The Common Information Model (CIM) is described using the Unified Modeling Language (UML). UML can also describe data model of cyber-physical power system components and networks. However, there are several difficulties to transform the data model into a strictly defined mathematical model. A strictly defined mathematical model is one for which all-differential algebraic and discrete model equations are explicitly defined [1] (i.e. the equations are written in human readable form). This is known as equation-based modelling [2], and it is utilized in many areas such as the automotive and aerospace industry [3]. The automated generation of an unambiguous equation-based model would allow performing time-domain simulations of cyber-physical power systems [4] and the assessment of textual requirements, which could be defined from the UML model directly [5]. This flexibility would allow adopting model-based systems engineering practices within the power industry, such as those used in process control. For the implementation of models in an equation-based language, the Modelica language [6] is the one of the best choices because it follows the Object Oriented Programming (OOP) notation, with a close relation with UML. Furthermore, the ModelicaML [5], an extended subset of the OMG Unified Modeling Language, enables integrated modelling and simulation of system requirements and design. Combining CIM, ModelicaML and Modelica models of cyber-physical power system components it is possible to automatically generate unambiguous mathematical models that can be used for simulation and requirements verification [7]. This CIM to Modelica Factory talk explores this possibility. One of the main challenges that we face with power systems models defined using the Modelica language is the initialization of the dynamic states (in equilibrium condition) of the components within a model [8, 9]. However, objects and components modelled in CIM standard contain attributes for storing a power flow solution. The purpose of the work described in this presentation is to develop a software tool capable to transform a CIM object model into a Modelica model that can be directly simulated using different Modelica engines. To this aim, we start from the CIM/UML representation of power system components and models, and exploit the ModelicaML profile to achieve a proper code representation of the power system in Modelica code. To confront issues with dynamic initialization, the power flow solution from CIM is linked to the Modelica component models and utilized within the initialization algorithms of the simulation engines. The result is a software tool that allows performing time domain simulations directly from a CIM/UML structure, while maintaining consistency in the resulting mathematical model within different simulation engines.

1. CIM 2 Modelica Factory
Automated Equation-Based Cyber-Physical Power System Modelica
Model Generation and Time-Domain Simulation from CIM
Francisco Gomez1, Prof. Luigi Vanfretti1,2, and Svein
H. Olsen2
luigiv@kth.se , fragom@kth.se
Electric Power Systems Dept.
KTH
Stockholm, Sweden
Luigi.Vanfretti@statnett.no
svein.harald.olsen@statnett.no
Research and Development Division
Statnett SF
Oslo, Norway

2. • Background & Motivation
– Modeling and Simulation
– Modelica
– CIM for Dynamics
• Modelica
– Language Description
– MetaModelica
– CIM/UML to Modelica
• CIM 2 Modelica
– Initial Conditions
– Simulations
Overview
Acknowledgments
• This work has been funded in part by the
EU funded FP7 iTesla project:
http://www.itesla-project.eu/ and Statnett
SF, the Norwegian power system operator.
• Work related to the iTesla Modelica power
systems library presented here is a result of
the collaboration between RTE (France),
AIA (Spain) and KTH (Sweden) within the EU
funded FP7 iTesla project:
http://www.itesla-project.eu/
• Special thanks for ‘special training’ and
support from
• Prof. Fritzson and his team at Linköping University
• Prof. Berhard Bachmann and Lennart Ochel, FH
Bielefeld

3. • Research
• Development of software architecture
supporting transformation from CIM,
implementing tools for either translating
from CIM to Modelica models
• Development of models of cyber-physical
power systems components,
communication network components, and
other components from other domains
• Application
• iTESLA: Innovative Tools for Electrical
System Security within Large Areas
• CIM provides standard format for power
systems data
• Use of data from TSO
– Description of data equipment, power systems
topology and measurements for model validation
Motivation

4. Modelica
• Modelica is an OOP for declarative
equation based mathematical language
• Non-proprietary language, suitable for
standardization and exchange of models
• Modelica tools, commercial and free of
charge
• Electric power steering and
controller model
[1] Andreas Deuring, Johannes
Gerl, Harald Wilhelm
“Multi-Domain Vehicle Dynamics
Simulation in Dymola”,
Modelica Conference, Dresden,
2011
• Thermodinamic Network
of the ICE model
[2] L. Morawietz, S. Risse, H.
Zellbeck, H. Reuss, T. Christ
“Modeling an automotive power
train and electrical power supply
for HiL applications using
Modelica”,
Modelica Conference, Hamburg,
2005
TU Dresden, University of
Stuttgart, BMW Group,
Germany.
Background

5. Common Information
Model
• Conceived for information exchange:
power systems topology, equipment,
measurements
• Using UML representation to design a
structured data model: Semantic
transformation from real world to a
model
• Standardization of the model diagrams
for cyber-physical components
• Generators
• Turbine Governors
• Capacitors
• Protections
• Measurements
• IEC61970 provides standard data
model for power systems components
Background
class Sy nchr onousMachineDy namics
Wir es::Sy nchr onousMachine
+ aVRToManualLag :Seconds [0..1]
+ aVRToManualLead :Seconds [0..1]
+ baseQ :ReactivePower [0..1]
+ condenserP :ActivePower [0..1]
+ coolantCondition :Float [0..1]
+ coolantType :CoolantType [0..1]
+ earthing :Boolean [0..1]
+ earthingStarPointR :Resistance [0..1]
+ earthingStarPointX :Reactance [0..1]
+ ikk :CurrentFlow [0..1]
+ manualToAVR :Seconds [0..1]
+ maxQ :ReactivePower [0..1]
+ maxU :Voltage [0..1]
+ minQ :ReactivePower [0..1]
+ minU :Voltage [0..1]
+ mu :Float [0..1]
+ operatingMode :SynchronousMachineOperatingMode [0..1]
+ qPercent :PerCent [0..1]
+ r :Resistance [0..1]
+ r0 :Resistance [0..1]
+ r2 :Resistance [0..1]
+ referencePriority :Integer [0..1]
+ satDirectSubtransX :PU [0..1]
+ satDirectSyncX :PU [0..1]
+ satDirectTransX :PU [0..1]
+ shortCircuitRotorType :ShortCircuitRotorKind [0..1]
+ type :SynchronousMachineKind [0..1]
+ voltageRegulationRange :PerCent [0..1]
+ x0 :Reactance [0..1]
+ x2 :Reactance [0..1]

6. CIM for Dynamics
• Dynamic models used in the
industry today use application
specific data format and are
embedded within the solver
(integration routine)
• No information on how the
model is implemented (i.e.
actual equations used)
• Dynamic models can be
represented in CIM, and
exchanged among utilities
• Need to extend CIM to support
dynamics models
• CIM should also support the
exchange not only of parameters
Background
doc Sy nchr onousGener ator MechanicalEquation

7. Extension for CIM
Dynamics
• Automatic model
transformation from CIM to
a well defined (equation
based) language
• Information exchange,
parameters and equations
with CIM and Modelica
The benefit / role of CIM:
Modelling of the network are done
separate from the analytic
Existing Steady-State Solver Engine
(SSSE) can be used to initialize the
transient study
Background
model gensal
…
parameter Real wbase = 2 * pi * 50 "system base speed";
parameter Complex Epqp = fpp + a * It;
parameter Real delta0 = arg(Epqp);
parameter Real Pm0 = p0 + (id0 * id0 + iq0 * iq0) * Ra;
Real delta "rotor angle";
Real w "machine speed deviation, p.u.";
…
initial equation
delta = delta0;
w = 0;
equation
…
der(w) = ((Pm0 - D * w) / (w + 1) - Te) / (2 * H);
der(delta) = wbase * w;
end gensal;

8. • Background & Motivation
– Modeling and Simulation
– Modelica
– CIM for Dynamics
• Modelica
– Language Description
– MetaModelica
– CIM/UML to Modelica
• CIM 2 Modelica
– Initial Conditions
– Simulations
Overview

9. Modeling Language
• Modeling language based
on equations, allow
specification of
mathematical models
• Multi-Domain modeling
• Visual Acausal Hierarchical
Component Modeling
• Physical structure
• No specification of data
flow direction load
EM
DC
G
R L
Electrical
Mechanics
model DCMotor
Modelica.Electrical.Analog.Basic.Resistor r1(R = 10);
Modelica.Electrical.Analog.Basic.Inductor i1;
Modelica.Electrical.Analog.Basic.EMF emf1;
Modelica.Mechanics.Rotational.Inertia load;
Modelica.Electrical.Analog.Basic.Ground g;
Modelica.Electrical.Analog.Sources.ConstantVoltage v;
equation
connect(DC.p,R.n);
connect(R.p,L.n);
connect(L.p,EM.n);
connect(EM.p,DC.n);
connect(DC.n,G.p);
connect(EM.flange_b,load.flange_a);
end DCMotor;
Modelica

10. Modeling Language
• Typed Declarative
Equation-based Textual
Language
• Object-Oriented Language
with class concept
• Reuse of classes
• Reuse of components
• Scalable and Modular
models
• Decoupling the model
from the solver
model GENROU
parameter Complex It=conj(S/VT) “Some comments here“;
parameter Complex Is = It + VT/Zs;
parameter Complex fpp = Zs*Is;
parameter Real ang_P=arg(fpp);
parameter Real ang_I=arg(It);
parameter Real ang_PI=ang_P-ang_I;
parameter Real psi = 'abs'(fpp);
equation
der(Epq) = (1/Tpd0)*(Efd0 -XadIfd);
der(Epd) = (1/Tpq0)*(-1)*(XaqIlq);
…
anglev =atan2(p.vi, p.vr);
Vt = sqrt(p.vr^2 + p.vi^2);
anglei =atan2(p.ii, p.ir);
I = sqrt(p.ii^2 + p.ir^2);
…
end GENROU;
Variable
declaration
DAE Equations
Modelica

11. Power Systems Library in
Modelica
• The FP7 iTESLA project develops
a high level library for modeling
power grid components
• Generators,
• Governors,
• Controls,
• Branches,
• Loads,
• Buses,
• Events
• The library makes available
standardized power systems
models usually available in
power system tools only
accessible through proprietary
(and expensive) licenses
Modelica

12. CIM / UML to Modelica
• Modelica provides data definition
and compilers for equation based
modeling
• ModelicaML is a tool to create UML
definition for Modelica models
• Design of classes, components and
models using a data model
representation:
• Definition of start values for
components and definition of
mathematical equations
• Code generation creates classes and
models with relation between classes
Modelica

13. CIM / UML to Modelica
• Semantic transformation for
automatic simulation directly
from CIM definition
Modelica

14. • Background & Motivation
– Modeling and Simulation
– Modelica
– CIM for Dynamics
• Modelica
– Language Description
– MetaModelica
– CIM/UML to Modelica
• CIM 2 Modelica
– Initial Conditions
– Simulations
Overview

15. Process flow design
• Automatic generation of Modelica code from
CIM/UML definition
• Manual design of CIM/UML definition and
Mapping
• Loading CIM/XML and Mapping
• Semantic transformation into Modelica code:
– Set initial values from load flow solution
– Set connection between classes
CIM 2 Modelica

16. CIM 2 Modelica Mapping
• Relation between CIM classes
and Power system library classes
• CIM Attributes and values ->
Modelica Variables and starting
values
• CIM relations between classes ->
Modelica connection between
components
or
• CIM relations between classes ->
Use of Modelica classes as objects
CIM 2 Modelica

17. Modelica Simulation Engine
(architecture)

18. Simulation Engine
• Open-source software for
cyber-physical system
simulation
• Plug-in different compilers and
solvers
• Enhancement to CIM:
• Integration with PMU
measurements or simulation ->
Harmonization with HDFS is an
alternative
• Include the Modelica library
code as part of the CIM standard
CIM 2 Modelica
Properties
Results
HDF5
JMOMC
PYTHON
JAVA
Dymola

19. Thank you! Questions?
luigiv@kth.se , fragom@kth.se
Electric Power Systems Dept.
KTH
Stockholm, Sweden
Luigi.Vanfretti@statnett.no
svein.harald.olsen@statnett.no
Research and Development Division
Statnett SF
Oslo, Norway