The Oil Viscosity Correlations: A Simulation Approach

Over the years, different authors have proposed many oil viscosity correlations for various crude oil mixtures (dead, saturated, or undersaturated) from all over the world. Authors tend to support their own correlations, which are developed for specific sets of hydrocarbon mixtures. When tested on other data sets, however, they do not perform as anticipated. The authors considered a total of 13 undersaturated correlations for a sharp review from a simulation perspective. They came to the conclusion of supporting the use of undersaturated viscosity correlations that use the exponential parameterized pressure differential form. Thus a new fine tuning parameter, which sets a sound basis for local data sets to be accounted for, has been proposed.     

The Artificial Neural Network's Prediction of Crude Oil Viscosity for Pipeline Safety

Abstract

Predicting crude oil viscosity is a challenge faced by reservoir engineers in production planning. Some early researchers have propounded some theories based on crude oil properties and have encountered various problems leading to errors in forecasted values. This article discusses work carried out with a model using an artificial neural network (ANN) for predicting crude oil viscosity of Nigerian crude oil. The model was started through adoption of a classical regression technique empirical method for dead oil viscosity as a function of American Institute for Petroleum (API) and reduced temperature. The Peng–Robinson equation of state and other thermodynamic properties are introduced, coupled with the Standing model for calculating bubble point pressure (P_b). The developed model was evaluated using existing measured real-life data collected from 10 oil fields within the Niger Delta region of Nigeria. Both the predicted and measured viscosities were plotted against each corresponding reservoir pressure to establish the model's level of reliability. The superimposition of the pressure-viscosity relationship shows that at each point, the viscosity model captures the physical behavior of viscosity variations with pressure. In each case, the ANN does not require a data relationship to predict the crude oil viscosity but rather relies on the field data obtained for training. For this reason, it is recommended that the ANN approach should be applied in oil fields for reduction in error, computational time, and cost of overproduction and underproduction.     

基于人工神经网络的油-水乳状液粘度智能预测研究

油水混合粘度受温度、剪切速率、含水率3种因素协同影响,很难用常规方程准确计算.提出了采用人工神经网络进行油水乳状液粘度预测的新方法.建立了三层结构BP神经网络模型,输入层有3个神经元,分别代表温度、剪切速率、含水率,输出层有一个神经元,代表油水混合物粘度,隐层神经元数目为30个.在实验室配置一定比例的油水乳状液,通过流...     

基于人工神经网络（BP）方法预测汽油辛烷值

本文基于人工神经网络（ＢＰ）方法,用毛细管色谱法预测汽油馏分的辛烷值,其预测最大绝对误差为０．２８,平均误差为０．１２２,比常用的线性回归数学模型法更能准确地预报辛烷值。     

The Prediction of Permeability Using an Artificial Neural Network System

Abstract

The authors studied the efficiency and accuracy of neural network model for prediction of permeability as a key parameter in reservoir characterization. So, some multilayer perceptron (MLP) neural network models with different learning algorithms of Levenberg-Margnardt, back propagation, improved back propagation (IBP), and quick propagation with three layers and different node numbers (3, 4, 5, 6, 7) in the middle layer have been presented. These models have been obtained by 630 permeability data from one of offshore reservoirs located in Saudi Arabia. The accuracy of models was studied by comparing the obtained results of each model with experimental data. So, the neural network with IBP learning method and five nodes in the middle layer has the most accuracy.     

基于人工神经网络混合油品粘度预测模型研究

在分析前向BP神经网络基本原理的基础上,对3种混油建立了人工神经网络混油粘度预测模型,该模型结构为1-7-1的三层BP网络模型。运用实测数据对BP网络进行训练和仿真。结果表明,三种模型预测误差全在2.5%以内,比前苏联学者提出的混油粘度计算公式——克恩达尔-莫恩罗埃公式和兹达诺夫斯基公式更具有计算精度高、适用性强的特点,可完全满足工程实际需要。     

神经网络构建稠油掺稀黏度预测模型研究

本文利用BP神经网络能够较好地在实验数据基础上建立稠油掺稀黏度预测模型,以对新疆塔河油田稠油掺入四种稀油的黏度预测为例,通过BP神经网络建立预测模型,并与四种传统基于线性回归的建模方法及进行改进的方法进行对比,结果表明：利用神经网络建立模型的最大误差为4.1%,黏度与温度、稀稠比的非线性关系能够较好拟合,对比基于线性回归方法的建模方法及其改进算法有着更高的拟合精度。     

Application of Artificial Neural Network in the Residual Oil Hydrotreatment Process

Abstract

Based on the industrial measured data of the residual oil hydrotreatment process, the artificial neural network (ANN) model was developed to determine metal, sulfur, nitrogen, and carbon residue content of hydrogenated residual oil. The established ANN model has seven input variables, four output variables, and 1 hidden layer with 15 neurons. The training results show that the agreement between predicted and industrial measured values is good. The mean relative errors of the testing data for the four output variables are less than 6%. It indicated that the developed ANN model has good predictive precision and extrapolative features. The model can provide reference for the further processing of hydrogenated residual oil. This kind of application can be easily developed in any other hydrotreatment process with available adequate historical data.     

基于模糊神经网络的建筑工程造价估计方法

提出了建筑工程造价估计的模糊神经网络方法,给和该方法进行建筑工程造 价估计的基本原理,网络模型及估价方法,计算实例表明,应用模糊神经网络估计工程造 价具有方便、准确的特点。     

基于神经网络的高含硫气田L360钢腐蚀速率预测

高含硫气田地面集输系统广泛使用L360钢,由于腐蚀因素的多样性及协同效应,其腐蚀速率预测一直是个难题。文章介绍了不同腐蚀因素对L360钢腐蚀速率的影响。随着H2S和CO2压力的增高,腐蚀速率先降后升,在H2S和CO2压力为1.00和0.67 MPa时达到最小值;随Cl-质量浓度的升高,腐蚀速率增大,但当Cl-质量浓度高于40 g/L后,腐蚀速率反而降低;随着温度的升高,腐蚀速率增大,当温度超过70℃后,腐蚀速率反而降低。建立了三层结构BP神经网络模型,输入层有6个神经元,分别代表H2S,CO2分压、Cl-质量浓度、温度、流速和沉积硫6种腐蚀影响因素,隐层神经元数目为8个,输出层神经元数目为1个,代表腐蚀速率。结果表明,L360钢在试验水中的平均腐蚀速率的预测最大误差在15.9%以内,可以满足工程应用要求。     

基于神经网络模型的混合原油黏度预测

采用BP(back propagation) 神经网络模拟了混合油配比和温度与黏度之间的映射关系,建立了混合油黏度神经网络预测模型.将鲁-宁输油管道混合油的预测黏度与实际黏度进行了对比,结果表明,利用BP神经网络进行混合油黏度预测是可行的,完全可以满足工程需要的精度要求.     

Oxidative Desulfurization of Fuel Oil: Modeling Based on Artificial Neural Network

Oxidative desulfurization of fuel oil was investigated using a process consisting of oxidation and distillation steps. In the oxidation step, various organic carboxylic acid/H₂ O₂ systems, especially acetic acid/H₂ O₂, were used as oxidant. They oxidize both easy and refractory sulfur compounds and convert them into oxidized sulfur compounds. The oxidized sulfur compounds are finally removed from fuel oil by distillation in the presence of water. The sulfur content of fuel oil was decreased to levels as low as 20 ppm (up to 90%) in a short contact time, ambient temperature, and atmospheric pressure. The results showed that applying this process did not have any deleterious influence on the distillation characteristic, composition, and content of fuel oil that was examined. An artificial neural network, using back propagation (BP), was also utilized for modeling oxidative desulfuration process of fuel oil. The comparison between the output of ANN modeling and the experimental data showed satisfactory agreement.     

The Prediction of Wax Precipitation by Neural Network and Genetic Algorithm and Comparison With a Multisolid Model in Crude Oil Systems

Abstract

This article introduce a new implementation of the neural network and genetic programming neural network technology in petroleum engineering. An intelligent framework is developed for calculating the amount of wax precipitation in petroleum mixtures over a wide temperature range. Theoretical results and practical experience indicate that feedforward networks can approximate a wide class of function relationships very well. In this work, a conventional feedforward multilayer neural network and genetic programming neural network (GPNN) approach have been proposed to predict the amount of wax precipitation. The introduced model can predict wax precipitation through neural network and genetic algorithmic techniques. The accuracy of the method is evaluated by predicting the amount of wax precipitation of various reservoir fluids not used in the development of the models. Furthermore, the performance of the model is compared with the performance of multisolid model for wax precipitation prediction and experimental data. Results of this comparison show that the proposed method is superior, both in accuracy and generality, over the other models.     

基于能量及原油物性定量分析的油-水混合液黏度预测模型

高含水油-水混合液往往不能形成稳定的乳状液,而是原油将其中一部分水乳化,形成了油包水(W/O)乳状液液滴和游离水的掺混体系.传统的乳状液黏度模型并不适用于这种非稳定乳化的油-水混合体系.采用搅拌测黏法测定并研究了搅拌转速、含水率及温度对油-水混合液表观黏度的影响.结果表明:油-水混合液的表观黏度随着搅拌速率的增大、含水...     

模糊神经网络推理系统定量计算储层参数

储层参数是地层评价的基础。自适应模糊神经网络推理系统是综合了模糊逻辑与人工神经网络两者优势的一种人工智能方法．该方法既可用于模式分类．又可进行连续计算。这是一种基于数据的建模方法,模糊隶属度函数及模糊规则是通过大量已知数据的学习得到的．而不是按经验或直觉信息给定的。这对于那些特性还不为人们所了解或者特性非常复杂的系统尤其重要．因此特别适于复杂储层评价中的定性解释与定量计算。研究结果表明,这种方法具有较高的计算精度。     

智能优化算法及人工神经网络在催化裂化模型分析中的应用进展

催化裂化是一个由多种高度非线性和相互强关联因素影响的复杂工艺过程,对其工艺过程和产品收率优化的数学建模分析一直是石油加工领域研究的热点和难点。集总动力学模型是机理分析层面最为常用的研究方法。选用合适而快捷的参数估算和求取方法,是集总动力学模型构建过程中的重要一环。遗传算法、粒子群算法和模拟退火算法等智能算法一定程度上克服了经典算法对初值依赖性,难寻找全局最优的问题,同时还保证了算法的收敛性,对于集总动力学模型的发展起到了极大的促进作用。此外,通过构建原料油性质、催化剂性质、操作条件和产品分布之间的神经网络模型,可以从统计学的角度找到产物分布的影响机制,分析得到常规集总分析方法忽略的一些因素,且可对产物分布进行进一步的预测,是构建催化裂化分析模型的一种新型且有效的手段。笔者对现有关于人工智能算法在催化裂化工艺模型构建中应用的研究成果做一整理,以期对后续的研究提供帮助。     

GM(1.1)与BP神经网络组合模型在原油产量预测中的应用

为了维持油田能够长期稳定高产，必须制订科学合理的生产方案。而常规预测方法对数据依赖大，预测精度不高。为此，在灰色预测理论的基础上引入BP神经网络模型，建立了GM（1．1）和BP神经网络组合模型。此组合模型兼有灰色预测和BP神经网络预测的优点，克服了原始数据少，数据波动性大对预测精度的影响，同时也增强了预测的自适应性。最后通过实例对比分析，说明了组合模型的有效性及可应用性。     

基于神经网络的海水淡化系统产水模拟及预测

海水淡化是一个涉及传热传质等诸多因素的复杂的非线性过程。为了克服传统数学建模方法在模拟仿真精度和实时性方面的不足,对大样本空间的神经网络构造及网络学习加速方法进行了大量探索及尝试,成功地将神经网络技术引入海水淡化产水过程中构造新的产水模型,建立了以空气入口干球温度、预冷器进口冷却水温度、海水喷淋温度和海水喷淋量作为输入参数的海水淡化系统神经网络模型。分析表明,该模型不仅具有较高的仿真精度,而且保持了系统原有的光滑性,能满足系统实时仿真模拟预测的要求。     

Prediction of the Kinematic Viscosity of Crude Oil Fractions

Abstract

A generalized equation based on modified Eyring's theory for predicting kinematic viscosity of petroleum fractions is proposed in this work. The equation uses two reference fluids including a pair of (C6 and C10), (C10 and C14), or (C14 and C20) for petroleum fractions of molecular weight higher than 70 and lower than 300.

Validity and accuracy of this equation have been confirmed by comparing the obtained results of this equation with experimental data. In contrast to other correlations that require so many specific parameters for oil viscosity prediction, this type of equation requires only molecular weight and true boiling point. The results obtained in this work are in agreement with experimental data with an average absolute deviation (AAD) of less than 5%.     

改进神经网络在预测评判油气田设备腐蚀分析中的应用

主要针对一般BP神经网络易陷入局部最小值、收敛速度慢、引起扬荡效应的缺点,提出用一种改进遗传算法对BP网络的权值、阈值进行训练,构建优化的混合算法神经网络模型。在华北油田某管道的腐蚀情况分析中,证明了该方法的正确性和优越性。