Competition between Decline Curve and Simulation: Practical Method to Forecast Unconventional Wells

On January 2, 2019, the Wall Street Journal (WSJ) released an article titled "Fracking’s Secret Problem—Oil Wells Aren’t Producing as Much as Forecast." The piece delves into a revelation uncovered by the authors who collaborated with a third-party energy consulting firm. Together, they re-evaluated forecasts for "16,000 wells operated by 29 of the biggest producers in oil basins in Texas and North Dakota." These updated projections were then compared to the original corporate forecasts, initially utilized to justify drilling campaigns and attract investors.

According to the WSJ authors, their investigation revealed that "two-thirds of projections made by the fracking companies between 2014 and 2017 in America’s four hottest drilling regions appear to have been overly optimistic."

In the figure below, sets of wells from the Bakken region are depicted, highlighting how the estimation of Arps decline curves can significantly overestimate predictions. Such inaccuracies can prove highly detrimental for investors and the company's expansion strategy. Despite doubts regarding the effectiveness of simulations by many operators, the decline curve remains the primary method for quantifying predictions and conducting financial evaluations

In this article, I aim to introduce a new module within Eushaw Dynamics, designed to efficiently replace conventional decline curve methods. This simulation module allows for the rapid modeling of individual wells, with the boundary pressure exposed to the well declining according to models typically employed in decline curve analysis. It is this reduction in boundary pressure that primarily accounts for the observed decline in unconventional wells.

The pressure reduction indeed leads to the expected rate reduction. Typically, in this analysis, the casing pressure is utilized as a Bottom Hole Pressure (BHP), although this assumption is not entirely accurate, it remains a common practice. While initial fracture performance, necessitating geomechanical effects consideration and fracture reservoir interactions, may be neglected, other aspects of the history are matched accurately.

In the following figure, we compare the simulation results with those obtained from the decline curve fitted exponentially. Of particular interest is the prediction of Gas-Oil Ratio (GOR) in the simulation, a parameter inaccessible in traditional decline curve estimation methods. It's worth noting that the pressure reduction and subsequent GOR rise are the primary drivers of rate decline. What sets this simulation apart is its ability to provide GOR insights in less than 5 seconds, making history matching straightforward and effective.

The simulated GOR closely tracks the cumulative GOR (cGOR), calculated based on cumulative gas production over cumulative oil production. However, the instantaneous GOR exhibits erratic behavior and cannot be accurately fitted with the simulation.

There are two key distinctions in this modeling approach. Firstly, the Estimated Ultimate Recovery (EUR), which is highly dependent on GOR, can be accurately modeled using this simulation. Secondly, the implementation of artificial lifts is another significant factor. Unlike decline curves, artificial lift implementations cannot be predicted accurately because the Bottom Hole Pressure (BHP) fluctuates, resulting in varied performance outcomes in the well.

As evident in the following figure, the harmonic method tends to overestimate, while the simulation accurately captures the decline attributed to pressure reduction and closely tracks the well's performance.

Indeed, one of the key advantages of this module is its ability to facilitate a comparison between the actual simulation results and the predictions obtained through decline curve methods. By juxtaposing these two approaches, they can complement each other, enhancing the reliability of the final prediction. This integrated analysis allows for a more comprehensive understanding of the reservoir behavior and improves the overall confidence in the forecasted outcomes.

Using this module, running sensitivity analyses on Pressure-Volume-Temperature (PVT) properties and relative permeability characteristics can significantly enhance the estimation of P10 (10th percentile) and P90 (90th percentile) outcomes. By varying these parameters within a range of plausible values, we can gain insights into the potential range of outcomes and better quantify uncertainty in the predictions. This approach allows for a more robust assessment of the reservoir's behavior and provides a clearer understanding of the associated risks and opportunities.

If you’d like to learn more about how Eushaw Dynamics Simulator reach out at?[email protected]?or give us a call at +1(403) 667-7293