Application of the Monte Carlo Simulation in Calculating HC-Reserves

Abstract

One of the key features of E&P companies is their proved reserves in hydrocarbon deposits. Reserve estimation requires knowledge of Initial Hydrocarbon in Place, technical reserves and economic conditions including annual cash flow estimation in the forecast period. Since all parameters used in evaluation procedure are burdened by rather more than less certainty. Therefore, in a sophisticated evaluation process, there should be determined not the expected values only (deterministic way), but errors/uncertainty of estimation as well (stochastic way) applying Monte Carlo simulation. The estimation procedure comprises three main stages (the third stage /economic modeling/ is not discussed in this paper). In the first stage, key input data (e.g., area, thickness, porosity, and so on) are treated as statistical variables, and the result of the simulation is probability distribution function of HCIIP. This is an input of next stage. In the second stage technical reserves (recoverable resources) should be estimated. There could be several assumptions for production procedure for a reservoir (as e.g., drive mechanism, hydrodynamic system, phase behavior of reservoir fluids, well spacing, water injection, presence of pressure barriers etc). Each regime (i.e. scenario) can be modeled applying input parameters as statistical variables. This method is named a multiscenario method in the literature. Simulation result for each scenario is a probability distribution function (PDF). While, expected value of PDF reconstructs the deterministic result and gives a basis for project evaluation, the "width" of PDF is proportional with uncertainty of the estimation. Estimating probability of each scenario a combined technical reserve PDF can be derived. Its first percentile can yield proved reserve for booking procedure after economic limit test. Authors show some case histories how to apply method after a brief theoretical summary referring to SPE-PRMS accepted. IntroductionThe fundamental issue in reserve estimation is the volume of hydrocarbon that can be economically recovered from the reservoir. This is a complex task. Experts of several disciplines should closely cooperate for reaching a good solution moreover, we will always have only limited amount of information. This is the reason why we focus on the Monte Carlo simulation (MCS) procedures in this paper (except the economic modeling). We expect that both experts of geo-sciences and petroleum engineers will be interested.As this paper will not cover profitability estimation, we can not speak about reserves according to international standards but only about recoverable resources. But we wish to highlight that the subject of our analysis is closer to reserves than to resources, therefore we will use the term of technical reserve as formerly used in the industrial practice: technical reserve is a resource, which can be economically recovered using regular production technologies by expectation of technical experts, but the profitability was not specifically analyzed.

Estimation of in Place Hydrocarbon Volume in Multilayered Reservoirs Using Deterministic and Probabilistic Approaches

“…Zhang and Srinivasan in 2003 presented a methodology for generating MCMC (Markov Chain Monte Carlo) model of a geological system and described it in mathematical expressions (Zhang and Srinivasan, 2003). In 2009 Z. Komlosi and J. Komlosi used Monte Carlo simulation to estimate the amount of reserves in three oil fields (one anticline-type and the other two lithologic-type) and concluded that "Monte Carlo simulation's accuracy is determined by the reliability of the geo-technical model, parameters and conditions forecasted" (Komlosi and Komlosi, 2009).…”

“…In 2009 Z. Komlosi and J. Komlosi used Monte Carlo simulation to estimate the amount of reserves in three oil fields (one anticline-type and the other two lithologic-type) and concluded that "Monte Carlo simulation's accuracy is determined by the reliability of the geo-technical model, parameters and conditions forecasted" (Komlosi and Komlosi, 2009).…”