Author(s): Jieyu Li
Linked Author(s):
Keywords: Flood control system; Joint operation of reservoirs; Risk analysis; Spatial and temporal evolution; Yellow River Basin
Abstract: The dynamic changes in the flood situation increase the complexity of real-time flood control operation and risk analysis for the multi-reservoir systems. This paper proposes a two-layer dynamic flood risk analysis model for the flood control system in the middle reaches of the Yellow River, facilitating risk identification and management across both temporal and spatial dimensions. First, a flood risk index system is established, and a fuzzy recognition model (FRM) is applied to assess risk levels at reservoirs and flood control points over time, achieving the first-layer risk identification. High-risk periods are then identified, and an optimal flood control operation model is developed to enhance mitigation potential. Subsequently, a Bayesian Network (BN) refines risk evaluation for optimized operation schemes during these critical periods, forming the second-layer analysis. Application to a five-reservoir system in the Middle Yellow River reveals that during FRM-identified high-risk periods, optimal schemes involved greater reservoir releases, yet resulted in more stable flow at the flood control point. The system exhibited divergent risk profiles: a gradual increase at the Hekoucun Reservoir, a rise-then-fall pattern at the Huayuankou point, and no risk elsewhere. The two-layer model thus proves to be a timely and accurate tool for flood control decision-making.
Year: 2026