Data-Driven Ice Blockage Estimation of Water Intake at Niagara Hydropower Station

Frazil ice and anchor ice could cause ice blockage at water intakes, which results in shutting down intake facilities and/or lowering the active power output of the hydropower plant. This paper proposes a data-driven ice blockage estimation method for the water intake. Specifically, the flow velocity coefficient of the conduit is calculated from water levels at the water intake and forebay, and water flows at hydropower units in conjunction with the water storage and leveraged as the index for ice blockage. Detection conditions and rejection conditions are showcased to extract only highly probable ice blockage events and exclude false positives, respectively. The ice blockage estimation performance was examined with eleven years of historical data at Niagara hydropower stations. System operation logs demonstrate that the anchor ice, lake ice, and shore ice were continuously observed 4–15 hours before the three detected occurrences, which serve as evidence of the validity of the proposed method.