Uncertainty evaluation of wellbore stability model predictions
Uncertainty evaluation of wellbore stability model predictions
复制标题
DOI：
10.1016/j.petrol.2014.09.033
复制
发表时间：
2013-10
影响因子：
--
通讯作者：
John Emeka Udegbunam;B. Aadnøy;K. Fjelde
中科院分区：
文献类型：
--
作者：
John Emeka Udegbunam;B. Aadnøy;K. Fjelde
石油和天然气行业对井壁稳定性问题的担忧日益加剧。随着作业人员向更具挑战性和严酷的环境发展，如超深水和高压高温(HPHT)领域，改善井眼作业的需求变得更加迫切。不同的不一致会影响许多以前的井筒稳定性分析，导致不正确的结果，或者是井身规划者无法扩展到其他井型的结果。典型的井筒破裂和坍塌模型提供了对地质压力的单点估计。模型输入数据可能是不确定的。这项工作的目的是用随机方法研究典型的断裂和坍塌模型相对于输入数据的不准确性。将评估输入数据中的不确定性，包括地应力、岩石强度数据和孔隙压力，以显示这些因素如何导致模型预测中的累积不确定性。在这种方法中，输入参数被分配了适当的概率分布。然后将这些分布应用到井筒稳定性模型中。通过蒙特卡罗模拟，传播不确定性并生成输出的直方图。应用两种类型的分布--三角形分布和均匀分布--来考察假设的输入参数分布类型对模型预测的影响，并进行了敏感性分析。这是为了确定最重要的输入因素，这些因素在很大程度上导致了临界压裂和坍塌压力的累积不确定性或变异性。所提出的方法可以帮助减少许多钻井问题，如井漏、卡钻和井塌。因此，该行业可能会节省大量非生产性时间。此外，油井规划者将拥有更好的信息来做出关键决策。
There is an increasing concern in the oil and gas industry regarding wellbore stability problems. The need to improve well operations becomes more imperative as the operators move towards more challenging and harsher environments such as ultra-deep waters and high-pressure and high-temperature (HPHT) fields.Different inconsistencies affect many previous wellbore stability analyses, resulting in incorrect results, or results that cannot be extended to other well configurations by well planners. Typical wellbore fracture and collapse models provide single point estimates of the geopressures. The model input data may be uncertain. Failure to capture these uncertainties has led to poor predictions.The purpose of this work is to investigate typical fracture and collapse models with respect to in accuracies in the input data with a stochastic method. Uncertainties in the input data, which include in-situ stresses, rock strength data, and pore pressure will be evaluated, to show how these contribute to the cumulative uncertainties in the model predictions.In this approach, the input parameters are assigned appropriate probability distributions. The distributions are then applied in the wellbore stability models. By means of Monte Carlo simulations, the uncertainties are propagated and the histograms of the outputs are generated. Two types of distributions – triangular and uniform – are applied, to see how the types of input-parameter distributions that are assumed influence the model predictions.Sensitivity analysis is also conducted. This is to ascertain the most significant input factors, which are largely responsible for the cumulative uncertainties or variabilities in the critical fracturing and collapse pressures.The proposed methodology can help in reducing many drilling problems such as circulation loss, stuck pipe, and well collapse. As a result, the industry may save much non-productive time. In addition, well planners will have improved information to make critical decisions.