采用光滑有机玻璃作为相似材料进行某渡槽水面线模型试验，发现模型试验测得渡槽最高内水位与均匀流公式计算出的水位相差较大。针对此问题开展模型试验、理论分析和数值模拟研究。分析影响模型试验中渡槽水位的各种因素；对模型试验所用的光滑有机玻璃材料进行糙率率定，基于率定试验结果和理论分析，建立一种糙率比尺不相似情况下的水位修正方法；开展渡槽水面线的三维有限元数值模拟研究，在渡槽模型试验实测数据验证数值模拟准确的基础上，证明针对糙率比尺不相似提出的模型试验渡槽水位修正方法的正确性。结果表明：试验糙率比尺不相似是导致渡槽实测水位与均匀流计算水位相差较大的关键因素，光滑有机玻璃板糙率取0.007 9~0.008 3较为合适，糙率值随流量的增大呈现减小的趋势，随坡度的增大呈现增大趋势。

To address the issue of variations in normal water levels across different segments during the operational phase of the canal characterized by a gentle slope and curved trajectory, a smooth organic glass material was utilized to simulate the water surface line in model tests conducted on the canal. The experimental approach adopted a normal distribution model and was formulated on the basis of gravity similarity principles, employing a geometric scale ratio of 1:40. The outcomes of the experiment indicated a notable disparity between the water levels recorded in the flume model test and those calculated using the uniform flow formula. Specifically, the measured water level surpassed the calculated water level, prompting an investigation into the underlying causes of the elevated measured water level and the attainment of precise water level estimations within the spillway. This study employed a tripartite methodology encompassing model tests, theoretical analysis, and numerical simulations. The analysis of various factors influencing water levels in the aqueduct model test revealed several key aspects. Firstly, the curvature of the aqueduct impacted the flow pattern, resulting in elevated water levels. Secondly, discrepancies in roughness ratio scales were observed, with the model's roughness being excessive, leading to heightened head loss and subsequently increased water levels. Lastly, the discrepancy between the actual aqueduct section, which featured a reduced water area, and the rectangular section utilized in the uniform flow calculation formula contributed to the rise in water levels. Subsequent roughness determination tests were conducted on the smooth organic glass material utilized in the model test, and the obtained roughness data were scrutinized. A water level correction method was later devised based on the calibration test results and the uniform flow calculation formula to rectify situations where roughness ratio scales were contradictory, thereby correcting the measured water levels. Furthermore, a three-dimensional finite element numerical simulation study was conducted to analyze the water surface line of the flume. The accuracy of the numerical simulation was verified against the data obtained from the flume model test, affirming the efficacy of the proposed water level correction method for addressing dissimilar roughness ratio scales. The study revealed that the water level at the bend of the fishway was 8-10 centimeters higher than that of the straight section. As the distance from the bend increased, the water level gradually decreased until it aligned with the calculated water level, indicating that the bend no longer influenced the water level. Additionally, the water level of the aqueduct's cross section design was 4-6 centimeters higher than that of a rectangular cross section. The primary factor contributing to the substantial disparity between the measured water level in the aqueduct and the calculated water level for uniform flow was the dissimilarity in roughness ratio scales. The roughness coefficient was not merely indicative of wall surface roughness, but was also influenced by hydraulic factors and water flow characteristics. The surface roughness of the polished organic glass plate ranged from 0.007 9 to 0.008 3. The data indicated a decline in roughness as flow rate rises, and conversely, an increase in roughness with steeper slopes.