Abstract: Constant T² cuttoff was mainly used in the determination of irreducible water saturation with NMR log information before. Some new methods, such as taper T²cuttoff, film model and frequency model, have been developed recently. As for the clastic reservoirs with variable lithology and complicated pore structure, the methods using conventional log information or using NMR log information based on T²cuttoff all can’t obtain the irreducible water saturation meeting the require of log evaluation and interpretation. In this instance, film model or frequency model could be used. Both models suggested that irreducible water was existed in all pore size, and the percentage of irreducible water varies with the size of pore. Bur there has not been a criteria for that of what the size intervals of pore should be used in the determination of irreducible water saturation with the two models in practical application. Some attempts about the problem had been done upon the clastic reservoirs with variable lithology and complicated pore structure in Hailaer Basin. Firstly the distribution of power of 2 is used in multi-exponential inversion of NMR echo trains to obtain the T² spectrum. Then nine pore volumes selected from T² spectrum of NMR, together with the difference of neutron and density porosity reflecting the volume of clay, were used to calculate the irreducible water saturation with regression. Preferable results had been achieved in its application of clastic reservoirs of Hailaer Basin, the irreducible water saturation calculated had been improved compared to the core analysis data. Especially in Beier area, the average relative error between the irreducible water saturation predicted and the core analysis data is reduced from 15.1% to 8.7%.

Key words: Hailaer Basin, irreducible water saturation, NMR, difference of neutron and density porosity