bahram habibnia - Associate Professor and Faculty member, Petroleum University of Technology
erfan hosseini - MSc Student in Petroleum Exploration, Petroleum University of Technology
majid nabi bidhendi - Full Profesor and Faculty member, Institute of Geophysics University of Tehran
siamak moradi - Assistant Professor and Faculty member, Petroleum University of Technology

The most commonly used data in reservoir description are well data and seismic data. Well data such as logs typically provide sufficient vertical resolution but leave a large space between the wells. 3D seismic data, on the other hand, can provide more detailed reservoir characterization between wells. However, vertical resolution of seismic data is poor compared to that of well data. Conventionally, seismic data has been used to delineate reservoir structure, however, there has been limited application of using them to help directly map reservoir properties such as P-wave velocity. Therefore we can combine these two types of data to obtain the reservoir parameters such as P-wave velocity. In fact, the desired parameters of the number of wells are available in the reservoir and seismic cube. And we are looking for calculate the desired parameter in total of reservoir cube. To do this, there are several methods including multiple linear regression and neural networks methods. Therefore, by determining the reservoir characteristics, detect and correct prediction of reservoir parameters can be done maximize the utilization of reservoirs with less number of wells and thus reduce the cost of exploration and production. In this article, will be explained the first to introduce seismic attributes and their calculation and then, in brief is expressed theoretical methods for combining data and the desired parameter estimation. In the next step, apply the above methods to obtain the P-wave velocity in total of reservoir cube on the three data series available of an oil field in south-west of Iran and finally, these methods are compared with each other. The results clearly show superiority of using the neural networks compared to previous method in the reservoir parameter estimation.

Reservoir, Well log data, Porosity, Neural network, Regression methods