Estimation of geomechanical rock characteristics from specific energy data using combination of wavelet transform with ANFIS-PSO algorithm

Abstract The geomechanical characteristics of a drill formation are uncontrollable factors that crucial to determining the optimal controllable parameters for drilling operation. In present study, data collected in wells drilled Marun oilfield southwestern Iran were used develop adaptive network-based fuzzy inference system (ANFIS) models parameters. specific energy (DSE) was calculated using such as weight-on-bit (WOB), rate penetration (ROP), rotational speed string (RPM), torque, bit section area, hydraulic factor, and power. A stationary wavelet transform subsequently decompose DSE signal fourth level. approximation values details each level served inputs ANFIS particle swarm optimization (PSO) algorithm genetic (GA). As model outputs, Young’s Modulus, uniaxial compressive strength (UCS), cohesion coefficient, Poisson’s ratio, internal friction angle compared obtained from petrophysical logs laboratory-developed empirical relationships. Both predicted modulus, UCS, coefficient with high accuracy, but lacked accuracy predicting ratio. root mean square error (RMSE) determination ( R 2 ) lower ANFIS-PSO than ANFIS-GA model, indicating presents higher better generalization capability model. readily available, proposed method can provide valuable information strategizing operation absence logs.