Final Report Davide Troiani Sea Waves and Hydropower

You might also like

Download as pdf or txt
Download as pdf or txt
You are on page 1of 15

Ph.D.

Course: Sea waves and hydropower


A.A. 2022/2023

EXERCISE
Numerical modelling of a single body wave energy

Introduction

When one has an idea of a new Wave Energy Converter, how to assess the annual energy production
of the WEC? The Annual Energy Production (AEP) is usually obtained by multiplying the expected
power output of each sea state (defined by 𝐻 , 𝑇 pairs) by its occurrence (in hours) and then summing
over all sea states (Fig. 1).

Power matrix Wave scatter diagram

Annual Energy
Production

Fig. 1. Annual energy production calculation

This means that it is necessary to characterize what is to be installed and the location where is to be
installed. Deployment sites are characterized by wave scatter diagrams, which are tables listing
occurrence of sea states in terms of significant wave height 𝐻 (𝑚) and wave peak period 𝑇 (𝑠).
These plots allow to achieve an overall understanding of wave resource and to identify the range of
sea states, which are expected at the site. Devices are characterized by the so-called power matrix,
which relates the average output power of the WEC to the wave field, characterized by significant
wave height and peak period.
Then, the question is: how to assess the power matrix of a WEC? To build the power matrix, we need
a numerical model of the device, to evaluate the absorbed power in function of given wave conditions.
It is a model of the WEC as a whole system, including different physics: wave mechanics,
electromagnetism and fluid-structure interaction. The model is based on the equation of motion of
the floating body, which is generally expressed as the ODE:

[𝑀]𝑋̈ = 𝐹 +𝐹 +𝐹 +𝐹 +𝐹 ( , ,… ) (1)

where 𝐹 is the force due to wave/structure interaction, including excitation and radiation force,
𝐹 is the force due to the power take off (PTO) system, which converts mechanical energy into
electricity, 𝐹 is hydrostatic restoring force, 𝐹 is the force due to mooring system and 𝐹
includes other forces, such as viscous damping force. Usually 𝐹 ≈ 𝐹 ≈𝐹 ≫𝐹 ≈
𝐹 .
CASE STUDY
The growing demand for sustainable and renewable energy sources has propelled the exploration of
innovative technologies to harness the untapped potential of different potential sources currently
unexploited. Among these pioneering technologies, Wave Energy Converters (WECs) have gained
significant attention for their capacity to generate electricity from the perpetual motion of sea waves,
which has been recognised to have the potential to become the largest source of energy. Indeed the
Intergovernmental Panel on Climate Change (IPCC) puts the potential annual global production at
29.500 TWh, almost ten times Europe’s annual electricity consumption.
The present report focuses on evaluating the wave energy potential of coastal regions of Alghero and
Mazara del Vallo, where the intricate interplay of geographical features and wave dynamics presents
a compelling opportunity for sustainable energy extraction. To provide a visual context, the map in
figure 1 showcases the distribution of wave energy along the Italian coast, emphasizing the strategic
significance of Alghero and Mazara del Vallo sites (circled in red). As it can be easily recognized by
the map those two specific points present the largest potential respectively between 51-110 MWH/m
for Alghero and between 41-50 MWh/m for Mazara del Vallo. This analysis endeavours to elucidate
the feasibility and potential of a single degree of freedom (heaving) Wave Energy Converters in these
coastal regions.

Fig 1: Map of the potential of wave energy along the Italian coasts, Rete Ondametrica Nazionale
In order to analyse more deeply the potential of these two locations, Wave Scatter Diagrams are
highly site-specific visual representation presenting the probability of occurrence of particular sea
states, providing useful information on the dominant sea conditions and thus resulting as a very
important instrument for the choice of the proper WEC for a specific location. In Fig. 2 the Wave

Davide Troiani 2
Scatter Diagram of Alghero (a) and Mazara del Vallo (b) are reported, with the occurrence in hours
per year of each sea state. Here it can be noticed how Alghero site is characterized by a broader cases
of sea states with the occurrence of high amplitude and long period waves, while in Mazara del Vallo
we have a more concentrated diagram.

Fig 2: Wave Scatter Diagram for the Locations of a) Alghero b) Mazara del Vallo
The same thing can be noticed if looking at the raw data given for the exercise and looking at another
representation of the scatter diagram. An interesting thing we can notice from Table 1 and Table 2 is
that the most frequent sea state in both cases results to be the same, that is

𝐻 = 0.75 𝑚; 𝑇 = 5.25 𝑠

Table 1: Alghero scatter diagram

Davide Troiani 3
Table 2: Mazara del Vallo scatter diagram
Computing the sea power for each sea state, (𝐻 , 𝑇 ) couple, as

𝜌𝑔 𝐻 𝑇 𝑘𝑊
𝑃 =
64𝜋 𝑚

It is possible to compute the wave power for unit of wave-crest length for each sea state, and plotting
the results it is possible to generate the sea power map. Obviously, the map below shows a redundant
information; it is clear that waves with higher wave height and period are the most energetic.

Fig 3: Sea Power

Davide Troiani 4
Then multiplying this value by the hours of occurrence registered in the wave scatter diagrams, it is
possible to compute the wave energy achieved for each analysed sea state in the specific locations
under examination, as reported in Fig. 4.

Fig 4: Wave annual Energy in the two locations of a) Alghero and b) Mazara del Vallo

Comparing Fig. 4, 3 and 2 some observations can be drawn. It is evident that the sea states exhibiting
the highest frequency of occurrence do not correspond to those with the highest energy levels. Hence,
the selection of an appropriate Wave Energy Converter for a specific location becomes paramount,
emphasizing the need to strike a balance between wave occurrence and energy levels.
The mean annual wave power is computed by summing the wave energy across all sea states (Annual
Wave Energy) and subsequently dividing the cumulative value by the total annual hours. The results
of this computation are reported in Table 1, where Alghero presents the higher mean annual wave
power, and the values in correspondence of both sites are in accordance with the results presented in
Fig. 1 obtained by Rete Ondometrica Nazionale.

Table 3: Annual Wave Energy and Mean Annual Power for Alghero and Mazara del Vallo sites

Davide Troiani 5
1 clc
2 clear all
3
4 %% Dati
5 rho = 1025;
6 g = 9.80665;
7
8 %% Mazara
9 filename = 'Mazara.txt';
10 data = importdata(filename);
11 wave_mazara = data(2:end, 2:end);
12 x = data(1, 2:end);
13 y = data(2:end, 1);
14 [X, Y] = meshgrid(x, y);
15
16 figure
17 subplot(1, 2, 2)
18 contourf(X, Y, wave_mazara)
19 title('Wave Scatter Diagram - Mazara')
20 xlabel('Wave period T_{p} [s]')
21 ylabel('Wave height H_{s} [m]')
22 set(gca, 'FontSize', 12)
23 h = colorbar;
24 set(h, 'ylim', [0 1200])
25
26 %% Alghero
27 filename = 'Alghero.txt';
28 data_alg = importdata(filename);
29 wave_alghero = data_alg(2:end, 2:end);
30 x = data_alg(1, 2:end);
31 y = data_alg(2:end, 1);
32 [X, Y] = meshgrid(x, y);
33
34 subplot(1, 2, 1)
35 contourf(X, Y, wave_alghero)
36 title('Wave Scatter Diagram - Alghero')
37 xlabel('Wave period T_{p} [s]')
38 ylabel('Wave height H_{s} [m]')
39 set(gca, 'FontSize', 12)
40 h = colorbar;
41 caxis([0, 1200])
42
43 %% Power Matrix
44 Sea_Power = ((rho * (g^2) * (y.^2) * x) / (64 * pi)) / 1000;
45
46 figure
47 contourf(X, Y, Sea_Power)
48 title('Sea Power [kW/m]')
49 xlabel('Wave period T_{p} [s]')
50 ylabel('Wave height H_{s} [m]')
51 colorbar;
52 caxis([0, 100]);
53
54 %% Annual Energy
55 E_mazara = Sea_Power .* wave_mazara;
56 E_alghero = Sea_Power .* wave_alghero;
57
58 cum_E_mazara = sum(E_mazara);
59 cum_E_alghero = sum(E_alghero);
60
61 figure
62 subplot(1, 2, 2)
63 contourf(X, Y, E_mazara)
64 title('Wave Annual Energy - Mazara [Kwh/m]')
65 xlabel('Wave period T_{p} [s]')
66 ylabel('Wave height H_{s} [m]')
67 set(gca, 'FontSize', 12)
68 h = colorbar;
69 caxis([0, 6000])
70
71 subplot(1, 2, 1)
72 contourf(X, Y, E_alghero)
73 title('Wave Annual Energy - Alghero [Kwh/m]')
74 xlabel('Wave period T_{p} [s]')
75 ylabel('Wave height H_{s} [m]')
76 set(gca, 'FontSize', 12)
77 h = colorbar;
78 caxis([0, 6000])
79

Davide Troiani 7
Numerical model of a single body heaving WEC

Let’s consider a simple wave energy converter composed of a floating vertical cylinder with one
degree of freedom (1 DOF) along the vertical axis (heave mode), connected to a power take off
system, which behaves as a linear damper (Fig. 2). This would resemble the Seabased’s wave energy
converter, developed by the Swedish Center for Renewable Electric Energy Conversion of Uppsala
University.
Assume also that sea states can be characterized by a single sinusoidal wave, having the same power
of the full wave spectrum. According to linear wave theory, a monochromatic wave of height 𝐻 and
period 𝑇 has the same power of a wave spectrum, defined by the couple 𝐻 and 𝑇 , if 𝐻 is equal to
the root mean square wave height (i.e. 𝐻 = 𝐻 = 𝐻 ⁄√2) and 𝑇 is equal to the energy period (i.e.
𝑇 = 𝑇 = 0.88𝑇 for Italian seas).

Buoy

Fig. 2. Sketch of a single body heaving WEC

Within the frame of the linear potential theory, the one-dimensional model of the device can be
written as the ODE:

𝑚 ∙𝑧̈ (𝑡) = 𝐹 (𝑡) + 𝐹 (𝑡) (2)

where 𝑚 is the total mass of the moving parts (i.e. buoy and PTO), 𝑧(𝑡) is the vertical coordinate at
time 𝑡 measuring deviations from the static equilibrium, 𝐹 (𝑡) is the resultant of the hydrodynamic
forces acting on the buoy and 𝐹 (𝑡) includes all the forces due to the PTO system.
Neglecting viscous effects, the hydrodynamic forces acting on the buoy are excitation, radiation and
hydrostatic force. Excitation force 𝐹 (𝑡) is the sea wave force acting on the buoy at rest and it is the
forcing of the system. Excitation force can be decomposed into two contributions: Froude-Krylov
force, due to undisturbed wave field, and diffraction force accounting for wave deformation due to
structure. For a regular wave of angular frequency 𝜔 and height 𝐻, it is given by:

𝐹 (𝑡) = 𝐶 _ (𝜔) 𝑐𝑜𝑠 𝜔𝑡 + 𝐶 _ (𝜔) (3)

where 𝐶 _ (𝜔) is force amplitude and 𝐶 _ (𝜔) is phase shift relatively to sea surface elevation.
Wave radiation force 𝐹 is due to waves propagated by the oscillating buoy in calm waters. It is
composed of two parts: one is proportional to buoy acceleration by means of added mass coefficient
𝐶 (𝜔) and the other is proportional to buoy velocity by means of radiation damping coefficient
𝐶 (𝜔). In case of a regular wave, it reads as follows:

𝐹 (𝑡) = −𝐶 (𝜔)𝑧̈ (𝑡) − 𝐶 (𝜔)𝑧̇ (𝑡) (4)

Finally, the hydrostatic restoring force due to buoyancy and gravity is given by:

𝐹 (𝑡) = 𝜌𝑔𝐴𝑧(𝑡) (5)


where 𝐴 is buoy waterplane area and 𝑔 is gravity.
The frequency-dependent coefficients of added mass 𝐶 (𝜔), radiation damping 𝐶 (𝜔) and excitation
force 𝐶 _ (𝜔), 𝐶 _ (𝜔) are usually calculated with commercial codes (e.g. WAMIT, ANSYS
AQWA), which solve the elliptic problems taking the advantage of a boundary element integration
method. They are estimated prior to model simulations and given as inputs to the ODE solver.
The simplest model for the PTO force 𝐹 is the linear damping:

𝐹 (𝑡) = −𝛾 𝑧̇ (𝑡) (6)

where 𝛾 is the PTO damping coefficient. 𝛾 is a parameter which can be optimized in real time
(i.e. continuous control) or for each sea state (i.e. slow control) or assumed constant for all sea states
(i.e. no control).
Concluding, in case of a regular wave of angular frequency 𝜔 = 2𝜋⁄𝑇 and height 𝐻 = 𝐻 ⁄√2 the
WEC dynamics can be estimated by solving the following equation (7):

𝑚 ∙𝑧̈ (𝑡) = 𝐶 _ (𝜔) 𝑐𝑜𝑠 𝜔𝑡 + 𝐶 _ (𝜔) − 𝐶 (𝜔)𝑧̈ (𝑡) − 𝐶 (𝜔)𝑧̇ (𝑡) − 𝜌𝑔𝐴𝑧(𝑡) − 𝛾 𝑧̇ (𝑡)

This equation is a second order ordinary differential equation (ODE), but available time integration
schemes are for first order ODE, only. To circumvent this limit we define the state vector:
𝑧
𝑍= (8)
𝑧̇

and then rewrite equation (7) as an ODE of the first order:

𝑍̇ = 𝑓(𝑡, 𝑍)
(9)
𝐻
𝑓(𝑡, 𝑍) = 𝑍(2); 𝐶 (𝜔) 𝑐𝑜𝑠 𝜔𝑡 + 𝛼(𝜔) − 𝐶 𝑍(2) − 𝜌𝑔𝐴𝑍(1) − 𝛾 𝑍(2) (𝑚 + 𝐶 )
2

This equation can now be solved by one of the MATLAB functions for time integration (e.g. ode15s,
ode23, ode45). For a given sea state, the simulation time span should be equal to at least 20 waves
(i.e. 20𝑇 ), in order to get a converged solution.
Once the equation is solved, the buoy velocity for the last simulated wave period can be used to
estimate the average power production:

〈𝑃〉 = ∫ 𝛾 [𝑧̇ (𝑡)] 𝑑𝑡 (10)

Simulating WEC motion for different sea states allows to derive the so-called power matrix. By
multiplying it with the wave scatter diagram of the selected location, the annual mean absorbed power
is estimated.

Davide Troiani 9
CASE STUDY

The Wave Energy Converter analyzed in this report resembles the


Seabased’s WEC developed by the Swedish Center for Renewable
Electric Energy Conversion of Uppsala University, whose schematic
is shown in Fig. 5. In this case the WEC consists of a buoy, placed on
the water surface, and a generator enclosed by a water tight
pressurized hull. The main motion exploited for energy production
purpose is the heaving, which, driven by the buoy, the translator is
moved inside the generator, and voltage is induced in the stator
windings.
The present report studies a more schematic WEC which resembles
the previously described Seabased technology. In order to do it we
will consider a simple wave energy converter composed of a floating
vertical cylinder with one degree of freedom (1 DOF) along the
vertical axis (heave mode), connected to a power take off system,
which behaves as a linear damper.
The study will be primarily focused on computing the power matrix
of the device, which is a bivariate matrix providing the device power
output as a function of significant wave height and wave period. The
comprehensive understanding of the power matrix constitutes a
pivotal aspect in the assessment of the potential energy output and
overall performance of each WEC, comprising device characteristics,
Fig 5: Seabased technology style and dimensions.
The power matrix is computed taking into account the different sea
states as inputs with its own couple (𝐻 , 𝑇 ), and once solved the ordinary differential equation of
motion of the device the power production can be computed.
In Fig. 6 the Power matrix of the considered single body heaving device is reported. As we can see
the produced power is higher for high waves with a small period.

Fig 6: 1D WEC Power Matrix

Davide Troiani 10
Below the computed numerical power matrix is reported

Table 4: Heaving WEC Power Matrix


The annual energy production of this specific Wave Energy Converter installed in the two locations
under analysis, Alghero and Mazara del Vallo, can be finally easily computed multiplying the power
matrix, characterizing the WEC, with the wave scatter diagram previously showed, characterizing the
site. Below the annual energy production matrix for Alghero and Mazara are reported, (expressed in
MWh) which take into account the hours of occurrence of each sea state.

Fig 7: Alghero annual energy matrix

Davide Troiani 11
Fig 8: Mazara del Vallo Annual Energy Matrix
Eventually, the annual mean absorbed energy can be simply computed summing the energy over all
of the considered sea states.

Table 5: Annual Energy Production


The two sites under analysis present a very similar annual energy production, around 15 MWh. As
we have seen in the first part of the report Alghero presents a higher value of annual wave energy but
considering the characteristics of the device and its ability to produce power, the installation is equally
favourable in both locations, otherwise it could be considered to change the type of device to exploit
the higher energy present in Alghero.

Davide Troiani 12
1 clc
2 clear all
3
4 g = 9.80665; % Accelerazione di gravità [m/s2]
5 rho = 1025; % Densità dell'acqua marina [kg/m3]
6 N_pT = 1000; % Numero di nodi per periodo d'onda, utilizzato per interpolare i
7 risultati delle ODE
8 N_waves = 20; % Numero di onde per simulazione (ossia, numero di onde per
9 raggiungere condizioni stazionarie) [-]
10
11 R = 2; % Raggio della boa [m]
12 A = pi * (2 * R)^2 / 4;
13 m = 5000; % Massa dell'intero sistema (boa + PTO) [kg]
14 gammaPTO = 20000; % Coefficiente di smorzamento PTO [N*s/m]
15 z0 = [-0.01; 0.01]; % Posizione e velocità iniziali della boa [m, m/s]
16
17 %% Alghero
18
19 filename = 'Alghero.txt';
20 data = importdata(filename);
21 wave = data(2:end, 2:end);
22 period = data(1, 2:end);
23 height = data(2:end, 1);
24 [X, Y] = meshgrid(period, height);
25
26 P = zeros(length(height), length(period));
27
28 for i = 1:length(period)
29 for j = 1:length(height)
30 Hs = height(j);
31 H = Hs / 2^0.5;
32 T = period(i);
33 Te = 0.88 * T;
34 w = 2 * pi / Te;
35
36 Cr = -199.62 * w^5 + 2933.2 * w^4 - 15212 * w^3 + 30997 * w^2 - 16110 * w +
37 2894.4;
38 Ca = 43.014 * w^5 - 982.58 * w^4 + 6782.3 * w^3 - 18415 * w^2 + 15112 * w +
39 17600;
40 Ce_mod = 438.29 * w^5 - 6138.3 * w^4 + 32288 * w^3 - 71801 * w^2 + 24263 * w +
41 121771;
42 Ce_ph = -0.0052 * w^5 + 0.0561 * w^4 - 0.2236 * w^3 + 0.1742 * w^2 - 0.0621 * w
43 + 0.0083;
44
45 tspan = [0 20 * Te];
46 options = odeset('RelTol', 1e-10, 'AbsTol', 1e-10);
47 [time, z] = ode15s(@(t, z) [z(2); (Ce_mod * H / 2 * cos(w * t + Ce_ph) - Cr *
48 z(2) - rho * g * A * z(1) - gammaPTO * z(2)) / (m + Ca)], tspan, z0, options);
49
50 t = (0:T/N_pT:20 * Te);
51 z_buoy(:, 1) = interp1(time, z(:, 1), t);
52 v_buoy(:, 1) = interp1(time, z(:, 2), t);
53
54 n_period = 1;
55 inizio = ceil(length(t) / 20 * n_period);
56 tn = t ./ max(t) * N_waves;
57
58 xn = tn(end - inizio:end) - (tn(end) - n_period);
59 pos = z_buoy(end - inizio:end);
60 vel = v_buoy(end - inizio:end);
61 P(j, i) = gammaPTO * mean(vel .* vel) * 10^-3; % Potenza assorbita media [kW]
62 clear vel pos tn z_buoy v_buoy inizio t time z w Te T H Hs tspan Ce_ph Ce_mod
63 Ca Cr
64 end
65 end
66
67 E_A_alghero = (P .* wave);
68 Fig1 = figure(1);
69 contourf(X, Y, E_A_alghero);
70 set(Fig1, 'Color', [1 1 1]);
71 set(gca, 'FontSize', 12);
72 title('Annual Energy - Alghero');
73 xlabel('Wave period T_{p} [s]');
74 ylabel('Wave height H_{s}[m]');
75 dim = [.65 .5 .2 .4];
76 str = sprintf('%s', filename);
77 h = colorbar;
78 caxis([0, 2000]);
79
80 Fig2 = figure(2);
81 contourf(X, Y, P);
82 set(Fig2, 'Color', [1 1 1]);
83 set(gca, 'FontSize', 12);
84 title('Power Matrix - Alghero [kW]');
85 xlabel('Wave period T_{p} [s]');
86 ylabel('Wave height H_{s}[m]');
87 clear data x y X Y filename
88 h = colorbar;
89
90 energia_annuale_Alghero = sum(E_A_alghero(:));
91
92 %% Mazara
93
94 filename = 'Mazara.txt';
95 data = importdata(filename);
96 wave = data(2:end, 2:end);
97 period = data(1, 2:end);
98 height = data(2:end, 1);
99 [X, Y] = meshgrid(period, height);
100
101 P = zeros(length(height), length(period));
102
103 for i = 1:length(period)
104 for j = 1:length(height)
105 Hs = height(j);
106 H = Hs / 2^0.5;
107 T = period(i);
108 Te = 0.88 * T;
109 w = 2 * pi / Te;
110
111 Cr = -199.62 * w^5 + 2933.2 * w^4 - 15212 * w^3 + 30997 * w^2 - 16110 * w +
112 2894.4;
113 Ca = 43.014 * w^5 - 982.58 * w^4 + 6782.3 * w^3 - 18415 * w^2 + 15112 * w +
114 17600;
115 Ce_mod = 438.29 * w^5 - 6138.3 * w^4 + 32288 * w^3 - 71801 * w^2 + 24263 * w +
116 121771;
117 Ce_ph = -0.0052 * w^5 + 0.0561 * w^4 - 0.2236 * w^3 + 0.1742 * w^2 - 0.0621 * w
118 + 0.0083;
119
120 tspan = [0 20 * Te];
121 options = odeset('RelTol', 1e-10, 'AbsTol', 1e-10);
122 [time, z] = ode15s(@(t, z) [z(2); (Ce_mod * H / 2 * cos(w * t + Ce_ph) - Cr *
123 z(2) - rho * g * A * z(1) - gammaPTO * z(2)) / (m + Ca)], tspan, z0, options);
124
125 t = (0:T/N_pT:20 * Te);
126 z_buoy(:, 1) = interp1(time, z(:, 1), t);
127 v_buoy(:, 1) = interp1(time, z(:, 2), t);
Davide Troiani 14
128
129 n_period = 1;
130 inizio = ceil(length(t) / 20 * n_period);
131 tn = t ./ max(t) * N_waves;
132
133 xn = tn(end - inizio:end) - (tn(end) - n_period);
134 pos = z_buoy(end - inizio:end);
135 vel = v_buoy(end - inizio:end);
136 P(j, i) = gammaPTO * mean(vel .* vel) * 10^-3; % Potenza assorbita media [kW]
137 clear vel pos tn z_buoy v_buoy inizio t time z w Te T H Hs tspan Ce_ph Ce_mod
138 Ca Cr
139 end
140 end
141
142 E_A_mazara = (P .* wave);
143 [X, Y] = meshgrid(period, height);
144 Fig3 = figure(3);
145 contourf(X, Y, E_A_mazara);
146 set(Fig3, 'Color', [1 1 1]);
147 set(gca, 'FontSize', 12);
148 title('Annual Energy - Mazara');
149 xlabel('Wave period T_{p} [s]');
150 ylabel('Wave height H_{s}[m]');
151 dim = [.65 .5 .2 .4];
152 str = sprintf('%s', filename);
153 h = colorbar;
154 caxis([0, 2000]);
155
156 Fig4 = figure(4);
157 contourf(X, Y, P);
158 set(Fig4, 'Color', [1 1 1]);
159 set(gca, 'FontSize', 12);
160 title('Power Matrix - Mazara [kW]');
161 xlabel('Wave period T_{p} [s]');
162 ylabel('Wave height H_{s}[m]');
163 clear data x y X Y filename
164 h = colorbar;
165
166 energia_annuale_mazara = sum(E_A_mazara(:))
167

Davide Troiani 15

You might also like