基于不平衡数据集的客户流失预测研究*

目前客户流失预测面临的主要问题之一就是类不平衡性(class imbalance)。针对这个问题,首先应用欠抽样法(undersampling)处理客户流失数据降低不平衡性,再应用C4.5D、C4.5N、RIPPER、NaiveBayes和RandomForest机器学习方法对客户流失进行预测。实验结果表明,欠抽样法是在牺牲负类样本预测精度的前提下,提高正类预测精度,于是采用重复抽样法(resampling)来弥补欠抽样法的缺陷,减少负类样本中含有大量有用信息的丢失,实验结果证明了这种方法的正确性和有效性     

基于SVR的组合预测模型及其应用

针对油田开发单一产量预测模型泛化能力低、中长期预测准确性差等问题,提出了一种基于支持向量回归机(svR)的组合预测模型.该模型可基于小样本建模并能综合不同单一预测模型的适用条件和优势,具有较强的泛化能力,对只可获得少量实验数据的油田产量预测问题具有较好的适应性.给出了SVR组合预测模型的结构设计和实现算法,对油田实际产量数据进行处理,取得了较精确的预测结果,验证了模型和方法的有效性.     

PCA-SVM在电力负荷预测中的应用研究

研究电力负荷预测问题,传统方法无法消除数据之间冗余及复杂特征,导致预测精度较低.为了提高电力负荷预测精度,提出一种基于主成份分析(PCA)的支持向量机(SVM)电力负荷预测方法(PCA-SVM).首先利用主成分分析对电力负荷的影响因素进行处理,消除各因素之间的高度冗余性,通过提取样本集的主成分完成数据预处理,有效地压缩样本集的维数,加快SVM学习速度并提高预测精度,然后利用支持向量机,对保留的主成成分进行建模预测.最后利用PCA-SVM模型对华东地区1978～1998的电力负荷进行了验证性测试和分析.实验结果表明,相对于各参比模型,PCA-SVM模型可以有效地降低样本集的维数,提高负荷预测精度.PCA-SVM是一种高效、高精度的电力负荷预测方法.     

交流变频电加热技术在高凝油生产中的应用

提出了一种在石油开采中应用的采油油管变频加热技术,它利用油井中的采油工作油管作为热源体,将电能转化成热能直接对油井油管内的液体进行加热,达到降凝化蜡目的.该项技术所需电能由三相配电变压器输出,经AC-DC-AC逆变电路输出,将电能直接输送到油井内的工作油管,经油管下部的油套管连接器连接到套管,形成一个完整的回路.该项技术有效地解决了电网扰动和油井采油工作套管严重结蜡的问题.     

智能完井技术测试与控制系统设计

智能完井技术是数字化智能化油藏开采的一项新技术,测控系统是其核心.对智能完井测控系统的硬件系统和软件系统进行了设计,通过井下温度和压力传感器采集温度和压力数据,利用温度、压力与流量的关系对井下每个层位的流量进行计算,开发了一套智能优化开采系统,对井下生产数据进行实时监测和分析,利用油藏优化开采理论,确定优化开采方案,自动控制井下各油层流量控制阀套,实现油井的优化开采.现场试验表明,该套智能完井测控系统可以实现油井的智能优化开采.     

工业无线WIA-PA技术在智能油田中的应用

油田开采过程数据是数字油田的基础与核心,基于无线通信的技术是实现数据采集和传输的手段,其安全性直接决定了未来石油生产的安全.因此,基于我国自主研发的WIA-PA技术来构建油井数据采集和传输的通道将极大提高数字油田的可控性,避免国外技术某些隐藏的"后门"造成关键数据外泄.WIAPA技术是一项面向大规模传感器信息采集和传输的智能组网和高可靠、高安全无线通信技术,填补了我国在物联网领域自主技术的空白.     

基于自组织模糊规则归纳的电子商务客户流失预测

为提高客户流失预测的精度,构建了基于自组织模糊规则归纳算法FRI(Fuzzy Rule Induction)的电子商务客户流失预测模型.该模型利用数据分组处理技术GMDH(Group method data handling) 从训练样本中自动地提取接近于人类自然语言描述的电子商务客户流失模糊规则,进而对测试样本客户流失状态进行判别.采用某网上商场的1500名客户样本进行电子商务客户流失预测实证研究,结果表明,该方法对测试样本预测精度达到了90%以上,是一种有效和实用的电子商务客户流失预测工具.     

潜油电泵智能选型系统的设计与实现

潜油电泵的智能选型对浅海油田的石油开采具有现实意义。在智能选型系统中建立相关的计算模型并完成电机表面流速、挂泵深度、井底流压等多项参数的计算。利用计算机进行数值计算、绘图及数据存储,将石油勘探、钻井、完井、采油过程中产生的数据有机地结合在一起,构建利用计算机平台处理的潜油电泵机组智能选型设计模型。从而提高工作效率、节省大量的资金,使浅海油田的石油开采从传统方式向智能化迈进一步。     

基于特征融合的抽油机井检泵周期预测

检泵周期是反映抽油机井工作情况的重要指标,准确预测检泵周期对提高油井产能和经济效益具有重要意义。针对油田检泵周期预测准确率低等问题,提出一种基于特征融合抽油机井检泵周期预测方法。该方法引入SVR提取油田数据的静态特征,利用卷积神经网络学习油田数据的动态特征,引入多模态压缩双线性池化对静态特征和动态特征进行融合,利用判别模型训练融合特征实现检泵周期的准确预测。实验结果验证了该模型的有效性和可行性。     

交流变频电加热技术在高凝油生产中的应用

提出了一种在石油开采中应用的采油油管变频加热技术,它利用油井中的采油工作油管作为热源体,将电能转化成热能直接对油井油管内的液体进行加热,达到降凝化蜡目的。该项技术所需电能由三相配电变压器输出,经AD-DC-AC逆变电路输出,将电能直接输送到油井内的工作油管,经油管下部的油套管连接器连接到套管,形成一个完整的回路。该项技术有效地解决了电网扰动和油井采油工作套管严重结蜡的问题。     

A semi-Markov model with Bayesian belief network based human error probability for availability assessment of downhole optical monitoring systems

Compelled by increasing oil prices, a research effort is underway for designing and implementing intelligent oil fields in Brazil, with a first pilot directed towards mature wells in the Northeast. One of the major benefits of this technology is the anticipation of oil production volumes and an improved reservoir management and control. Given the considerable steep investment on the new technology, availability is a key attribute: higher availability means higher production volumes. An important part of this effort is the development of pressure–temperature optical monitoring systems (OMS) and their availability assessment. Availability analysis of an OMS impose some complexities, where the most relevant aspects are: (i) the system is under a deteriorating process; (ii) the available time to complete the maintenance; and (iii) human error probability (HEP) during maintenance that is influenced by the available time and other factors (e.g., experience, fatigue) in returning an OMS to its normal operational condition. In this paper we present a first attempt to solve this problem. It is developed an availability assessment model in which the system dynamics is described via a continuous-time semi-Markovian process specified in terms of probabilities. This model is integrated with a Bayesian belief network characterizing the cause-effect relationships among factors influencing the repairman error probability during maintenance. The model is applied to a real case concerning mature oil wells.     

CBiA-PSL抽油井异常工况预警模型

油田生产过程中,油井受各种因素的影响容易发生泵漏、管漏等异常工况,会降低油井产出甚至导致躺井,对异常工况预警是油田智能化管理的重要任务.基于CNN-BiGRU联合网络,提出一种改进的网络结构CBiA-PSL模型(CNN BiGRU attention-positive sharing loss),用于油井异常工况预警....     

Evolving artificial neural network and imperialist competitive algorithm for prediction oil flow rate of the reservoir

Multiphase flow meters (MPFMs) are utilized to provide quick and accurate well test data in numerous numbers of oil production applications like those in remote or unmanned locations topside exploitations that minimize platform space and subsea applications. Flow rates of phases (oil, gas and water) are most important parameter which is detected by MPFMs. Conventional MPFM data collecting is done in long periods; because of radioactive sources usage as detector and unmanned location due to wells far distance. In this paper, based on a real case of MPFM, a new method for oil rate prediction of wells base on Fuzzy logic, Artificial Neural Networks (ANN) and Imperialist Competitive Algorithm is presented. Temperatures and pressures of lines have been set as input variable of network and oil flow rate as output. In this case a 1600 data set of 50 wells in one of the northern Persian Gulf oil fields of Iran were used to build a database. ICA-ANN can be used as a reliable alternative way without personal and environmental problems. The performance of the ICA-ANN model has also been compared with ANN model and Fuzzy model. The results prove the effectiveness, robustness and compatibility of the ICA-ANN model.     

Comparison of the ParTrack mud/cuttings release model with field data

The oil industry promotes the development of numerical models for prediction of impacts from their discharges to sea. A model for the simulation of the spreading and deposition of drilling mud and cuttings on the sea floor as well as the spreading of chemicals (and small-sized particles) in the water column has been developed. The simulation is based on a Lagrangian ‘particle’ approach, which means that the properties of the discharge are represented by moving ‘particles’ in the model domain. The initialization of the particles is based on the output from an Eulerian near field underwater plume model. In addition, the model applies external current fields for the horizontal advection of the particles.This paper presents a comparison between simulated and measured concentrations of barium (barite) in surface contaminated sediment in the vicinity of an oil production field. As a part of the regular surveillance of oil production sites on the Norwegian Continental Shelf, the barium content in surface sediments is measured. These data might therefore serve as an opportunity for comparing simulation results with measured depositions of barium (barite) on the sea floor.The paper explains details in the comparison made between the measured barium concentrations in the sediment and the simulated deposition on the sea floor from the drilling of three exploration wells and 18 production wells off the west coast of Norway.     

基于Spark的医疗设备运维信息挖掘分析研究

在大数据时代,医疗设备复杂的运行状态环境中,实现准确的医疗设备运维预测,是实现智慧医疗的必要前提,为保证医疗设备的正常运行,此系统将使用Scala语言在Spark平台进行并行化实现,采用K-means聚类算法计算预测模型,提高算法处理大数据的能力。以数据采集、数据存储、数据分析为主,以Tomcat作为Web服务器框架进行系统可视化展现,打造一个大数据一体化系统平台,实现医疗设备运维信息的预测。     

油水井自动化系统在冀东油田的应用

中国石油油气生产物联网系统（油水井自动化系统）即通过传感、射频、通讯等技术,对油水井、计量间、油气站库等生产对象进行全面的感知,实现生产数据、设备状态信息在生产指挥中心及生产控制中心集中管理和控制,搭建规范、统一的数据管理平台,支持油气生产过程管理,进一步提高油气田生产决策的及时性和准确性的系统。文章介绍了在冀东油田建立的一套覆盖油气生产、处理等全过程的油水井自动化系统,实现了生产数据自动采集、关键过程联锁控制、工艺流程可视化展示、生产过程实时监测的综合信启,平台,达到强化安全管理、突出过程监控、优化管理模式,以实现优化组织结构、提高效益的日标。     

基于改进灰狼算法优化多核支持向量回归机及其应用

为更好发现数据中的复杂规律,避免核函数选择的盲目性和局部最优等非线性优化问题,本文提出一种基于改进灰狼算法优化多核支持向量回归机算法.首先,基于全局核函数和局部核函数构建多核支持向量机采油速度预测模型;其次,利用基于云模型和二次插值算法改进灰狼优化算法对核函数权值和参数的选取进行优化;最后,应用灰色关联分析理论确定采油速度影响因素集,并作为多核支持向量回归机预测模型的输入.与6种采油速度预测方法进行对比,所提方法具有较好的全局寻优能力和较高的预测率的优点.     

NSGA-II-trained neural network approach to the estimation of prediction intervals of scale deposition rate in oil & gas equipment

Scale deposition can damage equipment in the oil & gas production industry. Hence, the reliable and accurate prediction of the scale deposition rate is critical for production availability. In this study, we consider the problem of predicting the scale deposition rate, providing an indication of the associated prediction uncertainty. We tackle the problem using an empirical modeling approach, based on experimental data. Specifically, we implement a multi-objective genetic algorithm (namely, non-dominated sorting genetic algorithm–II (NSGA-II)) to train a neural network (NN) (i.e. to find its parameters, that is its weights and biases) to provide the prediction intervals (PIs) of the scale deposition rate. The PIs are optimized both in terms of accuracy (coverage probability) and dimension (width). We perform k-fold cross-validation to guide the choice of the NN structure (i.e. the number of hidden neurons). We use hypervolume indicator metric to evaluate the Pareto fronts in the validation step. A case study is considered, with regards to a set of experimental observations: the NSGA-II-trained neural network is shown capable of providing PIs with both high coverage and small width.     

工业人工智能的关键技术及其在预测性维护中的应用现状

随着人工智能技术的快速发展及其在工业系统中卓有成效的应用, 工业智能化成为当前工业生产转型的一个重要趋势. 论文提炼了工业人工智能(Industrial artificial intelligence, IAI)的建模、诊断、预测、优化、决策以及智能芯片等共性关键技术, 总结了生产过程监控与产品质量检测等4个主要应用场景. 同时, 论文选择预测性维护作为工业人工智能的典型应用场景, 以工业设备的闭环智能维护形式, 分别从模型方法、数据方法以及融合方法出发, 系统的总结和分析了设备的寿命预测技术和维护决策理论, 展示了人工智能技术在促进工业生产安全、降本、增效、提质等方面的重要作用. 最后, 探讨了工业人工智能研究所面临的问题以及未来的研究方向.