Research On The Method Of Production Dynamics Early Warning In Polymer Flooding Block Based On H.I.S.

With the deepening of oil field production and development, the scale of industrialization of the tertiary oil has expanded year by year. Because of the factors that affect block production is various, thus the production analysis method and knowledge system of the polymer flooding is still to be improved and the maintenance costs of production unit are high, the contradiction of block development is becoming more and more prominent, the development and management of tertiary oil is difficult.Recently, the continual development of hybrid intelligent systems(HIS) provide a new method to auxiliary decision oil field production via intelligent means. This paper analyzes the working principle and application requirement of dynamic early warning for polymer flooding production, puts forward a polymer flooding dynamic early warning method based on the analysis of domestic and foreign production of early warning and hybrid intelligent technology.Firstly, the dynamic early warning model for the polymer flooding block based on HIS is designed, the problems in the model is analyzed. Secondly, aiming at these problems, the integration of dynamic data in polymer flooding block and the mining method for abnormal pattern of polymer flooding, along with the reasoning and matching mechanism of abnormal pattern of polymer flooding production were all in-depth studied. In order to seek concrete solutions to provide technical support for the overall model. Finally, combine with the actual work flow and business requirements of production management in polymer flooding block to design the production management platform of polymer flooding and analysis system practical work effect. The main research contents of this paper are as follows:1. The dynamic early warning model for the polymer flooding block based on HIS (HI-DEMP) is put forward.This paper research on the characteristics of polymer flooding production data, early warning application needs and existing problems. The work principle, process and key technology in production dynamics early warning of polymer flooding block are mainly studied. The feasibility and advantage of hybrid intelligent technology in the early warning of polymer flooding production are analyzed. The dynamic early warning process for the polymer flooding block based on HIS is designed. Using a hybrid intelligent technology to put forward a dynamic early warning framework, and describe the overall structure of the model, the function of each layer and the problem to be solved in detail. Reference to multi agent systems, uncertainty reasoning and other related technologies to achieve the new method to accurate prediction of anomalies.2. The integration and management of dynamic data, the mining method for abnormal pattern, along with the reasoning and matching mechanism of abnormal pattern in polymer flooding were all mainly studied.To begin with the method of integration and management of polymer flooding production data. By analyzing the data sample unit of pattern mining, the structure of abnormal case is designed. Combined with the process of production management and data characteristics of polymer flooding, the dynamic data model of polymer flooding block was built. Bringing in multi agent system, the establishment of agent monomer structure, multi agent system framework, communication mechanism and management strategy were all based on data model. The data integration management model is proposed to actualize the integration of polymer flooding data.Then, researching the pattern mining method on the polymer flooding abnormal(PFD-Him method). The method analysis and summarize the pattern mining process of different types of polymer flooding early warning, The mixed mode of stage type multi process (SMP mixed mode) is put forward according to the tight coupling mixed mode. Introducing time series to do the data preprocessing and finish the screening of influencing factors. Using time domain or frequency domain parameters, via inner distance difference of radial basis to improve the feature evaluation method, extract abnormal sensitive features. To accomplish the pattern mining and interpretation by designing the C-ECM through ANFIS.Ultimately, the turn is the reasoning and matching mechanism of abnormal pattern in polymer flooding. Constructing the knowledge base of abnormal pattern of production in polymer flooding, raising the idea of rule derived knowledge and using algorithm library as the execution entity of C-ECM structure. The storage mode and interpretation method of production abnormal pattern in knowledge are described, the inheritance relationship between knowledge and rules is established, and the knowledge representation structure based on multilevel framework is designed. Through the reflective reasoning technique and the Bias method and the certainty factor algorithm to research the pattern reasoning and matching, in order to bring out the uncertainty reasoning process of model. At the same time, the feasibility and application results of the early warning model are verified by experiments.3. Empirical analysis of production dynamics early warning method of polymer flooding block.The application background is the daily management and analysis of polymer flooding production, the basis is the HI-DEMP model, the purpose is improving the intelligent level of production warning and dynamic analysis of tertiary oil recovery. The polymer flooding development management platform is designed, the general situation of the platform, the overall design, key technologies and other aspects of the content were all described. Through the actual test and data analysis of real environment, application shows that the system can real-time, effective implementation of polymer flooding production dynamics early warning.The results show that the method of production dynamics early warning of polymer flooding block based on HIS is able to solve the problem of production dynamic analysis and prediction of the current polymer flooding. Besides, by extending the research and analysis. the method provides a new way for the application of early warning of water flooding.