| Oilfield water injection system is a complex hydrodynamic system,covering a wide area,the scale is huge,after decades of development and operation,many water injection units no longer meet the production needs of the new era,the system efficiency is low,some water injection units frequent failures and serious failures.In order to ensure the normal operation of crude oil production,it is often necessary to diagnose and modify the water injection unit,but only rely on the experience of the management personnel to carry out the transformation design,and often cannot propose an accurate and reasonable transformation plan.Under the background of the construction of digital oilfield and smart oilfield,the combination of oilfield water injection system and artificial intelligence technology and big data mining technology is an important direction for future oilfield development.In this paper,the main fault problems that need to be optimized and improved in the operation of the oilfield water injection system are summarized in detail by fault tree analysis.Then,taking the fault information affecting the production of the oilfield as the starting point,the characteristic parameters of the fault are analyzed layer by layer,and the fault theoretical knowledge is established as comprehensive and orderly as possible for the fault diagnosis and decision-making expert system,and finally the fault tree of the oilfield water injection system is established.Secondly,combined with the fault tree knowledge of the above oilfield water injection system,this paper adopts the knowledge expression method of combining generative rules and framework,the reasoning method of positive and negative hybrid reasoning,the inference strategy of deep search and breadth search,and uses the expert system tool CLIPS to realize the construction of the intelligent expert system reasoning machine and knowledge base.In addition,on the basis of the expert system inference machine and knowledge base is complete,the intelligent diagnosis and decision expert system is developed,and the connection between Python and the database is used to realize the invocation of production data,and the user interface is developed by using Python’s external tool PyQt5;Python’s self-contained CLIPS-based installation package,Pyclips,realizes the interaction between Python and CLIPS,and finally realizes the established functions of the intelligent diagnosis and decision expert system.Then,through EPANET software,a case study of a small simulated oilfield water injection system was conducted to demonstrate the correctness and practicality of the diagnostic and decision-making expert system.Finally,on the basis of EPANET software simulation demonstration,the on-site application of intelligent expert system is carried out for an oilfield pipe network system,and the oilfield pipe network system is optimized and transformed through the diagnosis and decision-making of the intelligent expert system,so that the production performance of an oilfield is significantly increased. |
