鄂尔多斯盆地DJ区块煤层气储量大,但受地质及工程条件的影响,区块单井产气量存在明显的差异.基于区块煤层气地质及工程生产数据,首次将随机森林(random forests,RF)与交叉验证(cross validation,CV)相结合应用于煤层气领域,并利用RF-CV从地质和工程两方面筛选了影响日均产气的主控因素.为进一步验证该方法的准确性,利用随机森林回归算法预测煤层气井的产气情况,测试井的决定系数R2=0.85,预测效果较好并验证了RF-CV方法的准确性.基于熵值法(entro-py method,EM)的逼近理想解排序法(technique for order preference similarity to ideal solution,TOPSIS)对DJ区块127口井的产气效果进行综合评价,选出相对贴近度前12的井,利用统计分析法对优选的井与全部井进行了对比分析,确定了工程主控因素的最优区间,研究结果对煤层气井的压裂规模优化和有效开发提供了借鉴.

研究煤层气的可压性,优选出与煤层气可压性相关的参数,用物元可拓模型和主客观组合权重法,对可压性评价准则进行统计分析参数,物元构造和确定待评物元,应用到煤层气可压性评价实例中,结论也验证了采用方法的可靠性.

Successful hydraulic fracturing (HF) operation requires an accurate model for simulating the propagation of fractures. In this paper, a novel mechanism for the formation of T-shaped fractures and a three-dimensional propagation simulation for predicting their dimensional parameters prior to HF treatment are presented. Furthermore, a pressure decline analysis model (PDA) is also developed for calculating the dimensional parameters after HF treatment. Experimental results show that the two models can accurately simulate the T-shaped fractures and provide a valid data for the optimization and evaluation of HF treatments.

在油气田的开发过程中，由于新区块的投产、规模化的开发调整和三次采油方案的实施等因素影响，会出现多峰的产量变化模式，而传统的HCZ模型属于单峰预测模型，不能适用于多峰的产量变化的预测。本文在已建立的HCZ多峰预测模型基础上，对HCZ多峰预测模型进行无因次化处理，建立了HCZ多峰预测模型无因次关系式，通过预测模型典型曲线的拟合求解，进而对油田的产量进行预测。通过实例的应用表明，本文提出的HCZ模型典型曲线在多峰预测中的应用是实用有效的。

Analytical solution of composite reservoir model which considers wellbore storage and skin effect is complex in its form and curve shocks occur in the conventional numerical solving process. Therefore,it is necessary to make researches on accuracy of numerical solution. Firstly,the real space solution is obtained by using the stehfest numerical inversion to the analytical solution of Laplace space dimensionless pressure. After that,by analyzing the form of the solution,we determine the solving of the Bessel function of imaginary argument is the key to numerical solution of the model. Then we present sectional numerical(SN)algorithm of Bessel function of imaginary argument whose calculation accuracy of results can be comparable to the Matlab. The curves of the infinite,constant pressure and closed outer boundary conditions obtained by the above algorithm are smooth and complete. What′s more,the program runs fast enough to well satisfy the requirements of well test analysis and is suitable for composite reservoir well testing analysis.

压后压降测试分析是目前获取煤层渗透率的有效方法.通过对压裂停泵后压降测试数据进行分析,在Nolte经典模型基础上,结合裂缝闭合后压力分析的线性流和径向流数学模型,运用Horner法产量等效原则及Duhamel叠加原理,创建了储层渗透率的计算方法.并利用逐步最小二乘法,实现井底压力的自动拟合,提高了储层渗透率的求取准确性和效率.新疆煤层气井的实例计算说明了该方法的准确性和可靠性,并证实了XX区块渗透率较高、物性较好,具有较大的开发价值.

水力压裂是煤层气增产改造中最为重要的措施,压裂选井选层是其中的一个重要环节.选择增产潜力大、压后效果好的井层进行压裂,既可节约开发成本又可获得理想效果.文中首先根据煤层性质初选了影响压裂效果的参数,并用灰色关联度对这些参数进行优选,然后按煤层气井的压后效果,将生产层划分为高、中、低产3大类,并结合多总体的距离判别分析建立判别模型,将所优选出的参数作为距离判别分析的输入指标,代入样本建立距离判别准则.用该准则对样本进行回判,正确率达到92％,说明该新方法具有较好的适应性和较高的准确性.该方法在山西省某区块的煤层气井层实际应用效果较好.

储层的含油气性识别是储层综合评价的难点和关键,文中以粗糙集理论为基础,利用布尔逻辑和粗糙集理论相结合的离散化算法对每个条件属性进行离散化处理;利用基于遗传算法的粗糙集理论提取具有一定决策概率的不精确判别规则;利用规则的支持度、置信度和覆盖度挑选有效规则,进行储层类型的识别.实例应用结果表明,该方法提高了储层识别的正确率,提取的判别规则具有可解释性且较易理解,能够有效挖掘勘探数据中的潜在关键信息,对储层开发方案的制定具有重要的指导意义.

为了解决遗传BP神经网络在储层油气识别中存在的问题,采用改进的遗传算法优化了RBF网络的连接权值及结构,不仅解决了神经网络易陷入局部最优的问题,而且提高了网络的泛化性能.针对储层性质差别大会影响油气识别精度的问题,给出基于马氏距离的模糊聚类方法,对原样本空间按储层性质聚类得到了新的样本空间,并以常规测井和录井资料作为网络的输入参数进行了油气识别.通过样本的聚类处理,提高了遗传神经网络映射的精度.

This paper is concerned with delay-dependent stability for continuous systems with two additive time-varying delay components. By constructing a new class of Lyapunov functional and using a new convex polyhedron method, a new delay-dependent stability criterion is derived in terms of linear matrix inequalities. The obtained stability criterion is less conservative than some existing ones. Finally, numerical examples are given to illustrate the effectiveness of the proposed method.

Theoretical model of fracture closure is developed based on mass balance principle rock mechanics and fluid dynamics theory.This paper analyzes the wellhead pressure variation during fracture closure and gives a calculation method of both time fracture closure and roppant settlement-distance.Analysis and application show that both the time of the artificial fractures nature closing time and the distance of roppant settlement have grown which make a serious pollution on both formation and fracture,and ultimately affect the fracturing effect at low permeability layer condition,therefore,reasonable choice of fluid flowback time is of great significance.

The production mechanism of CBM has decided the necessity of dewatering and depressurization of the coalbed methane wells. Reasonable and refined production system which control the coalbed methane wells output is one of the key factors to ensure productivity. This paper analyzes the production mechanism in brief and well drainage curves are divided into three categories. By analyzing onsite data, we can contribute the reasons of the low yield and high yield of coalbed methane wells to four factors, moreover, we can get ideal production of coalbed methane wells through adjustment and optimization.

The carbonate reservoir has a sound heterogeneity and low-permeability, which brings certain difficulties to the prediction of the remoulding of the reservoir. The effective comprehensive evaluation method of reservoir plays a significant role in the modification of the layers of well. Aiming at the characteristics of the reservoir geology, the paper builds up reservoir evaluation criteria based on the principles of geological evaluation and construct the membership functions and models of analytic hierarchy process, secondly, by the means of fuzzy analytic hierarchy process, the paper obtains the proportion of well logging, well testing, geology and geophysical prospecting to the importance of reservoir modification. Last, by the means of fuzzy comprehensive evaluation method, it gains the conclusion about whether the reservoir needs to be modification or not. The rebuilding results of reservoir shows that the above method can effectively resolve the consistency problem among various appraisal parameters, improve the accuracy of evaluation and the efficiency of the modification and reduce invalid production. This method has been successfully applied to the evaluation of low-permeability reservoirs in Tarim oilfield.

水平井压裂后形成的多条裂缝与水平井井筒夹角不尽相同,增加了压后水平井产能计算的复杂性,而水平井的产量预测分析是制定压裂方案的关键技术之一.从裂缝与井筒存在不同夹角的情形出发,结合数学理论和复位势理论、势能叠加原理等渗流理论,应用考虑裂缝干扰的产能计算模型,提出了一个好的数值计算方法.通过大量的实例计算和对比,分析了裂缝平面与水平井井筒不同夹角时的产能.结果表明:裂缝与水平井筒夹角越大,产量越高；裂缝与水平井井筒夹角的不同组合对产能影响较大；对于不同的裂缝布局,不同位置的裂缝应该尽量错开排列.研究结果对现场水平井压裂方案优化设计具有指导意义.

Considering the temperature variation of wellbore,using Cullender-Smith method and wellbore fluid-carrying equation,integrating flash calculation of fluid phase equilibrium thermodynamics and applying state equation modeling,we can predict the temperature,pressure,gas-fluid components,gas-fluid molecular fraction and accumulation state.Moreover,the wellbore dynamics of condensate gas well with high gas-fluid ratio in different producing periods can be grasped to guide the production.

According to analysis methods for the production system of condensate gas wells,the limit diameter to bring the production capacity of gas wells into normal play can be recognized.Based on the analysis of inflow performance curve and tube water-carrying curves of a condensate gas well,the limit diameter for producing effusion in shaft is explored.In light of two points above,the reasonable tube diameter of a condensate gas well can be obtained.Thereby,methods for calculating reasonable tube diameters have been established which have not only considered fluid phase behavior variations of a condensate gas well,but also taken the supply ability of formation energy,shaft lifting ability as well as shaft liquid-carrying ability into account.The method can better guide the production of a condensate gas well and has a broad application prospect.In accordance with this method,this paper predicted the reasonable tube diameter of the well WD7 and the results showed that the reasonable WD7 well diameter should be 50.8 mm.

将分形几何理论与渗流理论相结合,推导不等厚横向非均质油藏不稳定渗流的试井分析数学模型.储集层模型由m(m≥2)个环绕中心的环形区域组成,在不同的环形区域内,储集层和流体的性质不同,其孔隙度和渗透率可以具有不同的分形分布;地层厚度也可以不同,但在同一区域内看作近似不变.考虑井筒储存和表皮效应影响,建立了此类分形复合油藏模型的不稳定渗流有效井径数学模型.针对典型的3类外边界条件,应用Laplace变换求得拉氏空间的解析解;制作了2类两区复合油藏模型的典型压力曲线,分析了压力动态特征和参数影响.分形指数对晚期的压力动态有很大影响:压力曲线开始时合并为一,随着时间的增加而发散;分形指数值愈大(表明分形网络越扭曲复杂,地层的连通性越差),直线愈陡,反之愈平缓.地层厚度的变化对典型曲线也有一定的影响,晚期压力随着厚度增大(第二区逐渐变厚)而逐步减小.不等厚横向非均质油藏、均质复合油藏及分形油藏均是此模型的特例,其建模方法可以推广到双重介质分形复合油藏.图1参6(向开理摘)

将非线性分形几何理论应用于渗流力学,建立分形油藏低速非达西渗流问题的组合模型.该模型由两个同心圆(圆中心为一口定产量的生产井)的分形油藏组成,内域为低速非达西渗流的分形油藏,外域为达西渗流的分形油藏.在内边界定流量和考虑井筒储存、表皮效应影响情况下,建立了分形油藏低速非达西渗流问题的有效井径组合数学模型,应用拉普拉斯变换方法求出了在两区域内压力分布的解析解,应用Stehfest 数值反演方法求得井底的无因次压力,并分析了井底压力动态特征和参数影响.

将非线性的分形理论应用于渗流力学,在考虑井筒续流、表皮效应和井筒相再分布的影响下,建立了分形油藏不稳定渗流有效井径数学模型.用拉普拉斯变换和反演方法求出了无限大地层模型的解析解及长、短时渐近解和有界地层模型的Laplace空间的解析解.分析了地层压力动态特征和参数的影响.

The distribution of the low-permeability gas reservoirs in China is very wide and their reserves amount to 80% of the reserves of the total gas reservoirs in China according to an incomplete statistics. Obviously, developing this kind of gas reservoirs is of an important practical significance, but up to now,the low-permeability data have been interpreted by adopting the conventional gas reservoir theory in general and the ideal results have not been obtained often due to the particularities the low-permeability gas reservoirs have but the conventional ones have not, such as tight rocks, extremely low permeability and gas percolation flow with slippage effect, etc. In view of this situation, a corresponding mathematical model is set up on the basis of investigating and studying a vast amount of the foreign and domestic literatures and the practical calculation and starting from studying the specific property of the gas percolation flow in lowpermeability reservoirs. Through special processing of the coefficients and interior and exterior boundary conditions in the equation, a non-linear difference equation which solution is acquired by Newton′s iteration method is obtained. The model and the calculation method in the paper can be fit in better with the measured data of the gas well through processing the data of one well in central Sichuan. As a result of this work, an effective method for solving this kind of problems is found out and this achievement is easy to be spread to the multidimension and multiphase model.