reakt-d7.3

7/17/2019 REAKT-D7.3

http://slidepdf.com/reader/full/reakt-d73 1/30

1

Project number: 282862

Project name: Strategies and tools for Real Time Earthquake RiskReduction

Project acronym: REAKT

Theme: ENV.2011.1.3.1-1

Towards real-time earthquake risk reduction

Start date: 01.09.2011

End date: 31.12.2013 (24 months)

Deliverable: D 7.3 Final report for Feasibility studies on real-timerisk mitigation in the SINES Industrial Complex,Portugal

Version: Final

Responsible partner: IST

Month due: 36 Month delivered: 40

Primary author: Carlos Sousa Oliveira (IST) 23.12.2014_

Date

Reviewer: Carlo G. Lai (EUCENTRE) 25/09/2014

Date

Authorised: Paolo Gasparini (AMRA) 29.12.2014

Date

7/17/2019 REAKT-D7.3


2

Dissemination Level

PU Public X

PPRestricted to other programme participants (including the CommissionServices)

RERestricted to a group specified by the consortium (including the

Commission Services)CO Confidential, only for members of the consortium (including the

Commission Services)

7/17/2019 REAKT-D7.3


3

Acknowledgement

The research leading to these results has received funding from the European

Community’s Seventh Framework Programme [FP7/2007-2013] under grant

agreement n° 282862.

7/17/2019 REAKT-D7.3


4

INDEX

1. Introduction 5

2. The Sines Industrial Complex 5

2.1. The ZILS (Zona Industrial e Logística de Sines). Industrial & LogisticsPlatform 5

3. Geotechnical characterization 10

4. Risk assessment 11

4.1. QuakeIST seismic simulator. General description 11

4.2. QuakeIST® seismic simulator. Scenarios 14

4.3. Dependencies and interdependencies 15

5. Seismological context and Early warning developments 17

5.1. Early warning developments 18

5.1.1 Epicenters in the SW Iberia 19

5.1.2 Epicenter in the Lower Tagus Valley Region 21

5.1.3 Other early warning developments 22

6. Earthquake early warning survey 24

7. Cost-benefit preliminary analysis 27

References 29

7/17/2019 REAKT-D7.3


5

1. Introduction

The Industrial Complex of Sines (Portugal) with more than 13 km2, is one of the largest

in Europe, housing a significant number of National (Portuguese) and European Critical

Infrastructures. Located next to the Atlantic Coast, at about 180 km from two major

seismogenic sources (the Gorringe Bank and the Marquês de Pombal Fault) both ableto generate a 8.5 to 9 magnitude earthquake, leading to Peak Ground Accelerations of

about 0.20 g to 0.24 g in stiff rock soils, with the possibility of 0.4 g to 0.5 g in soft soils.

In this area, several major industries and services are present, namely a thermal power

generation plant, concrete, chemical and cement production plants, fuel refinery,

dangerous materials and fuel parks, mobile communications, pipelines and many other

critical infrastructures, interacting in a complex physical and functional dependency so,

prone to trigger chain reactions amplifying and propagating disastrous effects. Within a

5 km distance, a population of about 50.000 persons is present, with schools, hospitals,

and many other sensible targets.

Already identified as a risky zone, during the establishment of the National and

European Critical Infrastructures Protection Projects, the Sines Industrial Complex

constitutes an optimum feasibility case study in the field of Seismic Early-Warning. It is

no coincidence that several national and international major stake-holders have already

expressed their interest in this subject, and have already showed their strong

willingness in participating in this Project.

2. The Sines Industrial Complex

The exploitation of the port of Sines started in 1973. Sines assumed an important

geographical position in the world, being a privileged axis at the crossroads of maritime

routes due to its deep-sea water port, combined with good conditions for port

expansion and secured direct access to railway and road networks.

Last year, Sines handled some 25.8 million tonnes of cargo and almost 447,500

twenty-foot equivalent units (TEU). Liquid bulk cargo traffic amounted to 16.2 million

tonnes. (http://shipping.seenews.com/, June 2013).

The expansion of the Panama Canal in 2014, with a direct link from the Pacific to

Atlantic for larger ships, may lead to an increased flow of trade between the Pacific

basin, both coasts of North America, the Mercosur and Europe. Ports like Sines will be

strategic for Europe (Moreira, P.J.P, 2013).

2.1. The ZILS (Zona Industrial e Logística de Sines). Industrial &

Logistics Platform

In which concerns for Sines in terms of Industrial and logistic platform (Figure 1a) and

b)), provided the integrated development of the entire region and emerged on the

concentration of diversified industries and industrial facilities.

http://shipping.seenews.com/




7/17/2019 REAKT-D7.3


6

Figure 1a: Plan view of Sines and main infrastructures considered in this study. GIS platform.

7/17/2019 REAKT-D7.3


7

Figure 1b: Plan view of Sines and main partners involved in this study.

Some of the main infrastructures of the harbour are listed and briefly described below:

Sines container terminal “Terminal XX” - the Sines container terminal, called Terminal

XXI, started its operations in 2004 and is the largest container terminal in Portugal.

Figure 2: Sines container terminal “Terminal XX”

7/17/2019 REAKT-D7.3


8

Figure 3: Three-dimensional scheme of the intermodal platform at the container terminal

(Garcia, P.R., 2003)

Sines Petrochemical Complex (REPSOL) - mainly destined for export olefins and

polyolefins. The Port of Sines has a terminal dedicated to petrochemical products.

Figure 4: REPSOL Sines petrochemical complex(Source: http://www.repsol.com/pt)

Sines Refinery (Galp) - diesel production. It is constituted by a number of processing

units spread across two plants, known as Manufacturing I and Manufacturing II. It has a

large storage area with a capacity of approximately 3 million cubic meters for crude oil,

fuels and other final products and intermediate product.

The Sines refined products are: gasoline; diesel; LPG (liquefied petroleum gas); fuel

oil; naphtha (used in the petrochemical industry to produce polymers from which

7/17/2019 REAKT-D7.3


9

plastic, fibers for textiles and even bubble gum is produced); jet fuel (fuel for airplanes);

bitumen (for asphalt and insulate) and sulphur (for pharmaceutical products, farming

and pulp whitening).

Needs: road and sea transport for filling and dispatching every products. Reception of

crude ships

Figure 5: Sines refinery (Source: http://www.galpenergia.com)

Figure 6: The manufacturing I produce gases, gasolines, aviation fuel, diesel and fuel oil

Figure 7: The manufacturing II has several processing units, namely a vacuum distillation and a

visbreaker

http://www.galpenergia.com/




7/17/2019 REAKT-D7.3


10

Liquefied Natural Gas (LNG) (REN Atlântico) - The LNG terminal consists of a docking

station with a discharge capacity of 40,000 cubic meters to 165,000 cubic meters with

an average discharge time of 20 hours, two storage tanks each having a capacity of

115,000 cubic meters and five open rack vaporizers for regasification. Storage may

reach 390,000 m3 of liquefied gas.

Figure 8: Sines Liquefied Natural Gas (LNG) (Source: http://www.ren.pt)

3. Geotechnical Characterization

“In geological terms, it is characterized by arenaceous formations in various states of

alteration and covered with various land fill materials. The soil is of a poor grade. The

original soil has been greatly altered and is practically all paved. Mainly it is made oflandfill materials, associated with buildings and roads. As a consequence, these are

zones that have been rendered impermeable to a considerable degree, of no

agricultural or ecological value. The hydro graphic network of the project area is made

of small watercourses that drain directly into the ocean. The most important in the

proximity of the project area are Ribeira das Caraminheiras, on the north, and Ribeira

da Junqueira, on the south, both of which display torrential flow in open channels. No

watercourse is identified in the area of the refinery.”

http://www.galpenergia.com/EN/sustainability/corporate-

responsibility/environment/Documents/EN-RNT_ProjConvesao_RS.pdf

The soil map settled according to the Eurocode 8 classification is shown in Figure 9

and is the input to the earthquake scenario simulator - QuakeIST®. The classification

was based on a preliminary analysis of a geotechnical map from LNEC (1974) and the

geological map from Direcção Geral de Geologia e Minas (1986). Further studies in

this area are needed for a refined analysis.

http://www.ren.pt/

http://www.ren.pt/

http://www.ren.pt/

http://www.galpenergia.com/EN/sustainability/corporate-responsibility/environment/Documents/EN-RNT_ProjConvesao_RS.pdf





http://www.ren.pt/

7/17/2019 REAKT-D7.3


11

Figure 9: Soil classification.

4. Risk assessment

Earthquake simulators developed until now show direct physical damage in terms of

victims, buildings, essential facilities and transportation systems, without including

estimations of indirect losses or propagated effects (functional interdependencies).

The QuakeIST® is an integrated simulator developed by the Instituto Superior Técnico,

and used by the REAKT project to obtain disaster scenarios affecting large scale

systems and their interdependencies.

Simulation can test different sets of decisions for the same disaster scenario to find the

optimal solution for restoration without wasting time and money; finally it can help to

develop a strategy that can increase the resiliency of the critical infrastructures to an

urban area.

4.1. QuakeIST seismic simulator. General description

The QuakeIST® simulator is an integrated earthquake simulator, developed by IST and

written in C++, oriented to risk calculations and damage propagations; and opens up

new territory for earthquake science and engineering with the goal of reducing the

potential for loss of life and property. QuakeIST® using data stored in GIS environment

7/17/2019 REAKT-D7.3


12

can handle different ground motion scenarios provided by the user and compute a

number of output variables corresponding to the existing data under analysis.

The QuakeIST currently has the following characteristics:

i) Users can upload their own hazard, vulnerability and exposure models;

ii) Different types of assets can be modelled (e.g. buildings, lifelines,

population);

iii) Modelling of the cascade effects (disruption index) is considered;

iv) It can be used on a single processor laptop, as well as on a cloud GIS

computing infrastructure (QuantumGIS, etc.)

The project test area (Industrial Complex of Sines) with urban and industrial occupation

constitutes an optimum feasibility case study in the field of Seismic Early-Warning and

contains infrastructures that represent an actual interdependent system containing

enough interconnections for research on multiple infrastructure interdependencies.

In Sines test cases, we considered the following selected buildings, infrastructures and

elements:

- Buildings- Power transformers- Ng pipelines (ng - natural gas)- Water pipes- Port facilities

Each theme (infrastructure and so on) was obtained from the different stakeholders, in

different formats and units. An extensive data treatment was performed in order to

homogenise the information and introduced in the GIS platform and QuakeIST

simulator. A geometrical and mechanical property of each element of the infrastructure

was used to set typologies and corresponding vulnerability functions.

Each type of a structure and infrastructures has its own structural dynamics response

characteristics and hence a particular structural analysis is needed. The literature

review has addressed the issue of vulnerability or fragility relations for each component

subjected to ground shaking. In this step we want to assess the infrastructure

vulnerability level (Table 1).

Table 1. Elements at risk and procedure for vulnerability analysis

Element at risk Methodology Intensity

parameter

Buildings Giovinazzi & LagomarsinoMacroseismic method

EMS98

Power transformers REAKT WP5 PGA

Gas pipelines HAZUS, ALA PGV

Water pipes HAZUS, ALA PGV

In this earthquake loss estimation study, the simplified method for water systemnetwork performance evaluation applies to a distribution pipe network (HAZUS, 2007).

7/17/2019 REAKT-D7.3


13

Material K1 K2 K3

Aço 0.3 0.0001 2.25

Betão Armado 1 0.0001 2.25

Fibrocimento 1 0.0001 2.25

Ferro fundido dúctil 1 0.0001 2.25

FV 0.3 0.0001 2.25

Hidronil 0.3 0.0001 2.25

Não conhecido 0.3 0.0001 2.25

Polietileno de alta densidade 0.3 0.0001 2.25

Poliester reforçado a fibra de vidro 0.3 0.0001 2.25

PVC (Policloreto de vinilo) 0.3 0.0001 2.25

INX 0.3 0.0001 2.25

Damage algorithms for water pipes:

Repair Rate (repairs/km) =K1 x K2 x PGV(K3) [O'Rourke and Ayala (1993)]

Note that the diameter of pipe is not considered to be a factor. For ductile pipelines

(steel, ductile iron and PVC), the above relation is multiplied by 0.3 (i.e., ductile

pipelines have 30% of the vulnerability of brittle pipelines).

Therefore, due to PGV, the estimated number of leaks is 80% x Ʃ nr. Repairs (= RR x

length), and the estimated number of breaks is 20% x Ʃ nr. Repairs.

Acceptable damage rate evaluation:

According to ALA (2001), an acceptable damage rate should be about 0.1 to 0.2

breaks per km. In Repair rates means 0.02 to 0.04.

Authors (Grigoriu et al.) suggest that a break rate of 0.02/km corresponds to about

Intensity VII, and a break rate of 0.10/km corresponds to about Intensity VIII.

For the gas system elements the damage algorithm for gas pipes is:

Repair Rate (repairs/km) =K1 x 0.002416 x PGV [ALA (2001)]

Material K1

Welded steel 0.15

PVC 0.50

Data on location and material were collected and entered into a GIS database,

providing a powerful basis for evaluating earthquakes effects on lifelines, as well as the

consequences of these interactions on communities.

7/17/2019 REAKT-D7.3


14

4.2. QuakeIST ® seismic simulator. Scenarios

One earthquake scenario, 1755 earthquake, was used for simulation of damage and

serviceability of the main elements under study. As an example, we show the gas and

water networks results. For detailing of the ground motion attenuation (GMPE) used

and conversions from/to PGA/PGV/EMS-98 Intensity, soil influence, etc. (Mota de Sáet al., 2014).

Scenario 1: 1755 earthquake, M8.7

Figure 10: Peak ground motion (PGA) obtained for 1755 earthquake scenario (left) and

intensities (EMS-98) (right).

7/17/2019 REAKT-D7.3


15

Figure 11: Repair rates for gas and water networks under 1755 scenario.

The visualization of earthquake impacts which are predicted by such simulation

contributes to make recognition of earthquake disaster among population and urban

services or functions, but also the improvement of the engineering ability of local

government officials who are in charge of promoting earthquake disaster mitigation.

4.3. Dependencies and interdependencies

Since ports are indispensable nodes of supply chains involving many strategicstakeholders and activities interacting with each other, the main focus of IST research

is to provide a conceptual framework integrating the organizational relationships

between supply chain and port stakeholders.

Lifeline systems are interdependent, primarily by virtue of physical proximity and

operational interaction. The disruption or destruction to one infrastructural component

can rapidly cascade into damage to surrounding components, with system-wide

consequences as health, safety, security, economic or social well-being of people.

The earthquake, the aftershocks as well as the tsunami that can be generated can

exacerbate disasters. Damage to lifelines and industrial facilities take off line

immediately following an earthquake. And many of these remain off line from several

7/17/2019 REAKT-D7.3


16

months. These outages have a critical effect on region and national businesses and

overall quality of life.

In Sines cases, raw materials, suppliers, packaging, and shipping services are located

in the same region damaged by an earthquake and also not functioning. Thus, the

operation of modern industrial societies is highly interdependent and the success of aregion in carrying out business and industrial operations can be rapidly eroded by the

failure of a few key services or lifelines.

For a business to operate, it needs to be in a community that is functional following a

large earthquake. However, the damage to housing, schools, hospitals, commercial

structures, factories and infrastructure systems resulted in a wide spread economic and

social disruption.

Part of the overall aim of REAKT is to include in the damage scenarios developed by

QuakeIST®, the concept of Disruption Index (Oliveira et al., 2012; Ferreira, 2012;

Ferreira et al., 2013) and apply it to Sines for different levels of seismic action.

To develop the model, it remains essential to identify the system and their components

as well as the main dependencies and interdependencies. The Industrial disruption

index is based on a methodology developed by (Ferreira, 2012, Oliveira et al., 2012

and Ferreira et al., 2014) which intend to measure the urban disruption, quantifying

how a cascade effect contributes to the disruption of urban activities. A review of the

relevant literature and the contact with the Sines stakeholders possibly to draw the

industrial disruption index, identifying the main sources and infrastructures in an

industrial complex.

Figure 12 illustrate the main interdependencies among facilities and their equipments.

Figure 12: Industrial disruption index (IDI) (Ferreira et al., 2014)

7/17/2019 REAKT-D7.3


17

Table 2 refers that Energy depends on the operation of electric facilities and

components (as local transformers) and the functionality of port of Sines (once the coal

arrives by ship).

Table 2: Adjacency matrix (a “macro analysis”)

5. Seismological context and Early warning developments

Sines is located next to the Atlantic Coast, at about 180 km from various major

seismogenic sources (the Gorringe Bank, the Marquês de Pombal, Pereira de Sousa

and São Vicente Faults) capable to generate 8.5 to 9 magnitude earthquakes (Fig. 13).

These sources will generate peak ground accelerations of about 0.3 g to 0.50 g in stiff

rock soils, with the possibility of 0.4 g to 0.65 g in soft soils.

Another source of seismic activity that affects the Sines area is related to the faults

associated with the Lower Tagus Valley. Here the historical catalogue (Stucchi et al.,

2012) do not show values above M>6.5.

7/17/2019 REAKT-D7.3


18

Figure 13: Location of Sines and instrumental seismicity (1961-2013, from IPMA) and location

of the main historical earthquakes of the portuguese mainland and its adjacent margins. GB:

Gorringe Bank; TP: Tagus Plain; TS: Tore Seamount; FP: Ferradura Plain; AB: Ampere Bank;

FF: Ferradura Fault; GqF: Guadalquivir Fault; LTV: Lower Tagus Valley; MPF: Marquês de

Pombal Fault; NF: Nazaré Fault; MF: Messejana Fault; MVF: Moura-Vidigueira Fault; LF: Loulé

Fault; CAF: Cadiz-Alicante Fault. (adapted from Perreira et al., 2013).

The offshore seismic zones are also potentially tsunamigenic sources with great impact

along Atlantic coast south of Lisbon, including Sines. There will be a tsunami early

warning (TEW) of most importance and relevance for the Complex of Sines as well asfor many other low coastal regions. This subject is not addressed in REAKT.

5.1. Early warning developments

Stations that can incorporate the Early Warning Systems (EWS) in continental Portugal

belong to the networks in function (Figure 14). Many of them are not yet fully online.

But, they already can be used to study the feasibility of a EWS, mainly to assess the

possible “lead time” for the case these stations become fully online. There are broad -

band Seismological Stations and strong motion 18-24 bit Stations.

7/17/2019 REAKT-D7.3


19

Figure 14. Seismological Stations in continental Portugal and a cloud of epicentres: dark bluetriangles – IPMA; red triangles – SM, IST; light blue triangles – SM-IPMA. Left - detail stations

location SVI (São Vicente) and PFVI (Vila do Bispo). (courtesy: S. Custódio, 2012).

5.1.1 Epicentres in the SW Iberia

Table 3 shows a group of several earthquakes recorded in the period 2007-2013,

corresponding to the largest magnitudes over the last 10 years and to the ones

recorded with best equipment. These events are all M~2-3, with epicentres in the

southern part of the continental Portugal, the most active area. Two larger events,M>5.5 were also added. The arrival times of P and S waves to several stations were

reported in the EMSC-CSEM catalogue (2014).

Table 3: Earthquakes considered in the current analysis

For the purposes of early warning systems and mitigation, information on the arrival

times of P and S waves to different stations is useful. Data reported in Figure 15 shows

their time difference (S-P), for those stations against epicentral distance (obtained from

Table 3). We plotted the S-P arrival at each station and the S-P1, where P1 is the

arrival time at the station closer to the epicentre, IPMA (2014) Station PFVI (Figure 15),located at the southernmost point of continental Portugal.

7/17/2019 REAKT-D7.3


20

It is clear that if we only consider these times without any time for data treatment and

transmission, with the S-P1 we would increase the lead time in relation to one

calculated from the on-site S-P measurement of 15 sec for sites 200 km away from the

epicentre (Sines) and 22 sec for sites at 270 km (Lisbon). This corresponds to the time

difference between the two lines of Figure 15.

The total lead time without any delay time (the top line in Figure 16) is about 35 sec for

Sines if the detection is made in PFVI, or 38 sec if detection is made at SVI station (at

the São Vicente southwest most location).

Figure 15: Lead time in seconds for epicentres located SW of continental Portugal (Sines is

around 100 km from the southernmost location of continental Portugal where the Station PFVI is

located).

According with recent studies developed by Carranza et al. (2013), based on the

existing seismographic stations of IPMA (2014), the lead time for Sines for an event

with epicentre at about 289 km based on 5, 8 and 10 stations are 25 sec, 16 sec and

12 sec, respectively (Figure 16). Data treatments comprises the automatic analysis of

the accelerograms near the seismic source, leading to an estimation of the magnitude

and epicentral location, allowing the estimation of ground motions amplitude at any site

of interest. This can be done in 4-6 sec, depending on the algorithms. The decision of

issuing a warning will then be automatically made by means of comparing the

estimated values of a ground motion indicator with the threshold values of that indicator

to be defined as a function of the stakeholders needs.

However the above mentioned lead times present large differences that may be

reduced optimizing the location of the different stations, by means of concentrating a

large number of stations closer to the potential epicentre. With the current network

configuration, a minimum number of 10 stations needed to trigger the alert would mean

that Portimão would be within the blind zone that could not be alerted. Only 5 sec

would be available at Faro, 21 sec at Lisbon and 46 sec at the farthest city, Seville

(Spain).

7/17/2019 REAKT-D7.3


21

Figure 16: Lead time in seconds for an epicentre located SW of Continent (289 km from

Sines) (courtesy: M. Carranza et al., 2013)

The numbers presented in Figures 15 and 16 are not totally in agreement with each

other because Figure 16 was made with various epicentre locations closer to São

Vicente than the epicentre considered in Figure 14. Nevertheless, the values presented

in Figure 15 and 16 are quite similar in tendency. The more stations are used in the

computation the higher the reliability of the estimation and, of course, the smaller the

lead time. Perhaps, if stations are organized in a linear array in an L-shape along the

coast line, the gains might be higher.

Similar results were obtained by Pazos et al. (2014) and by Romeu et al. (2014) which,

using different software arrived for Lisbon to a lead time from 20 to 39 sec for two

epicentre locations, 100 and 200 km SW of PFVI Station.

Further studies are being done using only the arrival of the S waves at an advanced

location. This would present higher reliability on the estimations of S waves at Sines.

5.1.2 Epicentre in the Lower Tagus Valley Region

For the epicentres in the Lower Tagus river Valley (LTV) the situation is much different

from the first one. We do not have good information to understand how much lead time

we would have to send alerts to Sines. No event in the LTV zone has occurred in

recent times to allow an exercise similar to the one we did for events with epicentre

southwest of the continental Portugal where large magnitudes are expected.

The distance from LTV seismic sources to Sines vary quite significantly. If we

concentrate only in the faults to the north of Lisbon with an epicentral distance of 110

km, and considering the same model of Figure 15 and not discounting the time for data

treatment and transmission, the on-site S-P would be of just 12 sec, and if we use one

station 50 km ahead of Sines detecting the onset of P waves, the lead time would be

19 sec. For a feasible estimation, which would require more stations, the final lead time

for EEW would probably be very little.

For other offshore epicentral locations, at smaller distances (< 50 km) from Sines, the

lead time for EEW would result solely from on-site S-P arrivals and could be as much

as the epicentral distance would permit.

7/17/2019 REAKT-D7.3


22

5.1.3 Other early warning developments

Normally we look at times until the arrival of S-waves as the onset of important shaking

whose effects should be avoided or minimized. However, a structure takes some time

to respond to the input ground motion, depending essentially on the ratio of

“predominant” frequency and frequency of structure and on structural damping. In asimplified way and in case of a building (control installations) what we want is to launch

actions before the building attains a certain level of danger to the people inside. Actions

may be to escape from the building, to move to some shelter or safe place inside, to

proceed to open doors to the outside, etc. In other words, besides other considerations

dealing for instance with mobility under strong shaking, or at dark, or walk through

toppled objects, we are also interested to know how much time we may have since the

onset of S-waves to the moment the building attains the structural damage levels D2 to

D4 degree (Grünthal, 1998) in a scale of 5 levels. It is important to notice that the

effects of seismic actions in industrial equipment should be given not only in terms of

structural damage but also in terms of content release that may be activated during or

after the shaking (fire, leakage, toxic dispersion and so on) (Salzano et al., 2009).

Damages and losses of certain types of containments may be more important than the

direct damages inflicted. Also the effect of interdependences may be crucial to the

functionality of other equipment and facilities. In these cases short lead times might be

sufficient to block the cascade effect caused by interdependences or reduce factors

that contribute to leakage of toxic or dangerous products.

We present here the first steps to compute the time from the onset of S-wave to attain

various levels of structural performance, this is the time it takes for a structure to reach

several degrees of response related with the level of damage that the structure is

suffering. (Oliveira et al., 2014). First of all a collection of strong ground motionsrecords were selected to perform this analysis. Only earthquakes with M>6 and

especially M>8 where used: Izmit 1999 and Duzce 1999 (nearby) and Chile 2010 (far

away). We performed linear analysis for a group of single degree of freedom systems

with periods varying from 0.2 to 2.0 sec and a damping ratio of 1%, and non-linear

analysis for a Takeda hysteretic response using the commercial program CSI-

SAP2000®, (CSI, 2008) representing the most common building structures existing in

southern Portugal. This analysis is important to study not only the buildings that house

the control facilities.

We evaluated for these structures the time tDi for the structure to attain the level of

response capable of inducing D2 to D4 damage levels. The preliminary results indicate

that for near-field earthquakes the instant at which the PGA takes place is very close to

the instant of the maximum of the response, regardless of the period of the oscillator,

Figure 17. Usually, depending on the magnitude, the time from the onset of S-wave is

only a few seconds. This time enlarges for larger magnitudes, as it can be spotted in

Figure 18.

7/17/2019 REAKT-D7.3


23

a)

b)

Figure 17. Time from the onset of S-wave up to the attainment of various levels of structural

performance: a) time to maximum response in linear case (Vinas del Mar, Chile); b) time to tD2,

tD3 and tD4 for nonlinear case for building typologies constructed in the period 1960-1986

(Chile earthquake 2010). (T(s) – period; t(s) – onset time; Max past and Max prior PGA

represent the times to reach maximum values attained immediately past and prior to PGA).

Even though these preliminary results show that it might be possible to gain a few more

seconds from the onset of the S-wave, especially for the very large magnitudes, many

more examples should be carried out both using a larger selection of ground motion

records, and extending the type of structures under analysis.

7/17/2019 REAKT-D7.3


24

a)

b)

Figure 18. Time from the onset of S-wave to attain various levels of structural performance as a

function of the magnitude of event: a) time to maximum response in linear case (March 11,

2011 Tohoku records were added) ; b) time to tD2, tD3 and tD4 for nonlinear case for building

typologies constructed in the period 1960-1986. (Note: the magnitude scale cannot reach so

large values).

6. Earthquake early warning survey

A survey was developed and sent to stakeholders to test these assumptions andexplore the views of potential users in the following issues, demonstrating how variousstakeholders are positioning themselves in relation to early warning:

- how these stakeholders might use warnings of 12 and 25 seconds duration, theperceived benefits, costs and challenges in use of an earthquake early warningsystem;

0

5

10

15

20

5 6 7 8 9 10

t ( s )

Magnitude(Mw)

PeakAccelerationArrival

- starting from S-Wave

arrival

0.00

2.00

4.00

6.00

8.00

10.00

12.00

5 6 7 8 9 10 11

typology

1960/1986

Magnitude (Mw)

tD2(s)

tD3(s)

tD4(s)

7/17/2019 REAKT-D7.3


25

- determine what benefits the early warning system can realistically provide, andwhat is outside its capacities;

- analyse the ratio between lead time (early warning) and time necessary toperform some actions;

- understand people behaviour after they receive a warning? — particularly howthey prioritize different risks;

- identify actions that might be taken with 12 seconds warning and 25 seconds;- analyze the importance of false alarms, errors and missed events;- what do they think is the best balance between lead time and reliability, for their

equipment.

A lack of understanding of the uncertainty of estimations led some final users and thepublic in general to interpret some predictions that did not take place as wrongpredictions, and to believe that estimations could no longer be trusted. Statementssuch as "there is a 20 per cent chance that rainfall will be above the inter-annual mean"

present information in an unfamiliar language.

It is important to report, communicate and have an appropriate and effective interactionamong the main actors of the early warning process, such as the scientific community,stakeholders and decision makers. In addition, the scientific community messageshould communicate the level of uncertainty and the possibility of induced costs oftaking action after a warning that later is verified to be a false alarm. This requires themessage from the scientists to the final users to be stated in simple language so as tobe understood by those who receive it.

During the REAKT project a survey on user acceptability was conducted (March-June2014) to identify how the organizations might use warnings of 12 and 25 seconds, the

perceived benefits, and assess the factors that may influence organizationalacceptance and use of such a system.

A summary of the survey results is presented below:

1. Do your facilities were hit by some disaster (e.g., fire, floods, tornadoes, etc.) and didNOT receive any warning of their occurrence?

Yes: 0; No: 100%

2. How satisfied are you with the available warning systems on your installations?

Don’t know Very satisfied Satisfied Poor satisfied Not satisfied

Central phone 60% 40%

Sirens (light) 40% 60%

Sirens (sound) 40% 60%

Loudspeakers 40% 40% 20%

Radio/TV 100%

SMS 100%

Email 40% 40% 20%

Others 40%

7/17/2019 REAKT-D7.3


26

3. Do you consider early warning systems for fires, release of toxic substances, etc. asa strategy to effectively reduce the risk and vulnerability of your facilities/installationsand community?

Yes: 100%; No: 0

4. Knowing that your installation can be hit by strong earthquakes (and possibletsunamis), do you consider of utmost importance to install a warning system forearthquakes and tsunamis in your industrial facility?

Yes: 83%; No: 17%

5. Consider the occurrence of an earthquake and its vibration. Do you think that 12

seconds (with probability of success of 95%) is sufficient to take effective actions thatreduce the risk of fire/explosion/spill/other (e.g. equipment shutdown) and allow preparing an appropriate response?

Yes: 17%; No: 83%

6. Consider the occurrence of an earthquake and its vibration. Do you think that 25seconds (with probability of success of 70%) is sufficient to take effective actions thatreduce the risk of fire/explosion/spill/other (e.g. equipment shutdown) and allow

preparing an appropriate response?

Yes: 17%; No: 83%

7. What is the importance of a False Alarm (vibration) to your facility?

Don’tknow

Veryimportant

IndifferentLowimportant

Notimportant

In terms of security 33% 33% 17% 17%

In terms of costs (to restart

the system) 67% 23%

8. List the equipments that could benefit most from early warning system (vibration).

Valves connecting pipes and storage tanks as well as pressurized vessels containingliquefied gases, methanol, formaldehyde, paraxylene and acetic acid storage tanks (allof them are located at the Sines harbour), the respective pumps, the pipelineconnecting the Sines LNG Terminal to the natural gas transport network and all rotatingequipment.

9. List some advantages and disadvantages of implementing an early warning systemon your industrial installations.

7/17/2019 REAKT-D7.3


27

Advantages: Avoid casualties and damage and evacuate people from buildings. Thetsunami warning can save lives. For fixed equipment, little or nothing can be done;however mobile machinery and vehicles may be taken to a safe area, assuming anotice of at least 20 minutes.

The initial shutdown of pumping devices, piping transport, and similar actions can beactivated within the EEW alert. However, full shutdown will take more time.

Disadvantages: Warning time is not enough to take effective actions. In some cases aFalse Alarm can incur in high costs.

7. Cost-Benefit preliminary analysis

In this section, Fault Tree Analysis (FTA) will be applied to the analysis of an adverseevent associated with the failure of an alarm to generate the appropriate response.

FTA is a structured top-down deductive to failure analysis, starting with a potentialevent called a TOP event, and then determining all the ways it can happen.

Figure 19 presents a preliminary exercise made for Sines, considering the losses (if noEEWS is present) and the gains due to EEWS implementation in a life time of 50 yearswith probability of 39% of exceedance (RP=100 yrs). For the case presented there is aloss reduction (expected value) of about 34.2 M€, which corresponds to a saving of9.7% of total inflicted losses if no EEWS is implemented (E2-E1)/E2.

We have run other cases for checking the influence of other parameters in this savings.

Table 4 presents a sensitivity analysis for changing the Loss Reduction Factor (Lrf) for

two different Return Periods (RP).

Table 4: Sensitivity analysis for changing the Loss Reduction Factor (Lrf ) for two different Return

Periods (RP)

Time life -50 yr Time life -20 yr

Lrf (%) Reduction of Losses (%)IDI=1

Reduction of Losses(%) IDI=1

1 0.7 0.2

2 1.7 1.2

5 4.7 4.210 9.7 9.2

15 14.7 14.2

20 19.6 19.2

Note that Loss Reduction Factor (Lrf) is function of the effectiveness of thepreparedness plans implemented with EEWS and of the degree of IDI (IndustrialDisruption Index – Figure 12). All these results are very preliminary because most ofthe parameters have to be checked with end-users.

7/17/2019 REAKT-D7.3


28

Figure 19. Example of a FTA made for Sines.

E2-E1

E1

E2

7/17/2019 REAKT-D7.3


29

REFERENCES:

ALA (2001). Seismic fragility formulations for water systems. American Lifeline

Alliance, ASCE.

Carranza, M.; Buforn, E.; Colombelli, S. and Zollo, A. (2013). “Earthquake early

warning for southern Iberia: A P-wave threshold-based approach”. Geophysical

Research Letters, vol. 40, 1 –6, doi:10.1002/grl.50903.

CIP (2011). Critical Infrastructure Protection. Good Practices Manual for CIP Policies.

RECIPE EU Project.

CSI (2008) CSI-SAP2000®, Computers and Structures, Inc., Berkeley, Calif, USA.

EMSC-CSEM (2014). European-Mediterranean Seismological Centre:

http://www.emsc-csem.org/about: (consulted on June 2014).

Ferreira, M.A. (2012). “Risco sísmico em sistemas urbanos”. Ph.D Thesis. Instituto

Superior Técnico, Universidade Técnica de Lisboa. 295 pp (in portuguese).

Ferreira, M.A.; Mota de Sá, F. and Oliveira, C.S. (2014). “Disruption Index, DI: an

approach for assessing seismic risk in urban systems (theoretical aspects)”. Bulletin of

Earthquake Engineering, DOI 10.1007/s10518-013-9578-5

Ferreira, M.A.; Mota de Sá, F. and Oliveira, C.S. (2014). Industrial Disruption Index

work in progress)

Garcia, P.R. (2007). Life and death of the Lisbon waterfront. PhD thesis. Universidade

Portucalense Infante D. Henrique)

Grigoriu, M., O’Rourke, T., Khater, M. (1989). Serviceability of San Francisco Auxiliary

Water Supply System. Proceedings of 5th International Conference on Structural

Safety and Reliability, San Francisco,

Grünthal, G. (Ed.) (1998). “European Macroseismic Scale 1998 (EMS-98)”. Cahiers du

Centre Européen de Géodynamique et de Séismologie 15, Centre Européen de

Géodynamique et de Séismologie, Luxembourg, 99 pp.

HAZUS99 (1999). “Earthquake loss estimation methodology”. Federal Emergency

Management Agency, Washington D.C.

HAZUS-MH MR4 Technical Manual (2007). Department of Homeland

SecurityEmergency Preparedness and Response Directorate. FEMA

IPMA (2014) (Instituto Português do Mar e Atmosfera): (consulted on June 2014).

ISDR-PPEW 2005ª. http://www.unisdr.org/ppew/whats-ew/basics-ew.htm

Moreira, P.J.P. (2013). The port of Sines: contribution for the emergence of a regional

cluster. Resume from a Master Thesis in Portuguese Economy and InternationalIntegration, ISCTE-Business School, Lisbon, http://catalogo.biblioteca.iscte-iul.pt/2

http://catalogo.biblioteca.iscte-iul.pt/2




7/17/2019 REAKT-D7.3


Mota de Sá, F.; Ferreira, M.A. and Oliveira, C.S. (2014). “QuakeIST earthquake

scenario simulator”. Proceedings, 2ECEES, Istanbul, Turkey, 24-29 August.

Oliveira, C.S.; Lopes, M.; Camacho, V. and Mota de Sá, F. (2014) “Time from the onset

of S-wave to attain various levels of structural performance“ (work in progress).

Oliveira, C.S.; Ferreira, M.A. and Mota de Sá, F. (2012). The concept of a disruption

index: application to the overall impact of the July 9, 1998 Faial earthquake (Azores

islands). Bulletin of Earthquake Engineering. 10(1): 7-25.

O'Rourke and Ayala (1993). Pipeline damage due to wave propagation. Journal of

Geotechnical Engineering. ASCE: 119 (9), 1490-1498.

Pazos, A.; Romeu, N.; lozano L.; Colom, Y.; López Mesa, M.; Goula, X.; Jara, J.A. and

Cantavella J.V.(2014). "A regional approach for earthquake early warning in south west

Iberia: a feasibility study”, (Submitted to BSSA).

Perreira, N.; Carneiro, J.F.; Araújo, A.; Bezzeghoud, M. and Borges, J.F. (2013).

“Seismic and structural geology constraints to the selection of CO2 storage sites – the

case of the onshore Lusitanian basin, Portugal”, Journal of Applied Geophysics (in

press).

Romeu, N;, Goula, X.; Jara, J.A.; Colom, Y. and T. Susagna. (2014) “An Earthworm

based prototype of an EEWS for SW Iberia” (submitted to PAGEOPH).

Salzano, E; Agreda, A.G.; Di Carluccio, A. and Fabbrocino, G. (2009). “Risk

assessment and early warning systems for industrial facilities in seismic zones”.

Reliability Engineering and System Safety, vol. 94, no. 10, pp. 1577 –1584.

Stucchi et al., (2012) “The SHARE European Earthquake Catalogue (SHEEC) 1000–

1899”. Journal of Seismology, doi: 10.1007/s10950-012-9335-2.

SYNER-G (2013). Systemic Seismic Vulnerability and Risk Analysis for Buildings,

Lifeline Networks and Infrastructures. European Research Project funded by FP7.

Project reference: 244061. http://www.vce.at/SYNER-G/.(consulted on June 2014).