Do We Need a New Architecture for Simulating Power Systems?

Gary Howorth

and Ivana Kockar

Institute of Energy and Environment, University of Strathclyde, 99 George St., Glasgow, U.K.

Institute of Energy and Environment, University of Strathclyde, 204 George St., Glasgow, U.K.

Keywords: Agent based Modelling, Multi Agent Systems, Multiscale, Multilayer, Power, Simulation, Smartgrid.

Abstract: The delivery of electric power and organisation of electric power systems is an incredibly complex system,

which will become ever more complex with an increased penetration of Distributed Energy Resources (DER),

including electric vehicles and renewables. Optimization of such a system using conventional techniques is

difficult and fraught with a myriad of issues, so simulation provides a more holistic approach to understanding

the evolving issues. We argue that although power simulation frameworks exist they may be inadequate for

simulating a more complex and evolving smart power grid infrastructure. A brief overview of research on

existing systems is provided and this paper argues for the development of a distributed multi-scale, multi

layered hybrid ABM/MAS system.

1 INTRODUCTION

In his work, Watts (2003) characterised power

systems as complex systems. Others have followed

this work but have focused on using graph based

statistical measures to help understand their stability

e.g. cascades and other issues. However Holland

(1999) suggests that a graph based statistical or

mathematical approach gives us no insight as to how

patterns e.g. price or power flows in our instance

arise. It tells us nothing about emergence in what is a

Complex Adaptive System (CAS). This emergence in

our context is an important element to understand and

enable us to develop (i) policy rules for future system

operation and (ii) appropriate short-term technical

control measures. Currently only a simulation

approach appears to provide us with a route to

understanding complex dynamics in these systems.

However, Agent Based (ABM) and Multi Agent

System (MAS) modelling could also provide us with

an important arsenal in discovering these patterns.

Holland also suggests that as we add stochastic

mobile agents to the system the potential for new

rules and patterns to emerge tends to grow rapidly,

resulting in many more persistent states as a result of

system emergence. In our case, electric vehicles

(EV’s) are the stochastic mobile agents of Holland’s

discussions. Understanding this complex system

behaviour seems to be an important ideal for us to

achieve. Essentially, our model of the power system

has to represent its physical and commercial aspects,

including the grid, the market, the generators, the

demand, DERs etc. The model also needs to include

an ability to represent and study adapting participant

behaviours. Therefore, it is now recognized that a

more holistic approach is required, one that ultimately

requires a more sophisticated simulator. This short

position paper presents initial scope/ideas for an

ABM/MAS system that meets our needs for an

electrical power simulator and can look at complex

interactions in a future power system (Smart Grid). It

is therefore a concept paper and is not meant to solve

and give detailed architectural designs for our

framework at this stage.

The paper is organised as follows. In Section 2,

relevant examples of academic research on ABM and

MAS are presented. Section 3 focusses on

ABM/MAS systems used in electrical engineering,

while section 4 proposes the conceptual design of a

new power system simulator. Finally, Section 5

concludes this paper.

2 OVERVIEW OF ABM/MAS

LITERATURE

There have been many surveys on ABM and MAS

systems (Heath et al., 2009, Kantamneni et al., 2015,

Leon et al., 2015, McArthur et al., 2007) plus many

190

Howorth, G. and Kockar, I.

Do We Need a New Architecture for Simulating Power Systems?.

DOI: 10.5220/0006917801900197

In Proceedings of 8th International Conference on Simulation and Modeling Methodologies, Technologies and Applications (SIMULTECH 2018), pages 190-197

ISBN: 978-989-758-323-0

dedicated webpages and tutorials outlining what they

can do in terms of Agent Communication Languages,

openness, programming language choice and so on.

Typically, the systems focus on a particular area of

research i.e. they are specialised to analyse particular

features or behaviour, or have other limitations. Only

a few specifically focus on power system

applications. There are now over 70 ABM/MAS

systems in existence with a typical lifespan of 4-5

years. Only a few of these systems /designs (e.g. Jade

(Bellifemine et al., 2007) , Repast (Cardoso, 2009)) )

have lifespans in excess of 10 years.

Few researchers propose aggregation of systems

or components or simple reuse. However, Cardoso

(2015) in his paper on SAJas proposes the use of an

API to join Repast to Jade, with a justification that

“multi-agent based system simulations (MABS)

focus on applying MAS to model complex social

systems typically involving a large agent population.

Several MAS frameworks exist, but they are often not

appropriate for MABS”. In a similar vein, Gormer et

al. (2011) propose the JRep framework for simulating

an agent-based airport scenario, linking Repast

(ABM) to Jade (MAS). As they note, “existing agent

frameworks focus on either the macro or the micro

perspective”, but don’t combine the two. The

objective of the proposed combination of an ABM

and MAS frameworks will allow us to better

understand power system organisation/operation.

3 POWER MODELS IN ABM/MAS

Although there are over 70 ABM/MAS systems in

existence few have been developed to address power

systems, with categorisation and analysis of MAS

applications in power provided in (Sujil et al., 2016).

Table 3 in that review summarizes papers that have

focused on lower level distributed simulation, while

Table 4 provides a useful breakdown of papers that

deal with specific power issues e.g. markets,

generators etc.

For example, AMES (Repast) (Li and Tesfatsion,

2009), ECMAS (Repast) (Conzelmann et al., 2005),

EMLab (Agent Spring) (De Vries et al., 2013) are

specific modelling environments that have power

system implementations, while “lower level

modelling” of multi agent systems using Presage2 is

presented in (Chen et al., 2016, Macbeth, 2015).

Furthermore, Anylogic (AnyLogic, 2018), which is

not specifically designed for power systems, is a

proprietary system that could be used and has an

architecture design whose logic allows analysts to

model not only agents, but also discrete events and

also use system dynamics. The agent behaviours can

be modelled using JavaScript, but this is too limiting

for our purposes, as we require a fully-fledged

asynchronous Object Orientated Programming

(OOP) language to model complex interactions. More

sophisticated agents using Java and Neural Nets

which are linked into AnyLogic using a Java Archive

file (JAR) have been developed in (Wallis and Paich,

2017). This method requires a greater degree of

programmer intervention to link in the various

components and is less flexible than our proposed

conceptual design.

EMLab has based their system on the Neo4J graph

database (Merkl Sasaki et al., 2017), rather than build

a Relational Database Management System (RDMS).

This is an open source/commercial system used by

many to analyse Twitter feeds and relationships. It is

a very efficient and can be used to store knowledge

maps, power networks and, most importantly, the

relationships between agents in different layers and

between agents on the same layer. It can in the right

circumstances be a faster than a normal database

(RDMS). It can also be used to quickly analyse

networks and identify problem nodes for example.

Due to its features which also fit nicely with the

typical representations of power grids (i.e. they are

node based) it may be a useful base for a future power

simulator framework.

Furthermore, the Mosaik platform (Rohjans et al.,

2013) is designed to allow reuse of components like

Matlab and other “simulators to create large-scale

Smart Grid scenarios”. It is written in Python,

provides an API for connecting these different

simulators including MATLAB, and uses JSON to

communicate between packages.

4 AN IDEAL ABM/MAS POWER

SIMULATION SYSTEM

An ideal ABM/MAS power simulator would provide

the following functionality:

 A design which allows us to understand how

different behaviours of the various power

system agents (generators aggregators,

consumers and policy makers etc.) will affect

the system technically and commercially. For

example, how will power flows across the

system change? How will prices change at

various nodes in the system? Will they be too

high? Will resulting power flows cause

congestion in the system and require new

investment?

Do We Need a New Architecture for Simulating Power Systems?

191

 Test out new policy rules or potential control

algorithms;

 Try out different agent behaviour techniques

e.g. policy agent rules, different agent learning

paradigms;

 Understand how voltage extremes (which

would result in system performance

degradation or failure) might be generated by

behaviours of participants in the system, which

in turn might be driven by policy makers rules;

The framework will also need to have the following

features:

 Has both a market and network physical (i.e.

power flow) layers that are incorporated into

market clearing and bidding representations;

 Is extensible and is truly layer based allowing

us the ability to switch in and out different

layers and change agent behaviours as needed;

 Easy to use;

 Could be solved in distributed manner so to

allow analysis of large scale networks;

 Plug in based while using a Component

modularity architecture;

 Have models of agents representing various

actors such as generators, loads, electric

vehicles, aggregators, storage, atomic and

temperature controlled loads, system operators,

regulators, and companies;

 To reuse existing software components where

ever possible;

4.1 Proposed Architecture

These requirements appear to drive us to a conceptual

design that would include the following elements

shown in Figure 1.

We discuss the main elements of this proposed

architecture in the sections below.

4.1.1 Asynchronicity

A review of ABM surveys and systems shows that

ABM (macro level) models are typically

synchronous, whereas micro level systems such as

Jade (MAS) are typically asynchronous. Youssefmir

and Huberman (1997) have investigated the impact of

asynchronicity on system performance. Figure 2

below shows such an impact from this paper. Note

that the system is usually stable and is punctuated

with periods of instability.

Youssefmir and Huberman's paper considers

multiple agents who take decisions and act on the

system simultaneously in an asynchronous manner to

improve their utility. Our problem domain has exactly

these characteristics and will incorporate learning and

adaptive agent behaviour. Initial experiments on a

simple power system model show similar patterns of

punctuated bursts of activity followed by “stability”.

Figure 1: Proposed high-level architecture.

SIMULTECH 2018 - 8th International Conference on Simulation and Modeling Methodologies, Technologies and Applications

192

Figure 2: Asynchronicity impacts on system dynamics.

Cornforth et al. (2005) discusses MAS agent

update strategies using a cellular automata (CA)

framework as a case study. They examine updating

strategies associated with some real life systems

where the agents behave with different asynchronous

update schemes and compare this with synchronous

updating. The paper provides results on the dynamics

of the CA system, under the different update schemes

and shows that the outputs can be significantly

different. Again initial experiments on a power

system show similar behaviours.

In a future smart grid system, peer to peer (P2P)

bargaining/interactions will have an important impact

both on local conditions and further afield in the

wider power system. By their nature these

transactions are asynchronous, but other parts of the

system will have synchronous interactions e.g. like

market clearing. Designing and testing the

interactions between these types of system e.g. P2P

and the system operator, will allow us to understand

how the system might perform in the future. A

simulator with the ability to try out different

synchronous/asynchronous protocols does not

currently exist in the power domain and would be a

useful addition to the power engineers toolkit.

In finance (Jacobs et al., 2004) developed the

“JLM stock market simulator” to look at the effect of

asynchronous investments on price patterns.

Although written nearly 15 years ago, few authors

and simulator designers have taken this approach,

which as they argue is more realistic. All of the power

system simulators that we have investigated assume

synchronous investments. We know from our own

domain experience, that in the real world, power

investors do not act synchronously. In our context,

earlier large capital investments in infrastructure may

“lockout” later investments, so timing is crucial.

Modelling asynchronous behaviour in our context

is therefore extremely important, so we would argue

that models that ignore asynchronicity would find it

difficult to predict future system states accurately.

Our domain has elements of all the examples we cite

above. We therefore propose that any future power

simulation environment provide a mechanism to

switch between modes of synchronization.

4.1.2 Multi-Scale (Equation Free Modelling)

From a CAS perspective, emergence occurs when

events in one scale (micro) are propagated to another

scale (macro) and vice versa. Capturing those effects

(Holland, 1999), is key to identifying and

understanding emergent behaviour in systems. In the

context of the power domain, we suggest that it is

important that system modellers investigate, these

phenomena, so that they can design appropriate

mitigation strategies. The multi-scale architecture

allows us to model these propagation effects. It also

fits well with the idea that we need to combine ABM

(macro) and MAS (micro) architectures.

However, developing models that can simulate a

combination of events that occur at both the second

(for generators, EV’s) and the years’ timescale (for

investments in infrastructure), are typically

computationally inefficient. We require some kind of

glue or bridge to join these timescales.

There have been many papers on multi-scale

simulations, in recent years, and this provides a

potential solution for our specific problem area.

However, as we discussed above, our systems

Do We Need a New Architecture for Simulating Power Systems?

193

typically are “stable” for large periods and are

punctuated with bursts of activity. Equation free

modelling (DeAngelis and Yurek, 2015, Kevrekidis

et al., 2004, Kevrekidis et al., 2003, Kevrekidis and

Samaey, 2009, Kevrekidis, 2004, Le Maître and

Mathelin, 2010) provides a promising

viewpoint/solution for this particular aspect and we

believe warrants further investigation, particularly in

the methodology to trigger the micro level simulation.

In this regard, there have been far less papers focused

on this specific aspect, especially in recent years. This

approach has not been implemented in the area of

power systems simulation and therefore

corresponding techniques need to be developed.

4.1.3 Multi-Level

Although there has been growing interest in

developing models on multi levels and multi time

scales, there still only a few concrete examples

(Sarjoughian et al., 2001, Ferreira et al., 2015). The

layered approach is discussed in many papers, but

typically as a conceptual model, rather than used as a

programming paradigm. This layer or multi-level

model also fits well with the conceptual model

presented by SGAM (Santodomingo et al., 2014) for

Smart Grid interactions in power.

We would propose that any new conceptual

design adopts a multilayer structure so that it can

capture different views of the system represented as

layers in a model (see Figure 3), such as a physical

layer (devices power nodes, flows, congestion),

market layer (prices) etc. We also propose that any

new conceptual design allow users to easily add,

define or remove layers, to allow experimentation

with different designs. This would be easier using a

graph database structure as the links in the database

would define the layers and their interconnections.

Figure 3: Multi-layer concept.

Figure 4: Representation of a power system.

SIMULTECH 2018 - 8th International Conference on Simulation and Modeling Methodologies, Technologies and Applications

194

4.1.4 Distributed and Scale

Many of the models we have investigated have been

developed using smaller problems and therefore may

be difficult to scale. To provide a realistic simulation,

our problem space requires that a potential model can

simulate tens of thousands of power nodes with

thousands of power consuming and producing

devices attached to each of those nodes. This drives

us to a potential distributed architecture.

4.1.5 Power Simulation Specifics

A power system can be represented as a cause and

effect diagram, as shown in Figure 4, which indicates

flows, interactions and relationships between

actors/agents e.g. generators. Note that that this

diagram will evolve through time, e.g. 5-10 years ago

EV’s would not have existed on this diagram.

Considering this, one of the necessary features of

a new proposed architecture is that it can easily allow

changes of the relationships and flows.

We take as given that in the power domain that

any model would also need to represent power flows

and be able to “clear” the market on a large scale. This

would necessitate that any framework have a

methodology and a database structure/design suitable

for power and particularly for designs associated with

the evolving smart grid area and its new participants.

Links to existing power system simulators e.g.

MATPOWER (Murillo-Sánchez et al., 2013,

Zimmerman et al., 2011) should be considered.

4.2 The Suitability of Existing Systems

Using a “traffic light” ratings approach for gap and

needs analysis we have created a “Navajo blanket” of

our power research problem to help us understand

how four systems stack up against the requirements

for an “ideal” power system simulator (see Figure 5).

For brevity we only show one of the “Navajo”

blankets” analysed.

Extensive experimentation with

the various systems discussed above has also been

performed and forms the basis of the scores presented

in Figure 5.

Each row represents a potential need or

requirement for our ideal simulator. The four columns

represent the four currently available systems that we

are comparing. Scores from 1 – 10 have been given

to each cell, with 10 representing that the system

meets that current need. This is the equivalent of dark

green in the figure. Zero represents that the system

does not currently have that functionality. Colours are

provided automatically by conditional formatting in

Figure 5: Navajo blanket of an ideal ABM/MAS power

simulator.

Excel 2016 using a graduated green - yellow – red

colour scale. Although models that are circled in

Figure 5 can be regarded as most comprehensive in

class, gaps for our requirements remain. Therefore,

the above analysis indicates that there is a need for a

hybrid ABM/MAS simulator to model electricity

systems that fills these gaps.

It is clear from the proceeding sections that there

are many useful ideas and components in the existing

systems that can be reused, and therefore we would

not advocate the complete redesign of a simulation

system, but the reuse of large parts of existing

simulators (e.g. EMLab, AMES).

5 CONCLUSIONS

This paper introduced current research and advocates

for the development of a hybrid ABM/MAS system

for power simulation. It presented an initial scope and

ideas for this potential simulator. The addition of a

multitude of electric vehicles and DER’s to the power

grid will exacerbate the complexity of the system and

suggests that we should develop a distributed multi-

scale, multi- layer architecture that will make use of

reusable components. We are not advocating that we

should build a completely new system, but see a

solution in reusing and linking existing simulators,

while some changes to their designs to accommodate

better asynchronous modelling capabilities will be

required. We have looked at a number of systems

applicable to our problem domain and found that they

lacked the following:

Do We Need a New Architecture for Simulating Power Systems?

195

 An easy way to address asynchronicity and its

effects on power systems especially at scale i.e.

millions of agents;

 A system that models both the micro levels

interactions and communications that are likely

to occur between Consumers, EV’s and system

mangers in the longer term without the need to

simulate every second for every agent;

 Systems that adequately link OPF to the

ABM/MAS environment at scale;

Our proposal therefore adds the following main

features to existing ABM/MAS power simulators in

the power domain; An asynchronous/synchronous

base applied to an existing ABM and MAS simulator;

the use of an equation free modelling technique or

some other alternative to simulate at the micro scale

when required, and an existing large scale OPF

model. We are currently looking to design and build

this simulator and are developing agent

methodologies and detailed interaction protocols

using an agent orientated design methodology.

Investigations into multiscale methodologies

(particularly equation free modelling) are ongoing.

ACKNOWLEDGEMENTS

This research is funded by the UK Engineering and

Physical Science Research Council (EPSRC) and the

International Strategic Partner (ISP) Research

Studentship.

REFERENCES

AnyLogic. 2018. Any Logic Platfrom:Simulation Modeling

Software Tools & Solutions for Business [Online].

Available: https://www.anylogic.com/ [Accessed].

Bellifemine, F. L., Caire, G. & Greenwood, D. 2007.

Developing multi-agent systems with JADE, John

Wiley & Sons.

Cardoso, H. L. 2009. Introduction to RepastJ. University of

Portugal NIAD&R – Distributed Artificial Intelligence

and Robotics Group.

Cardoso, H. L. 2015. SAJaS: enabling JADE-based

simulations. Transactions on Computational Collective

Intelligence XX. Springer.

Chen, M., Syed, M., Sansano, E. G., McArthur, S. D., Burt,

G. M. & Kockar, I. Distributed negotiation in future

power networks: Rapid prototyping using multi-agent

system. PES Innovative Smart Grid Technologies

Conference Europe (ISGT-Europe), 2016 IEEE, 2016.

IEEE, 1-6.

Conzelmann, G., Boyd, G., Koritarov, V. & Veselka, T.

Multi-agent power market simulation using EMCAS.

Power Engineering Society General Meeting, 2005.

IEEE, 2005. IEEE, 2829-2834.

Cornforth, D., Green, D. G. & Newth, D. 2005. Ordered

asynchronous processes in multi-agent systems.

Physica D: Nonlinear Phenomena, 204, 70-82.

De Vries, L. J., Chappin, É. J. L. & Richstein, J. C. 2013.

EMLab-Generation An experimentation environment

for electricity policy analysis.

DeAngelis, D. L. & Yurek, S. 2015. Equation-free

modeling unravels the behavior of complex ecological

systems. Proceedings of the National Academy of

Sciences, 112, 3856-3857.

Ferreira, A., Leitao, P. & Vrba, P. 2015. Simulating smart

grid using a two-layer multiagent framework. 2982-

2987.

Gormer, J., Homoceanu, G., Mumme, C., Huhn, M. &

Muller, J. P. Jrep: Extending repast simphony for jade

agent behavior components. Proceedings of the 2011

IEEE/WIC/ACM International Conferences on Web

Intelligence and Intelligent Agent Technology-Volume

02, 2011. IEEE Computer Society, 149-154.

Heath, B., Hill, R. & Ciarallo, F. 2009. A survey of agent-

based modeling practices (January 1998 to July 2008).

Journal of Artificial Societies and Social Simulation,

12, 9.

Holland, J. H. 1999. Emergence: From Chaos to Order,

Perseus Publishing.

Jacobs, B. I., Levy, K. N. & Markowitz, H. M. 2004.

Financial market simulation. Journal of Portfolio

Management, 142-152.

Kantamneni, A., Brown, L. E., Parker, G. & Weaver, W.

W. 2015. Survey of multi-agent systems for microgrid

control. Engineering Applications of Artificial

Intelligence, 45, 192-203.

Kevrekidis, I. G., Gear, C. W. & Hummer, G. 2004.

Equationfree: The computeraided analysis of

complex multiscale systems. AIChE Journal, 50, 1346-

1355.

Kevrekidis, I. G., Gear, C. W., Hyman, J. M., Kevrekidid,

P. G., Runborg, O. & Theodoropoulos, C. 2003.

Equation-free, coarse-grained multiscale computation:

Enabling mocroscopic simulators to perform system-

level analysis. Communications in Mathematical

Sciences, 1, 715-762.

Kevrekidis, I. G. & Samaey, G. 2009. Equation-free

multiscale computation: Algorithms and applications.

Annual review of physical chemistry, 60, 321-344.

Kevrekidis, Y. G. Equation-Free Modeling For Complex

Systems USFOE-Symposia, 2004.

Le Maître, O. & Mathelin, L. 2010. Equationfree model

reduction for complex dynamical systems. International

Journal for numerical methods in fluids, 63, 163-184.

Leon, F., Paprzycki, M. & Ganzha, M. 2015. A review of

agent platforms. Technical Report, ICT COST Action

IC1404, Multi-Paradigm Modelling for Cyber-Physical

Systems (MPM4CPS).

SIMULTECH 2018 - 8th International Conference on Simulation and Modeling Methodologies, Technologies and Applications

196

Li, H. & Tesfatsion, L. 2009. Development of open source

software for power market research: The AMES test

bed. The Journal of Energy Markets, 2, 111.

Macbeth, S. 2015. Presage2 Software [Online]. Available:

https://github.com/Presage/Presage2 [Accessed].

McArthur, S. D., Davidson, E. M., Catterson, V. M.,

Dimeas, A. L., Hatziargyriou, N. D., Ponci, F. &

Funabashi, T. 2007. Multi-agent systems for power

engineering applications—Part I: Concepts,

approaches, and technical challenges. IEEE

Transactions on Power systems, 22, 1743-1752.

Merkl Sasaki, B., Chao, Joy & Howard, R. 2017. Graph

Databases for Beginners Neo4J. Neo4J.

Murillo-Sánchez, C. E., Zimmerman, R. D., Anderson, C.

L. & Thomas, R. J. 2013. Secure planning and

operations of systems with stochastic sources, energy

storage, and active demand. IEEE Transactions on

Smart Grid, 4, 2220-2229.

Rohjans, S., Lehnhoff, S., Schütte, S., Scherfke, S. &

Hussain, S. mosaik-A modular platform for the

evaluation of agent-based Smart Grid control.

Innovative Smart Grid Technologies Europe (ISGT

EUROPE), 2013 4th IEEE/PES, 2013. IEEE, 1-5.

Santodomingo, R., Uslar, M., Goring, A., Gottschalk, M.,

Nordstrom, L., Saleem, A. & Chenine, M. SGAM-

based methodology to analyse Smart Grid solutions in

DISCERN European research project. Energy

Conference (ENERGYCON), 2014 IEEE International,

2014. IEEE, 751-758.

Sarjoughian, H. S., Zeigler, B. P. & Hall, S. B. 2001. A

layered modeling and simulation architecture for agent-

based system development. Proceedings of the IEEE,

89, 201-213.

Sujil, A., Verma, J. & Kumar, R. 2016. Multi agent system:

concepts, platforms and applications in power systems.

Artificial Intelligence Review, 49, 153-182.

Wallis, L. & Paich, M. Integrating artifical intelligence with

anylogic simulation. Simulation Conference (WSC),

2017 Winter, 2017. IEEE, 4449-4449.

Watts, D. J. 2003. Small Worlds: The Dynamics of

Networks between Order and Randomness Princeton

University Press

Youssefmir, M. & Huberman, B. A. 1997. Clustered

volatility in multiagent dynamics. Journal of Economic

Behavior & Organization, 32, 101-118.

Zimmerman, R. D., Murillo-Sánchez, C. E. & Thomas, R.

J. 2011. MATPOWER: Steady-state operations,

planning, and analysis tools for power systems research

and education. IEEE Transactions on power systems,

26, 12-19.

Do We Need a New Architecture for Simulating Power Systems?

197