Fracturing Optimization Research Based On Knowledge Discovery In Database

Hydraulic fracturing is one of the important measures to increase oil and gas production.It helps improve the development in low-permeability,undeveloped oil-bearing reservoirs and shale gas reservoirs.This paper is to study the optimization of hydraulic fracturing parameters.This paper is based on the idea of knowledge discovery.In combination of knowledge discovery and oil and gas development theories,historical hydraulic fracturing data of one oil extraction factory is analyzed to guide the following hydraulic fracturing with similar condition.Firstly,data needs to be pre-treated.Raw data,including reservoir and engineering parameters which have something to do with fracturing performance,needs to be chosen,sorted out and qualitative analyzed.Gray theory is then introduced to choose the most relevant factors.Secondly,with optimized SVM,whose penalty coefficient and kernel parameters are modified by GA,target wells and layers are chosen.By comparison with pure SVM method,optimized one results in better performance.Thirdly,with optimized BP neural network,whose network connection weights,thresholds and the structure are modified by PSO,parameters such as injection rate and proppant concentration are optimized.By comparison with pure BP neural network method,optimized one results in better performance.At last,program is developed to guide workers on the site.