Advanced Production Data Analysis (PDA) Techniques Using Stochastic Single-Well History Matching in Coalbed Methane (CBM) Wells

Abstract

This paper presents a case study in the application of an advanced production data analysis (PDA) technique for coalbed methane (CBM) through the use of stochastic single-well history matching (SWHM) as a method that enables the exhaustive identification of the range of reservoir parameters underpinning the historical production, with flexibility to handle multilayered completions, and handle the transient effects present in CBM production responses. Through the application of this method, we present how to perform the quantification of uncertainties in an efficient and timely manner. It is envisioned that this technique be used as part of a wide range of methods in an effort to effectively understand complex and variable well performance seen in CBM wells. SWHM provides many advantages to traditional PDA (straight line methods) by the introduction of simplified physics coupled with material balance or radial numerical modelling, whilst retaining a significant speed advantage over traditional 3D fullfield numerical simulation. The PDA SWHM technique is applied to strengthen existing reservoir characterisation workflows, guide appraisal and data acquisition planning, and accelerate traditional static and dynamic modelling workflows. This paper presents a case study based on the single-well history matching of over 20 CBM pilot wells located in the Surat Basin. The findings presented within this paper include: connected areas and volumes drained by the wells, reservoir quality variability from well to well, and layer-wise, uncertainty ranges over forecasted outcomes (ultimate recovery, recovery factor) as a function of the available history. The effect of uncertainties in bottom-hole pressure and relative permeability on reservoir property solutions and forecasts are also presented. Introduction: Stochastic single-well history matching workflow (SWHM) as a PDA tool The development of coalbed methane (CBM) plays highlights many challenges that necessitate the use of new tools and workflows to facilitate appraisal and development decisions. CBM reservoirs present the following challenges:Coals are highly heterogeneous: Reservoir properties (e.g. isotherm, gas content etc.) typically exhibit a high degree of variability both laterally and vertically and are influenced by geological controls such as coal purity (ash content), geological age, thermal maturity (coal rank). Extrapolation of properties derived from logs away from the well level is difficult given the variability.Naturally fractured reservoir: Key reservoir properties such as permeability and porosity are also dependant on fracture density, aperture, in-situ stresses and fracture network extent which are often difficult to predict. As a result, permeability has been observed to vary several orders of magnitude (1–100mDs) between wells in the same field.Static modelling: data acquisition methods such as seismic and log data provides insufficient information to adequately characterise reservoir. Consequently, conventional 3D modelling workflows produce static models that are insufficiently constrainedLarge well sample size: CBM fields typically have hundreds of wells over large areas (100s – 1000s km2). Gas and water production responses greatly vary well-to-well. Whilst heterogeneity in the underlying reservoir properties is a large contributor, a complex mix of storage mechanisms (gas adsorption vs. free gas) and flow mechanisms (diffusion vs Darcy flow) are often difficult to predict [2]