Dr. Keith Sunderland1, Dr. Gerald Mills2, Prof. Michael Conlon1
1 School of Electrical & Electronic Engineering, Dublin Institute of Technology, Ireland 2 School of Geography, Planning and Environmental Policy, University College Dublin, Ireland
12th June, 2014
‘The application of boundary layer
climatology and urban wind power
potential in smarter electricity networks’
AMERICAN METEOROLOGICAL SOCIETY CONFERENCE
Overview
Aims and Objectives
Research Context/ Motivation
Methodology
Findings
Future Work
Conclusions
AMERICAN METEOROLOGICAL SOCIETY2014
Overview
Aims and Objectives
Research Context/ Motivation
Methodology
Findings
Future Work
Conclusions
AMERICAN METEOROLOGICAL SOCIETY2014
Aims & Objectives
Research Aim:
To develop novel modelling capability that is inclusive of the power engineering complexities associated with urban (electricity) network integration of small/micro wind generation, and informed by urban climate research
Background Reference
Urban Classification
Backgrounduu
STDuUrban
DistributionNetwork
Cable Parameters/Configurations
Network Structure
Consumer configuration/loading
Earthing Configurations
surface roughness parameterisation
z ,0 zd
T.I.
Wind turbine 'pure' (zero-turbulence)
power characteristic
AMC Power Flow
Algorithm
1.
Radial
Mesh
Turbulence normalised network node voltage/voltage unbalance profile
Network node voltage/voltage unbalance profile based on the 'mean' urban wind resource
LCOE evaluation
Pu(turbulent)
u(mean)P
2. Observations
Slide 1 K. Sunderland (z0, zd)
Aims & Objectives
Research Aim:
To develop novel modelling capability that is inclusive of the power engineering complexities associated with urban (electricity) network integration of small/micro wind generation, and informed by urban climate research
Background Reference
Urban Classification
Backgrounduu
STDuUrban
DistributionNetwork
Cable Parameters/Configurations
Network Structure
Consumer configuration/loading
Earthing Configurations
surface roughness parameterisation
z ,0 zd
T.I.
Wind turbine 'pure' (zero-turbulence)
power characteristic
AMC Power Flow
Algorithm
1.
Radial
Mesh
Turbulence normalised network node voltage/voltage unbalance profile
Network node voltage/voltage unbalance profile based on the 'mean' urban wind resource
LCOE evaluation
Pu(turbulent)
u(mean)P
2. Observations
Synergetic application of urban climatology (BLT) and electrical power engineering
Empirical wind mapping/modelling cognisant of urban heterogeneity – in conjunction with urban wind observations
Slide 1 K. Sunderland (z0, zd)
Aims & Objectives
Research Aim:
To develop novel modelling capability that is inclusive of the power engineering complexities associated with urban (electricity) network integration of small/micro wind generation, and informed by urban climate research
Background Reference
Urban Classification
Backgrounduu
STDuUrban
DistributionNetwork
Cable Parameters/Configurations
Network Structure
Consumer configuration/loading
Earthing Configurations
surface roughness parameterisation
z ,0 zd
T.I.
Wind turbine 'pure' (zero-turbulence)
power characteristic
AMC Power Flow
Algorithm
1.
Radial
Mesh
Turbulence normalised network node voltage/voltage unbalance profile
Network node voltage/voltage unbalance profile based on the 'mean' urban wind resource
LCOE evaluation
Pu(turbulent)
u(mean)P
2. Observations
Slide 1 K. Sunderland (z0, zd)
Overview
Aims and Objectives
Research Context/ Motivation
Methodology
Findings
Future Work
Conclusions
AMERICAN METEOROLOGICAL SOCIETY2014
Research Context
CEN
TRA
LISE
DD
ISTR
IBU
TED
RENEWABLE ENERGYFOSSIL FUEL
PAST FUTURE
Increasing need for Localised Supply within urban load centres
Time
Ener
gy F
low
Bi-
Dir
ecti
on
al E
ner
gy F
low
DISTRIBUTED GENERATION
GENERATION
TRANSMISSION
DISTRIBUTION
CONSUMPTIONCONSUMPTION
DISTRIBUTION
TRANSMISSION
GENERATION
25kV
110/220/400 kV
38/20/10 kV
400/230 V
38/20/10kV &
400/230V
Slide 2 K. Sunderland
Research Context
CEN
TRA
LISE
DD
ISTR
IBU
TED
RENEWABLE ENERGYFOSSIL FUEL
PAST FUTURE
Increasing need for Localised Supply within urban load centres
Time
Ener
gy F
low
Bi-
Dir
ecti
on
al E
ner
gy F
low
DISTRIBUTED GENERATION
GENERATION
TRANSMISSION
DISTRIBUTION
CONSUMPTIONCONSUMPTION
DISTRIBUTION
TRANSMISSION
GENERATION
25kV
110/220/400 kV
38/20/10 kV
400/230 V
38/20/10kV &
400/230V
Voltage/VAR control testing
Micro Gen programme
Grid behaviour at high wind
Smart
Generation
Research
De
plo
ym
en
t
Demonstration
De
ve
lo
pm
en
t
Smart
Pricing
Smart meter pricing trial
Future smart pricing options
Smart
Networks
Grid 25 Strategy
Voltage conversion of MV Networks
Self healing networks pilot Smarter
Wind Farm operations
Smart
Operations
Smart Grid Ireland
Smart meter technology trialElectric
Vehicle Incentives
Smart Meter user trial
Smart
Users
DSM programmes
Smart commercial applications
Slide 2 K. Sunderland
Research Motivation
Micro/Small Wind Electricity Generation
Urban Locations
Heterogeneity, Complex
Building Morphology
Chaotic, Turbulent Airflows
Energy Capture Capability?
Rural Locations
Uninhibited Air Flows
Statistically Predictable
Airflows
Maximum Energy
Optimisation
Slide 3 K. Sunderland
Research Motivation
Micro/Small Wind Electricity Generation
Urban Locations
Heterogeneity, Complex
Building Morphology
Chaotic, Turbulent Airflows
Energy Capture Capability?
Rural Locations
Uninhibited Air Flows
Statistically Predictable
Airflows
Maximum Energy
Optimisation
Slide 3 K. Sunderland
Research Motivation
Micro/Small Wind Electricity Generation
Urban Locations
Heterogeneity, Complex
Building Morphology
Chaotic, Turbulent Airflows
Energy Capture Capability?
Rural Locations
Uninhibited Air Flows
Statistically Predictable
Airflows
Maximum Energy
Optimisation
Slide 3 K. Sunderland
WHY URBAN WIND?
Population Centres
Transmission/Distribution losses
Green solutions must include wind
Smarter energy diversification must be
inclusive of wind within urban centres BUT
solutions predicated on the resource and not
specifically the technology are needed
Research Motivation
Slide 4 K. Sunderland
Smart Cities…. Smart Grids
o An amalgamation of communication and electrical capabilities that allow utilities to understand, optimize, and regulate demand, supply, costs and reliability.
Facilitating electrical providers to interact with the power delivery system and determine whether electricity is being used and from where it can be drawn during the time of crisis and peak demand. On the demand side – the smart grid empowers the consumer to become a ‘prosumer’…
Research Motivation
Slide 5 K. Sunderland
Why is a Smart Grid needed?
o Future grid networks must be competitive and supportive of environmental objectives and sustainability
o Reliability, flexibility, accessibility and cost-effectiveness are the primary objectives
o Should accommodate both central and dispersed generation
o Options for end-users to be more interactive with both market and grid; promoting the concept of a ‘prosumer’
Research Motivation
Slide 5 K. Sunderland
Why is a Smart Grid needed?
o Future grid networks must be competitive and supportive of environmental objectives and sustainability
o Reliability, flexibility, accessibility and cost-effectiveness are the primary objectives
o Should accommodate both central and dispersed generation
o Options for end-users to be more interactive with both market and grid; promoting the concept of a ‘prosumer’
Therefore the means of applying the primary energy resource (Wind) in this regard within urban centres must be achieved
Overview
Aims and Objectives
Research Context/ Motivation
Methodology
Findings
Future Work
Conclusions
AMERICAN METEOROLOGICAL SOCIETY2014
Urban Effects & Wind Modelling Research Methodology
(Electrical) Distribution Network considered in terms of both local climate zones
Two zones categorised in terms of local heterogeneity: (Urban and Suburban)
Urban Environmnet
Suburban
Urban
Slide 6 K. Sunderland
Urban Effects & Wind Modelling Research Methodology
(Electrical) Distribution Network considered in terms of both local climate zones
Two zones categorised in terms of local heterogeneity: (Urban and Suburban)
Urban Environmnet
Suburban
Urban
[SL]
CITY
[C]
[CH]
[SH]
Urban Effects & Wind Modelling Research Methodology
(Electrical) Distribution Network considered in terms of both local climate zones
Two zones categorised in terms of local heterogeneity: (Urban and Suburban)
Urban Environmnet
Suburban
Urban
[SL]
CITY
[C]
[CH]
[SH]
Urban Effects & Wind Modelling Research Methodology
(Electrical) Distribution Network considered in terms of both local climate zones
Two zones categorised in terms of local heterogeneity: (Urban and Suburban)
Urban Environmnet
Suburban
Urban
Slide 6 K. Sunderland
Urban Effects & Wind Modelling Research Methodology
Slide 6 K. Sunderland
AIRPORT
– Rural reference
Urban Effects & Wind Modelling Research Methodology
Slide 6 K. Sunderland
AIRPORT
– Rural reference
Boundary Layer Climatology in terms
of surface roughness classification
Urban Effects & Wind Modelling Research Methodology
Slide 6 K. Sunderland
AIRPORT
– Rural reference
Boundary Layer Climatology in terms
of surface roughness classification
zdz1
z0201z
Roughness Sub-Layer
Canopy Sub-Layer
Urban Boundary Sub-Layer
Inertial Sub-LayerLogarithmic Extrapolation
(Rural) Airport Background
Internal Boundary Layer
(Statistically) Laminar Airflow based on a 'constant' frictional velocity
H(z )
(z*)
1u
Logarithmic Extrapolation
uRural
u(z*)
u(z )UBL
Wieranga, Bottema approximation and a Logarithmic extrapolation based on fitted surface roughness parameterisation
Urban Effects & Wind Modelling Research Methodology
Slide 6 K. Sunderland
AIRPORT
– Rural reference
Boundary Layer Climatology in terms
of surface roughness classification
zdz1
z0201z
Roughness Sub-Layer
Canopy Sub-Layer
Urban Boundary Sub-Layer
Inertial Sub-LayerLogarithmic Extrapolation
(Rural) Airport Background
Internal Boundary Layer
(Statistically) Laminar Airflow based on a 'constant' frictional velocity
H(z )
(z*)
1u
Logarithmic Extrapolation
uRural
u(z*)
u(z )UBL
z
ROUGHNESS Sub-Layer
INERTIAL Sub-Layer
d
Hmz
z0
CANOPY
z*
Logarithmic
Profile
u(z*)
Transitionary
Profile
Mac
Dona
ld
COST
Exponential
Profile
Wieranga, Bottema approximation and a Logarithmic extrapolation based on fitted surface roughness parameterisation
Slide 7 K. Sunderland
(Standardised) Distribution
Network analysis
o Single-phase 4-Wire
(and Ground)
o Complex/unbalanced
(consumer) load
configurations
DwG & DN Implications Research Methodology
Slide 7 K. Sunderland
(Standardised) Distribution
Network analysis
o Single-phase 4-Wire
(and Ground)
o Complex/unbalanced
(consumer) load
configurations
Energy flow - Mono-
directional Power Flow
DwG & DN Implications Research Methodology
Slide 7 K. Sunderland
(Standardised) Distribution
Network analysis
o Single-phase 4-Wire
(and Ground)
o Complex/unbalanced
(consumer) load
configurations
Energy flow - Mono-
directional Power Flow
DwG & DN Implications Research Methodology
25/16sq, (Concentric Neutral) L3
25/16sq, (Concentric Neutral) L2
25/16sq, (Concentric Neutral) L1
4xcore 185sq, XLPE
4xcore 70sq, XLPE
38
SUBSTATION
B
CDEFH
J
I
G
1234
6
7
5
9 81011
121314
15161718
2927 28
26252423
37363534
333231305453
51 525049
47 48
73
72
71
70
74
64
63
61
62
60
69
68
67
66
65
59
58
57
56
55
21 22
46
45
44
43
42
41
40
39
2019
A
4xcore 70sq, Al
Slide 7 K. Sunderland
(Standardised) Distribution
Network analysis
o Single-phase 4-Wire
(and Ground)
o Complex/unbalanced
(consumer) load
configurations
Energy flow - Mono-
directional Power Flow
DwG & DN Implications Research Methodology
25/16sq, (Concentric Neutral) L3
25/16sq, (Concentric Neutral) L2
25/16sq, (Concentric Neutral) L1
4xcore 185sq, XLPE
4xcore 70sq, XLPE
38
SUBSTATION
B
CDEFH
J
I
G
1234
6
7
5
9 81011
121314
15161718
2927 28
26252423
37363534
333231305453
51 525049
47 48
73
72
71
70
74
64
63
61
62
60
69
68
67
66
65
59
58
57
56
55
21 22
46
45
44
43
42
41
40
39
2019
A
4xcore 70sq, Al
Consumer
Earthing Connection
Pillar Earthing
Connection
Substation
Transformer
(Mini) Pillar
Consumer
Connection/Load
V4
V3
V2
V1
Pillar (i) Pillar (i+1)
Transformer
Earthing
Connection
Slide 7 K. Sunderland
(Standardised) Distribution
Network analysis
o Single-phase 4-Wire
(and Ground)
o Complex/unbalanced
(consumer) load
configurations
Energy flow - Mono-
directional Power Flow
DwG & DN Implications Research Methodology
25/16sq, (Concentric Neutral) L3
25/16sq, (Concentric Neutral) L2
25/16sq, (Concentric Neutral) L1
4xcore 185sq, XLPE
4xcore 70sq, XLPE
38
SUBSTATION
B
CDEFH
J
I
G
1234
6
7
5
9 81011
121314
15161718
2927 28
26252423
37363534
333231305453
51 525049
47 48
73
72
71
70
74
64
63
61
62
60
69
68
67
66
65
59
58
57
56
55
21 22
46
45
44
43
42
41
40
39
2019
A
4xcore 70sq, Al
Consumer
Earthing Connection
Pillar Earthing
Connection
Substation
Transformer
(Mini) Pillar
Consumer
Connection/Load
Ig(i)
Icons(i)[g]
Icons(i)[n]
cons(i)[c]
In(i+1)
Ic(i+1)
Ib(i+1)
Ia(i+1)
V4
V3
V2
V1
Ic(i)
Ib(i)
Ia(i)
In(i)
Pillar (i) Pillar (i+1)
Transformer
Earthing
Connection
7 8
13
…. Enhanced modelling cognisant of the inherent complexities associated with connectivity and unbalanced load/generation integration at final consumer level
Slide 7 K. Sunderland
Embedded Generation Issues
o Bi-directional power flow
o Network Power Quality
management
o Safety implications
DwG & DN Implications Research Methodology
Slide 7 K. Sunderland
Embedded Generation Issues
o Bi-directional power
flow
o Network Power Quality
management
o Safety implications
DwG & DN Implications Research Methodology
Slide 7 K. Sunderland
Embedded Generation Issues
o Bi-directional power flow
o Network Power Quality
management
o Safety implications
DwG & DN Implications Research Methodology
3
2
1.u.A.ρ.cP rotorairpMech
Overview
Aims and Objectives
Research Context/ Motivation
Methodology
Findings
Ongoing Work
Conclusions
AMERICAN METEOROLOGICAL SOCIETY2014
Surface Roughness Parameterisation
Results & Findings
Urban Observations & Modelling
Slide 8 K. Sunderland
Surface Roughness Parameterisation
Results & Findings
Urban Observations & Modelling
Slide 8 K. Sunderland
For each 300 sector, surface roughness was estimated by varying iteratively until the best fit was achieved so as to minimise the error between the predicted wind speed, based on the background climate, and the observed wind speed
CH SH
Observed Wieranga
Model Bottema
Model Log-
Model Observed Wieranga
Model Bottema
Model Log-
Model
Roughness length (z0)
-- 1.15 1.15 0.8713 -- 0.55 0.55 0.5171
Friction velocity
ratio -- 1.0 1.3312 1.7022 -- 1.0 1.2636 1.5512
uM [m/s] 4.5992 4.9728 3.2281 4.6165 4.4401 4.9804 3.5795 4.3940
uS [m/s] 2.1288 2.2497 1.4604 2.0885 2.1712 2.2269 1.6005 1.9647
MAE [m/s] -- 0.7113 1.4248 0.6133 -- 0.9392 1.0635 0.7594
RMSE[m/s] -- 0.9790 1.6878 0.8651 -- 1.2202 1.3873 1.0479
Observation/Modelling: high-platform observations
Results & Findings
Urban Observations & Modelling
Slide 9 K. Sunderland
CH SH
Observed Wieranga
Model Bottema
Model Log-
Model Observed Wieranga
Model Bottema
Model Log-
Model
Roughness length (z0)
-- 1.15 1.15 0.8713 -- 0.55 0.55 0.5171
Friction velocity
ratio -- 1.0 1.3312 1.7022 -- 1.0 1.2636 1.5512
uM [m/s] 4.5992 4.9728 3.2281 4.6165 4.4401 4.9804 3.5795 4.3940
uS [m/s] 2.1288 2.2497 1.4604 2.0885 2.1712 2.2269 1.6005 1.9647
MAE [m/s] -- 0.7113 1.4248 0.6133 -- 0.9392 1.0635 0.7594
RMSE[m/s] -- 0.9790 1.6878 0.8651 -- 1.2202 1.3873 1.0479
Observation/Modelling: high-platform observations
Results & Findings
Urban Observations & Modelling
Slide 9 K. Sunderland
0 2 4 6 8 10 12 14 16 18 200
5
10
15
20
[B]
u(modelled) [m/ s]
u(o
bs
erv
ed
) [m
/s
0 2 4 6 8 10 12 14 16 18 200
5
10
15
20
[A]
u(modelled) [m/ s]
u(o
bs
erv
ed
) [m
/s]
Raw Data
y = 0.8987x + 0.4831
r2 = 0.916
Raw Data
y = 0.7931x + 0.8724
r2 = 0.8765
0
5
10
15
20
25
C(C
ap
ac
ity
Fa
cto
r) [
%]
7 8 9 10 11 12 13 14 15 16 170
500
1000
1500
2000
2500
C(E
ne
rgy
) [k
Wh
]
z [m]
0
5
10
15
20
S(C
ap
ac
ity
Fa
cto
r) [
%]
5 6 7 8 9 10 11 120
500
1000
1500
2000
2500
S(E
ne
rgy
) [k
Wh
]
z [m]
89.6
29.524.4
20.4
17.9
13.1
38.5
[A]
[B]
95.9
50.9
34.1
25.2
20.2
54.2
14.3
16.9
14.3
15.7
Urban Comparison
Suburban Comparison
Observation vs. Modelling
Urban Comparison
Suburban Comparison
Scattergram comparison of high-platform
observed and modelled wind speeds
(Nov. 2010 –to– Jan 2011)
Energy implications with respect to height
variation for a wind generator at both sites
(Nov. 2010 –to– Jan 2011)
Results & Findings
Urban Observations & Modelling
Slide 10 K. Sunderland
Results & Findings
Distribution Network Reaction
Slide 11 K. Sunderland
SUBSTATION
B
CDEFH
J
73
72
71
70
7469
68
67
66
65A
1234
79 8
10111213
141516
171829
27 282625
24233736
35343332
3130545351 52
504947 48
G
6 5
21 22
46
45
44
43
42
41
40
39
2019
38
I
64
63
61
62
60
59
58
57
56
55
Typical Mean Year of Wind Speed (Markov Chain)
Results & Findings
Distribution Network Reaction
Slide11 K. Sunderland
SUBSTATION
B
CDEFH
J
73
72
71
70
7469
68
67
66
65A
1234
79 8
10111213
141516
171829
27 282625
24233736
35343332
3130545351 52
504947 48
G
6 5
21 22
46
45
44
43
42
41
40
39
2019
38
I
64
63
61
62
60
59
58
57
56
55
B 1 2 3 4 5 6 C D E F GH I J 1 2 3 4 5 6 7 8 910
1.03
1.04
1.05
1.06
1.07
1.08
1.09
Pillar/Customer
Vo
lta
ge
[p
u])
Line-1
B 1 2 3 4 5 6 C D E FGH I J 1 2 3 4 5 6 7 8 910
1.02
1.04
1.06
1.08
1.1
Pillar/Customer
Vo
lta
ge
[p
u])
Line-2
B 1 2 3 4 5 6 C D E F GH I J 1 2 3 4 5 6 7 8 910
1.04
1.06
1.08
1.1
Pillar/Customer
Vo
lta
ge
[p
u])
Line-3
B 1 2 3 4 5 6 C D E F GH I J 1 2 3 4 5 6 7 8 910
0
1
2
3
4
x 10-3
Pillar/Customer
Vo
lta
ge
[p
u])
Neutral
Overview
Aims and Objectives
Research Context/ Motivation
Methodology
Findings
Future Work
Conclusions
AMERICAN METEOROLOGICAL SOCIETY2014
Future Work
Slide 13 K. Sunderland
J. T. Millward-Hopkins, et al., "Estimating Aerodynamic Parameters of Urban-Like Surfaces with Heterogeneous Building Heights," Boundary-Layer Meteorology, vol. 141, pp. 443-465, 2011/12/01 2011.
J. T. Millward-Hopkins, et al., "Aerodynamic Parameters of a UK City Derived from Morphological Data," Boundary-Layer Meteorology, vol. 146, pp. 447-468, 2013/03/01 2013
To be applicable from the ISL into the RSL, neighbourhoods of homogeneity need to be identified – distinctly different surfaces can be considered separately
Future Work
Slide 13 K. Sunderland
Rastered Digital Elevation Model (DEM) - - building footprints (Dublin)
Divide the city into distinct neighbourhood regions – Adaptive Grid
Geometric Parameterisation: Employing an adaptive grid to calculate the geometric parameters
Future Work
Slide 13 K. Sunderland
Rastered Digital Elevation Model (DEM) - - building footprints (Dublin)
Divide the city into distinct neighbourhood regions – Adaptive Grid
Geometric Parameterisation
Morphemetric Model
z0
Overview
Aims and Objectives
Research Context/ Motivation
Methodology
Findings
Future Work
Conclusions
AMERICAN METEOROLOGICAL SOCIETY2014
Conclusions
In the context of smart cities and smarter (electricity) grids, this type of research is essential if renewable energy is to facilitate a cultural shift towards an era of prosumers.
In terms of the limits available to wind energy extraction in an urban context., the analyses illustrated limited opportunities below a height 2 → 4 x zHm
By linking urban wind observations to a background reference, an empirical logarithmically matched profile was possible. (Analytical linkages to observations within the canopy suggested that knowledge of the background resource in this regard is of limited value)
Analyses of a fully described 4-wire unbalanced section of Dublin city network, in respect of increasing levels of prosumer (with a grid-tied commercially available DwG), illustrated that for exemplar consumer load and a typical mean year of wind speed, voltage tolerance breaches are unlikely and of marginal concern (
Thank you