Transcript Modelling Electric Vehicles at Residential Low Voltage Grid by

Frankfurt (Germany), 6-9 June 2011
MODELLING ELECTRIC VEHICLES AT RESIDENTIAL
LOW VOLTAGE GRID BY MONTE CARLO SIMULATION
W.Du
TU Delft
The Netherlands
W.Du – the Netherlands – Session 5 – 1225
Frankfurt (Germany), 6-9 June 2011
Structure
Problem, Objective and Methodology
 Base Model (By Monte-Carlo Simulation)
 An Example of Simulated Low Voltage Grid
 EV Appliance and EV Scenarios
 DNOs’ Pricing Strategies
 Model Results, Validation and Conclusion

W.Du – the Netherlands – Session 5 – 1225
Frankfurt (Germany), 6-9 June 2011
Problem and Objective

Problem:


Uncertain EVs’ impacts on low voltage electricity grid
Objectives:



Supporting DNOs in grid asset capacity planning
Finding out more accurately probabilities of possible
overloads caused by EVs
Analyzing DNOs’ pricing strategy in influencing EV
charging behaviors
W.Du – the Netherlands – Session 5 – 1225
Frankfurt (Germany), 6-9 June 2011
Method
DNOs’ Pricing Planning
strategies
EV Scenarios
·
·
charging patterns
penetration degrees
Monte Carlo Simulation
(Base Household behaviors)
W.Du – the Netherlands – Session 5 – 1225
Capacity
Frankfurt (Germany), 6-9 June 2011
Base Model (Monte-Carlo Simulation)
Model represents regular household
behaviors of using electrical appliances
 Stochastically generate load profiles




more realistic then using aggreated and
deterministic simulatainty factor
especially true for low voltage grid
Aiming at finding more accurate the
impacts of having EVs at households
W.Du – the Netherlands – Session 5 – 1225
Frankfurt (Germany), 6-9 June 2011
Base Model (Monte-Carlo Simulation)-Cont.

Household Types (HTs)



Electrical appliances




Eight HTs by size, age and working status
Percentages for HTs are based on Dutch statistic data
Commonly used 25 types in Dutch households are stored
Information including highest power, duration of use time
Penetration degrees based on available statistics, market data
and related researches
Usage of appliances


Three types distinguished in relation to use time: constant use;
instant use and semi-constant use
Probability distributions are pre-defined in power usage,
frequency to be used and duration time of each use.
W.Du – the Netherlands – Session 5 – 1225
Frankfurt (Germany), 6-9 June 2011
An Example of Simulated Residential Grid
W.Du – the Netherlands – Session 5 – 1225
Frankfurt (Germany), 6-9 June 2011
EV Appliances

Treated as household appliances



Type: instant use
Only full battery-engined EV is considered
Attributes:




Power usage:in triangular distribution (unit :watt)
Duration of charging:in triangular distribution (time
unit : minute)
penetration degree: in percentage
frequency of being charged in different charging
periods:in triangular distribution (unit : integer)
W.Du – the Netherlands – Session 5 – 1225
Frankfurt (Germany), 6-9 June 2011
EVs Scenarios

Starting Time:



Charging Rate:


low rate of 3kW and high rate of 10kW.
Charging Time




Start charging immediately after arriving at home
Charging after 11:00pm
EV efficiency of 5km/kWh is assumed
EVs are assumed to be charged at households until full
Distance driven pre-defined based on Dutch statistics
Penetration Degrees

0.1 to 1 in step of 0.1
W.Du – the Netherlands – Session 5 – 1225
Frankfurt (Germany), 6-9 June 2011
DNOs’ Pricing Strategy

Shifting Charging Time



Shifted to after 11:00pm
from 10% of households onward up to 100%
in step of 10%
Reducing Charging Time

Reduced in 10% to 50% of previous settings
W.Du – the Netherlands – Session 5 – 1225
Frankfurt (Germany), 6-9 June 2011
Simulation Results
Highlighted EVs’ charging load in a single household load profile
W.Du – the Netherlands – Session 5 – 1225
Frankfurt (Germany), 6-9 June 2011
Simulation Results – Cont.
Aggregation of transformer loads
W.Du – the Netherlands – Session 5 – 1225
Frankfurt (Germany), 6-9 June 2011
Simulation Results – Cont.
Aggregated peak load and probabilities of overloads
W.Du – the Netherlands – Session 5 – 1225
Frankfurt (Germany), 6-9 June 2011
Validation

Two Steps:



Validation of the base model
Validation of EVs’ charging profiles
Data Comparition


Aggregations of household loads will be
compared with empirical transformer loads
provided by Enexis B.V., NL
The EVs’ charging results will be validated with
real sampled data at individual households
W.Du – the Netherlands – Session 5 – 1225
Frankfurt (Germany), 6-9 June 2011
Conclusion




Modelling EVs as household appliances by a
Monte-Carlo simulation
Aiming at analysing EVs’ charging impacts on
residential low voltage grid capacity
Charging scenarios are generated stochastically
with different penetration degrees and charging
patterns
DNOs’ pricing strategies also are estimated in
their influences on EVs’ charging patterns
W.Du – the Netherlands – Session 5 – 1225
Frankfurt (Germany), 6-9 June 2011
Thank you for your attentions!
W.Du – the Netherlands – Session 5 – 1225