department of engineering …chatziva.com/presentations/griful_pscc2016_presentation.pdfdepartment...

22
DEPARTMENT OF ENGINEERING AARHUS UNIVERSITY AU 23 J, 2016 HL C D G D R 19 th P S C C (PSCC) G,I S RG 1 , S. C 2 , R. H. J 1 , E. M. S 3 , J. M D 4 M. W 3 1 Aarhus University, Denmark 2 Massachuses Institute of Technology, USA 3 Lawrence Berkeley National Laboratory, USA 4 Universidad Pontificia Comillas, Spain

Upload: truongminh

Post on 09-Jun-2018

214 views

Category:

Documents


0 download

TRANSCRIPT

DEPARTMENT OF ENGINEERING

AARHUS UNIVERSITY AU

23 J���, 2016

H���������������L���C������������ �� D�����������G��� ��� D����� R�������19th P���� S������ C���������� C��������� (PSCC)G����, I����

S���� R������G�����1, S. C����������������2, R. H. J�������1,E. M. S������3, J. M������ D������4 ���M. W�����3

1Aarhus University, Denmark2Massachuse�s Institute of Technology, USA3Lawrence Berkeley National Laboratory, USA4Universidad Pontificia Comillas, Spain

Introduction Co-simulation with HiL Case Studies Conclusions

DEPARTMENT OF ENGINEERING

AARHUS UNIVERSITY AU

O������

1 Introduction

2 Co-simulation with Hardware-in-the-Loop

3 Case Studies

4 Conclusions

H�L C������������ ��� D����� R������� S���� R������G����� 23 J���, 2016 2 / 14

Introduction Co-simulation with HiL Case Studies Conclusions

DEPARTMENT OF ENGINEERING

AARHUS UNIVERSITY AU

O������

1 Introduction

2 Co-simulation with Hardware-in-the-Loop

3 Case Studies

4 Conclusions

H�L C������������ ��� D����� R������� S���� R������G����� 23 J���, 2016 3 / 14

Introduction Co-simulation with HiL Case Studies Conclusions

DEPARTMENT OF ENGINEERING

AARHUS UNIVERSITY AU

M���������� High penetration of renewables (42% of the

consumption in Denmark in 2015 from windturbines)

� Flexibility required for generation-consumptionbalancing

� Demand response as a solution⌅ Help renewables integration⌅ Create market opportunities at distribution level

� Need tools for comprehensive assessment of newdemand response opportunities

Source: Energinet.dk (2015)H�L C������������ ��� D����� R������� S���� R������G����� 23 J���, 2016 3 / 14

Introduction Co-simulation with HiL Case Studies Conclusions

DEPARTMENT OF ENGINEERING

AARHUS UNIVERSITY AU

M���������� High penetration of renewables (42% of the

consumption in Denmark in 2015 from windturbines)

� Flexibility required for generation-consumptionbalancing

� Demand response as a solution⌅ Help renewables integration⌅ Create market opportunities at distribution level

� Need tools for comprehensive assessment of newdemand response opportunities

Source: Energinet.dk (2015)H�L C������������ ��� D����� R������� S���� R������G����� 23 J���, 2016 3 / 14

Introduction Co-simulation with HiL Case Studies Conclusions

DEPARTMENT OF ENGINEERING

AARHUS UNIVERSITY AU

C������������ E����������Virtual Grid Integration Laboratory (VirGIL)

FMI FMI

FMIFMI

Master Algorithm

Ptolemy IICommunication

Networks SimulatorOMNet++ FMU

Control ModelsSimulator

Modelica FMU

Building ModelsSimulator

Modelica, EnergyPlusFMU

Power Systems SimulatorPowerFactory

FMU

FMI standard: Functional Mockup InterfacePtolemy II (actor-oriented simulation so�ware)

Source: Chatzivasileiadis et al. (2016)H�L C������������ ��� D����� R������� S���� R������G����� 23 J���, 2016 4 / 14

Introduction Co-simulation with HiL Case Studies Conclusions

DEPARTMENT OF ENGINEERING

AARHUS UNIVERSITY AU

C�����������

FMI

Hardware inthe Loop

Ptolemy II

FMI

FMIFMI

Ptolemy II

Master Algorithm

Ptolemy IICommunication

Networks SimulatorOMNet++ FMU

Control ModelsSimulator

Modelica FMU

Building ModelsSimulator

Modelica, EnergyPlusFMU

Power Systems SimulatorPower Factory

FMU

Extension

Coupling of co-simulation environment (VirGIL) with aHardware-in-the-Loop (HiL) infrastructure, ventilation sy-stem of large building, for demand response assessment

H�L C������������ ��� D����� R������� S���� R������G����� 23 J���, 2016 5 / 14

Introduction Co-simulation with HiL Case Studies Conclusions

DEPARTMENT OF ENGINEERING

AARHUS UNIVERSITY AU

T��� B��: G������� D�������� L��

� Student residence in Aarhus (Denmark)� 12 floors, 159 apartments, 200 students� 3,400 sensors (5 s resolution)

⌅ indoor climate conditions⌅ electricity consumption⌅ domestic water⌅ district heating⌅ others

Source: www.vpp4sgr.dkH�L C������������ ��� D����� R������� S���� R������G����� 23 J���, 2016 6 / 14

Introduction Co-simulation with HiL Case Studies Conclusions

DEPARTMENT OF ENGINEERING

AARHUS UNIVERSITY AU

O������

1 Introduction

2 Co-simulation with Hardware-in-the-Loop

3 Case Studies

4 Conclusions

H�L C������������ ��� D����� R������� S���� R������G����� 23 J���, 2016 7 / 14

Introduction Co-simulation with HiL Case Studies Conclusions

DEPARTMENT OF ENGINEERING

AARHUS UNIVERSITY AU

C��S��������� ���� H�L � C��� S����

Grundfos Dormitory Lab

Distribution GridModel

Curtailment Service Provider

Ext

erna

l Dat

a P

rovi

ders

H�L C������������ ��� D����� R������� S���� R������G����� 23 J���, 2016 7 / 14

Introduction Co-simulation with HiL Case Studies Conclusions

DEPARTMENT OF ENGINEERING

AARHUS UNIVERSITY AU

C��S��������� ���� H�L � C��� S����

Grundfos Dormitory Lab

Distribution GridModel

historical nodes load

PowerFactory

Curtailment Service Provider

Optimization & Control

node loads, line loads...

Ptolemy II

Exte

rnal

Dat

a P

rovi

ders

electricity price,wind speed...

H�L C������������ ��� D����� R������� S���� R������G����� 23 J���, 2016 7 / 14

Introduction Co-simulation with HiL Case Studies Conclusions

DEPARTMENT OF ENGINEERING

AARHUS UNIVERSITY AU

C��S��������� ���� H�L � C��� S����

Grundfos Dormitory Lab

OpenADR Client

OpenADR Server DR shed

DR capability,fan load, DR yes/no

OMNeT++

Distribution GridModel

historical nodes load

PowerFactory

Curtailment Service Provider

Optimization E Control

DR shedfan load

node loads, line loads...

Ptolemy II

Exte

rnal

Dat

a P

rovi

ders

electricity price,wind speed...

H�L C������������ ��� D����� R������� S���� R������G����� 23 J���, 2016 7 / 14

Introduction Co-simulation with HiL Case Studies Conclusions

DEPARTMENT OF ENGINEERING

AARHUS UNIVERSITY AU

C��S��������� ���� H�L � C��� S����

Controller

Ventilation Fans Model

Indoor CO2 Model

Grundfos Dormitory Lab

airflow

static pressure

OpenADR Client

external CO2

DR capability

fan load

indoor CO2 generation

DRshed

OpenADR Server DR shed

DR capability,fan load, DR yes/no

OMNeT++

+

historical apartments load

all buildingloadDistribution Grid

Model

historical nodes load

PowerFactory

Curtailment Service Provider

Optimization z Control

DR shedfan load

node loads, line loads...

Ptolemy II

fan load

Exte

rnal

Dat

a P

rovi

ders

electricity price,wind speed...

HiL

H�L C������������ ��� D����� R������� S���� R������G����� 23 J���, 2016 7 / 14

Introduction Co-simulation with HiL Case Studies Conclusions

DEPARTMENT OF ENGINEERING

AARHUS UNIVERSITY AU

D����������� G���M��������� PowerFactory (integrated

in VirGIL via FMI standard)� Grid data provided by DSO� Test bed electricity

consumption from sensors� Assumptions:

⌅ 3 times more densearea

⌅ all buildings similarpower consumption

H�L C������������ ��� D����� R������� S���� R������G����� 23 J���, 2016 8 / 14

Introduction Co-simulation with HiL Case Studies Conclusions

DEPARTMENT OF ENGINEERING

AARHUS UNIVERSITY AU

H���������������L��� C��������� Ventilation fans interaction

Secured API (HTTP-Modbus)Controller Air Handling Unit API

HTTPS POST (SupplyFanPressure)

HTTPS POST (ExhaustFanPressure)

HTTPS 201

HTTPS GET (SupplyFanPower)

HTTPS GET (ExhaustFanPower)

HTTPS GET (Airflow)

HTTPS GET (SupplyFanPressure)

HTTPS 201

HTTPS 200

HTTPS 200

HTTPS 200

HTTPS 200

HTTPS GET (ExhaustFanPower)

HTTPS 200

� CO2 sensor data: MongoDB query in Ptolemy IISource: Rotger-Griful et al. (2016)

H�L C������������ ��� D����� R������� S���� R������G����� 23 J���, 2016 9 / 14

Introduction Co-simulation with HiL Case Studies Conclusions

DEPARTMENT OF ENGINEERING

AARHUS UNIVERSITY AU

O���� C���������� Communications Networks

⌅ OMNet++⌅ Demand response protocol: OpenADR

� Demand response controller:⌅ Ptolemy II⌅ Multiple control strategies

� Master algorithm⌅ Ptolemy II

� Prior execution time analysis⌅ simulation time < real time

H�L C������������ ��� D����� R������� S���� R������G����� 23 J���, 2016 10 / 14

Introduction Co-simulation with HiL Case Studies Conclusions

DEPARTMENT OF ENGINEERING

AARHUS UNIVERSITY AU

O������

1 Introduction

2 Co-simulation with Hardware-in-the-Loop

3 Case Studies

4 Conclusions

H�L C������������ ��� D����� R������� S���� R������G����� 23 J���, 2016 11 / 14

Introduction Co-simulation with HiL Case Studies Conclusions

DEPARTMENT OF ENGINEERING

AARHUS UNIVERSITY AU

L��� F�������� ��� O������ RES I����������� Strategy: 1-minute load reference based on measured

on-site wind speed

� Observations:⌅ Building CO2 level OK⌅ Capable of following signal

H�L C������������ ��� D����� R������� S���� R������G����� 23 J���, 2016 11 / 14

Introduction Co-simulation with HiL Case Studies Conclusions

DEPARTMENT OF ENGINEERING

AARHUS UNIVERSITY AU

A�������� S������� � P������ R�������� Strategy: Primary reserves for up-regulation! reduce

consumption

� Observations:⌅ Inner fan controller limits response time⌅ Aggregation required for market participation (� 300

systems)H�L C������������ ��� D����� R������� S���� R������G����� 23 J���, 2016 12 / 14

Introduction Co-simulation with HiL Case Studies Conclusions

DEPARTMENT OF ENGINEERING

AARHUS UNIVERSITY AU

L��� A���������� ���� L��� F��������� Strategy: 15 ventilation systems following 1-minute varying

reference (white noise added)

� Observations:⌅ 2% transformer load change in few seconds⌅ Communication protocol e�ect (tuning time between

polling events)H�L C������������ ��� D����� R������� S���� R������G����� 23 J���, 2016 13 / 14

Introduction Co-simulation with HiL Case Studies Conclusions

DEPARTMENT OF ENGINEERING

AARHUS UNIVERSITY AU

O������

1 Introduction

2 Co-simulation with Hardware-in-the-Loop

3 Case Studies

4 Conclusions

H�L C������������ ��� D����� R������� S���� R������G����� 23 J���, 2016 14 / 14

Introduction Co-simulation with HiL Case Studies Conclusions

DEPARTMENT OF ENGINEERING

AARHUS UNIVERSITY AU

C����������� Propose co-simulation with Hardware-in-the-Loop (HiL) to

support demand response development� Extend the VirGIL co-simulation platform with HiL of

ventilation system� Proposed HiL approach can be deployed to other systems� Case studies:

⌅ Ventilation system can respond to fast demandresponse requests

⌅ Ancillary services regulations should reduceparticipation barriers for demand response resources

⌅ Communication protocol parameters can a�ect theloading of power components

H�L C������������ ��� D����� R������� S���� R������G����� 23 J���, 2016 14 / 14