DAM BREAK ANALYSIS OF NALGAD HYDROPOWER PROJECT IN NEPAL USING HEC-RAS

Abstract

This study mainly focuses on dam breach analysis of Nalgad Hydro-Power dam. Nalgad Hydro-Power Dam is a concrete faced rock-fill dam with clay core of 200 m height and 546 m length and designed to generate hydroelectric power, which has crucial part in development of Nepal. The results of dam breach analysis are useful in minimizing the loss of life and properties due to likely dam break flood in the event of failure. In Nepal, there is no law existing and practice of studying dam break floods, and thus, this study is perhaps the first for Nepal. This work focusses on predicting the breach outflow hydrograph of Nalgad Hydro dam and its routing through downstream valley employing a computer software for determination of consequences of dam failure. The computer program of U.S. Army Corps of Engineers HEC-RAS 5.05 software models such floods in two ways namely 1-D and 2-D. In 1-D dam breach flood analysis, the essential parameters involved in reservoir and river routing techniques were estimated manually externally. Dam breach parameters include time of failure of breach, side slope of breach, bottom breach width, manning roughness coefficient, shape of breach and boundary condition. The unsteady hydraulics of the dam breach due to overtopping failure mode was modeled using 1-D approach. The model results show a peak flow of 140350.40 m3/s at the dam site for overtopping mode of failure and it is 39.84 times greater than PMF of 3523 m3/s. The attenuation of discharge from head to tail is 27.07% and the peak discharge reaches the tail end, at 47.66 km, after 1.1 hour of attaining peak discharge at distance 0.51 km d/s of dam. To route the downstream valley both hydrologic and hydraulic routings were undertaken. Hydrologic routing employs the continuity equation and an analytical or an empirical relationship between storage within the reach and discharge at the end whereas hydraulic routing employs the continuity equation and both energy and momentum balances to calculate open channel flow profiles. 1-D dam break scenario is further compared with two other 1-D scenarios namely natural flow, i.e. no dam case and without dam break case. All the results computed by dam break case is higher than other two cases but time of arrival of peak discharge. The arrival time of peak natural flood computed by HEC-RAS 1-D model is 6 hour, 6.10 hour, 6.30 hour, 6.4 hour, 6.4 hour, 7.30 hour, 8.40 hour, 9.20 hour, 10.22 hour and 11 hour at the selected distance 0.51 km, 1..98 km, 11.31 km, 12.84 km, 14.37 km, 22.91 km, 31.63 km, 35.59 km, iv | P a g e 43.12 km and 47.66 km d/s respectively. Similarly, the arrival time of dam break peak flood is 7.2 hour, 7.2 hour, 7.3 hour, 7.4 hour, 7.4 hour, 7.5 hour, 8 hour, 8.10, hour, 8.20 hour 8.30 at the same selected distance. This shows that the arrival time computed in natural flood case is earlier than the dam break flood case. up to distance 22.91 km (at station no. 24875.10). After this distance, the arrival time of dam break flood is earlier than the natural flood case. Flood plain mapping for the downstream reach of Nalgad Hydro Dam was performed using maximum water surface elevations on the XS cut lines, within the limits of the bounding polygon and flow affected settlement area is found with help of remote sensing and GIS. The sensitivity analysis was performed using time of failure, bottom breach width, side slope of beach and relative effect of one parameter on the resulting peak discharge. The model results show that breach formation time is more sensitive than bottom breach width and side slope of breach.In 2-D analysis, the peak hydrograph of the first station from dam axis computed by 1-D is given as u/s boundary condition and friction slope of tail end reach is given as d/s boundary condition and then model is allowed to simulate unsteady flow. The computed peak discharge varies from 140001 m3/s to 102600 m3/s, i.e. the attenuation of discharge from head to tail is 26.71% and the peak discharge reaches the tail end, i.e. at 47.66 km after 2 hours of attaining peak discharge at distance 0.51 km d/s. When river flows in steep slope up to distance 14.36 km, the flow affected settlement area computed by 1-D is slightly higher than 2-D and also water depth is decreasing gradually but after 14.36 km, when river meets flatter slope, the flow affected settlement area computed by 2-D is higher than 1-D. The total flood affected settlement area computed by 1-D is 3.16 km2, and by 2-D it is 4.3 km2 which is 35.13 % higher than computed by 1-D. The depth of water computed by 2-D is abruptly increasing in flatter slope at distance 22.91 km and 47.66 km from dam axis which is not reliable. Hence, for steep reaches 1-D model performs better than 2-D. Steep streams are gravity driven and have small overbanks. It means rivers and flood plains where the dominant flow directions and forces follow the general river flow path. Such type of nature of river is obtained in hilly zone and in such cases 1-D modeling is preferable.