Site 1, an approximately 3,000 ML dam from the Breacher Validation Study, was selected for a detailed case study utilising the Breach Hydro.
Comprehensive mode was initially run with Breacher to check for a similar dam with a historical failure, Bradfield (Figure 24) was identified.
Comprehensive mode also provided feasible breach ranges (peak flow, width and time) according with the preprogrammed literature (Figure 25). From this, a breach bottom width of 30 to 55m and time of 0.4 to 0.8 hrs was selected.
1,000 simulations were then run with the following randomised parameter ranges:
· Breach_Time = [0.4, 0.8]
· Failure_Elev = [38, 63.5]
· Breach_Bot_Width = [30, 55]
· Weir_Cd = [1.2, 1.6]
· Orifice_Cd = [0.2, 0.5]
Resulting hydrographs from the 1,000 runs were extracted in Breacher-Post (Figure 26). The Breacher-Post, Binned Data tool was used to the critical failure elevation (elevation producing on average, or median, the highest peak flow). It was found for this dam and the SDF, R.L. 49 m produced the highest peak flow (Figure 27).
Breach failure elevation was fixed to R.L. 49 m (Failure_Elev = 49), previous runs and the summary file were cleared, and 100 new simulations ran.
Binning orifice coefficient of discharge (Orifice Cd) found no obvious correlation between selected Orifice Cd and peak flow (Figure 28). As such a value of 0.3 was adopted.
A similar approach was adopted for weir coefficient of discharge, which aided in identifying a Weir Cd of 1.4. With only 2 randomised parameters remaining, Breach Time and Breach Bottom Width, results were once again cleared and a final 1,000 simulations ran.
Peak flow distribution for the 1,000 runs is provided in Figure 29and peak flow distribution relative to breach time and bottom width is provided in Figure 30.
Interestingly, the median peak flow (4,355 m3/s) is higher than that predicted in the comprehensive literature assessment (between ~ 2,000 to 3,000 m3/s). This isn’t unexpected, as historical larger dam failures carry substantial uncertainties regarding the estimate of breach time and peak flow, this was well documented in (Azmi and Thomson, 2024).
The above being said, the median peak flow is less than 50% above the upper range estimated in the literature assessment and 80% above the matched historical dam (Bradfield).
While some can consider this a reasonable validation of the median peak flow, ultimately the judgement of the modeller is required at this point to select the most appropriate peak flow, corresponding parameters and be satisfied of it’s suitability for use.