Accurate estimation of formation permeability by means of petrophysical logs

Permeability can be considered as the one of the main petro-physical parameters that plays an important role in commercial production of reservoir. On the other hand, measuring the permeability is actually a principal challenge for investigators. Inasmuch as, taking core samples from every well and also surveying well-tested data require a large amount of time and capital, using an economical process is more interesting and it is the main cause to utilize electronic logging as a repeatable method. Artificial intelligence-based methods and especially least squares support vector machines (LSSVM) are reliable and accurate models. In the present work, the LSSVM has been trained by the Cuckoo optimization algorithm to predict permeability by means well-logging data including five different types of logs as input data. The correlation coefficient between the model prediction and the relevant real data is found to be about 0.99602 that can be nominated as an accurate yield.