عنوان مقاله [English]
نویسندگان [English]چکیده [English]
Determination of porosity distribution is important in hydrocarbon reserve estimation, facies variations, optimized planning for field development and decrease in drilling risks and costs. Porosity is one of the most important parameters, which is considered as a fundamental factor in reservoir engineering. By knowing this parameter, specialists are able to design and manage, effectively, the process of oil and gas field development. Seismic attributes and well logs are the data available in most of the reservoir studies. Seismic attribute analysis is generally done through correlating multi attributes to the reservoir characteristics. A good interpreter needs to observe several maps with certain information to prepare optimal drilling points. Such a process is long and exhausting with high probability of error. We present in this paper a method to ease the interpreter’s task of analyzing dozens of seismic attributes by integrating all the information into just one map, this map, the similarity map, shows the resemblance of the seismic response of each region of the whole study area with respect to a selected location in the field. In this paper, 3D seismic data in the study area are interpreted using well data. In addition, seismic inversion was conducted in order to estimate the porosity distribution based on the acoustic impedance within the study area. Moreover, an attempt was made to predict the effective porosity by designing a probabilistic neural network (PNN) and simultaneously using seismic attributes and effective porosity logs in the reservoir window. This was done by deriving a multi-attribute transformation between an optimum subset of seismic attributes and effective porosity logs. Seismic traces close to the well locations were used to generate seismic attributes. Effective porosity logs at the reservoir area were the target logs in this study. A set of seismic attributes were generated using HRS software and a forward stepwise regression process was used to determine an optimum subset of attributes to be utilized in the training of neural networks. Ultimately, we obtained a porosity map of the studied area. The inputs of the similarity analysis included a set of uncorrelated seismic attribute maps, the coordinates of the control point, and the radius around the control point that circles an area (the reference zone) of nearly constant attribute response. Four different attribute volumes generated were then used in the study: instantaneous amplitude, instantaneous phase, instantaneous frequency, and acoustic impedance. A horizon-slice at the reservoir was extracted from each of the attribute volumes. First, Well 08-08 (a high producing well) was chosen as a reference well. The selection of the reference well could be the highest production well, the lowest production well, a dry well, or any other classification depending on the objective of the analysis. The objective was to map the reservoir of the field based on the reference point 08-08 for possible high production areas. A radius value around the well was then chosen to calculate the mean and the standard deviation of the reference point within the radius from the extracted horizon slice for each of the attributes. The output of the first step was (N) different reference means and reference standard deviation for the same reference point; (N=4) is the number of attributes that were used in the study. The next step was to calculate a zero-one matrix from the extracted horizon-slice for each attribute based on a statistical criterion that would assign either zero or one to every node for a given horizon-slice. Finally, zero-one maps were integrated into one single map. The four attributes revealed different information and their zero-one maps showed different distributions that help the interpreters correlate each map to other types of information such as production or geologic information. The final map was obtained by integrating the zero-one maps. Studying the results obtained from the “similarity map” and “porosity map” in reservoir zone presented a convincing correlation between the two maps found through different methods each having specific information for the interpreters and helping them make more reliable decision to choose a prospective point, with less drilling risks.