CD Tesis
Pemodelan Tekanan Bawah Permukaan Dengan Menggunakan Metode Time Series Dalam Proses Injeksi Slurry Di Sumur Disposal Duri Field
Oily sand and water contaminated with oil is part of the results of exploration and exploitation of petroleum categorized as hazardous and toxic waste (B3), it is necessary to carry out special processing into slurry fluid and injection into disposal wells. The main problem in the injection process is the increase in bottom hole pressure during the injection process which causes well plugging problems and hampers oil production activities in the Duri field. The purpose of this study is to select the best model for forecasting bottom hole pressure for several future time periods in the slurry injection process based on the injection flow rate variable (m3/min) and slurry concentration (%) using the Arima Software 'R' time series method, determining the average bottom hole pressure from the forecasting and whether the Arima modeling can be applied to the slurry injection process. Forecasting results obtained that the best model from modeling well A with an injection flow rate of 2.1 m3/minute and a slurry concentration of 20% waste and 80% water is model 3 Arima (0,1,1) with the smallest value of Akaike Information Criteria (AIC) 3,452.1 and Schwarzt Bayesian Information Criteria (SBC) 3,459.64, the average bottom hole pressure forecasted is 1,021.47 psi and the average bottom hole pressure from the field data is 1,026.94 psi. The validation of the forecasting model shows that the percentage of model error compared to field data is 0,27%, Root Mean Square Error (RMSE) 3,41% and model error using gradient pressure of injection well A 0,32%. While the best model for forecasting Arima in well B with an injection flow rate of 2,3 m3/minute and a slurry concentration of 25% waste and 75% water is Arima model 3 (1,1,0) with the smallest value of Akaike Information Criteria (AIC) 2,773,98 and the smallest Schwarzt Bayesian Information Criteria (SBC) is 2,781, the average bottom hole pressure predicted is 1,256.4 psi and the average bottom hole pressure from the field data is 1,247.54 psi. Validation of the forecasting model for well B that the percentage of model error compared to field data in well B is 0.37%, Root Mean Square Error (RMSE) 4,85 and model error using gradient pressure of injection well B 0,4%. From the validation results that the error factor obtained is less than 5% so that Arima time series modeling can be applied to predict bottom hole pressure based on the injection strategy in the injection process to the disposal well.
Key words : Forecasting, bottom hole pressure, injection flow rate, slurry composition, software R, Arima
Tidak tersedia versi lain