A robust fuzzy logic-based model for predicting the critical total drawdown in sand production in oil and gas wells

Abstract

Sand management is essential for enhancing the production in oil and gas reservoirs. The critical total drawdown (CTD) is used as a reliable indicator of the onset of sand production; hence, its accurate prediction is very important. There are many published CTD prediction correlations in literature. However, the accuracy of most of these models is questionable. Therefore, further improvement in CTD prediction is needed for more effective and successful sand control. This article presents a robust and accurate fuzzy logic (FL) model for predicting the CTD. Literature on 23 wells of the North Adriatic Sea was used to develop the model. The used data were split into 70% training sets and 30% testing sets. Trend analysis was conducted to verify that the developed model follows the correct physical behavior trends of the input parameters. Some statistical analyses were performed to check the model’s reliability and accuracy as compared to the published correlations. The results demonstrated that the proposed FL model substantially outperforms the current published correlations and shows higher prediction accuracy. These results were verified using the highest correlation coefficient, the lowest average absolute percent relative error (AAPRE), the lowest maximum error (max. AAPRE), the lowest standard deviation (SD), and the lowest root mean square error (RMSE). Results showed that the lowest AAPRE is 8.6%, whereas the highest correlation coefficient is 0.9947. These values of AAPRE (<10%) indicate that the FL model could predicts the CTD more accurately than other published models (>20% AAPRE). Moreover, further analysis indicated the robustness of the FL model, because it follows the trends of all physical parameters affecting the CTD.

Deep Learning Approach for Robust Prediction of Reservoir Bubble Point Pressure

“…Additionally, trend analysis can also identify and remove unnecessary parts of the model structure . An input parameter can be changed between the minimum value and the maximum value, while the others are fixed at their constant mean values. ,− Graphs are plotted for the input parameter values as the x -axis against P b as the y -axis for the previous models and the LSTM model.…”

A reservoir bubble point pressure prediction model using the Adaptive Neuro-Fuzzy Inference System (ANFIS) technique with trend analysis

“…other inputs at their constant mean values. The studied input, such as Rs, is plotted as the xaxis and the output P b as the y-axis [27,[48][49][50]. The TrA is an essential part of this work, as some researchers used ANFIS, but they have not applied the trend analysis [40].…”

Determination of the Gas–Oil Ratio below the Bubble Point Pressure Using the Adaptive Neuro-Fuzzy Inference System (ANFIS)

“…Nowadays, AI-based models have become a hot topic in engineering applications and are efficiently applied in many petroleum engineering calculations. − Deep learning and gradient boosting methods were successfully conducted to determine complex carbonate rock’s permeability, capillary pressure, relative permeability, and the optimum operational conditions for CO 2 foam enhanced oil recovery. − Adaptive neuro-fuzzy inference systems (ANFIS), artificial neural network (ANN), fuzzy logic, and group method of data handling techniques have been effective in obtaining the mineralogy of organic-rich shales, the oil formation volume factor, the fractured well productivity, the natural gas density of pure and mixed hydrocarbons, the breakdown pressure of unconventional reservoirs, and the critical total drawdown for the sand production. − …”