Back propagation genetic and recurrent neural network applications in modelling and analysis of squeeze casting process

Today, in competitive manufacturing environment reducing casting defects with improved mechanical properties is of industrial relevance. This led the present work to deal with developing the input-output relationship in squeeze casting process utilizing the neural network based forward and reverse mapping. Forward mapping is aimed to predict the casting quality (such as density, hardness and secondary dendrite arm spacing) for the known combination of casting variables (that is, squeeze pressure, pressure duration, die and pouring temperature). Conversely, attempt is also made to determine the appropriate set of casting variables for the required casting quality (that is, reverse mapping). Forward and reverse mapping tasks are carried out utilizing back propagation, recurrent and genetic algorithm tuned neural networks. Parameter study has been conducted to adjust and optimize the neural network parameters utilizing the batch mode of training. Since, batch mode of training requires huge data, the training data is generated artificially using response equations. Furthermore, neural network prediction performances are compared among themselves (reverse mapping) and with those of statistical regression models (forward mapping) with the help of test cases. The results shown all developed neural network models in both forward and reverse mappings are capable of making effective predictions. The results obtained will help the foundry personnel to automate and précised control of squeeze casting process.     

Modeling and optimization method featuring multiple operating modes for improving carbon efficiency of iron ore sintering process

Iron ore sintering is one of the most energy-consuming processes in steelmaking. Since its main source of energy is the combustion of carbon, it is important to improve the carbon efficiency to save energy and to reduce undesired emissions. A modeling and optimization method based on the characteristics of the sintering process has been developed to do that. It features multiple operating modes and employs the comprehensive carbon ratio (CCR) as a measure of carbon efficiency. The method has two parts. The first part is the modeling of multiple operating modes of the sintering process. K-means clustering is used to identify the operating modes; and for each mode, a predictive model is built that contains two submodels, one for predicting the state parameters and one for predicting the CCR. The submodels are built using back-propagation neural networks (BPNNs). An analysis of material and energy flow, and correlation analyses of process data and the CCR, are used to determine the most appropriate inputs for the submodels. The second part of the method is optimization based on a determination of the optimal operating mode. The problem of how to reduce the CCR is formulated as a two-step optimization problem, and particle swarm optimization is used to solve it. Finally, verification of the modeling and optimization method based on actual process data shows that it improves the carbon efficiency of iron ore sintering.     

Performance and Modeling of an Up-flow Anaerobic Sludge Blanket(UASB) Reactor for Treating High Salinity Wastewater from Heavy Oil Production

In this study,an up-flow anaerobic sludge blanket(UASB) reactor was applied to treat the high salinity wastewater from heavy oil production process.At a HRT of ≥24 h,the COD removal reached as high as 65.08% at an influent COD ranging from 350mg/L to 640mg/L.An average of 74.33% oil reduction was also achieved in the UASB reactor at an initial oil concentration between 112mg/L and 205mg/L.These results indicated that this heavy oil production related wastewater could be degraded efficiently in the UASB reactor.Granular sludge was formed in this reactor.In addition,two models,built on the back propagation neural network(BPNN) theory and linear regression techniques were developed for the simulation of the UASB system performance in the oily wastewater biodegradation.The average error of COD and oil removal was-0.65% and 0.84%,respectively.The results indicated that the models built on the BPNN theory were wellfitted to the detected data,and were able to simulate and predict the removal of COD and oil by the UASB reactor.     

遥感图像中无水桥梁识别新算法

 给出了边缘密度和线段复杂度的定义,并提出一种遥感图像中无水桥梁的识别新算法.首先对图像进行边缘提取,计算像素点的边缘密度,根据边缘密度进行图像分割,接着采用Hough变换提取直线,利用线段复杂度等确定疑似桥梁区域,然后计算疑似桥梁区域像素点的纹理特征,并构成一个特征矢量,最后将此特征矢量送入BPNN进行分类,统计该区域所有像素点的分类结果以判决是否属于桥梁.实验结果表明,该算法能够较好地识别出遥感图像中的无水桥梁目标.     

基于量子微粒群的BPNN在转炉炼钢静态模型中的应用

针对转炉炼钢静态模型终点命中率较低的问题,首先分析了影响转炉炼钢终点命中率的各种因素,确定了BP神经网络(BPNN)的拓扑结构,并依此建立了转炉炼钢静态模型.然后把量子微粒群算法(QPSO)应用于BP网络的学习中,并比较了QPSO、基本微粒群优化算法(PSO)、梯度下降法的学习性能.最后,基于某炼钢厂的历史数据进行了仿...     

GRNN与BPNN的函数逼近性能对比研究

为了研究GRNN和BPNN非线性函数的逼近能力,从数学角度详细阐述了GRNN和基于LM优化算法改进的BPNN的学习过程,编程建立了GRNN和BPNN,并分别用两种神经网络对指定的非线性函数进行逼近实验。仿真结果表明,在训练样本数量相等且中小规模网络的条件下,相对于BPNN而言,GRNN的逼近精度更高、收敛速度更快,具有很好的逼近能力,为解决非线性函数的逼近问题提供了良好的解决手段。     

A dynamic meta-learning rate-based model for gold market forecasting

In this paper, an improved EMD meta-learning rate-based model for gold price forecasting is proposed. First, we adopt the EMD method to divide the time series data into different subsets. Second, a back-propagation neural network model (BPNN) is used to function as the prediction model in our system. We update the online learning rate of BPNN instantly as well as the weight matrix. Finally, a rating method is used to identify the most suitable BPNN model for further prediction. The experiment results show that our system has a good forecasting performance.     

基于IAFSA-BPNN的短期风电功率预测

为提高短期风电功率预测精度,提出一种基于IAFSA-BPNN的短期风电功率预测方法。该方法通过改进的人工鱼群算法来优化BP神经网络的权值和阈值,从而提高BP神经网络的收敛速度和泛化能力。利用2014年上海某风场实测数据对新算法进行检验。试验结果表明,改进的人工鱼群算法一定程度上克服了原算法后期搜索的盲目性较大,收敛速度减慢,搜索精度变低的缺陷。IAFSA-BPNN混合算法在预测的稳定性和精度、收敛速度等方面优于BPNN、AFSA-BPNN算法。IAFSA-BPNN算法不仅能提高短期风电功率预测的精度,而且改善了预测结果稳定性。     

基于PSO优化BPNN估计光伏阵列MPPT控制系统研究

为了充分利用光伏阵列转换能量,提高光伏阵列的发电效率,在分析光伏阵列的伏安特性及最大功率点跟踪(MPPT)原理的基础上,提出了一种基于粒子群算法优化BP神经网络(PSO-BPNN)的建模方法,并用这种改进的神经网络构建了光伏阵列的动态模型.通过PSO-BPNN模型拟合光伏阵列输出功率与输出电压的非线性关系,实现了对光伏阵列的最大功率点跟踪.Matlab/Simulink仿真及在线测试结果表明:基于PSO-BPNN估计的光伏阵列MPPT控制系统能快速、精确地跟踪光伏阵列的最大功率点,改善了BP神经网络收敛速度慢,易陷入局部极值,建模精度不高的缺点,提高了系统的稳定性和能量转换效率,是研究光伏发电这个复杂非线性系统的一个可行办法.     

基于DBN网络与BP神经网络PID控制的永磁同步电机调速策略比较研究

针对BP神经网络自身存在的学习速率固定、记忆不稳定等缺点,设计了一种基于DBN网络PID的永磁同步电机控制器,通过Matlab/Simulink对基于BP神经网络PID控制器的电机调速策略和基于DBN网络PID控制器的电机调速策略进行建模仿真分析,探讨两者对于PMSM调速策略中控制鲁棒性和稳定性的优劣。仿真结果表明,基于DBN网络PID的永磁同步电机调速控制策略训练效果更佳,具有更好的稳定性和鲁棒性。