Exploitation
Tapan Dey; Gopinath Samanta
Abstract
Accurate grade prediction is an important step in the mining planning process. Various methods, namely the Inverse Distance Method and Kriging, are widely used. The application of machine learning is a new development in the grade estimation technique. The present study focused on the application of ...
Read More
Accurate grade prediction is an important step in the mining planning process. Various methods, namely the Inverse Distance Method and Kriging, are widely used. The application of machine learning is a new development in the grade estimation technique. The present study focused on the application of XGBoost, Random Forests (RFs), Multi Layer Perceptron (MLP), and Gradient Boosting Regression (GBR) models to predict iron ore grades in an Indian mine. An ensemble model was also applied to obtain a more stable grade prediction in the deposit. Models were trained using 4,112 sample data, which have spatial coordinates (east, north, and altitude) and iron grades. The dataset was divided into two parts: 80% (3,289 samples) of the data was used for model training, and 20% (823 samples) was used for model testing. The performance of the models was assessed through the coefficient of determination (R²), Mean Squared Error (MSE), and Mean Absolute Percentage Error (MAPE). The results show that the XGBoost model performs better in the estimation process when compared with other methods, such as RFs, GBR, and MLP. The XGBoost model produced R² of 0.77, MSE of 2.87, and MAPE of 1.8%. The findings indicate that the XGBoost model is effective for predicting iron ore grades in this type of deposit. However, considering geological uncertainty, the application of an ensemble model may be beneficial for grade prediction in an iron deposit.
Sh. Maleki; H. R. Ramazi; M. J. Ameri Shahrabi
Abstract
Shear wave velocity (Vs) is considered as a key parameter in determination of the subsurface geomechanical properties in any hydrocarbon-bearing reservoir. During a well logging operation, the magnitude of Vs can be directly measured through the dipole shear sonic imager (DSI) logs. On a negative note, ...
Read More
Shear wave velocity (Vs) is considered as a key parameter in determination of the subsurface geomechanical properties in any hydrocarbon-bearing reservoir. During a well logging operation, the magnitude of Vs can be directly measured through the dipole shear sonic imager (DSI) logs. On a negative note, this method not only is limited to one dimensional (1D) interpretation, it also appears to be relatively costly. In this research work, the magnitude of Vs is calculated using one set of controversial petrophysical logs (compressional wave velocity) for an oil reservoir situated in the south part of Iran. To do this, initially, the pertinent empirical correlations between the compressional (Vp) and shear wave velocities are extracted for DSI logs. Then those empirical correlations are deployed in order to calculate the values of Vs within a series of thirty wells, in which their Vp values are already recorded. Afterwards, the Kriging estimator along with the Back Propagation Neural Network (BPNN) technique are utilized to calculate the values of Vs throughout the whole reservoir. Eventually, the results obtained from the two aforementioned techniques are compared with each other. Comparing those results, it turns out that the Kriging estimation technique presents more accurate values of Vs than the BPNN technique. Hence, the supremacy of the Kriging estimation technique over the BPNN technique must be regarded to achieve a further reliable magnitude of Vs in the subjected oil field. This application can also be considered in any other oil field with similar geomechanical and geological circumstances.
