In order to solve the problem that logging curve identifying complex reservoir lithology has multiplicity and uncertainty

this paper uses logs to build principal component variables and takes the principal components with the cumulative variance greater than 85% as input information

ensuring that input variables are minimum and information loss is minimal. and then simplifies the data structure to identify lithology. Based on measured values of gamma ray (GR)

deep resistivity (RES_BD)

resistivity (RES_BM)

shallow resistivity (RES_BS)

neutron porosity (TNPH)

bulk density (RHOB) and interval transit time (DTCO) in deep-water reservoir of the Qiongdongnan Basin

seven integrated variables were constructed. Principal component 1 and principal component 2

with the total explained variance of 85.11%

can effectively explain the comprehensive information of seven variables. Based on lithologic data obtained from sidewall core

siltstone

mudstone

and fine sandstone can be identified on the cross plot of principal component 1 and principal component 2. Additionally

the multi-fractal method was used to analyze the spectral function curves. The sandstone corresponds to spectral function curve leaning right and mudstone corresponds to spectral function curves leaning left

which can be used as valid evidence to distinguish sandstone and mudstone.