In this study, a methodology combining multi-source data fusion, hydrological model improvement and data assimilation is proposed for estimating daily seamless root zone soil moisture (RZSM) at a spatial resolution of 30 meters. First, we produced more accurate seamless fused surface soil moisture data by triple collocation and least square merging methods. Then, the Richards equation-based soil moisture module was integrated into the SWAT (Soil and Water Assessment Tool) model. The modified model greatly enhanced the simulation of RZSM compared to the original SWAT model. Finally, the fused product was assimilated into the modified SWAT model using the ensemble Kalman filter algorithm to obtain the RZSM at 30 meters spatial resolution. The results showed that the assimilation further improved the soil moisture simulation capability of the model.

To accurately predict the remaining useful life (RUL) of rotating machinery while continuously providing the task data, a novel continuous RUL prediction methodology was proposed. The methodology comprises a self-guided attention convolutional neural network (SGACNN) and memory consciousness adjustment (MCA) mechanism. First, a multihead focal channel-wise self-attention (MFCWSA) mechanism was implemented to effectively capture the degradation information across all the channels and achieve the attentional focus. Next, the SGACNN was constructed using the MFCWSA, squeeze-and-excitation mechanism, and convolutional block attention module. A new network gradient direction was synthesized by leveraging the gradients from both the previous task and the current task. Further, a weight constraint loss term based on the gradient magnitude was designed to constrain the learning process of important parameters. With the new network gradient direction and weight constraint loss, a novel MCA mechanism was proposed and integrated into the SGACNN for implementing the continuous RUL prediction tasks. Finally, various RUL prediction experiments on the life-cycle bearing and gear data sets were carried out, and its outcomes were compared to those of the advanced methods of the same kind. The comparative results validated the superiority of the proposed methodology.

This paper proposes a flow-path network (FPN) model to simulate complex surface flow based on a drainage-constrained triangulated irregular network (TIN). The TIN was constructed using critical points and drainage lines extracted from a digital terrain surface. Runoff generated on the surface was simplified as ‘water volumes’ at constrained random points that were then used as the starting points of flow paths (i.e. flow source points). The flow-path for each ‘water volume’ was constructed by tracing the direction of flow from the flow source point over the TIN surface to the stream system and then to the outlet of the watershed. The FPN was represented by a set of topologically defined one-dimensional line segments and nodes. Hydrologic variables, such as flow velocity and volume, were computed and integrated into the FPN to support dynamic surface flow simulation. A hypothetical rainfall event simulation on a hilly landscape showed that the FPN model was able to simulate the dynamics of surface flow over time. A real-world catchment test demonstrated that flow rates predicted by the FPN model agreed well with field observations. Overall, the FPN model proposed in this study provides a vector-based modeling framework for simulating surface flow dynamics. Further studies are required to enhance the simulations of individual hydrologic processes such as flow generation and overland and channel flows, which were much simplified in this study.