Summary of the results and conclusions

Today, I read the following part of the paper and summarize the conclusions and results.

Summary of the results and conclusions

The training process of the neural network is conducted using a back propagation technique.

Once models are developed, Key Performance Indicator process (KPI) is performed in order to investigate the influence of each parameter in data-driven models.

R-squared calculated for training, calibration and verification segments, is also an indicator of accuracy of the data-driven models in performing results as output.

The one of the figures of actual versus generated well logs is shown below:

For well#5, in Figure 9, the generated data is not close to the actual and it might be because of the location and the producing payzone depth differences.

The figure of well#5 is shown below:

The name of the figure is not right because it should be well#5.

In conclusion, the author demonstrates the application of AI&DM as a reliable tool in performing accurate results for generating synthetic geomechanical logs. In simple terms, they used conventional well logs to generate geomechanical properties and create distribution maps of geomechanical properties for the entire asset.

Synthetic geomechanical logs include following properties

Poisson’s Ratio: the negative value of the ratio of transverse strain over axial strain.

Total Minimum Horizontal Stress: a hydraulic fracture will propagate perpendicular to the minimum principal stress.

Bulk Modulus: the ratio of the infinitesimal pressure increase to the resulting relative decrease of the volume .

Shear Modulus: the ratio of shear stress to the shear strain .

Young’s Modulus: the relationship between stress (force per unit area) and strain (proportional deformation) in some linear elastic solid material .

After know the meanings of these definitions, I find that they are all properties of rocks.

R-squared: coefficient of correlation.

Anyway, I still do not find how the ANN works in the paper. Maybe I can find something new and more in the references 8 and 10 of the paper and others from Prof. Siddharth Misra.

Tomorrow, I will read more papers on ANN.