APPEA Journal 2012—1 SECOND PROOF—MAHADIK 27 MAR 12 M.K. Mahadik , H. Bahrami, M. Hossain and P.A. Tsar Mitchel Department of Petroleum Engineering 613 (rear), level 6, ARRC Curtin University 26 Dick Perry Avenue Kensington WA 6151 M.Mahadik@postgrad.curtin.edu.au ABSTRACT Exponential decline curve analysis is widely used to esti- mate recoverable reserves due to its simplicity. In most cases, however, an exponential model cannot provide a satisfactory match of overall production history. The generalised form of a hyperbolic decline model is more powerful in matching production history than the other decline models, but it is difficult to apply in practical production data analysis since it requires predicting two unknowns as decline constants. Although a hyperbolic model may provide a good fit to early-time production decline data; it may overestimate the late-time production, especially for hydraulic fractured wells in a tight-gas reservoir. In fact, the exponential decline model might be more reliable for forecasting the late-time production. This paper presents a practical approach to production decline analysis for non-fractured and fractured wells in a tight-gas reservoir using numerical simulation. Some pro- duction rate functions and type curves are introduced to obtain the best matching values of hyperbolic, exponential and harmonic production decline constants. The simulated production rate decline data for various well and reservoir parameters are used to indicate the op- timum time duration of use of each decline model and to show the time when the production decline starts following the exponential model. The proposed approach is applied in production data analysis and forecasting for a tight-gas field in WA. The results showed good agreement with the produc- tion forecast obtained from a reservoir simulation. KEYWORDS Production decline analysis, production forecasting, tight-gas reservoirs, decline models, rate derivative function, effect of well and reservoir parameters, decline constant, decline exponent. INTRODUCTION Reservoir production performance analysis and forecasting are fundamental responsibilities for a reservoir engineer. There are several techniques that help to predict the production pro- file; one of the most common methods used is known as decline curve analysis. Typically, a decline curve analysis is begun by plotting stable initial production rates in detail against a short period of time. The extrapolation of the existing early-time data is generally carried out with three decline trends: exponential; hyperbolic; and, harmonic. Nevertheless, at the beginning of the well’s production decline trend, extrapolating forward in- volves much uncertainty. Decline curve analysis techniques are a substitute for the volumetric and material balance method. At times, based on the type, quantity and quality of the data available, decline curve analysis is the only method available for measuring re- serves (Rushing et al, 2007). For example, material balance re- quires estimation of stabilised shut-in bottomhole pressures (BHPs); however, in low permeability reservoirs the BHPs are not available due to the long period of time needed for stabilisa- tion (Lee and Wattenbarger, 2008) Although decline curve analysis is an excellent tool used to estimate the production in conventional oil and gas reservoirs, it should be applied under proper conditions. Normal applica- tion of decline curve analysis has the following assumptions (Lee and Wattenbarger, 2008). • Type curves are generated with constant bottom-hole flow- ing pressure. • The permeability and skin of the reservoir is kept constant, thus any changes in reservoir parameters due to stimulation strategies will affect the reserve estimates. • If there is no true boundary dominated flow then there is no theoretical basis and the future predictions can be inac- curate. • Late-time production data in a reservoir located off-center may not represent true boundary-dominated flow. • Decline curve assumes for a volumetric reservoir, meaning there is no energy from outside, such as pressure from an aquifer. If any of these assumptions are violated, it may result in sig- nificant errors in the reserve estimates. These are even greater in tight-gas reservoirs due to the complexity of geological and petrophysical systems, as well as the large scale of heterogene- ities (Rushing et al, 2007). PRODUCTION DECLINE ANALYSIS A method presented by Arps in the 1950s has been used due to its simplicity, and since it is based on empirical methods it does not require any reservoir parameters. The Arps equation is used to curve-fit the production data and to get the best fit by changing the decline constants. The general Arps rate/time equation represents hyperbolic decline behaviour as shown in Equation 1. (1) Where q i is the initial production rate, t is elapsed time, D i is the initial decline constant, and b is the decline exponent. In the Arps equation, if b equals zero, it represents exponen- tial decline behaviour, as shown in Equation 2. (2) If b equals one, it corresponds to harmonic decline, as shown in Equation 3. PRODUCTION DECLINE ANALYSIS AND FORECASTING IN TIGHT-GAS RESERVOIRS Lead author Mujeeb Mahadik q q bD t t i i b = + ( ) 1 1 q q e t i Dt i = -