Expect the values of these parameters to vary slightly withĮach simulation. This analysis yields the statistical parameters desired for prospectĪnalysis. P10, P50, P90 and any other desired statistical parameters. Re-running the simulation with a larger number of passes. Lacking a smooth distribution necessitates Probability function (Cumulative Probability vs. Of each value by dividing the sample number by the total number of samples Assuming a 1000 pass simulation run, results are processed asįrom 1 to the total number of samples (e.g., 1000). Once a Monte Carlo simulation run concludes, analysis of the resultsįollows. Typically, 1000 or more passes comprise a single Monte Carlo simulation.Īs a qualitative rule, smooth CDF of output variables is indication of The number of input variables, the larger the minimum number of passes Generally, enough runs are needed toĮnsure that the entire domain of input variables is examined. The minimum number of passes required depends on the number of Number of times, a cumulative distribution for the EUR emerges. Of EUR generated by a pass is meaningless. The three independent variable sampled values to yield a sample reserveĪn individual calculation or run to estimate a prospect size in thisįashion is known as a "pass." By itself, the individual value Respective random number to determine the "sampled" value for The cumulative probability distribution for each input variable at their Probability) for each of the three input variables. The following describes one pass in the Monte Carlo simulation procedure:Ī random number between zero and one (representing the value of the cumulative That would result from drilling a large number of similar prospects. "most likely outcome," and the "mean" level of reserves ThisĪnalysis approach is superior to the single-valued deterministic approachīecause of the valuable insight gained into the "upside," "downside," The MonteĬarlo method can make use of these distributions to arrive at an overallĬumulative probability distribution (overall uncertainty) for EUR. Thereby describing the uncertainty in each of these variables. Independent cumulative probability distributions can somehow be defined, Suppose that for each of the three input variables, A, Simultaneously, which is generally very unrealistic. The deterministicĪpproach assumes that the most likely value of every input is encountered The deterministic approach would simply multiply the "best estimate"įor each of these quantities to obtain a single value of EUR. The algebraic expression of this simple model is: Stbo) calculated as the product of the prospect area ( A,Īcres), average net hydrocarbon thickness ( h, Very simple, a trivial example provides the easiest route to developingĪn understanding of the Monte Carlo simulation procedure.Ĭonsider the simple estimation of prospect reserves ( EUR, This uncertaintyĬannot be directly modeled using analytical solutions. Situations where there is uncertainty in the input variables. Monte Carlo simulation is a stochastic modeling method to simulate real-world Mean and standard deviation of the underlying normal distribution). Is the standard deviation of the natural log of the data set (i.e. Is the mean of the natural log of the data set, and σ Of a set of data that follows a lognormal distribution follows a normalĭistribution. Variable whose natural logarithm follows a normal distribution. Normal DistributionĪ log-normal distribution is the probability distribution of a random (PDF), Cumulative Distribution Function (CDF), mean, median, mode andįor each of the Probability Distribution Types. Theĭefinitions include the equations for the Probability Density Function Types that are used as a part of Harmony's Risk Analysis feature. In this section, we will define the different Probability Distribution These PDFs are set up, the more realistic the estimate of the output parameterĪs calculated by the Monte Carlo simulation. That describe the likely values of an input parameter. The inputs to these deterministicĬalculations are randomly drawn from probability density functions (PDFs) Monte Carlo analysis involves a large number of runs, each of Hypercube sampling) is an established approach of performing RiskĪnalysis. The uncertainty in the output also provides a On output variables, and to determine a range of possible outcomes, as opposed to a single deterministic solution. Risk Analysis is a technique to quantify the impact of these uncertainties There may be uncertainty in net pay or drainage area while estimating The best estimate of a value, probabilistic methods are used. In the presence of uncertainty in input data required for determining You are here: Theory and Equations > Analysis Method Theory > Risk Theory Risk Theory