Novel hybrid machine learning optimizer algorithms to prediction of fracture density by petrophysical data

Abstract One of the challenges in reservoir management is determining fracture density (FVDC) rock. Given high cost coring operations and image logs, ability to predict FVDC from various petrophysical input variables using a supervised learning basis calibrated standard well extremely useful. In this study, novel machine approach developed 12-input variable well-log based on feature selection. To FVDC, combination two networks multiple extreme machines (MELM) multi-layer perceptron (MLP) hybrid algorithm with genetic (GA) particle swarm optimizer (PSO) has been used. We use MELM-PSO/GA that never used before, best comparison result between MELM-PSO-related models performance test data RMSE = 0.0047 1/m; R 2 0.9931. According accuracy analysis, are MLP-PSO < MLP-GA MELM-GA MELM-PSO. This method can be other fields, but it must recalibrated at least one well. Furthermore, provides insights for reduce errors avoid overfitting order create possible prediction prediction.