Font Size: a A A

Production Forecast For Daliudi.S1 Gas Reservoir

Posted on:2012-10-13Degree:MasterType:Thesis
Country:ChinaCandidate:J T XueFull Text:PDF
GTID:2210330338967855Subject:Oil and gas field development geology
Abstract/Summary:PDF Full Text Request
The prediction of production is a common problem in the gas field development. The result of production forecast will affect the subsequent development of well development programs. Accurate gas production forecasts will help the adjustment and optimization of work systems. it is a guiding role for the use of acid fracturing and other technological measures. The methods of production are established, combined with specific geological background. the methods are suitable for specific gas reservoir.Daniudi gas field belongs to the low permeability, low pressure gas reservoir. The mechanism of fluid flow is complex. So the accuracy of traditional methods is very low. In order to improve accuracy of production forecasting, this paper uses data mining methods to analyze. The gas well production forecasting model is established, combined gas field geological background and gas reservoir. There are two aspects to analyze gas well production capacity, including qualitative and quantitative. Predicted from the qualitative point of view, means that the capacity of gas wells belongs to high,middle or low wells. Predicted from the quantitative point of view, means that the capacity of gas wells. It is used related methods of data mining whether from qualitative point of view or from the quantitative point of view. The methods of data mining include:decision tree, neural network, time series methods, svm, and knn methods. Using these methods, production forecasting models are built, including: decision tress model, neural network model, support vector machine model, knn model, combined with production data. Production is predicted using the established production forecasting models. It can be determined whether the model is right or wrong by comparing with forecast data and actual data. The best forecasting model is determined ultimately. The decision tree model is the optimal qualitative prediction model; the neural network model and time series model are the optimal quantitative prediction models.Production forecasting model can be built by analyzing Daliudi gas field specific geological characteristics. It is can be used to predict production capacity. The production forecasting model provide reasonable reference to ensure the economic interests of the Daniudi gas field.
Keywords/Search Tags:Decision tree, Support vector machine, KNN, Neural network, Time series
PDF Full Text Request
Related items