2. Dimitris Labridis (AUTH) - Presentation of the Theoretical Concepts and Modelling Aspects of the Cassandra Platform

Theoretical Concepts
and Modelling
Aspects of the
Cassandra Platform
Dimitris Labridis
Georgios Andreou
Apostolos Milioudis
Aristotle
University of
Thessaloniki
(AUTh)

Power Distribution Network
Cassandra Project 1st Workshop September 11, 2013
Small Industry
MV/LV Power
Transformer
Commercial
MV/LV Power
Transformer
MV/LV Power
Transformer
Residential
Residential
Commercial
Commercial
Overhead or Underground
Power Lines
HV/MV Power
Substation
Residential

Power Distribution Network
• Consumer Side (Demand Side)
• Demand Side Management
– Consumption optimization
– Energy efficiency
– Profit
• Problems
– Which is the profile of each
consumer?
– How to cluster/categorize
consumers?
– Which DSM program/incentive to offer
to each consumer?

Cassandra as a solution
• A helping hand in order to:
– Simulate the consumption of consumer populations,
– Understand behaviors that drive consumption,
– Cluster consumers into groups for more efficient incentivisation,
– Simulate consumer response to specific Demand Response programs
• Two modeling approaches:
– Input based on demographic data → aggregate results for a consumer
population
– Input based on consumer level measurements → bottom-up approach
with results regarding individual consumers and/or consumer
populations

Basic modeling concepts
• Reality:
– Installation (the aggregation of all the appliances installed within the
premises of a consumer)
– Person(s) performing activities by the use of appliances according to
their utility
– Person(s) have habits (recurrent behaviour)
– Person(s) receive incentives to modify consumption, and choose either
to respond or not according to their utility

Basic modeling concepts
• Modeling:
– An Installation consists of its Appliances
– Measurements of the installation are used in order to determine its
Appliances, as well as the consumer Activities (Disaggregation
procedure)
– Appliances are described by Consumption Models
– Consumer utility is modelled through Utility Models
– Consumer response to specific incentives is modelled through Response
Models
– Auxiliary Models:
• Pricing Schemes Models
• CO2 Emission Calculation
• Prosumer Generation

Functionality according to quality of input
• Basic simulation rule: The more information you offer as input,
the more accurate results you will obtain
• Demographic data input: The simulation will be based on statistics,
thus it will not produce accurate individual consumer results
• Measurement data input:
– The Cassandra modeling is based on per minute measurements of
active and reactive power for an installation (loss of information)
– Accuracy of results depends on information compensation

Appliance Consumption Models
• They define mathematically the active and reactive power
demand of an appliance for a specified time of operation
• Example: Refrigerator active power
𝑐 𝑝 𝑡 =
𝑘=1
𝑁
99.8 ∙ 𝑢 𝑡 − 𝑘2520 − 𝑢 𝑡 − 620 − 𝑘2520 , 𝑡 < 𝑑0

Activity Models
• They describe mathematically the occurrence, duration and
composition of an activity
𝐴 → 𝒂, 𝑓𝑠 𝑡 , 𝑓𝑑 𝑡 , 𝑓𝑁 𝑛
A set of appliances
associated with the
activity
The probability density
function (pdf) of the activity
start time(s) within a day
The pdf of the number of
times the activity is
performed during a day
The pdf of the activity
duration
𝑓𝑎 𝑖
𝑡 𝐴 : The probability that an appliance will be triggered at a
time t after the activity initiation (may differ according to
day type and/or season)

Activity Models
• Procedure:
– Correlation of Appliances and Activities (installation modelling phase):
Each Appliance may be correlated with a single Activity
– Determination of Appliances and respective Activities within
measurements (Disaggregation phase)
– Training of the Activity Modelling procedure

Response Models
• The aim of the CASSANDRA platform in this case is to estimate
whether a consumer will respond to a specific incentive or not
• Factors affecting behavioural change are:
– Pricing
– Environmental impact
– Awareness and Sensitivity
• Time shifting models describe the changes of the probability
density function fs(t) of the activity start time(s) within a day, as a
result of the change of price
• Consumers may also respond by reducing the number of times
they perform some activities

Disaggregation Methodology
• Aims to provide information regarding the appliances that
produce a given set of active and reactive power curves.
• Step 1: Determination of background demand

Disaggregation Methodology (cont.)
• Step 2: Determination of general events

• Step 3: Determination of repeated events

• Step 4: Determination of individual appliances
– Slope Percentage Vector (SPV) calculation from active power
measurements vector Px
– Combination of Points of Interest (POIs) leads to POI vector
– Aggregated consumption curved is decomposed into events

B. Disaggregation Methodology (cont.)
• Step 5: Identification (clustering) of individual appliances
– Classification according to specific parameters:
• Active power
• Reactive power (power factor)
• Differentiation of active/reactive power during operation
• Typical (average) time of use for each appliance
– Matching in appliance library

Results in Pilot Cases (Lulea – 40 days)
Background
45,86 %
Refrigeration 9,79%
Cooking
18,79 %
Entertainment 6,67 %
Cleaning
9,31 %
Lighting
0,65 %
Unidentified 8,93 %
Percentage Energy Per Activity for Installation A

Results in Pilot Cases (Lulea – 40 days)
Background
6,36 %
Refrigeration
13,76 %
Cooking
16,17 %
Entertainment
3,78 %
Cleaning
48,08 %
Lighting
0,32 %
Unidentified
11,53%
Percentage Energy Per Activity for Installation B

Load identification - Lulea installation B
0
200
400
600
800
1000
1200
1400
1600
1800
2000
TV1 TV2 Refrigerator Freezer Extra Freezer PC Cooker Microwave
Oven
Vacuum
Cleaner
TimesofUSe
Installed Appliance

Related Conference Papers
1. A Framework for the Implementation of Large Scale Demand Response
G. T. Andreou, A. L. Symeonidis, C. Diou, P. A. Mitkas and D. P. Labridis, IEEE
International Conference on Smart Grid Technology, Economics and Policies (SG-TEP
2012), December 3-4, 2012, Nürnberg, Germany.
2. Redefining the Market Power of Small-Scale Electricity Consumers through
Consumer Social Networks
K. C. Chatzidimitriou, K. N. Vavliakis, A. L. Symeonidis and P. A. Mitkas, 10th IEEE
International Conference on e-Business Engineering (ICEBE 2013), September 11-13,
2013, Coventry, UK.
3. Event Detection for Load Disaggregation in Smart Metering
A. N. Milioudis, G. T. Andreou, V. N. Katsanou, K. I. Sgouras and D. P. Labridis, 4th IEEE
European Innovative Smart Grid Technologies (ISGT 2013) Conference, October 6 - 9,
2013, Copenhagen, Denmark.
4. Agent-based Small-Scale Energy Consumer Models for Energy Portfolio
Management
A. Chrysopoulos, C. Diou, A. L. Symeonidis and P. A. Mitkas, 2013 IEEE/WIC/ACM
International Conference on Intelligent Agent Technology (IAT-13), November 17-20,
2013, Atlanta, Georgia, USA.

Professor Dimitris P. Labridis
Director of Power Systems Laboratory (PSL)
Department of Electrical and Computer Engineering
Aristotle University of Thessaloniki (AUTh)
P.O. Box 486
54124 Thessaloniki, Greece
tel: +302310996374
fax: +302310996302
e-mail: labridis@auth.gr
Skype: labridis
URL: http://users.auth.gr/labridis/index_en.htm
PSL: http://power.ee.auth.gr/index.php?page=intro&action=lnchange&lnid=124&lnsym=EN
AUTh: http://www.auth.gr/
Authors affiliations

Lecturer Georgios T. Andreou
Power Systems Laboratory (PSL)
P.O. Box 486
tel: +302310996118
fax: +302310996302
e-mail: gandreou@auth.gr
URL: http://power.ee.auth.gr/index.php?page=andreou
Authors affiliations (cont.)

Dr. Apostolos N. Milioudis
Power Systems Laboratory (PSL)
P.O. Box 486
tel: +302310996379
fax: +302310996302
e-mail: amilioud@auth.gr
URL: http://power.ee.auth.gr/index.php?page=milioudis
Authors affiliations (cont.)

2. Dimitris Labridis (AUTH) - Presentation of the Theoretical Concepts and Modelling Aspects of the Cassandra Platform

More Related Content

Viewers also liked (7)

Similar to 2. Dimitris Labridis (AUTH) - Presentation of the Theoretical Concepts and Modelling Aspects of the Cassandra Platform (20)

More from Cassandra Project (7)

Recently uploaded (20)

2. Dimitris Labridis (AUTH) - Presentation of the Theoretical Concepts and Modelling Aspects of the Cassandra Platform