Spatio-temporal correlations of available wind power and impact on transmission power flows

K.R.W. Bell, D.C. Hill, D. McMillan, D.G. Infield, G.W. Ault, F. Li, R.W. Dunn

Research output: Contribution to conferencePaper

5 Citations (Scopus)

Abstract

This paper presents a description of a number of points of debate concerning the possible impact of future wind power development on power system planning and operation. It is noted that firm conclusions cannot be reached without adequate modelling of available power. Whilst this would normally require many years of experience of wind farm operation across a wide geographical area, in Britain such data are currently unavailable. However, more extensive coverage via national meteorological centre data may be available to facilitate a synthesis of future patterns of available wind power. These can be used by a transmission planner to assess the distribution of possible flows across main transmission boundaries. To be useful, the approach must respect the correlations of available wind power at different locations on the system. Furthermore, trends in wind speed through a day and through a year must be reliably reproduced captured so that the relationship to annual and diurnal load variations can subsequently be studied. A wind synthesis methodology is described. By being based on many years of wind speed data it permits the estimation of long-term risks associated with more extreme wind conditions. After dealing with gaps in the original wind speed dataset, a vector autoregression (VAR) approach is used to model wind speeds. Wind speeds generated by the model are converted to those that may be observed at 'typical' wind farm locations at 10m above ground level for different terrain types in different zones and then converted to hub height. The 'per unit' available wind power is then calculated by use of a wind speed to power curve. After having specified the total wind generation capacity in each terrain in each zone for the scenario they wish to study, the user of the tool is then able to calculate the total available power. When combined with scenarios describing load demand and the availability and dispatch of conventional generation, future capacity margins and power flows can be studied, e.g. for identification of future system reinforcement requirements. An extension of the approach is presented that permits study of available power and power flows through a year of operation. This also uses vector autoregression (VAR) to model wind speeds but is applied after a careful detrending process to allow diurnal and seasonal effects to be correctly treated. Finally, future developments of the approach are outlined and it is suggested that the same approach may be useful not only in power system planning but also in support of power system operation.

Conference

Conference43rd International Conference on Large High Voltage Electric Systems 2010, CIGRE 2010
CountryFrance
CityParis
Period21/08/1026/08/10

Fingerprint

Power transmission
Wind power
Farms
Planning
Reinforcement
Availability

Cite this

Bell, K. R. W., Hill, D. C., McMillan, D., Infield, D. G., Ault, G. W., Li, F., & Dunn, R. W. (2010). Spatio-temporal correlations of available wind power and impact on transmission power flows. Paper presented at 43rd International Conference on Large High Voltage Electric Systems 2010, CIGRE 2010, Paris, France.

Spatio-temporal correlations of available wind power and impact on transmission power flows. / Bell, K.R.W.; Hill, D.C.; McMillan, D.; Infield, D.G.; Ault, G.W.; Li, F.; Dunn, R.W.

2010. Paper presented at 43rd International Conference on Large High Voltage Electric Systems 2010, CIGRE 2010, Paris, France.

Research output: Contribution to conferencePaper

Bell, KRW, Hill, DC, McMillan, D, Infield, DG, Ault, GW, Li, F & Dunn, RW 2010, 'Spatio-temporal correlations of available wind power and impact on transmission power flows' Paper presented at 43rd International Conference on Large High Voltage Electric Systems 2010, CIGRE 2010, Paris, France, 21/08/10 - 26/08/10, .
Bell KRW, Hill DC, McMillan D, Infield DG, Ault GW, Li F et al. Spatio-temporal correlations of available wind power and impact on transmission power flows. 2010. Paper presented at 43rd International Conference on Large High Voltage Electric Systems 2010, CIGRE 2010, Paris, France.
Bell, K.R.W. ; Hill, D.C. ; McMillan, D. ; Infield, D.G. ; Ault, G.W. ; Li, F. ; Dunn, R.W. / Spatio-temporal correlations of available wind power and impact on transmission power flows. Paper presented at 43rd International Conference on Large High Voltage Electric Systems 2010, CIGRE 2010, Paris, France.10 p.
@conference{3f78012a838b4872aaf54412c4f344ee,
title = "Spatio-temporal correlations of available wind power and impact on transmission power flows",
abstract = "This paper presents a description of a number of points of debate concerning the possible impact of future wind power development on power system planning and operation. It is noted that firm conclusions cannot be reached without adequate modelling of available power. Whilst this would normally require many years of experience of wind farm operation across a wide geographical area, in Britain such data are currently unavailable. However, more extensive coverage via national meteorological centre data may be available to facilitate a synthesis of future patterns of available wind power. These can be used by a transmission planner to assess the distribution of possible flows across main transmission boundaries. To be useful, the approach must respect the correlations of available wind power at different locations on the system. Furthermore, trends in wind speed through a day and through a year must be reliably reproduced captured so that the relationship to annual and diurnal load variations can subsequently be studied. A wind synthesis methodology is described. By being based on many years of wind speed data it permits the estimation of long-term risks associated with more extreme wind conditions. After dealing with gaps in the original wind speed dataset, a vector autoregression (VAR) approach is used to model wind speeds. Wind speeds generated by the model are converted to those that may be observed at 'typical' wind farm locations at 10m above ground level for different terrain types in different zones and then converted to hub height. The 'per unit' available wind power is then calculated by use of a wind speed to power curve. After having specified the total wind generation capacity in each terrain in each zone for the scenario they wish to study, the user of the tool is then able to calculate the total available power. When combined with scenarios describing load demand and the availability and dispatch of conventional generation, future capacity margins and power flows can be studied, e.g. for identification of future system reinforcement requirements. An extension of the approach is presented that permits study of available power and power flows through a year of operation. This also uses vector autoregression (VAR) to model wind speeds but is applied after a careful detrending process to allow diurnal and seasonal effects to be correctly treated. Finally, future developments of the approach are outlined and it is suggested that the same approach may be useful not only in power system planning but also in support of power system operation.",
author = "K.R.W. Bell and D.C. Hill and D. McMillan and D.G. Infield and G.W. Ault and F. Li and R.W. Dunn",
year = "2010",
language = "English",
note = "43rd International Conference on Large High Voltage Electric Systems 2010, CIGRE 2010 ; Conference date: 21-08-2010 Through 26-08-2010",

}

TY - CONF

T1 - Spatio-temporal correlations of available wind power and impact on transmission power flows

AU - Bell, K.R.W.

AU - Hill, D.C.

AU - McMillan, D.

AU - Infield, D.G.

AU - Ault, G.W.

AU - Li, F.

AU - Dunn, R.W.

PY - 2010

Y1 - 2010

N2 - This paper presents a description of a number of points of debate concerning the possible impact of future wind power development on power system planning and operation. It is noted that firm conclusions cannot be reached without adequate modelling of available power. Whilst this would normally require many years of experience of wind farm operation across a wide geographical area, in Britain such data are currently unavailable. However, more extensive coverage via national meteorological centre data may be available to facilitate a synthesis of future patterns of available wind power. These can be used by a transmission planner to assess the distribution of possible flows across main transmission boundaries. To be useful, the approach must respect the correlations of available wind power at different locations on the system. Furthermore, trends in wind speed through a day and through a year must be reliably reproduced captured so that the relationship to annual and diurnal load variations can subsequently be studied. A wind synthesis methodology is described. By being based on many years of wind speed data it permits the estimation of long-term risks associated with more extreme wind conditions. After dealing with gaps in the original wind speed dataset, a vector autoregression (VAR) approach is used to model wind speeds. Wind speeds generated by the model are converted to those that may be observed at 'typical' wind farm locations at 10m above ground level for different terrain types in different zones and then converted to hub height. The 'per unit' available wind power is then calculated by use of a wind speed to power curve. After having specified the total wind generation capacity in each terrain in each zone for the scenario they wish to study, the user of the tool is then able to calculate the total available power. When combined with scenarios describing load demand and the availability and dispatch of conventional generation, future capacity margins and power flows can be studied, e.g. for identification of future system reinforcement requirements. An extension of the approach is presented that permits study of available power and power flows through a year of operation. This also uses vector autoregression (VAR) to model wind speeds but is applied after a careful detrending process to allow diurnal and seasonal effects to be correctly treated. Finally, future developments of the approach are outlined and it is suggested that the same approach may be useful not only in power system planning but also in support of power system operation.

AB - This paper presents a description of a number of points of debate concerning the possible impact of future wind power development on power system planning and operation. It is noted that firm conclusions cannot be reached without adequate modelling of available power. Whilst this would normally require many years of experience of wind farm operation across a wide geographical area, in Britain such data are currently unavailable. However, more extensive coverage via national meteorological centre data may be available to facilitate a synthesis of future patterns of available wind power. These can be used by a transmission planner to assess the distribution of possible flows across main transmission boundaries. To be useful, the approach must respect the correlations of available wind power at different locations on the system. Furthermore, trends in wind speed through a day and through a year must be reliably reproduced captured so that the relationship to annual and diurnal load variations can subsequently be studied. A wind synthesis methodology is described. By being based on many years of wind speed data it permits the estimation of long-term risks associated with more extreme wind conditions. After dealing with gaps in the original wind speed dataset, a vector autoregression (VAR) approach is used to model wind speeds. Wind speeds generated by the model are converted to those that may be observed at 'typical' wind farm locations at 10m above ground level for different terrain types in different zones and then converted to hub height. The 'per unit' available wind power is then calculated by use of a wind speed to power curve. After having specified the total wind generation capacity in each terrain in each zone for the scenario they wish to study, the user of the tool is then able to calculate the total available power. When combined with scenarios describing load demand and the availability and dispatch of conventional generation, future capacity margins and power flows can be studied, e.g. for identification of future system reinforcement requirements. An extension of the approach is presented that permits study of available power and power flows through a year of operation. This also uses vector autoregression (VAR) to model wind speeds but is applied after a careful detrending process to allow diurnal and seasonal effects to be correctly treated. Finally, future developments of the approach are outlined and it is suggested that the same approach may be useful not only in power system planning but also in support of power system operation.

UR - http://www.scopus.com/inward/record.url?scp=84876747429&partnerID=8YFLogxK

M3 - Paper

ER -