Experimental evaluation and modeling of sedimentation basin performance using skimmer-type dewatering control devices

The effectiveness of four sediment removal/water control treatments was evaluated in a 155-m3 sedimentation basin using 100-m 3 inflow hydrographs (35-mm, 2-yr runoff events) that contained 454 kg of sediment, which simulated runoff from a 0.4-ha denuded construction site. The four treatments were two alternate configurations of the floating riser, or Faircloth skimmer, and two modifications of the skimmer that were designed to minimize the outflow from the basin during the inflow event. Each treatment was designed to dewater the basin in 24 h. The first configuration of the skimmer had the control orifice positioned at the base of the device; it discharged 19.2 of the 454 kg of sediment injected. The second configuration of the skimmer had the control orifice positioned at the top of the arm; it discharged 14.7 kg of sediment. One modified skimmer, called the “Skimmer-Chute-Tube”, had an elongated arm and a tube positioned perpendicular to the arm; it discharged 15.4 kg of sediment. The other modified skimmer, called the “Skimmer-Chute”, had an elongated, parallel second arm that enhanced flow at shallow water depths; it discharged 15.7 kg of sediment.; Two reactor theory-based models were used to predict the hydraulic and sediment loss responses of the sedimentation basin using each of the skimmer-type dewatering control devices. One sedimentation basin model was based on plug-flow reactor theory, and the other was based on continuous-flow stirred tank reactors in series (CSTRS) reactor theory. Both models were developed before the use of the skimmer as a dewatering control device, and the model results were compared to observed data for the basin using the four sediment removal/water control treatments. Neither of the models was able to accurately predict the sediment loss from the basin using the skimmer-type devices: the error ranged from 17 to 184%. A calibration factor was derived from the results predicted by the plug-flow reactor model and used to correct the results from the CSTRS model. The calibrated model was able to adequately predict the sediment loss from the basin: the error was less than 12%.