Research On Knowledge-driven Mechanical Specific Energy Analysis Model

With the continuous deepening of oil field exploration,cracks and hole storage layers have become an important reservoir for the exploration of Daqing Oilfield.The distribution of oil and gas and water in this type of reservoir is more complicated.Technical problems.Mechanical specific energy is an important evaluation and intermediate indicator of drilling work.In the process of storage material evaluation,the mechanical specific energy is of great significance.However,currently recording data exploration data is relatively complicated,showing fragmented,and the use of oil field technicians to use the use of machinery models mainly depends on manual enrollment data.It calculates formulas through artificial experience.At the same time Data use.In order to solve the problems such as the uncertainty of the mechanical specific energy caused by the complicated environment during the current record operation process,the mechanical specific energy analysis model under knowledge-driven is applied,and the intelligent method is applied to the mechanical specific energy evaluation and storage.In the process of layers,the research content is as follows:Introduce the knowledge base of mechanical specific energy model to analyze the relevant knowledge of mechanical specific energy models in the recording reservoir data,combined with relevant information and books,on-site data managers’ experience,historical solution cases,expert knowledge,etc.The extraction is performed,and the relevant knowledge architecture is constructed on the basis of the mechanical specific energy model and stored in.For the actual characteristics of the mechanical specific energy model,select case-based reasoning(CBR)and rules-based reasoning(RBR)based on case-based reasoning(CBR).Analysis,so as to quickly analyze the optimal mechanical specific energy model suitable for different wells and well sections,to replace artificial choices,and provide a new method for the determination of mechanical specific energy models.For the implementation of the specific reasoning method,the case-based reasoning uses a hierarchical analysis method and combines the K near adjacent algorithm to find similar cases in the case library.For the situation of no similar cases,the rules-based reasoning adopts gradient to improve the decision-making tree algorithm(Gradient Boosting Decision Tree,GBDT)analyzes the hidden rules in the data,generates corresponding decision trees,and continuously iterates the final result.Through multiple experiments,verify the results of the experiment,and compare the integrated reasoning with random forests and case reasoning,indicating that the feasibility and accuracy of the analysis of mechanical specific energy is conducive to the feasibility of fusion of CBR and RBR.Finally,in the actual records of the field exploration project,based on the mechanical specific energy model driven by knowledge,the mechanical specific energy mud-logging system is designed and developed to evaluate the reservoir property system,and the overall design,process analysis,and main modules of the system are explained in detail,and the main modules are in the system.Using and analyzing the results under the real well-recording data,it is found that the system performance is better,and the effectiveness of the research of this project is verified.