Figure 3 - uploaded by Daoyuan Zhai
Content may be subject to copyright.
Geographic map indicating the three closest injectors included in each of the five independent producercentric models.
Source publication
Liu, et al [5], [6] developed an Extended Kalman Filter (EKF) for a producer-centric reservoir modeled as a collection of continuous-time impulse responses that convert injection rates of all contributing injectors into a production rate. Their EKF model estimates two parameters for each producer-injector pair, which can be used to compute an Injec...
Context in source publication
Context 1
... oilfields, i.e., we selected the three geographically closest Figure 2. Average daily estimated IPR values for the five independent producer-centric models, assuming exact knowledge of the contributing injectors to the producers. Table I. Designed IPRs for the simulated oilfield in Fig. 1. injectors to a producer as its contributing injectors. In Fig. 3, the injectors having arrows pointing to a producer indicate that they are included in that producer's ...
Similar publications
The current paper presents a computer-aided analysis of the way in which a natural gas storage created in a depleted gas deposit and the associated gas collection network are operated, both during the extraction phase and during the injection phase. In a first step, an analysis is carried out on a two-dimensional, nonstationary model for the gas st...
Citations
... These models are often referred to as high fidelity or physics-based models. In addition to rigorous physics-based models, empirical or reduced order models are also used [35][36][37][38][39][40][41][42]. The form of some empirical models are determined solely from data while others are structured sets of equations with unknown or uncertain parameters. ...
... For fields under waterflood or EOR, where reliable geologic data is difficult to obtain, reduced order proxy modeling presents a method for determining the relationship between controllable reservoir system inputs and oil production. Various data driven models have been applied to the history matching problem including finite impulse response [79], autoregressive models, [36,37,42], and the capacitance resistance model [38][39][40][41]. These models have been coupled with various history matching algorithms including gradient based methods [38] and EKF [80]. ...
This paper presents a review of history matching and oil field development optimization techniques with a focus on optimization algorithms. History matching algorithms are reviewed as a precursor to production optimization algorithms. Techniques for history matching and production optimization are reviewed including global and local methods. Well placement, well control, and combined well placement-control optimization using both secondary and tertiary oil production techniques are considered. Secondary and tertiary recovery techniques are commonly referred to as waterflooding and enhanced oil recovery (EOR), respectively. Benchmark models for comparison of methods are summarized while other applications of methods are discussed throughout. No single optimization method is found to be universally superior. Key areas of future work are combining optimization methods and integrating multiple optimization processes. Current challenges and future research opportunities for improved model validation and large scale optimization algorithms are also discussed.
... Estimating the Injector-Producer-Relationship (IPR) values between a producer and its contributing injectors using only injection and production rates is more practical and economical than reservoir simulation [4], [5], [6] and tracer testing [1], [3], [12], [14] because well rates are the most regularly available data in an oilfield. Many researchers have made contributions in this area: Albertoni and Lake [2] model production rate as a weighted sum of the delayed injection rates; Lee et al. [8] characterize the reservoir between multiple producers and injectors as a standard linear FIR-MIMO system; Liu et al. [9], [10] and Zhai et al. [20], [21] introduced strategies based on Extended Kalman Filter (EKF) and Inequality-Constrained EKF; Yousef et al. [18] developed a Multiple Linear Regression method based on a Capacitance Model; and Sayarpour et al. [15] improved this method by introducing a Capacitance-Resistance Model. The performance of these well-rate-based approaches are all dependent on the accuracy and frequency of the actual measured injection and production rates. ...
... Such low frequency data can have serious negative impacts on the continual estimation of the IPRs. In this paper, we focus on the EKF approach that was studied in [9], [10], [20], [21]. We shows in Section V that, by directly using the low frequency data, it takes a long time for the EKF to reach correct IPR results, which makes it impossible to detect any dynamic changes of the system in a timely manner. ...
... In previous works that are related to this paper [9], [10], [20], [21], a two-parameter Auto-Regressive model [also referred to as the Liu-Mendel Model (LMM)] has been used to characterize the inter-well connectivity between the producer and and its j th injector, i.e.: ...
Estimating inter-well connectivities based on injection and production rates has become increasingly popular, because it is an economical way to detect preferential subsurface flow trends. The techniques that adopt this idea all rely on the quality (accuracy and frequency) of the actual measured rates; however, in many oilfields, the production rates are measured infrequently, e.g. every eight days on average, which is significantly lower than the frequency required to produce reliable outcomes.
This paper proposes an Iterative Extended Kalman Filter and Smoother to dynamically interpolate missing data between two available measurements. It shows, by using reservoir simulator data, and assuming one measurement every eight days, that the interpolated production rates lead to almost equivalent inter-well connectivity estimates as those obtained when production rates are measured daily.
It appears that signal processing can be used to replace costly measurements. Such "virtual" measurements can be utilized by different well-rate-based techniques to improve their performances.
