As seismic reflection waves pass through the different layers and formations ofthe Earth, they are affected by the attenuation phenomenon that occurs after passingthrough each layer. One of the most effective and important criteria that can be used inthe assessment of attenuation is to check the amount of the Q-value. This value can beused to monitor the amount of attenuation. A key point to remember is that thecalculation of Q is always associated with various computational and operationalchallenges; in other words, the value of Q cannot be calculated in all of the wells that arein a hydrocarbon field.The purpose of this paper is to present an approach to the problem of estimatingthe Q-factor by using the latest artificial intelligence method, which is deep learning. Byusing the CUDNNLSTM algorithm in this paper, we were able to estimate the Q-factoraccurately. we achieved an accuracy of 98.5 and a validation loss of 1.3 in estimatingthe Q-factor. With our Q-factor estimating by deep learning, along with speeding upcalculations, we will be able to resolve the problem of lacking suitable VSP seismic datato calculate the Q-factor, as well.
展开▼
