The Grid: Stronger, Bigger, Smarter?: Presenting a Conceptual Framework of Power System Resilience

TL;DR: In this paper, a conceptual framework of power system resilience, its key features, and potential enhancement measures is discussed, with a focus on the resilience of critical power infrastructures to high-impact, low-probability events.

Abstract: INCREASING THE RESILIENCE of critical power infrastructures to high-impact, low-probability events, such as extreme weather phenomena driven by climate change, is of key importance for keeping the lights on. However, what does resilience really mean? Should we build a stronger and bigger grid or a smarter one? This article discusses a conceptual framework of power system resilience, its key features, and potential enhancement measures.

Summary

1. INTRODUCTION

• The design and operation of the critical power infrastructure has been traditionally driven by the key reliability principles of security and adequacy.
• The US northeastern states were struck by Hurricane Sandy in 2012, which destroyed over 100,000 primary electrical wires; in addition, several substation transformers exploded and numerous substations were flooded.
• Hence, a power infrastructure that can maintain high levels of performance under any condition should be reliable to the most “common” blackouts, but also resilient to much less frequent disasters.

2. CONCEPTUALIZING POWER SYSTEMS RESILIENCE

• C.S. Holling first defined resilience in 1973 as a measure of “the persistence of systems and of their ability to absorb change and disturbance and still maintain the same relationships between populations or state variables”.
• Since this foundational definition, the concept of resilience has evolved remarkably in several systems, such as safety management, organizational, social-ecological and economic ones.
• There have been several attempts by organizations worldwide in the power and energy engineering communities, such as the UK Energy Research Center and the Power Systems Engineering Research Center , USA, to define resilience and distinguish it from the concept of reliability.
• This framework consists of the “4Rs”: robustness, redundancy, resourcefulness and rapidity.
• Some key resilience characteristics that differentiate it from the concept of reliability are shown in Table I, which will be discussed in detail throughout this article.

2.1. A conceptual resilience curve associated to an event

• The illustrative conceptual resilience curve of Fig.1 shows the resilience level as a function of time with respect to a disturbance event.
• The resourcefulness, redundancy and adaptive self-organization are key resilience features at this stage of the event, as they provide the corrective operational flexibility necessary to adapt to and deal with the evolving conditions (that are possibly never experienced before).
• The post-restoration resilience level Rpr may or may not be as high as the pre-event resilience level Ro, i.e. Rpr < Ro.
• The undergrounding of an overhead corridor might improve the capability of the system to withstand events, but then if the cable is damaged it may take much longer to repair it than an overhead line.

2.2. A Conceptual long-term resilience framework

• The resilience definition by the National Infrastructure Advisory Council (NIAC), USA, takes the infrastructure resilience framework a step further, as it additionally considers the long-term adaptation as a key feature for achieving resilience.
• The adaptation capacity, which enables the long-term resilience planning, is thus a critical resilience feature as it provides the capacity to deal with unforeseeable and continuously changing conditions.
• Some of these measures are more resilience-efficient than others, and some measures are more cost-efficient than others.
• In the UK, for instance, the North Sea storm in December 2013 resulted in the flooding of 2,600 homes, but approximately 800,000 homes had been protected from flooding.

2.3. Quantifying resilience

• Quantifying resilience is not a straightforward process (quite the opposite, actually, as it may prove the most challenging task within a resilience analysis framework) because, as discussed earlier, resilience is a multidimensional, dynamic concept with several intrinsic complexities.
• Both short-term and long-term resilience metrics are needed accordingly (Fig. 3).
• Finally, the time dimension needs to be incorporated explicitly in the assessment, so as to capture the capability of the system of both slowly 8 degrading from and fast recovering back to the original pre-event state.
• These curves express the failure probability of power system components as a function of a weather parameter, e.g. wind speed or rain intensity.
• Following this, as previously discussed, resilience enhancement measures can be applied if necessary.

3. BOOSTING THE RESILIENCE OF FUTURE POWER SYSTEMS

• The majority of electrical utilities worldwide have recognized the necessity of taking actions to boost the grid resilience to high-impact low-probability events.
• Evaluating and enhancing resilience to weather events using fragility curves Fig. 6 10 illustrates conceptually how hardening measures might (also depending on the resilience metric used) generally be more effective than operational ones, but they are also likely to come at a higher cost, also known as 9 Fig. 4.

3.1. Making the Grid Stronger and Bigger

• Hardening measures may refer to topology and structural changes in order to make the network less vulnerable to severe events.
• This is mainly because of the 11 complicated nature of these systems and the inability of the repair crews to visually detect the damaged components.
• Targeted or selective undergrounding of overhead lines could thus be a more viable solution than a total conversion, following a proper risk and cost/benefit analysis.
• The T-pylons are shorter than the traditional towers, have less impact on the environment and, more importantly, are considered more robust.
• Elevating substations, relocating facilities or re-routing transmission lines to areas less prone to extreme weather help provide protection against flood damage and any other type of damage caused by weather events, for instance tower collapses due to extreme winds and snowfalls.

As aforementioned, the term “smart” here refers to a broad set of operational actions that can be taken to improve

• The observability, controllability and operational flexibility of a power system, particularly in response to an extreme event.
• This is critical in building resilience as it provides the system (and system operators) with monitoring and control assets for dealing with the unfolding disaster in a timely and efficient way.

Distributed Energy Systems and Decentralized Control

• Decentralized energy systems with large scale deployment of distributed energy resources (and distributed generation and storage, in particular) and decentralized control can play a key role in providing resilience to external shocks.
• In fact, generating, storing and controlling energy locally without the need of long transmission lines can make the network less vulnerable to disasters and the response to an emergency much faster and more efficient.
• Restoration times can also be improved in smaller balancing areas.
• Localized protection and control assets are however required for achieving a more resilient decentralized operation, which is to be considered in the wider picture of smart grid evolution.

Microgrids

• A microgrid can be simply defined as the subset of the grid (typically at low voltage and medium voltage levels) that can be islanded and can still supply all or part of their customers during emergencies, thus intrinsically enhancing system resilience.
• A microgrid requires the smart technologies mentioned above to continue delivering power to the customers in islanded mode.

Adaptive Wide-area Protection and Control Schemes

• The majority of the existing wide-area protection and control schemes are event-based, which means that they will operate once the pre-determined criteria are fulfilled.
• They usually follow the logic of “if A AND B is true, then apply C”, where A and B are the electrical events that the scheme is designed to provide protection against and C are the protection and control actions to be implemented.
• Nevertheless, adaptive protections have not been widely implemented yet due to concerns about the reliability of these schemes themselves.
• Advanced Visualization and Situation Awareness Systems Electrical utilities often have a set of incomplete information on the state of their own network, resulting in delayed and inefficient responses.
• It can thus be seen that human resilience also plays a key role in preserving power system resilience.

Disaster Response and Risk Management

• The smart and operational measures discussed above can improve those emergency and preparedness procedures that enhance disaster response and risk management.
• Recovering from a state of degraded performance and resilience (Rpe, see Fig. 1) requires an effective 14 post-disaster restoration process.
• There are mainly two aspects that drive the development of this procedure: the time required to restore each of the damaged components and the criticality of each component in restoring resilience.
• The former is strongly related to the infrastructure resilience and depends on several factors, such as availability of backup components, accessibility to the affected areas, and number and location of repair crews.
• If a component is ranked first of the most critical components in restoring operational resilience, but under specific circumstances it may be very difficult or lengthy to restore, then it might not be highly ranked in the priority list.

3.3. Hybrid Grids: Stronger, Bigger and Smarter

• It can be clearly seen from the discussions in the previous sections that understanding and enhancing grid resilience is a very open challenge.
• Hardening/reinforcement schemes may come at a significantly higher cost than the smart/operational measures.
• On the other hand, operational measures without sufficient strengthening of the network may not be enough for keeping the lights on in the face of a disaster.
• The term “hybrid” can be interpreted here in two different, but related, ways.
• Such a hybrid system would thus offer the advantages of both bigger and more robust networks as well as more operational flexibility and security.

4. CONCLUSIONS

• Building a power infrastructure that is reliable to known and credible threats, but also resilient to the highimpact low-probability events is very challenging.
• Resilience is not a static concept, but it is a dynamic, ongoing procedure for adapting (and possibly transforming) the structure and operation of power systems to be better prepared to external, unforeseeable shocks.

Figures

Fig. 3: Quantifying short-term and long-term resilience

Full text

IEEE Power and Energy Magazine, May 2015
1
Foreword – Increasing the resilience of critical power infrastructures to high-impact low-probability
events, such as extreme weather phenomena driven by climate change, is of key importance for keeping the
lights on. However, what does resilience really mean? Should we build a stronger and bigger grid, or a
smarter one? This article discusses a conceptual framework of power system resilience, its key features, and
potential enhancement measures.
1. INTRODUCTION
The design and operation of the critical power infrastructure has been traditionally driven by the key reliability
principles of security and adequacy. These allow dealing with known and credible threats so as to guarantee high
quality power supply to end users on a nearly continuous basis, with few interruptions over an extended time period.
It cannot be doubted that this has led to the development of one of the most reliable (and complex!) infrastructures
of the last century.
However, it is becoming more and more apparent that further considerations beyond the classical reliability-
oriented view are needed for keeping the lights on. This is evidenced by several catastrophes that occurred
worldwide in the last decade or so. For example, the US northeastern states were struck by Hurricane Sandy in 2012,
which destroyed over 100,000 primary electrical wires; in addition, several substation transformers exploded and
numerous substations were flooded. This altogether led to the disconnection of approximately 7 million people. Over
the 2010-2011 summer, Australia’s second largest state, Queensland, was affected by widespread flooding that
resulted in significant damage to six zone substations and numerous poles, transformers and overhead wires.
Approximately 150,000 customers experienced power disruptions. In 2008, China was hit by a severe ice storm,
which resulted in the failure of 2,000 substations and in the collapse of 8,500 towers leading to power interruptions
in 13 provinces and 170 cities.
Mathaios Panteli, Member, IEEE, and Pierluigi Mancarella, Senior Member, IEEE
A Stronger, Bigger or Smarter Grid?
Conceptualizing the Resilience of Future Power Infrastructure

IEEE Power and Energy Magazine, May 2015
2
These are only a few examples of the effect of weather-driven high-impact low-probability events that a power
infrastructure can experience. These events also illustrate that we need to distinguish blackouts from disasters. A
blackout occurs when a large proportion of a power grid is disabled by a combination of unplanned contingencies,
which result in a temporary power interruption. A reliable and well-designed power system should be capable of
minimizing the amount of power disruption and of recovering very quickly from a blackout. On the other hand, a
disaster, which usually includes a blackout, refers to severe and rapidly changing circumstances possibly never
before experienced. A disaster can cause the incapacitation of several and often large parts of a power grid, which
may last for a long period depending on the extent of the disaster. Hence, a power infrastructure that can maintain
high levels of performance under any condition should be reliable to the most “common” blackouts, but also
resilient to much less frequent disasters.
Resilience (or resiliency) comes from the Latin word “resiliō”, which literally refers to the ability of an object to
rebound or return to its original shape or position after being stressed (e.g., bent, compressed, stretched, etc.). In the
context of power systems, it refers to the ability of a power system to recover quickly following a disaster, or more
generally to the ability of anticipating extraordinary and high-impact low probability events, rapidly recovering from
these disruptive events, and absorbing lessons for adapting its operation and structure for preventing or mitigating
the impact of similar events in the future. Adaptation thus refers to the long-term planning and operational measures
taken to reduce the vulnerability to external sudden shocks.
As power engineers, how can we build a network that is both reliable and resilient? The most obvious way is
building a bigger and stronger (more redundant and robust) network; however, how cost efficient is this approach? A
more cost efficient solution could be investing more into “smart” operational measures, but how robust is this
approach? More insights in the concept of resilience can help address this issue.
2. CONCEPTUALIZING POWER SYSTEMS RESILIENCE
C.S. Holling first defined resilience in 1973 as a measure of “the persistence of systems and of their ability to
absorb change and disturbance and still maintain the same relationships between populations or state variables”.
Since this foundational definition, the concept of resilience has evolved remarkably in several systems, such as
safety management, organizational, social-ecological and economic ones. After Holling, numerous interpretations of

IEEE Power and Energy Magazine, May 2015
3
resilience have been developed, resulting in many different definitions and a lack of a universal understanding of
what resilience really is.
In the context of power systems as critical infrastructures the picture is even more blur, as the concept of
resilience has only emerged in the last decade or so. There have been several attempts by organizations worldwide in
the power and energy engineering communities, such as the UK Energy Research Center (UKERC) and the Power
Systems Engineering Research Center (PSERC), USA, to define resilience and distinguish it from the concept of
reliability. According to the UK Cabinet Office, resilience encompasses reliability and it further includes resistance,
redundancy, response and recovery as key features. Another pioneer definition comes from the Multidisciplinary
and National Center for Earthquake Engineering Research (MCEER), where a generic resilience framework has
been developed that is applicable to any critical infrastructure, including power systems. This framework consists of
the “4Rs”: robustness, redundancy, resourcefulness and rapidity.
The list of power system resilience definitions is endless, but the majority of these definitions focus on the
ability to anticipate, absorb and rapidly recover from an external, high-impact low-probability shock. Although a full
comparison is outside the scope of this work, some key resilience characteristics that differentiate it from the
concept of reliability are shown in Table I, which will be discussed in detail throughout this article.
TABLE I
RELIABILITY VS RESILIENCE
Reliability
Resilience
High-probability, low-impact
Low-probability, high-impact
Static
Adaptive, ongoing, short- and long-term
Evaluates the power system states
Evaluates the power system states and transition
times between states
Concerned with customer interruption time
Concerned with customer interruption time and the
infrastructure recovery time
2.1. A conceptual resilience curve associated to an event
The illustrative conceptual resilience curve of Fig.1 shows the resilience level as a function of time with respect
to a disturbance event. This figure is used here for demonstrating the key resilience features that a power system
must possess for coping effectively with the evolving conditions associated to an event, for instance, a heavy storm
moving across the system.

IEEE Power and Energy Magazine, May 2015
4
R
Time
Operational Resilience
R
o
R
pe
R
pr
t
o
t
e
t
r
t
pe
t
pr
Robustness/
Resistance
Resourcefulness/Redundancy/
Adaptive Self-organization
Response/
Recovery
Robustness/
Resistance
Resilient
State
Event
progress
Post-event degraded state
Restorative
state
Post-
restoration
state
Infrastructure
Resilience
Infrastructure
Recovery
t
ir
t
pir
Fig. 1: Conceptual resilience curve associated to an event
Before the event occurs at t
e
, a power system must be robust and resistant to withstand the initial shock. A well-
designed and operated power system should demonstrate sufficient resilience (indicated here with R
o
, where R is a
suitable metric associated to the resilience level of the system – see also further below) to cope with any type of
events. The capability of preventive operational flexibility is highly critical here, as it provides the operators with the
assets to configure the system in a resilient state.
Following the event, the system enters the post-event degraded state, where the resilience of the system is
significantly compromised (R
pe
). The resourcefulness, redundancy and adaptive self-organization are key resilience
features at this stage of the event, as they provide the corrective operational flexibility necessary to adapt to and deal
with the evolving conditions (that are possibly never experienced before). This helps minimize the impact of the
event and the resilience degradation (i.e., R
o
- R
pe
) before the restoration procedure is initiated at t
r
.
The system then enters the restorative state, where it should demonstrate the restorative capacity necessary for
enabling the fast response and recovery to a resilient state as quickly as possible.
Once the restoration is completed, the system enters the post-restoration state. The post-restoration resilience
level R
pr
may or may not be as high as the pre-event resilience level R
o
, i.e. R
pr
< R
o
. In particular, while the system
may have recovered from the point of view of fully returning to its pre-event operational state (thus showing a
certain degree of operational resilience), the infrastructure may take longer to fully recover (infrastructure
resilience), i.e. (t
pir
- t
ir
) > (t
pr
- t
r
). This would depend on the severity of the event, as well as on the resilience
features that the power system will demonstrate before, during and after the external shock. It is interesting to notice
how some measures might make the system more resilient operationally but less from an infrastructure perspective.

IEEE Power and Energy Magazine, May 2015
5
For instance, the undergrounding of an overhead corridor might improve the capability of the system to withstand
events, but then if the cable is damaged it may take much longer to repair it than an overhead line. This might
become a critical issue if a new event were to arrive relatively soon (for instance, settling waves following a major
earthquake wave).
It is important to highlight that for a full understanding and assessment of system resilience, which is by
definition a multi-dimension concept, both the resilience levels of and the transition times between the power system
states associated to an event are needed. Referring to Fig. 1, in fact, the system resilience is not only characterized
by the levels R
o
, R
pe
and R
pr
associated to different states, but also by the transition time between states (i.e., t
pe
– t
e
,
t
pr
– t
r
, and t
pir
– t
ir
, respectively). In particular, actions to increase resilience should aim at (i) reducing the resilience
level degradation during the event (R
o
- R
pe
); (ii) achieving a relatively “slow” and possibly controlled degradation
(t
pe
– t
e
), thus also mitigating the degree of cascading; and (iii) reducing the recovery time (both from operational
point of view, t
pr
– t
r
, and infrastructure point of view, t
pir
– t
ir
). As also indicated in Table I, this “time dimension” is
an important feature that distinguishes resilience from reliability.
2.2. A Conceptual long-term resilience framework
The resilience definition by the National Infrastructure Advisory Council (NIAC), USA, takes the infrastructure
resilience framework a step further, as it additionally considers the long-term adaptation as a key feature for
achieving resilience. This resilience feature refers to the ongoing process of resilience-building using the
information and experiences from past events in order to evaluate existing resilience measures and regularly update
resilience planning and decision making. Fig. 2 shows the framework for conceptualizing this infinite procedure of
evaluating and improving power systems resilience, which is depicted by the resilience enhancement circle.
The adaptation capacity, which enables the long-term resilience planning, is thus a critical resilience feature as
it provides the capacity to deal with unforeseeable and continuously changing conditions. As can be seen in Fig. 2,
the first step towards this goal is to perform vulnerability and adaptation studies using the input from past
experiences and/or simulations. This would help detect the vulnerabilities of a power system at the different stages
associated to an event, i.e. before, during, and after, and develop the adaptation strategies necessary for improving
the key resilience features and enhancing the response of the power system to the evolving conditions during a
similar event that were to occur in the future.

Frequently asked questions

Q1. What are the contributions mentioned in the paper "A stronger, bigger or smarter grid? conceptualizing the resilience of future power infrastructure" ?

This article discusses a conceptual framework of power system resilience, its key features, and potential enhancement measures. 

Q2. What are the future works mentioned in the paper "A stronger, bigger or smarter grid? conceptualizing the resilience of future power infrastructure" ?

A resilient network must thus be robust and operational flexible, but must also possess the adaptation capacity to plan, facilitate and implement the actions and measures required for preparing to similar or new events in the future. 

Q3. What are the key resilience features at this stage of the event?

The resourcefulness, redundancy and adaptive self-organization are key resilience features at this stage of the event, as they provide the corrective operational flexibility necessary to adapt to and deal with the evolving conditions (that are possibly never experienced before). 

Q4. What are some examples of how to make a network more resilient?

Undergrounding distribution and transmission lines - Upgrading poles and structures with stronger, more robust materials - Elevating substations - Relocating facilities to areas less prone to extreme weather - Re-routing transmission lines to areas less affected by weather - Redundant transmission routesHardening measures may refer to topology and structural changes in order to make the network less vulnerable to severe events. 

Q5. What is the definition of a hardening measure?

Hardening measures are denoted as infrastructure reinforcement actions for making the power system less susceptible to extreme events. 

Q6. What are the main aspects that drive the development of this procedure?

There are mainly two aspects that drive the development of this procedure: the time required to restore each of the damaged components and the criticality of each component in restoring resilience. 

Q7. How can the resilience curves be used to measure the impact of a weather event?

By mapping the time-series profile (thus considering the event’s inter-temporal dimension) of the weather event at different locations of the power system (thus considering the event’s inter-spatial dimension) to these fragility curves, the components’ weatherrelated failure probabilities and therefore the resilience implications can be quantified using suitable multidimensional metrics (for instance, energy not supplied, duration of interruptions, and time to full infrastructure recovery). 

Q8. What is the role of communication and information systems in preserving resilience?

In addition, the reliability and functionality of the relevant communication and information systems is critical to enable effective information exchange and coordination between system operators and field/repair crews. 

Q9. What are the main objectives of the efforts?

These efforts mainly aim to achieve system adaptation, which refers to the measures taken to reduce the impact of future events, and system survivability which refers to the ability to maintain an adequate functionality during and after the event.9Fig. 

Q10. What are the key reliability principles of the critical power infrastructure?

The design and operation of the critical power infrastructure has been traditionally driven by the key reliability principles of security and adequacy. 

Q11. What is the role of the situation awareness tool in preserving resilience?

The development of adequate situation awareness tools that enables the effective and timely decision-making could thus play a key role in preserving resilience during emergencies. 

Q12. How can the resilience assessment be done?

In order to do so, the spatial-temporal influence of the event on the resilience of the power infrastructure needs to be adequately modelled. 

Q13. What is the relationship between the latter and the infrastructure resilience?

The former is strongly related to the infrastructure resilience and depends on several factors, such as availability of backup components, accessibility to the affected areas, and number and location of repair crews. 

Q14. What is the way to achieve a resilient network?

A resilient network should thus be able to achieve a resilience level that is close or equal to Ro (see Fig. 1) as quickly as possible following the disaster by possessing adequate operational and infrastructure resilience features. 

Q15. What is the example of resilience enhancement?

An example of resilience enhancement is shown in the fragility curve of Fig. 4, in which the components are made more robust to higher intensities of the weather event.