means++algorithm.Each feedstock oil type was then used as an input variable for modeling.An XGBoost model and a back propagation(BP)neural network model with a structure of 20-15-15-2 were developed to predict the combined yield of gasoline and propylene,as well as the coke yield.In the test set,the BP neural network model demonstrated better fitting and generalization abilities with a mean absolute percentage error and determination coefficient of 1.48%and 0.738,respectively,compared to the XGBoost model.It was therefore chosen for further optimization work.The genetic algorithm was utilized to optimize operational variables in order to increase the combined yield of gasoline and propylene while controlling the growth of coke yield.Seven commercial test results in the MIP unit showed an average increase of 1.39 percentage points for the combined yield of gasoline and propylene and an average decrease of 0.11 percentage points for coke yield.These results indicate that the model effectively improves the combined yield of gasoline and propylene while controlling the increase in coke yield.
文摘The quantization algorithm compresses the original network by reducing the numerical bit width of the model,which improves the computation speed. Because different layers have different redundancy and sensitivity to databit width. Reducing the data bit width will result in a loss of accuracy. Therefore, it is difficult to determinethe optimal bit width for different parts of the network with guaranteed accuracy. Mixed precision quantizationcan effectively reduce the amount of computation while keeping the model accuracy basically unchanged. In thispaper, a hardware-aware mixed precision quantization strategy optimal assignment algorithm adapted to low bitwidth is proposed, and reinforcement learning is used to automatically predict the mixed precision that meets theconstraints of hardware resources. In the state-space design, the standard deviation of weights is used to measurethe distribution difference of data, the execution speed feedback of simulated neural network accelerator inferenceis used as the environment to limit the action space of the agent, and the accuracy of the quantization model afterretraining is used as the reward function to guide the agent to carry out deep reinforcement learning training. Theexperimental results show that the proposed method obtains a suitable model layer-by-layer quantization strategyunder the condition that the computational resources are satisfied, and themodel accuracy is effectively improved.The proposed method has strong intelligence and certain universality and has strong application potential in thefield of mixed precision quantization and embedded neural network model deployment.