CASCADE DAMS OPERATION-DAM SAFETY
GO Kadapawo1, G Okello2, and D Kimera3
1. O&M Energy Uganda Limited-GPG, Uganda
2. Ministry of Water & Environment, Uganda
3. Busitema University, Tororo, Uganda
PRESENTER: KADAPAWO GERALD OPOLOT
ABSTRACT
High dams have been/are being constructed along the R.Nile for various uses thus creating a cascade
arrangement in the same flow valley. It is because of this setup that the possibilities of hazards resulting
from exceedingly excess discharge from one upstream Dam due to different reasons, puts the
downstream structures at danger as well as the environment.
Particularly, this study was prompted considering the prevalence of the Nalubaale Hydropower Dam
wall which surpassed its design life. This dam also has the largest reservoir in the world of 270 BCM.
This paper gives insights of real time management of a cascade. It covers design requirements,
modelling, simulation and instruments involved and their relevance in the automation of the spillways to
suit a safe existence of Cascade dams.
Implementation improved the operation and management of Hydro Power Plants in the following ways:
Reduce the chances of over topping in any of the dams in the cascade, Reduce chances of over flooding
of any one reservoir, Reduce the risk of dam wash away and failure of the structures, Reduce the
chances of unnecessary spilling of water in the reservoirs, Helps to guide decision making during
capacity tests and improves Emergency Preparedness and Response actions for Cascade dams.
2.0 METHODS AND APPROACHES USED
2.1 Conceptual design
To study the cascade system, one has to zero to two dams at a time. In this method, focus was given
to the first two Dam walls (taking Nalubaale and Kiira as one since they are in parallel) while basing at
Bujagali HPP.
1
Figure 1. Uganda Cascade dam system
2.2 River Nile flow model
The objective of this task was;
•
•
•
To develop flow model to the River Nile cascade of dams, assuming operations are to
maximize hydropower production and minimise the effect of catastrophe caused due to
anomalies of upstream dams.
To calibrate the model to match the output (flows in the reaches) from the existing actual data
of the respective dams in cascade for the most recent year available. Perform sensitivity
analyses on reservoir operation. Use HEC model to help identify most impactful reservoirs and
hot spots.
For simplicity of the Nile cascade, it is important to develop an ideal channel that gives similar
output as that generated from Inundation of the Nile and its reservoirs using HEC-RAS tool.
2.3 Data collection
2.3.1.1 Geometric Data
Data collection using remote sensing was mainly employed using GIS software packages as illustrated
below,
Figure 2. Geometric Data acquisition using Google earth V7
2.3.2
Hydraulic and Hydrologic Flow data
Hydraulic and Hydrologic information of Flow data describing the water and flow characteristics was
obtained through retrieval from the SCADA system display on the Historian.
Where,
Mean Annual Discharge, Q from upstream = 950 cumecs, Reservoir capacity at Bujagali Dam
=54,000,000hec cubic meters, Time of impact for normal flow
=45 minutes, Spillway capacities
and operation modes. Fitchner, (2015).Design manual for BHPP.
Since the Nile cascade system is an arrangement in series, flow from the three dams is expected to
almost be uniform, basing on Bujagali Dam as a reference point, the average flows provided from the
2
data base of the monthly data, there was characterisation of other discharges from the respective
cascade dams.
With reference to the continuity equation. (Conservation of mass)
Qentering = Qleaving
→ u1A1 = u2A2
Equation 1. Continuity equation/ energy balance
2.3.3 Chanel characteristics
Google earth software was used to obtain cross sectional and longitudinal contours of the channel and
respective reservoirs remotely. Data acquired from remote sensing using Google will have to be
substituted by actual field measurements.
River cross sections are given as sequences of station abscissas, 𝑠𝑖 , and bottom elevations 𝑏𝑖 , given
and a general cross section is approximated. Flow rate 𝑄𝑖 , in a trapezoidal element is
5
5
(𝑏𝑖+1−𝑏𝑖 )3 5
(𝑠𝑖+1− 𝑠𝑖 )3 {𝑧𝑖 −[
]}3
5
2
1
1
3
𝐵 𝐴𝑖
𝐵
2
2
2
𝑖, 𝑛
2 ]1/3
2 +(𝑏
−
−
𝑏
)
𝑠
)
[(𝑠
𝑛
𝑖𝑃 3
𝑖
𝑖+1
𝑖+1 𝑖
𝑖
𝑖
𝑄 =
𝑆 =
𝑆
Equation 2. Manning’s equation for composite trapezoidal channel
A simulation was done to an equivalent ideal trapezoidal channel of known parameters that would
behave in a similar way as the actual river Nile between the cascade dams. The simulation is simpler
and outputs as the actual channel.
The DEM of the watershed was developed using ARC GIS Version 10.1 within which the river Nile
channel flows, where validation and matching was taken for the simulated channel shown below.
The following assumptions were considered.
I.
Adopted a trapezoidal channel whole left and right banks are inclined at 45 (m=1) based on the
average contours and profiles obtained using Google Earth V2016.
II.
The channel is uniform and consistent.
III.
Negligible sedimentation levels
IV.
The wetted perimeter consists of earth material i.e. Manning’s value, n of 0.04
An excel sheet to calibrate and match the following parameters with real system data was developed.
Equation 3. Hydraulic trapezoidal channel equations
As the system works based on measured flow rate, the instrument measures the value of hydraulic head
H, of the channel which is an input to the Tool. Based on the value of H, all other relevant hydraulic
parameters are automatically computed. Which includes time (t) for before the effect is received on the
downstream dam reservoir, channel cross section (A), Wetted perimeter (P)
From Chezzy, V=C√𝑅𝑆
From Continuity, Q=AV
𝐿
Therefore, Time before effect received on the downstream reservoir T= Where L- distance from point
𝑉
of instrumentation to downstream reservoir, V- velocity of discharge.
Equation 4. Equations for simulated channel instrumentation
2.3.4 Characteristics and dimensions of the respective reservoirs.
Remotely obtained the generalised dimensions of the respective reservoirs and also requested for the
design values from one dam for matching remote data of the reservoirs. As shown in the figures. This
step once again must be substituted with actual field measurements for real implementation of the
project.
3
Figure 3. Nalubaale & Bujagali Reservoir geometries
Figure 4. Nalubaale & Bujagali Reservoir geometries
Basing on the known reservoir capacity of Bujagali and a computed theoretical capacity, It was possible
to compute the theoretical capacity for Nalubaale Dam reservoir which is contained within Lake Victoria,
also extrapolated to obtain its actually volume. A more detailed simulation using HEC HMS would
provide better data.
The volume was used to obtain a potential maximum flood generated once the dam wall collapsed given
the following assumptions;
The collapse would have occurred on a sunny day such that no rainfall inputs from run off
and direct rain will be added volume.
The worst case scenario is considered i.e. the dam wall collapsed abruptly as a whole
block as characteristic of gravity concrete dams. (not in phases)
A value of discharge is computed from the given channel characteristics and used as an input in HEC
RAS software simulation.
2.4 Simulation of the reservoir and channel
(Using HEC RAS software V5.03)
2.4.1
Hydraulic Modelling Using HEC-RAS
Had to follow seven Steps for Simulation using HEC-RAS Analysis as provided in the software users’
manual.
There were two main types of data to enter: Geometric Data, describing the river channel setting and
the Reservoir structure, and Steady & unsteady Flow data describing the water and flow characteristics
Then the Model is run and its outputs are viewed.
The processes can be done over and over again under different steady flow conditions until a desired
logical output is got.
4
Figure 5. HEC RAS simulation windows
Hydraulic computations are divided into three major parts.
1. Channel and Reservoir computations.
These include computations of the various channel parameters from an initial input of channel
hydraulic head H, whose final output is Inflow to the downstream dam Q . Other necessary
hydrological parameters can be calculated for analysis such as: Volume, Cumulative volume,
Time before impact is received downstream.
As shown in the table below.
2.4.2
Computation and simulation of the spillways hydraulics.
Studies were taken of the operation hydraulics of the spillway options including generating and nongenerating spillways, computed the maximum potential discharge for all spillway options and the
respective levels for which to optimise their functionality.
Optimisation of the necessary actions to be taken in case of anomalies in inflow from the point of
instrumentation upstream was done. This actions are real time.
2.5 Worst case scenario: flood from dam collapse
The purpose of this was to simulate the flood generated by the hypothetic collapse of an upstream
concrete dam, and investigate the effect on downstream dams in the cascade.
If the reservoir emptied its 270 billion cubic meters of water within an hour. What would happen?
2.5.1 Investigation of stresses and stability of Cascade dam, 0
Being a composite dam wall (Earth and gravity concrete), it is important to compute the various forces
acting on immediate downstream cascade dam. As shown in the steps below:
As obtained from the design engineering drawing of the Dam structure, from which a Model and a Free
Body Diagram obtained.
These model assumptions were considered,
1. Only four major forces acting on the Dam structure were considered i.e. Self-Weight (W), Water
pressure Ph, Uplift pressure FU, and Wave pressure Fwave
2. Wind pressure, Ice loads, Thermal loads were considered negligible
3. Siltation levels in the reservoir negligible.
4. The Dam was assumed to be a unit homogeneous block.
5. Only the 1-meter section of the concrete dam was considered.
5
Figure 6. 3D model of dam-Flood wave action on concrete section
Using the following equations, an excel spread sheet was developed programmed to give response to
various inputs on the forces acting on the Dam. For simplicity, no dam modelling tool was used here but
instead adopted a theoretical approach.
The main stabilizing forces of weight, water pressures and uplift in gravity dams was computed.
2.5.2
Investigation of flood generated from upstream dam collapse
In this case one upstream dam collapse was considered for simulation:
1. A flood wave obtained from sudden release of water from the reservoir of a collapsed dam was
analysed and its parameters used as inputs for water pressure on the Dam Structure Stability
analysis. The effects were studied.
2. Possible operational actions/remedies to minimise hazard/ impact will be generated.
Assumptions taken
A sudden simultaneous collapse
As the real case is, not all the emptied reservoir volume flows within the channel.
Uniform flow is assumed
If the reservoir emptied its 270 billion cubic meters of water within a given period of time. The collapsed
dam creates a moving wall of water in the channel. Behind this moving wall of water will be more water.
2.6 Development of the tool/program for SCADA
Designing a tool (desktop software) for optimum operation of the dams in cascade of river Nile in
Uganda.
1. The hydraulic data generated from the simulation and modelling of the cascade system was
matched with the generation characteristics for normal and abnormal conditions.
2. A code was developed based on the simulations and various operation actions taken at various
dams on cascade regarding generation.
3. A program was developed which is ready be incorporated into the SCADA systems at various
dams.
4. The program has the ability to process and advice an operator in any of the three cascade dams
about actions to consider under prevailing conditions in the cascade, especially during
emergency.
6
5. The program gives chance to the operator to take action within the first 2minutes and if no
response, then it initiates automatic operation of the spillways to relieve the structures as well as
prepare the reservoir for receipt of the flood wave.
The program mainly works based on the principle of General volume balance of the hydrological cycle.
Given below as,
[
𝜕𝑆
] = 𝐼(𝑇) − 𝑂(𝑇)
𝜕𝑇
Equation 5. General volume balance of the hydrological cycle
2.7 Results
Graphs comparing the action of a particular spillway as selected by the application in response to a
given hydraulic head, will show how effective the system is. The graph validates the accuracy of the
system since the line for cumulative volume discharged by the given spillway shall be analyzed and the
convergence or divergence shall be studied.
2.8 Project implementation
• validating :From the input data, a plot of cumulative volume for inflow at upstream and
outflow from the spillways is obtained and analysed
•
Testing: Different values of channel head, H are manually fed into the system, Its functionality
and capabilities are observed and Sensitivity of the program is observed. Values of out put
studied.
7
1.0 RESULTS AND DISCUSSIONS
This gives the details of the outcomes of the design and simulation project.
1.1 Programmable Flow model
Hydraulic modelling (Flow model developed). The Flood wave resulting from upstream Dam collapse
has the impact on the Dam analysed as shown in the graph below;
1. The Flow model was developed from HEC RAS.
2. The Ideal channel was developed, it is easier to program using an ideal simulated channel of
the real conditions.
3. Comparison of sensitivity between HEC RAS River Nile channel and Ideal River channel.
Figure 7. HEC RAS simulation
Figure 8. Ideal channel simulation-cross section
The figure of varying wave height for different inputs upstream was used. The effect analyzed as shown
below;
8
Plot of flood wave height against eccentricity e,over turning
FSO, sliding FSS and shear friction factor
SFF.
6
5
4
3
2
safe against over turning
1
safe against sliding
0
-1
12
13
14
15
16
17
18
19
20
21
-2
-3
Acting towards the upstream
WAVE HEIGHT, H
Eccentricity, e
over turning
Figure 9. Flood wave effect analysis for Gravity dam
The dam collapse upstream results into a flood wave whose impact is expressed inform of wave height
H, of the water upstream of the dam.
3.2. Validation and testing
From the graph below where action of the spillways as selected by the tool in response to an upstream
event, show that they efficiently managed to relieve the structure. The difference in the curves is time
between the upstream event and response.
cummulative volume
MASS FLOW CURVES
2000000
1800000
1600000
1400000
1200000
1000000
800000
600000
400000
200000
0
flap gate response to H=7m
inflow upstream(Cumecs-seconds)
time interval0.5
2
3.5
5
6.5
8
9.5 11 12.5 14 15.5 17 18.5 20 21.5 23 24.5 26 27.5 29
Figure 10. Mass flow curve for flap gate response to upstream flow variation
9
4.0 CONCLUSIONS AND RECOMMENDATIONS
4.1 conclusions
•
•
•
•
For programming purposes, an ideal Hydraulic model of the Nile River channel which gives
similar output as the Nile itself. (H=7, Q=955) had to be created. More than 98% efficient
calibrated within the region of most frequent anomalies
The Upstream dam collapse of Nalubaale Dam wall results into hydraulic failures. i.e. seepage,
breach/ wash away. It has no statically significant impact because even exposure of the wave
is not for a long spell. (Not statics failure) on the wall.
The Major focus should be on precision of water level control for dam safety against hydraulic
failures using real time system. The study has noted that quick response of the spillways saves
the structure from overtopping. This can only be achieved through the use early flood detection
systems as opposed to monitoring from rising levels at the reservoir.
The cascade dams’ application/ System is valid and implementable
4.2 Recommendations
•
•
Incorporate the application into automatic operation mode of spillways.
Link the tool with SCADA-Supervisory, Control and Data Acquisition system, Install instruments
that communicate with the PLCs at the plants through cloud.
•
System be customised for each dam at the cascade
•
On implementation, high tech level sensors be used for purposes of achieving precision
•
Since Dam safety has a function of settlement directly relating to classification of dams, flood
inundation to identify prone areas should be incorporated to reduce risks of life loss and
property/environment damage.
5.0 ACKNOWLEDGEMENTS
The authors thank Bujagali Energy Limited for their permission to publish this paper. The opinions and
views presented in this paper are, however, those of the authors and do not necessarily reflect those of
Bujagali Energy Limited.
6.0 REFERENCES
Howard, BC (2015). 4 Hidden causes of Dam failures.
Maria A, n.d (2012). Modelling and simulation. Book
Alley, T.WSV&T (2013). The catastrophic failure of dams in china 1975. Journal
Tribune, i.H (2006). Britain had secret to cut flow of Nile River- newly opened official file. s.l., s.n.. Book
Vision, N., (2015). Evolution of the 250MW Bujagali Dam.
Anon. s.l.:s.n, (2006). basics of hec ras river analysis system.
Anon. In: s.l.:s.n, (2016). Busitema Lecture notes for Hydropower Development & Planning.., p. 56.
Anon., n.d. (2017).From Wikipedia, the free encyclopaedia Nalubaale power station.
Prof.Anund Killingtveit, N. s.l., s.n, (2018). Dam Safety Management.
Lious Hattingh. s.l., s.n, (2018). Dam Safety Management-surveillance principles.
10
The Observer, m. y., (2016). Isimba hydro plant to start power generation in 2016. The Observer
(Uganda). Issue 23,pp 7-18
v.sundararajan, n.d. (2015).what is modelling and simulation engineering. pune 411007.
Vision, N., (2012). 0wen Falls Dam. Powering Uganda for Five Decades.
6.0 THE AUTHORS
Kadapawo Gerald Opolot graduate in Water Resources Engineering from Busitema University.
Experience in Dam safety management with publications. Currently a Civil and Maintenance engineer
for GPG-Naturgy (O&M Uganda ltd) the operators of Bujagali HPP and Achwa HPPs. Individual Member
of SANCOLD-2-300.
Eng. Okello Geatano PE MASCE MUIPE RE a Professional Civil Engineer - High Dams specialist
(MSc) HIACE; Sofia Bulgaria, Sanitary Engineering IHE-DELFT the Netherlands. Member of
American Society of Civil Engineers MASCE, A Corporate Member of Uganda Institution of Professional
Engineers MUIPE. Works with Ministry of Water and Environment as an Assistant Commissioner, Water
Use Planning and Allocation. Served as a Lecturer at the Civil Engineering department of Makerere
University; Head of Water Engineering department of the then Uganda Polytechnic Kyambogo now
Kyambogo University and most recently Lecturer at the department of Water Resources Engineering of
Busitema University.
David Kimera is an engineering graduate with a bias in maintenance and operations of marine systems
and structures. Working with Busitema University and Elite Consulting Engineers.
11