Adaptive identification of system lifecycle by the method of integrated phenomenological models with variable parameters

Abstract

The relevance of the discussed issue is caused by the need to increase information content of identification methods and predicting evolutionary processes of the system lifecycle based on phenomenological models under uncertainty with limited amounts of initial data. The main aim of the study is to develop the models and algorithms for adaptive identification in evolutionary processes of the system lifecycle based on phenomenological models with variable parameters and with a priori information. The methods used in the study are theoretical and practical developments in system analysis, system modeling, and identification with additional prior information, optimization methods of functions and linear algebra. The problem was solved theoretically based on the filed data and expert estimates of performance indicators of gas fields exploitation objects in Russia. The results. The authors have proposed to develop the models and algorithms for adaptive identification in evolutionary processes of the system lifecycle based on non-linear integrated systems of phenomenological models with variable parameters, additional a prior information and expert estimates. To solve the identification problems the authors used the local approximation function method. Adaptive process of phenomenological models of the system lifecycle is introduced in the form of two optimization problems solutions by determining phenomenological models parameters and control parameters. The proposed adaptive method, taking into account a priori information, allows synthesizing a wide range of the known and new algorithms for adaptive identification of linear and nonlinear phenomenological models of system lifecycle in the condition of a priori uncertainty with small amount of input data. It is shown that under uncertainty with limited field data the development of phenomenological models with variable parameters, identification algorithms, prediction of annual gas production and recoverable reserves of gas fields are more accurate and stable in comparison with phenomenological models with constant parameters measured on the basis of their identification algorithms and prediction.

Keywords

• identification
• adaptation
• integrated systems of phenomenological models with variable parameters
• system lifecycle
• a priori information
• prediction