AbstractAn industrial scale simulation method (ISSM) has been proposed to predict the contamination evolution and oil quality distribution in multiproduct pipelines. The proposed ISSM consists of two subalgorithms: computational domain tracking method (CDTM) and spatial domain conversion method (SDCM). The computational effort can be greatly reduced by CDTM. SDCM is implemented for flow simulation in industrial scale pipelines. The basic concept of SDCM is adaptively adjusting the spatial mesh resolution according to the spatial distribution of contamination. The communication between spatial meshes with different resolutions is achieved by machine learning technology in SDCM. The validation of training and predicting process in SDCM indicates the artificial neural networks (ANN)-based method could efficiently construct the implicit function between the contamination concentration and spatial coordinates in pipelines. Then, the detailed implementation of ISSM, which is a hybrid of CDTM and SDCM, is given. The accuracy and efficiency are validated by both the direct numerical simulation results and practical experiments in industry multiproduct pipelines. Compared with the direct numerical method, ISSM could greatly improve the computational efficiency. Finally, two applications, mixing dynamics of miscible oil interface and oil quality evolution within long-distance pipelines, are carried out. The results reveal that ISSM can capture the asymmetric distribution phenomenon of the contamination concentration along the pipeline. Furthermore, the quality index of oil product can be tracked within the present framework of ISSM, which could be potentially used to design a quality-based scheduling scheme for industrial multiproduct pipelines.