1887

Abstract

History matching is an inseparable part of reservoir characterisation which is a highly nonlinear inverse problem and suffers from ill-posedness. Different regularisation methods such as Tikhonov regularisation and Bayesian framework have been used to overcome the ill-posedness using prior knowledge. In this study, the application of a multi-objective genetic algorithm (GA) in the history matching as a regularisation and an optimisation is introduced. In this approach, two separate objective functions likelihood and prior are defined. In Tikhonov and Bayesian approach, the mentioned objectives are defined with one weighted function. In the Bayesian framework, covariance matrixes are utilised as weighting factors for each parameter, but there is no constant to join likelihood and prior objectives. However Tikhonov relates the objectives with a weighting factor, it is a challenging task to find the optimum value for the constant in the history matching. Consequently, in these two regularisation methods, it is potential that one objective dominates the other one. To validate the approach, ECLIPSE is coupled with MATLAB. A synthetic 3 dimensional 3 phase reservoir is constructed. Gaussian noise is added to the history. After that, different approaches are used to match the history and reconstruct the reference case. Bayesian and Tikhonov regularisation with different optimisation methods, real-valued genetic algorithm and nonlinear least square Levenberg-Marquardt algorithm optimisation are used. Then, their results are compared with a multi-objective GA. The outcomes demonstrate that the proposed method converges quicker than other methods and more importantly the results are realistic. In multi-objective systems, each objective has effect on the other one. Hence, optimising a system without considering all the objectives together leads to unrealistic outcomes. Using a multi-objective GA, it would be feasible to consider all objectives togather and provide the Pareto front.

Loading

Article metrics loading...

/content/papers/10.3997/2214-4609-pdb.293.H024
2012-06-04
2021-10-22
Loading full text...

Full text loading...

http://instance.metastore.ingenta.com/content/papers/10.3997/2214-4609-pdb.293.H024
Loading
This is a required field
Please enter a valid email address
Approval was a Success
Invalid data
An Error Occurred
Approval was partially successful, following selected items could not be processed due to error