Flushing culverts

Phase I (2016):

Laboratory measurements were conducted in a wave flume at the Hydrotechnical Laboratory of the Faculty of Civil Engineering, University of Zagreb. Inside the wave flume with a wave generator, a breakwater physical model with embedded flushing culverts is placed. The wave generator is a "piston" type generator that can produce regular and irregular waves. The dissipation chamber is located at the opposite end of the roughly 18.5 m long and 1 m wide channel in order to minimise the wave reflection of transmitted waves on the back plate of the flume.

The breakwater physical model was constructed with a seaward slope of 1:1.5 and covered with boulders with an average diameter of 10 cm. Inside the breakwater, flushing culverts of various lengths and diameters were placed and tested. The culvert diameters correspond to 60 and 100 mm with three different lengths of 92 mm, 142 mm and 192 mm. Each geometrical shape of the culvert was subdued to a set of 27 wave conditions. There were 3 different water levels inside the wave flume and for each water level a set of 9 irregular waves were tested. These water levels were at the top of the culvert (W1), at the axis of the culvert (W2) and at the bottom of the culvert (W3). A total of 162 tests were conducted with spectral waves and breakwaters with one embedded flushing culvert.

Figure 2. Photo gallery of the wave flume and breakwater scale model at the University of Zagreb

Additionally, tests with groups of 4, 6 and 8 culverts were also conducted. The four-culvert model was constructed so that all four culverts were in one row while in the six and eight-culvert models, culverts were positioned in two rows.

Phase II (2017):

Field measurements were conducted in ACI marina Opatija in northern Croatia, near the city of Rijeka (Figure 3). The marina primary rouble mound breakwater has 8 parallel 1 m diameter culverts placed at the southern part of the marina with the intention to improve water renewal inside the basin. The culverts were placed so the top of the circular culvert would correspond to the mean sea level, while the average depth inside the marina is approximately 5 m. The protected basin encompasses approximately 350 m alongshore and 150 m cross-shore, forming an area of about 40 000 m2.

Figure 3. Photo gallery taken at ACI marina Opatija during measurement device deployment/retrieval

Two measuring campaigns were conducted, one during the winter and one during the summer. The first measuring period was done between February 2nd and March 31st and the second measuring campaign was performed between July 4th and August 31st. Five Acoustic Doppler Current Profilers (ADCPs), one portable flow measurement system for pipes (PCM), one anemometer, one CTD probe and a time lapse camera were placed in ACI marina Opatija during the measuring periods.

Underwater review of the measuring devices near flushing culverts were conducted after every deployment and separate CTD vertical salinity and temperature measurements were conducted at the positions of all the measuring instruments (Video 2). A sanity check of various natural influences (e.g. waves and rain) could be also performed utilizing the visual observations recorded with the time lapse camera (Video 3).

Phase III (2018):

Numerical models will be utilized to investigate the flushing culvert performance in several conditions, namely: the impact of the layout of the port and the effect of horizontal positioning of the flushing culverts during varying wind and wave conditions. As a comparative parameter for determining the quality of a particular flushing culvert solution, residence time will be used. Initially, a value of 1 will be assigned to the concentration of a passive tracer inside the enclosed basin and a value of 0 outside the basin (Video 4). Due to seawater exchange, a reduction of the initial passive tracer concentration is observed. By analysing the results of each simulation carried out, corresponding "e-folding" times are obtained which represent the time required for the concentration of the passive tracer to reduce to 37% of its initial value. The residence time is actually equal to the "e-folding" time according to Kreek, 1983.

Further numerical modelling will be conducted in order to optimize the flushing culvert design in terms of cross section, number of rows and columns in a flushing culvert group, etc. It is in the best interest to maximize the flow rate through the flushing culvert and minimize the potential negative effects of overwhelming wave energy penetration. A 3D volume of fluid model will be utilized in order to assess the criteria for the flushing culvert.