Development of a new rate of penetration model using self-adaptive differential evolution-artificial neural network

The rate of penetration (ROP) is one of the key factors that affect the drilling costs. Optimizing the ROP is a big challenge as it depends on many factors such as revolutions per minute (RPM), weight on bit (WOB), torque (T), horsepower (HP), and uniaxial compressive strength (UCS) of the drilled rocks. In addition, drilling fluid properties have a major effect on ROP. The main goal of this study is to develop a new ROP model using an artificial neural network (ANN) combined with the self-adaptive differential evaluation (SaDE) technique. The model was built using different drilling mechanical parameters and drilling fluid properties. A new ROP empirical correlation was developed by extracting the weights and biases of the optimized SaDE-ANN model. The optimized ANN architecture based on SaDE is 5-30-1, where five input parameters were used in the input layers to predict the ROP which are drilling fluid density to plastic viscosity ratio, RPM, WOB/D, T/UCS, and HP. The optimized number of neurons was 30 and the output layer consists of one output parameter which is ROP. The data was divided into 60% training and 40% testing. The developed ROP model based on SaDE-ANN showed high accuracy where the correlation coefficient (R) was 0.98 and the average absolute percentage error (AAPE) was 5%. The new ROP empirical correlation outperformed the previous ROP models.