The origin of formation water salinity variation in Chang 9 stratum, Jiyuan oilfield, Ordos basin is studied here. 91 formation water samples show that water salinity is characterized by a wide range and a complex plane distribution. In order to find out the main cause of such distribution complexity and reveal the relationship between formation water and evolution of reservoir traps, core data, chemical analysis result of formation water and log data are analyzed from perspectives of diagenesis and tectonism. And then, their characteristics are presented as the followings. In high salinity area, tuffaceous mudstone interlayer is found growing. Besides, the condition of Na++K+ is opposite to that of Ca2+, for its rate of concentration increase slows down with total salinity accumulating. In low salinity area, while, with fracture and faults developing, some formation water of CaCl2 type turns into MgCl2, NaHCO3 or Na2SO4 type. The cause is thus proposed to be composed of two aspects. One covers tuff alteration and later diagenesis for the high salinity. To be specific, montmorillonite, developed from tuff alteration, absorbs cation selectively and then ions migrate, during which more Na++K+ get lost, while more Ca2+ reserved. Afterwards, those reserved Ca2+ get released with montmorillonite transforming to illite, which results in a loss of Na++K+ and accumulation of Ca2+. Lots of ions are released into formation water during that process and later diagenetic process, which leads to the high water salinity. The other aspect is the development of faults and fractures, through which, the upper low salinity formation water gets connected. And that is the main cause of low salinity. At last, geological significance is discussed from two angles. Firstly, tuff alteration and later diagenesis are pivotal to reservoir reconstruction; and secondly, faults and fractures play an important role in oil transportation and storage.
The proliferation of IoT devices has led to an unprecedented integration of machine learning techniques, raising concerns about data privacy. To address these concerns, federated learning has been introduced. However, practical implementations face challenges, including communication costs, data and device heterogeneity, and privacy security. This paper proposes an innovative approach within the context of federated learning, introducing a personalized joint learning algorithm for Non-IID IoT data. This algorithm incorporates multi-task learning principles and leverages neural network model characteristics. To overcome data heterogeneity, we present a novel clustering algorithm designed specifically for federated learning. Unlike conventional methods that require a predetermined number of clusters, our approach utilizes automatic clustering, eliminating the need for fixed cluster specifications. Extensive experimentation demonstrates the exceptional performance of the proposed algorithm, particularly in scenarios with specific client distributions. By significantly improving the accuracy of trained models, our approach not only addresses data heterogeneity but also strengthens privacy preservation in federated learning. In conclusion, we offer a robust solution to the practical challenges of federated learning in IoT environments. By combining personalized joint learning, automatic clustering, and neural network model characteristics, we facilitate more effective and privacy-conscious machine learning in Non-IID IoT data settings.