Developing a New Mathematical Model to Estimate Fracture Gradient Using Genetic Algorithms and with Gene Expression Programming Approach in one of the Fields in Persian Gulf

Accurate evaluation of pore pressure and fracture pressure are key parameter in well planning and drilling operation. Estimation of these parameters has an important role in reduction in well costs. An inaccurate estimation of fracture gradient may jeopardize whole drilling operation and cause serious problems such as loss circulation and blowout. In this paper, the method of Gene Expression Programming (GEP) has used for developing a mathematical model for prediction of formation fracture pressure. Overburden pressure, pore pressure and Poisson ratio are independent variables in this model. Studied areas include two oil-bearing formations: Kangan and Upper-Dalan in an onshore field which is located near Persian Gulf. A separate model is presented for each formation by using well logging data from two wells. Models developed by using more than 4,300 well logging data from a well. Verification of these models is done with 6,000 data from second well in same formation. Statistical analyses confirm that the models are suitable for fracture gradient prediction so we can use them for estimation of fracture pressure.