Exercises from book: Ch.

6 Hydropower
Wolf-Gerrit Früh

1 Hydro exercises § 6.2.4

1.1 Energy Density and specific energy

Calculate the energy density and specific energy of water in a reseroir 300 m above a turbine house.
edens <- rho*g*Z
espec <- g*Z

• Energy density E/V = mgZ/V = ρgZ = 2.943 × 106 J/m3

• Specific energy E/m = gZ = 2943 J/kg

1.2 Flow rate required

Calculate the flow rate of water requried to generate 270 MW from a static head of 300 , and also calculate
the flow rate of water required to generate 1 GW from a static head of 7 m.
Q1 <- P1/(rho*g*H1)
Q2 <- P2/(rho*g*H2)

• High head flow rate 92 m3 /s

• Low head flow rate 14562 m3 /s

1.3 Reservoir and pipe

Imagine a pipe with a length of 500 m and a diameter of 0.8 m leading from a reservoir 75 m above a turbine
house to a discharge valve.

1.3.1 Closed valve

Calculate the pressure at the valve if the valve is closed.

# Hydrostatic pressure
p <- rho * g * Z

The pressure is 7.3575 × 105 Pa = 7.4 bar gauge pressure (ie pressure difference to atmospheric pressure) or
8.3575 × 105 Pa = 8.4 bar absolute.

1.3.2 Opened valve

Calculate the velocity, volume flow rate, and power in the water jet if the valve is opened to the atmosphere
to a degree such that the outlet diameter is 0.4 m.
# Jet velocity from Bernoulli's equation
Vjet <- sqrt(2/rho * p)
# alternative way: this is the same as Vjet <- sqrt(2*g*Z)
D2 <- D/2

1
# Flow rate from continuity equation
Q <- pi * D2^2/4 * Vjet
# Power from hydraulic power
# Note capital P for power and lower case p for presssure:;
# also: p*Q is same as rho*g*Z*Q
P <- p * Q

• Jet velocity 38.4 m/s

• Flow rate 5 m3 /s
• Power 4 MW
Repeat the same when the valve is fully opened to an outlet diameter of 0.8 m.
D2b <- D
Qb <- pi * D2b^2/4 * Vjet
Pb <- p * Qb

• Jet velocity 38.4 m/s

• Flow rate 19 m3 /s
• Power 14 MW

1.3.3 Pipe friction

Assuming that the pipe is made from concrete, estimate the frictional loss of energy density, specific energy,
and head from the flow through the pipe.
First, calculate the head losses for the fully open valve (ie, velocity in pipe is same as velocity in jet because
cross-sectional area is the same). We assume here that the surface roughness of the pipe inner wall is 1 mm:
# using Colebrooke's formulay (eq. 6.13), we calculate the Fanning friction factor
f <- 1 /16/ (log10(eps/(3.7*D)))^2
# Calculating the nondimensional frictional losses
fLD <- 4*f*L/D
# energy density : scale by volume: 1/2 rho V^2
ploss <- fLD * 0.5 * rho *Vjet^2
# specific energy : scale by mass: 1/2 V^2
eloss <- fLD * 0.5 *Vjet^2
# energy density : scale by weight (mg): V^2 / (2g)
hloss <- fLD * 0.5/g *Vjet^2

• Fanning friction factor f = 0.0051868

• non-dimensional loss coefficient 4f L/D = 13
• energy density (pressure) loss 9.540433 × 106 Pa
• specfic energy loss 9540 J/kg
• head loss 973 m
Clearly, this tells us that the losses are more than the resource in the first place (so, in real life, the water
will never flow this fast. . . )
Repeating this with the valve half-closed, so that the flow rate is 4.8204768 m3 /s.
# Calculating the velocity in the pipe
Vpipe <- 4*Q/(pi*D^2)
# energy density : scale by volume: 1/2 rho V^2
ploss <- fLD * 0.5 * rho *Vpipe^2
# specific energy : scale by mass: 1/2 V^2
eloss <- fLD * 0.5 *Vpipe^2

2
# energy density : scale by weight (mg): V^2 / (2g)
hloss <- fLD * 0.5/g *Vpipe^2

• Fluid velocity Vpipe = 9.59 m/s

• energy density (pressure) loss 5.96277 × 105 Pa
• specfic energy loss 596 J/kg
• head loss 60.8 m
This is a bit better but still unrealistic.

1.4 Connect reservoir to reaction turbine

Now connect the pipe to a turbine and calculate the velocity of water with a dyanmic head of 7 % of the
elevation difference. From that, work out the volume flow rate and the flow rate of kinetic energy. The work
out the frictional head loss. Finally owrk out the reminaing pressure and static head ast the inlet to the
turbine, and the hydraulic pwer entering the turbine.
Hdyn <- reldynhead * Z
Vpipe <- sqrt(2 * g * Hdyn)
Q <- Vpipe * pi * D^2 / 4
Pkin <- 0.5 * rho *pi*D^2/4 * Vpipe^3
HL <- fLD * Vpipe^2/(2*g)
H <- Z - Hdyn - HL
PH <- rho * g * H * Q

• Dynamic head 5.3 m

• Water velocity in pipe 10.15 m/s
• Volume flow rate 5 m3 /s
• Power from kinetic energy (flow rate of kinetic energy) 263 kW
• Head loss 68.1 m
• Availabe head 1.7 m
• Hydraulic power 84 kW

1.5 Torque on turbine shaft

rpm <- 600

Calculate the torque on a turbine shaft rotating at 600 rpm.

T <- PH / (rpm/30*pi)

Torque is 1.3 kNm

On reflection, a pipe with diameter of 0.8 m has a far too small diameter, so that water flowing at a decent
volume flow rate has to flow at a high velocity, meaning that both, a lot of the available energy has to be put
into kinetic energy, and a lot of the initially available energy is lost through friction. . . .

2 §[Link] Pelton Wheel

Design one (or several identical) Pelton turbine(s) for a level difference of 202 m, linked to generators turning
at 250 rpm, wiht an available flow rate of 32 m3 /s.

3
Assuming that we do not want to lose more than 5 % from the available resource to friction in a penstock of
typical surface roughness 3 mm and that the length of the penstock is somewhat longer than the vertical
drop, for example, L = 300 m, we can estimate the required diameter of the penstock:
# using Colebrooke's formulay (eq. 6.13) and an initial guess for the
# penstock of 1 m, we estimate the Fanning friction factor
f0 <- 1/16/(log10(eps/(3.7)))^2
D5 <- 4 * f0 * L * 8 * Q/(g * pi^2)
D <- D5^(1/5)
f <- 1/16/(log10(eps/(3.7 * D)))^2
D5 <- 4 * f0 * L * 8 * Q/(g * pi^2)
D <- D5^(1/5)
f <- 1/16/(log10(eps/(3.7 * D)))^2
HL <- 4 * f * L/D * 8 * Q/(g * pi^2 * D^4)
# Available head:
H <- Z - HL
# Hydraulic power
Presource <- rho * g * H * Q
# expected power output
Pout <- Presource * eta
# Jet Velocity
Vjet <- sqrt(2 * g * H)
# Optimum Pelton Wheel diameter
DWheel <- Vjet/omega
# Power per jet from specfic speed
Pjet <- rho * (KN/omega)^2 * (g * H)^(5/2)
Njet <- round(Pout/Pjet)
NWheels <- ceiling((Njet/6))
NjetperWheel <- round(Njet/NWheels)

With a penstock of diameter 1.83 m, we get a head loss of 1 m, and an available head of 201 m. This gives a
hydraulic power of 63.1 MW or, with a typical turbine efficiency of 90 %, an expected power output of 56.8
MW.
From the available head, we can also calculate the jet velocity as 62.82 m/s. From this, we get the optimum
Wheel diameter of 2.4 m.
Using the expected nondimensional specific speed of KN = 0.15 for a Pelton Wheel jet, the expected power
per jet is therefore Pjet = 5.68 MW. This means that we can have in total around 10 water jets, arranged as
2 Pelton Wheels, each with a nominal power output of 28.4 MW and each with 5 jets.

3 §6.5.6 Suitable turbine selection

For that, download the turbine selection Excel spreadsheet from VISION and put those numbers in. . . .

4 §6.6.1 Pumped storage power station

A pumped-storage hydropower station has a level difference of 320 m between two reservoirs. The penstock
for a pump-turbine has a diameter of 3 m, and the design flow rate is 25 m3 /s.
In turbine mode, the efficiency is 85 %, while in pump mode, the efficiency of the pump-turbine is 70 %.

4
Estimate the round-trip efficiency of the installation (that is how much energy is needed for pumping a
certain amount of water into the upper reservoir compared to hwo much energy is extracted when that same
amount of water flows down again through the turbine).
Also calculate the volume capacity required from the reservoir to have enough water stored for one week’s
worth of continuous generation.

4.1 Solution

Solution process:
• calculate the head losses, assuming a surface roughness of 3 mm and a length of 480 m
• calculate the hydraulic power and power output
• within the choice of Francis-type machines (as they are the machines which work well as pump-turbines),
find a sensible specificaion for generator and number of turbines based on typical specific speed options;
here we try KN = 0.3, 0.5, 0.7, respectively. One rarely finds generators with fewer than 4 pole pairs.
On the other hand, increasing the number of pole pairs, increases the size and cost of the generator; so
let us look for generator with 4 to 6 pole pairs. Also, let us only explore up to 4 turbines.
# Fanning friction factor using Colebrooke's formula
f <- 1 /16/ (log10(eps/(3.7*D)))^2
# Head loss
HL <- 4*f*L/D * 8*Q^2/(g*pi^2*D^4)
# Available head
H <- Z - HL
# Hydraulic power
PH <- rho * g * H * Q
# Power output
Pout <- eta_g * PH
# determine number of pole pairs based on range of KN and number of turbines
np <- round( (2*pi*fel/ designKN * sqrt(Pout/rho) /(g*H)^(5/4)) %o% (1/sqrt(nTurb)))
# only select the ones within the chosen range
idx <- which(np >=npmin & np <= npmax,[Link] = TRUE)
nchoices <- dim(idx)[1]
select <- array(dim = c(nchoices,2))
for (i in 1:nchoices){select[i,] <- c(nTurb[idx[i,2]],np[idx[i,1],idx[i,2]])}
rpm <- 60*fel/select[,2]
KN <- rpm/30*pi * sqrt(Pout/select[,1]/rho)/ (g*H)^(5/4)
choices <- cbind(select,rpm, round(KN,3))
colnames(choices) <- c("N_turbines","N_pp","rpm","KN")

• Fanning friction factor 0.0049089

• Head loss 2 m
• Available head 318 m
• Hydraulic power 78 MW
• Power output 66.3 MW
Turbine selection:
## N_turbines N_pp rpm KN
## [1,] 1 5 600 0.694
## [2,] 2 5 600 0.491
## [3,] 2 4 750 0.613
## [4,] 3 4 750 0.501
## [5,] 4 6 500 0.289

5
To analyse the reverse flow, we need to make an assumption about the power consumption or flow rate. Since
the head losses are very small, let us assume that we drive the pump-turbine at the same power as the power
output in its generation mode. The result is that the reverse flow rate will be less, due to the frictional losses
and the machine efficiencies. That would mean that the head losses will be even smaller. However, with
current headlosses of 1 % of the head, reducing that will be a small effect, which we can correct in a second
step.
Then, to calculate the round-trip efficiency, we note that the energy stored in the reservoir is Estored =
M gZ = ρV gZ = ρgZQp Tp , where Qp is the pumping flow rate and Tp is the time duration of the pumping.
With the pump efficiency and the head losses, the energy input into the pump is Ein = ρg(Z + HL,p Qp Tp /ηp
Likewise, the energy released from the same amount of water in the form of volume flow rate Qg over a
period of Tg results in an energy production of Eout = ρg(Z − HL,g Qg Tg × ηg .
Since the volume is the same, V = Qp Tp = Qg Tg , the round trip efficiency is

ρg(Z − HL,g )Qg Tg × ηg Z − HL,g

ηrt = = ηg ηp
ρg(Z + HL,p )Qp Tp /ηp Z + HL,p

# initial assumption of reverse flow rate

Qpump0 <- Pout*eta_p / (rho * g * (Z + HL))
# resulting head losses
HLp <- 4*f*L/D * 8*Qpump0^2/(g*pi^2*D^4)
# corrected pumping flow rate
Qp <- Pout*eta_p / (rho * g * (Z + HLp))
# round trip efficiency
eta_rt <- eta_g * eta_p * (Z-HL)/(Z+HLp)

• initial flow rate: 14.69 m3 /s and corrected 14.75 m3 /s

• pumping head loss 0.69 m
• round trip efficiency 59 %
The volume to generate for a week is V = Qg Tg = 1.512 × 107 m3 .

5 §6.9 Major Hydropower stations

There is no single correct answer but a range of good ones (and obviously a range of not so good ones). I will
sketch possible way using some typical values and choices, and only one generator type and only a single
suitable option.
My choices (and you can and should explore other choices):
CL = 0.7
Peak_to_average = 2
np <- 6
fel <- 50
rpm = fel/np*60
V_av = 0.7
L <- 1.5 * Z
eta_g <- 0.9
relHL <- 0.1

• intended use of the plant for round-the-year load-following plant with an expected load factor of 70%
• peak load twice the average load
• generator chosen with 6 pole pairs for 50Hz mains frequency (500rpm)
• available volume from rainfall is 70%

6
 • Length of penstock 1.5 times level difference
• typical turbine-generator efficiency 90%
• estimated head losses 10%

5.1 Flow rates and expected installed capacity

This means
H_av = (1-relHL)*Z
V_annual = V_av*A_catch*h_rain
Q_average = V_annual / (365*24*60*60)/CL
Q_peak = Peak_to_average * Q_average
Capacity_estimate = eta_g * rho * g * H_av * Q_peak

• Available head 338 m

• Annually averaged flow rate 165 m3 /s
• Peak flow rate 330 m3 /s
• Estimated installed capacity 983 MW

5.2 Turbine type and rating

From the head, a standard turbine choice would be a Francis turbine. Choosing an appropriate specific value,
it is possible to get an estimate of the power rating of a single machine, using

p  2
PT /ρ KN 5/2
KN = ω 5/4
→ PT = ρ (gH)
(gH) ω

KN = 0.5
P_turbine = round(rho * (KN / (rpm/60*2*pi))^2 * (g*H_av)^2.5/1e6) *1e6
N_T = round(Capacity_estimate/P_turbine)

• Specific Speed $K_N = 0.5

• Turbine rating (rounded to the nearest MW) 58 MW
• Number of turbines N_T

5.3 Penstock

Q_penstock = Q_peak/N_T
f_Fanning = 0.007
H_f = round(relHL*Z/2,1)

Allowing for a penstock per turbine, each penstock would need to be able to take 19.3989316 m3 /s. With a
typical Fanning friction factor of fF = 0.007, ad assuming that the friction losses are about half of the total
losses, HL = 37.5 m and Hf = 18.8 m. Using

L 8Q2p 32f LQ2p Hf

Hf = 4f → D5 =
D gπ 2 D4 gπ 2
D5 = 32 * f_Fanning * L * (Q_peak/N_T)^2 / (g*pi^2*H_f)
D = ceiling(10*(D5^0.2))/10

7
The penstock for each turbine must have a diameter of around 2 m.
If we then were to run the scheme at average flow rate, we would turn off half of the turbines and run the
remaining at their rated power (because they are most efficient then). So, the output from the scheme is
simply the number of turbines running at their design point (Number of turbines times 58 MW)

5.4 One penstock for a pair of turbines

If we had chosen to pair turbines, so that a penstock feeds two turbines, and that one of each pair is turned
off when the plant is running at average flow rate, then would have had to size the penstock for operation at
full capacity at:
D5 = 32 * f_Fanning * L * (Q_peak/N_T*2)^2 / (g*pi^2*H_f)
D = ceiling(10*(D5^0.2))/10

The penstock for each pair of turbines would have a diameter of around 2.6 m.
The frictional losses at the average flow rate in a penstock would then be, and the resulting power output
from a turbine would then be
Hf_av = 4 * f_Fanning * L /D * 8 *(Q_average/N_T*2)^2 / (g*pi^2 * D^4)
HL_av = 2*Hf_av
P_av = round(eta_g * rho * g * (Z - HL_av) * Q_average/N_T*2/1e6)*1e6

• frictional losses 4.1 m

• head loss 8.2 m
• Single turbine output 63 MW
Compare this with the initial sizing . . .

5.5 Simple Payback period

For the payback period, I revert back to the one penstock per turbine choice and assume operating costs of
2% of the investment costs.
p_el = 0.5/100
Eout = CL*N_T*P_turbine/1000 * 365*24 # kWh
Sales = p_el * Eout
CAPEX = N_T*P_turbine
OPEX = 0.02*CAPEX
SPBP = CAPEX / (Sales - OPEX)

With a load factor of 70% and 17 turbines with a rated output of 58 MW, the scheme produces 6046 GWh.
• CAPEX 986 m£
• OPEX 19.7 m£
• Gross Income 30.2 m£
• Net Income 10.5 m£
• Simple Payback Period 94 years