在当前的油气开采作业中，水平井的重要地位日益凸显，那么应该如何提高其利用效率呢？

编译 | 惊蛰 TOM?

井筒水平段长度和均匀性对于生产的优化非常重要，直接决定整个生产周期内的持液率和生产率。然而，受限于当前的井筒设计水平以及高时间成本，井筒结构可能对后期开采产生负面影响，限制油藏的整体采收率提高潜力。以Midland区块为例，通过研究横向长度和轨迹对井产量的影响，可以确定最佳的水平段长度和轨迹设计。

操作人员通常使用一个通用标准来确定在某口井水平段长度对于后期产量的贡献。一般来讲，这种方法会评价水平段长度与产量增加的关系、水平段长度增加与产量提高的比例关系（即是否均匀），以及各种井眼偏差和狗腿度等是否能够转化为具有相同标量的数据，直接应用于估算生产状况。

为此，一个研究团队调查了水平段长度最终采收率(EUR)的影响。EUR计算结果是根据15个开发商提供的的日产量数据计算得出的，然后根据水平段的顶部到底部射孔段长度进行归一化，利用改进的双曲线方法进行衰退分析。结果表明，简单地使用某一个标量并不能衡量EUR与水平段长度之间的关系。

在具有多级水力压裂的非常规油藏井筒中，水平井段产量随着时间延长生产率逐渐下降，这主要是由于这些人造裂缝的导流能力随着生产和时间而抓紧衰减。因此，为了充分了解井眼轨迹和横向长度对生产的影响，该团队开发了一个特殊的工作流程，以深入分析井筒水力学的瞬态效应和非常规油藏性能随时间的变化。

为了了解随着开采深度储层开采性能的变化，研究人员根据之前的地质和储层数据开发了数值模型。在此次研究中，主要针对该区块的Spraberry和Wolfcamp B两口井，数值模型首先通过历史压力和生产数据进行校准。研究团队使用了几个生产指标来比较在井筒开采周期的不同时间内，水平段的不同位置对于产量的贡献，测量数据包括生产压差、水平段压降和液体滞留结果。

分析结果

研究团队首先通过假设所有射孔簇对称排布，为每口井建立单簇双孔隙度模型，然后用相应的生产历史校准。对于裂缝半长的分析，首先通过速率瞬态分析（RTA），然后通过历史匹配进一步调整。沿着井眼的网格厚度按照对数方式改变，以优化井眼及压裂裂缝周围的高流体流速。地层孔隙率通过岩石物理分析估算得到，基质和裂缝的渗透率最初通过RTA和岩石物理分析确定，然后通过历史匹配压力和生产数据进行调整。最后通过使用历史生产数据（即井底流动压力，产油率，产水率和气/油比）校准模型。

井筒流动状况模拟

井眼流动模拟的目的是通过结合任何给定时间的储层生产率，来研究整个水平段的持液率和压降水平。套管和油管参数与实际候选井的数据一致，但合成案例的轨迹除外。横向长度的管状粗糙度与水平段的新生产套管一致，研究团队可以进行摩擦压力损失分析。在每次模拟中一致的是，模拟器中的计算都以垂直段800英尺为标准进行建模。

在油藏数值模拟的输出结果，研究小组将预计的油藏压力，油气比和含水率直接导入井筒流动模拟模型。为了准确地研究水平段横向压降，首先用预测产量和压降压力曲线计算地层产能，之后通过候选井的完整横向长度标准化地层生产率。在接下来的一步中，研究团队针对同一地层中所有合成案例井的原始横向长度，使用归一化生产力指数（PI）计算。其中并没有考虑可能存在的长水平段完井复杂性影响，而是假设特定时间步长具有相同的PI。 因此，在研究中，水平段长度不低于15000英尺的井筒具有相同的归一化PI值。

相对于井筒顶部状况，模拟器需要设定压力边界条件。在这些简略的案例中，“顶部”一般指的是垂直段的顶部。而顶部压力的获得一般根据井底流动压力，通过储层模拟器计算输出得到。在此基础上，研究人员减去静水压力和显着的摩擦压力损失，获得垂直截面顶部压力的准确估计。

为了研究横向长度和轨迹对油气生产的影响，研究团队通过整合油藏性能模型和水平井筒流动模型开发了全新的工作流程。根据研究人员获得的数据，从良好生产的角度来看，模拟得到的上斜井眼轨迹似乎能够始终如一地达到最佳的油液产量，而波动的水平井筒轨迹在两组流体压力/体积/温度的作用下，油气生产状况最差。

与之前研究的结果相似，这项研究表明，横向长度并不是能够衡量井筒生产性能的决定性因素。然而，在有限的案例研究结果基础上，通过假设均质储层性质和完井均匀性，确定了上述比例因子对于EUR估算的有效性。目前来看，我们还需要进行更多的案例，以涵盖广泛的储层流体PVT条件和其他储层性质。

假设对于相同横向长度所有轨迹的钻井和完井时间一致，完美的上斜水平井将产生更高的EUR。然而，为了确定最佳横向长度和轨迹，不仅要考虑基于数值模型的EUR，还要考虑地层地质、钻井和完井时间以及相应的成本、增加水平段长度带来的完井复杂性、地面作业限制和其他经济因素。

总之，水平段的轨迹可以决定其对井筒产量的贡献，在轨迹上斜的情况下，井筒的大部分产量贡献可能来自水平段的侧向射孔段；而在水平段轨迹下斜的情况下，井筒的大部分贡献可能来自位置较高的一端。而通过建立完全均匀的水平段轨迹，可实现在开采早期限制水平段持液量，优化整个水平段侧面的井眼贡献。

Wellbore lateral lengths and uniformity are important for production optimization, affecting liquid holdup and productivity throughout well production life. However, wellbore construction designed for holding leases and completed under time constraints could negatively affect production and limit operators’ potential for their acreage. This paper demonstrates a work flow to determine optimal lateral lengths and trajectories in the Midland Basin by studying the effect of the lateral length and trajectory on well production.

?Introduction

Operators often use a scalar to determine wellbore contributions for wells with an atypical lateral length on their leases. The work flow presented here provides realizations to justify use of this scalar, evaluating whether lateral lengths proportionally increase production, whether wellbore contributions are uniform throughout the lateral length of the wells, and whether various wellbore deviations and tortuosities should have the same scalar applied for estimating well production.

A team investigated the effect of lateral length on well estimated ultimate recovery (EUR). The EURs were calculated from daily production data from 15 operators and then normalized by well lateral length from the top perforation to the bottom perforation. The modified hyperbolic method was used to perform the decline-curve analysis. Simply applying a scalar for the lateral length among the wells would not yield proportionate EURs. When drilled and completed in unconventional reservoirs with multiple-stage hydraulic fracturing, horizontal wells experience a decline in productivity with time because the conductivity of those man-made fractures deteriorates with production and time. Therefore, in order to understand the effect of wellbore trajectory and lateral length on production fully, the team developed a work flow to incorporate an in-depth analysis on transient effects of wellbore hydraulics combined with unconventional reservoir performance over time.

Methodology

To develop an understanding of reservoir performance over time, numerical models are developed on the basis of previous knowledge, geology, and reservoir data. With respect to the Lower Spraberry and Wolfcamp B wells considered in this study, case studies for which are presented in the complete paper, the numerical models were first calibrated with historical pressure and production data. Fig. 1 displays the work flow that the team developed to combine long-term reservoir performance with wellbore hydraulics.

The team used several production metrics to compare wellbore contributions throughout the lateral affected by the trajectory with respect to different times throughout the well life, including drawdown pressure, pressure decline throughout the lateral, and liquid-holdup results.

Analysis of Well Performance

The team first built a single-cluster dual-porosity model for each well by assuming that all clusters drain symmetrically and then calibrated the models with corresponding production history. The fracture half-length was first estimated from rate transient analysis (RTA) and then further tuned through history matching. The grid thicknesses along the wellbore are logarithmically changed to optimize for high fluid velocities around the wellbore and hydraulic fractures. The porosity of the matrix was estimated from petrophysical analysis. Permeabilities of the matrix and fracture were determined through RTA and petrophysical analysis initially and later tuned through history matching pressure and production data. The models were calibrated by use of historical production data (i.e., bottomhole flowing pressure, oil-production rate, water-production rate, and gas/oil ratio).?

Wellbore-Flow Simulator

The objective of the wellbore-flow simulator is to study the level of drawdown throughout the lateral and liquid holdup by combining reservoir productivity for any given time. Casing and tubing parameters were consistent with the real candidate setups, with the exception of synthetic cases’ trajectories. The tubular rugosity for the lateral length is consistent with the standard rugosity for new production casing along the lateral, allowing the team to investigate friction pressure losses. For each case, wellbore trajectory with an 800-ft vertical section from the heel was modeled in the flow simulator.

On the basis of reservoir numerical-simulation outputs, the team imported the forecast reservoir pressure, gas/oil ratio, and water cut directly into the wellbore-flow-simulation models. To investigate the pressure drop along the lateral accurately, the formation productivity first was calculated with the forecast production and predetermined drawdown-pressure profile. Then, the formation productivity was normalized by the completed lateral length for the candidate wells. The team then used the normalized productivity index (PI) with respect to the original lateral length of the well for all of the synthetic cases in the same formation. The team did not consider the effect of possible completion complexity for longer lateral lengths, instead assuming the same PI for a given timestep. Therefore, the normalized PI value would be the same for cases that have 15,000-ft or greater lateral lengths in the same formation discussed in this paper.

The simulator requires pressure-boundary conditions with respect to the top of the wellbore. In these truncated cases, the “top” would be the top of the vertical section. These pressures were determined by using the flowing bottomhole pressures from the reservoir-simulator outputs as a reference. Then, the group subtracted the hydrostatic and significant frictional pressure losses to obtain an accurate estimate of the pressures at the top of the vertical section.

Conclusions

To investigate the effect of lateral length and trajectory on oil and gas production, the team developed a work flow by integrating a reservoir-performance model and a horizontal-wellbore-flow model.

From a well-production perspective, the toe-up wellbore trajectory appears to result in the best oil and liquid production consistently, while the undulated wellbore trajectory results in the worst oil and liquid production given the two sets of fluid pressure/volume/temperature (PVT) data and reservoir properties that the team studied.

Similar to the findings from previous studies, this study revealed that scaling well performance is not conclusive through lateral length. However, on the basis of the limited case-study results, the team determined the effectiveness of the aforementioned scaling factors to estimate EUR by assuming homogenous reservoir properties and homogenous completion. More case studies are needed to cover wide ranges of reservoir--fluid PVT conditions and other reservoir properties

Assuming consistent drilling and completion times for all trajectories for the same lateral length, perfect toe-up trajectories will yield higher EURs. However, to determine the optimal lateral length and trajectory, one should consider not only the numerical-model-based EURs but also formation geology, drilling and completion time and corresponding cost, increased completion complexity for longer laterals, surface footprint, and other economic factors.

Trajectories could dictate wellbore contributions and drawdown per lateral foot. With toe-up-trajectory cases, the majority of the wellbore contribution could come from the toe-side perforations; with the toe-down-well cases, the majority of the wellbore contribution could come from heel-side perforations. A perfectly uniform trajectory could optimize wellbore contribution throughout the lateral by limiting liquid holdup at earlier times, with an exception at the heel.