Numerical Simulation Of Hydraulic Characteristics And Fish-passing Effect Of Yujiang Laokou Shipping Hub Fishway Project

In this paper,the newly built Yujiang Lao Kou shipping hub fishway project was taken as the research object.The three-dimensional numerical simulation was used to study the hydraulic characteristics of the three local models of the Lao Kou fishway.The velocity range and turbulent kinetic energy range in the mainstream of the fishway suitable for fish tracing in the Yongjiang reach were analyzed;Through the introduction of the advanced fish counting system River Watcher,a fishway monitoring system was constructed,which combined with artificial net catching methods to collect fish data and analyzed the passing effect of the Lao Kou fishway.Finally,a large number of fish image samples establish a fish identification model to realize the automatic identification and classification of fishes,and improve the efficiency of fishway monitoring.The main research results are as follows:(1)Through three-dimensional numerical simulation,the Lao Kou fishway has good flow state,moderate flow rate and low turbulent kinetic energy,indicating that the engineering design part of the fishway is reasonable;The velocity range in the mainstream area is 0.3?1.2m/s and the turbulent kinetic energy range is 0.013?0.055J/kg,it has basic conditions suitable for fish tracing.(2)Through the field test of the Lao Kou fishway,according to the River Watcher fish counting system,a total of 4148 fish were traced back.After manual statistics and identification,there were 15 species of fish,which belonged to 4 orders,7 families and 14 genera.The main fishes are Perciformes(37.76%),Cypriniformes(34.83%)and Siluriformes(22.58%).The main species are Linnaeus(body length is 11?24.3cm,36.9%)and Xenocypris davidi Bleeker(body length is 16.6?40.8cm,29.14%).(3)After field experiments,the optimal gate opening of Lao Kou fishway is 1.3m,corresponding to the flow rate in the fishway is 0.426m3/s;the time period during which Lao Kou fishway passes the largest number of fish during the day is from 8:00?10:00,In the evening at 18:00?20:00;The body length of the fish with the largest number of backtracking ranges from 20 to 50 cm(76.03%).(4)Based on the MATLAB software system,a fish recognition model based on support vector machine was established.After fish image feature extraction and model training,the fish in the Lao Kou fish pass were identified.After inspection,the accuracy of the model was 95.6%,combined with the fish pass monitoring system,can realize the automatic identification and classification of retrospective fish,and improve the efficiency of fish pass monitoring.