选煤厂原煤智能配比控制系统设计

针对选煤厂配煤入选过程中通过理论计算的原煤定值配比难以满足现场生产要求,而依据人工经验调节原煤配比随机性大、精煤质量难以保证、人工劳动强度大等问题,设计了一种选煤厂原煤智能配比控制系统。利用最小二乘支持向量机建立原煤智能配比预测模型,并采用粒子群优化算法进行模型参数优化。该系统以原煤灰分实测值、原煤硫分实测值、每小时原煤平均入选量、分选密度、精煤灰分实测值、精煤灰分目标值、精煤硫分实测值和精煤硫分目标值作为模型输入变量,经模型计算得出相应的原煤配比预测值;通过胶带秤测量给煤机的给煤量得到原煤配比实测值,并与预测值比较得出原煤配比偏差量;PID控制器根据偏差量控制给煤机的变频器频率,实现原煤配比精准调节。该系统应用后精煤灰分和硫分波动范围明显减小,精煤质量稳定性良好;精煤灰分与灰分目标值差值控制在±0.2%,精煤硫分与硫分目标值差值控制在±0.15%,提高了精煤质量。     

人工神经网络用于煤直接液化反应模拟的研究

采用人工神经网络的方法模拟了煤直接液化过程中各变量之间的非线性关系。选用ｌｅａｖｅ－ｏｎｅ－ｏｕｔ的交叉验证（ｃｒｏｓｓ－ｖａｌｉｄａｔｉｏｎ）方法确定网络的最佳拓扑结构,达到对催化条件下煤直接液化过程最大限度的拟合。结果表明,尽管部分样本的预报绝对误差大于１０％,人工神经网络可以较好地模拟反应条件与产物产率之间的关系,并能准确预报催化剂浸渍量对煤的总转化率和沥青烯轻质产物产率的影响趋势。由于实验     

两段式气化工艺流程的ASPEN PLUS软件模拟

针对目前Shell、Texaco气化技术的不足,本文提出了一种新型的两段气化工艺,以达到充分利用煤气显热的目的.借助Aspen Plus过程模拟平台,运用Gibbs自由能最小化方法,模拟两段式水煤浆进料的气化工艺流程,同时建立二段固定床气化模型.考察了加入二段同定床气化炉内的煤虽及蒸汽煤比对出口煤气组分、温度及整个气化工艺的冷煤气效率、有效气产率等气化指标的影响.结果两段式气化工艺能有效利用一段炉煤气中的显热,提高气化炉的冷煤气效率.同时通过研究,还得到了不同工艺下的最佳条件.本论文所得结论为进一步研究该新型组合式气化工艺提供参考.     

石脑油裂解过程中一次反应选择性系数的调整方法研究

在分析石脑油裂解炉辐射段的Kumar分子反应动力学模型的基础上,提出了一种应用数据融合技术对Kumar模型一次反应选择性系数进行在线调整的方法。首先,利用进化的遗传算法,以改进的石脑油裂解模型计算出的产率与实际工况产率的误差作为目标函数,对选择性系数进行离线调整,进而建立原料性质与选择性系数相匹配的标准样本数据库。其次,利用数据融合技术中的Dempeter-Shafer证据理论建立模型,计算待估物料与标准样本数据库中样本物料的匹配度,进而对待估原料的一次反应选择性系数进行估算。研究结果表明,利用估算出的一次反应选择性系数建立的裂解工艺模型具有较高精度,将仿真结果与实际工况产率进行对比,证实了该方法的有效性和可行性。     

多种基于Aspen Plus煤干馏工艺模型的建立及对比分析

煤干馏工艺作为煤化工技术的重要环节,是煤炭工业生产中的关键组成部分,是煤化工工业的基础。本文选用了Suuberg模型、脱挥发模型及半焦分解模型共三种建模方法,基于化工流程模拟软件Aspen Plus建立煤干馏工艺模型,针对每一种模型的优缺点及适用情况进行了讨论,以期为实际工业流程模拟提供借鉴。通过研究得出结论:该三种模型在煤干馏工艺模拟中均适用,并在一定程度上避免了模型的复杂性,计算简便,大大减小了工作量。Suuberg模型基于实验数据可推算煤干馏过程的产物组成,但不同煤种间通用性差,有一定局限性;脱挥发模型是以大量实验数据为基础、通过反复比对调整模型常数得出,普适性更强;而半焦裂解模型为虚拟模块,适用于无需关注煤干馏具体反应过程,较前两种模型更加简洁。     

基于GMDH的卷烟工艺参数-指标关系模型研究

通过对烟草加工中工艺参数与质量指标之间的关系研究,提出采用自组织数据挖掘方法建立相应的关系模型,并利用该模型预测质量指标取值。通过与多元线性回归模型的预测值对比,证明了该方法的有效性。     

硫醇类硫化物与H+相互作用的分子模拟

为了研究催化裂化、加氢等工艺过程中,硫醇类硫化物与固体酸性催化剂的B酸中心之间的相互作用机理,采用动力学和半经验量子力学方法,对几种典型硫醇类硫化物与H+之间的相互作用进行了研究.以2-戊硫醇、环己硫醇和硫酚为模型化合物,模拟计算了各硫醇类硫化物的最高占据轨道(HOMO)以及H+进攻硫位及不同碳位时所形成中间体的生成热及键长变化,进而讨论H+的优先进攻位置以及反应中间体的稳定性.结果表明,H+的进攻位置对形成的中间体的稳定性有很大影响.当H+进攻硫醇类硫化物中的碳位时,反应中间体的生成热远大于或接近于H+直接进攻硫位所形成的中间体的生成热;对于反应中间体的生成热同进攻硫位所形成的中间体的生成热相近的碳原子上的质子化反应,其实质与直接进攻硫原子时相一致,并生成硫化氢,这说明H+将优先进攻硫醇类硫化物中的硫原子.由此可以认为,硫醇类硫化物在固体酸性催化剂的作用下,由于B酸中心的给质子性质,较容易在硫位上发生质子化反应,以生成硫化氢的形式转化脱除.     

基于粒子群优化算法的乙烯裂解炉全周期建模及参数优化

针对乙烯裂解炉目标产品产率最大化和延长清焦周期之间存在矛盾,建立了一种严格的机理模型,并将粒子群优化算法(Particle Swarm Optimization,PSO)应用于该模型,通过选取和建立产率和利润综合最优目标函数,获得最优的乙烯裂解炉入口参数.PSO算法是一种群集智能方法,通过粒子之间的合作与竞争及进化实现对多维复杂空间的高效搜索.属于一类随机全局优化技术,已成功应用于各科学和工程领域.该方法应用于KTI公司的GK-V型裂解炉辐射段入口参数优化的研究,结果表明该方法效果良好,可以对目标产品产率和结焦速率进行合理调整,并使利润最大化.     

工程信息结构的语义模型

本文提出一种新的语义模型，该模型能够支持ＣＩＭ工程中相互联系的各类数据的结构化管理。首先定义了问题范围内的对象与对象类；在研究对象间联系的基础上建立起对象、对象类之间的语义关系；然后定义了作用于语义关系和对象上的操作，从而得出一种面向工程信息结构的语义模型。模型描述了一个项目的层次结构。通过在共享对象上建立等价语义关系，能够将项目的层次结构图转化为项目树，进而可用树结构实现对ＣＩＭ工程中产品及各类     

面向虚拟环境的装配体模型研究

人机交互是虚拟装配的本质特性和技术瓶颈，能反映交互过程的装配体模型构造，是虚拟装配领域的一个关键问题，给出一种时序的装配体模型，对时序的聚合关系、时序的约束关系以及时序的行为结构给出定义和分析，为实现该模型，还给出了零件的中间模型和对象复杂行为构造，同时对虚拟装配直接操作的交互过程、交互模式及装配工艺知识给出了时序的形式描述，这是连接人机交互与装配体模型以及模型与工艺生成的桥梁．一致的表示模型减少了交互装配信息的冗余性，有利于提高交互信息的复用，为交互生成工艺提出了新方法。     

Matlab神经网络工具箱在煤层界面插值中的应用

人工神经网络方法通过大量的样本学习和训练,能够对实际问题进行最佳逼近。在分析煤层界面建模数据的基础上,设计插值煤层界面的人工神经网络结构。利用Matlab的神经网络工具箱建立对煤层界面插值的人工神经网络模型。以某矿的钻孔数据为样本对人工神经网络进行训练,实现了对煤层界面网格的插值;通过Matlab实现了对煤层界面的插值结果的三维显示。     

人工神经网络在煤氧化性预测方面的研究

提出了利用人工神经网络技术进行煤氧化速率定量预测的新方法．根据对煤自燃的实际特点和基本规律的综合考虑，研究了影响因素的选取、煤氧化速率预测模型的建立等问题。采用BP神经网络算法对煤氧化速率进行了建模。结果表明，用神经网络模型对煤氧化速率进行模拟预测，具有理论上的可行性和现实意义，说明人工神经网络技术在煤氧化速率预计领域中具有实用价值。     

Improvements in the accuracy of an Inverse Problem Engine's output for the prediction of below-knee prosthetic socket interfacial loads

The monitoring of in-service loads on many components has become a routine operation for simple and well-understood cases in engineering. However, as the complexity of the structure increases so does the difficulty in obtaining an acceptable understanding of the real loading. It has been shown that it is possible to solve these problems by interfacing traditional analysis methodologies with more modern mathematical methods (i.e. artificial intelligence) in order to create a hybrid analysis tool. It has, however, been recognised that an Artificial Neural Network (ANN) predicts poorly in the high and low ranges of the envelope of which it is trying to predict. This paper presents results of research to develop the ANN Difference Method to improve the accuracy of the Inverse Problem Engine's output. This method has been applied to accurately predict the complex pressure distribution at the residual limb/socket interface of a lower-limb prosthesis. It has been shown that application of the ANN Difference Method to the output of a backpropagation neural network can reduce inherent errors that exist at the low and high ends of the ANN solution envelope. This powerful approach can offer load information at high speed once the relationship between the loading and response of the component has been established through training the ANN. Utilising an experimental technique combined with an ANN can provide in-service loads on complex components in real time as part of a sophisticated embedded system.     

Neural Network and Classification Approach in Identifying Customer Behavior in the Banking Sector: A Case Study of an International Bank

The customer relationship focus for banks is in development of main competencies and strategies of building strong profitable customer relationships through considering and managing the customer impression, influence on the culture of the bank, satisfactory treatment, and assessment of valued relationship building. Artificial neural networks (ANNs) are used after data segmentation and classification, where the designed model register records into two class sets, that is, the training and testing sets. ANN predicts new customer behavior from previously observed customer behavior after executing the process of learning from existing data. This article proposes an ANN model, which is developed using a six‐step procedure. The back‐propagation algorithm is used to train the ANN by adjusting its weights to minimize the difference between the current ANN output and the desired output. An evaluation process is conducted to determine whether the ANN has learned how to perform. The training process is halted periodically, and its performance is tested until an acceptable result is obtained. The principles underlying detection software are grounded in classical statistical decision theory.     

Calibrating artificial neural networks by global optimization

Artificial neural networks (ANNs) are used extensively to model unknown or unspecified functional relationships between the input and output of a “black box” system. In order to apply the generic ANN concept to actual system model fitting problems, a key requirement is the training of the chosen (postulated) ANN structure. Such training serves to select the ANN parameters in order to minimize the discrepancy between modeled system output and the training set of observations. We consider the parameterization of ANNs as a potentially multi-modal optimization problem, and then introduce a corresponding global optimization (GO) framework. The practical viability of the GO based ANN training approach is illustrated by finding close numerical approximations of one-dimensional, yet visibly challenging functions. For this purpose, we have implemented a flexible ANN framework and an easily expandable set of test functions in the technical computing system Mathematica. The MathOptimizer Professional global-local optimization software has been used to solve the induced (multi-dimensional) ANN calibration problems.     

基于人工神经网络组合预测油田产量

油田原油产量的准确预测可以对油田的生产管理进行合理的指导。该文探讨了应用神经网络组合方法预测油田产量，对开井数、含水率、动用储量以及往年产量同未来产量之间的复杂关系建立模型。采用了两层预测系统：第一层包含两个神经网络，一个多层前馈网络和一个函数链接网络；第二层是把第一层的两个网络输出进行组合。研究了五种不同的组合算法：平均法、最小平方回归法、模糊逻辑法、自适应前馈神经网络法和自适应函数链接神经网络法。根据油品类型分为稀油、热采稠油、常规稠油和总产量四组数据，对上述方法进行了测试，结果表明应用人工神经网络的组合预测方法优于其他的预测方法，而且适用范围广。     

煤与瓦斯突出的粗神经网络预测模型研究

将粗集方法作为BP神经网络的前端处理器,通过对煤与瓦斯系统属性特征的提取和影响因素的约简,较好解决了预测输入特征的“维数灾”问题,构建了粗集与神经网络相结合的煤与瓦斯突出预测模型。仿真实验表明,验证了该方法的有效性,模型学习速度更快、精确度更高,对提高瓦斯突出预测时效性有重大意义。     

Online implementation of an adaptive calibration technique for displacement measurement using LVDT

The paper presents the design and validation of an online intelligent displacement measurement technique with Linear Variable Differential Transformer (LVDT) using Artificial Neural Network (ANN). The objectives of the proposed work are to design a calibration technique using an optimised neural network model such that it a) produces an output which is linear for the full scale of input range, b) makes the output independent of the variations in supply frequency, the physical parameters of the LVDT, and ambient temperature. The output of an LVDT is converted to a DC signal by using a rectifier circuit. The rectified output is further amplified using a differential amplifier. This voltage signal is acquired onto a computer for further processing using an ANN. The optimisation of the ANN is carried out to find the minimum number of hidden layers along with the number of neurones in each layer to give least Mean Square Error (MSE) and Regression (R) nearing to one. This optimisation is done considering various schemes of ANN, training algorithms, and the transfer function of neurones. Once the ANN model is designed, it is subjected to test with both simulated data and experimental validation. The results confirm the successful achievement of the objectives of this paper and thus avoiding the need for repeated calibration.     

BP神经网络在证券分析预测中应用

介绍了人工神经网络的基本概念和组成后,讨论了它与传统分析预测方法的区别和优势。对证券分析和预测提出了一种基于BP算法的人工神经网络模型,结合计算实例和结果的误差分析,给出了改进的方向。     

A filter model for feature subset selection based on genetic algorithm

This paper describes a novel feature subset selection algorithm, which utilizes a genetic algorithm (GA) to optimize the output nodes of trained artificial neural network (ANN). The new algorithm does not depend on the ANN training algorithms or modify the training results. The two groups of weights between input-hidden and hidden-output layers are extracted after training the ANN on a given database. The general formula for each output node (class) of ANN is then generated. This formula depends only on input features because the two groups of weights are constant. This dependency is represented by a non-linear exponential function. The GA is involved to find the optimal relevant features, which maximize the output function for each class. The dominant features in all classes are the features subset to be selected from the input feature group.