Inverse analysis is commonly used in identifying geomechanical parameters based on the monitored information such as displacement or stress. Conventional inverse analysis method is incapable of recognizing non-linear relationship involving displacement, stress and mechanical parameters effectively. A hybrid model which combined Multi-output-Support Vector Machine (MSVM), Artificial Bee Colony (ABC) and numerical analysis has been proposed to estimate the in situ stress and rock mechanical parameters based on borehole fluid pressure. MSVM is used to represent the non-linear relationship between parameters of numerical model and borehole fluid pressure. ABC is used to search the set of unknown recognized parameters based on the objective function. Numerical analysis of hydraulic fracturing is used to create the necessary training and testing samples for the hybrid MSVM-ABC model. Results of numerical experiments demonstrate that a hybrid MSVM-ABC model for inverse analysis can effectively identify in situ stress and rock mechanical parameters based on wellbore fluid pressure in the hydraulic fracturing process.
Geomechanical parameters such as Young's modulus and in situ stress in the field of petroleum are important to reservoir simulation, borehole stability analysis and production of petroleum (Gokceoglu et al. 2004, Juliusson 2012). However, it is difficult to obtain those parameters accurately and efficiently using the traditional laboratory test and in-situ test because of the complex, nonlinear and uncertainty of rock mass characters (Zhao & Yin 2009). Inverse analysis method provides a good way to get these parameters by combining the field-observed information with numerical simulation. Inverse analysis are commonly used in rock mechanics and engineering such as tunnel, underground engineering and rock slope etc (Sakurai & Takeuchi 1983, Gioda & Maier 1980, Miranda et al. 2011, Feng & Hudson 2011).
In the petroleum industry, hydraulic fracturing is the most common technique to perform well stimulation and earth stress characterization of hydrocarbon reservoirs, especially for unconventional reservoirs such as shale gas, tight gas and coal bed methane. In this paper, geomechanical parameters were estimated by combining inverse analysis and hydraulic fracturing test.