Using Neural Network Models to predict Equivalent CirculationDensity (ECD)

The most important criterion for reducing the cost of drilling is prediction of ROP from the currentavailable data. ROP performs rock bit interaction which appertain rock compressive strength and bitaggressively. ROP prediction is complex process because of too many variables are included, their inputparameters are often not readily available, and their relationships are complex and not easily modeled. So,the application of Neural Network is suggested in this study. To predict the rate of penetration Some newmethodology has been developed like using the Artificial Neural Network (ANN). Application of the newnetwork models would then be used for selecting the best parameters for an optimal drilling strategybased on field data. Rock bit interactions in the field as a function of rock mechanical property parameterswas achieved by predicting ROP which relates to rock compressive strength and bit aggressively; as wellas TWR which relates to rock abrasiveness and wear resistance. Based on field data, the prediction ofrock mechanical property parameters can be accomplished by the use of a neural network as analternative prediction and optimization method.

Using Neural Network Models to predict Equivalent CirculationDensity (ECD) Keywords:

ECD , Rate of penetration , drilling cost , bit , ANN