Assessing Emulated Inertial Response from Wind

Generation in the GB Power System

Lei Wu David Infield

Institute for Energy and Environment

Department of Electronic and Electrical Engineering

University of Strathclyde, Glasgow, UK

Presented at EWEA Conference, Vienna, 7th Feb. 2013

Introduction

System modelling

Assessing aggregate inertial response

Simulation results

Conclusions

Outline

Introduction

How best should power system frequency response in the context of high wind

penetration be provided:

From wind turbines ? From conventional plant? From demand side management?

Introduction

The rate of change of system frequency will depend on The power mismatch

Total power system inertia

Total power system inertia Synchronous generators and some system load

Variable speed wind turbines

Deliver frequency support from wind plant Inertial response

Droop response

Wind turbine modelling4 sub-systems:

1. Rotor aerodynamics

The power delivered by the rotor of a wind

turbine is given by: ),(322

1 pCvRP

2. Drive train

3. Turbine controller

above rated power/wind speed

below rated power/wind speed

4. Electrical sub-system

The principle is to modify the demanded torque in response to a change in system

frequency by adding an extra torque term.

The modified demanded torque is then given by:

Delivering inertial response from wind

Based on a lower-order System Frequency Response model first proposed by P. Anderson in 1990.

GB Power system modelling

Aggregate inertial response from a wind farm

Example for mean wind speed of 10 m/s

Describe the wind through a family of wind ramps ranges

Calculate the power response corresponding to each wind ramp case

Calculate the joint probability of each wind ramp case

Calculate the expected aggregate inertial response by summing up all the power response corresponding to each wind ramp weighted by the appropriate probabilities [1]

0-3.53.5-6.5

6.5-9.59.5-12.5

12.5-15.5>15.5

0-3.53.5-6.5

6.5-9.59.5-12.5

12.5-15.5>15.5

0

0.1

0.2

0.3

0.4

Start wind

Join

t pro

babi

lity

End wind

[1] Towards an Assessment of Power System Frequency Support from Wind Plant – Modelling Aggregate Inertial Response, accepted by IEEE Transactions on Power Systems.

Assess aggregate inertial response from wind generation in the GB power system

To develop a decentralised approach,

the GB power system is divided into 17

study regions.

The wind capacity in each study zone

is estimated for the study year of 2020.

A Vector Auto Regressive (VAR) model

is applied to synthesise wind speed data

across the power system.

Example wind speeds for regions

Simulation results

02 04 06 08 10 12 14 16 18 20 22 2425

30

35

40

45

Time (hr)

Lo

ad

(G

W)

02 04 06 08 10 12 14 16 18 20 22 240

5

10

15

Time (hr)

Win

d S

pe

ed

(m

/s)

Region 1

Region 2

Region 3Region 4

Region 5

System load on a typical British summer day

Frequency minimum (nadir) following transient

Simulation results

Maximum rate of change of frequency (ROCOF) immediately following transient

2 4 6 8 10 12 14 16 18 20 22 2449.1

49.2

49.3

49.4

49.5

49.6

49.7

Time (hr)

Fre

qu

en

cy m

inim

um

(H

z)

Frequency minimum without support from wind

Frequency minimum with support from wind

2 4 6 8 10 12 14 16 18 20 22 240.05

0.1

0.15

0.2

0.25

0.3

Time (hr)R

OC

OF

(H

z/s

)

ROCOF without support from wind

ROCOF with support from wind

Conclusions

A probabilistic approach to assessing the aggregate inertial response available from

wind generation has been proposed.

The impact of frequency support from wind on the power system as a whole can be

quantified with some degree of confidence.

Provision of frequency support from wind plant can help reduce the rate of change of

system frequency (ROCOF) and improve frequency minimum.