B
ui
ld
ings
Ab
ove
Nor
m
Figure 5.4
Graphs show the Local Personal Risk associated with the production scenario “Basispad Kabinet” for average
(green), cold weather and warm weather years, and the Reference Scenario (24 Bcm/year) (black), and for the
period 2018 to 2027. Each cold (blue) and warm (red) year is followed by average temperature years.
Right graph:
number of buildings exceeding the norm mean LPR larger than 10
-5
/year
Left graph:
number of buildings exceeding the norm mean LPR larger than 10
-4
/year
Seismic Risk Assessment for Production Scenario “Basispad Kabinet” for the Groningen field - June 2018
68
B
ui
ld
ings
Ab
ove
Nor
m
Figure 5.5
Graphs show the Local Personal Risk associated with the production scenario “Basispad Kabinet” for average
(green), cold weather and warm weather years, and the Reference Scenario (24 Bcm/year) (black), and for the
period 2018 to 2027. The cold and warm weather years have been gathered as the blue and red line
respectively.
Right graph:
number of buildings exceeding the norm mean LPR larger than 10
-5
/year
Left graph:
number of buildings exceeding the norm mean LPR larger than 10
-4
/year
2020, CS1
2020, CS2
2020, CS3
2020
Seismic Risk Assessment for Production Scenario “Basispad Kabinet” for the Groningen field - June 2018
70
2022, CS1
2022, CS2
2022, CS3
2022
Seismic Risk Assessment for Production Scenario “Basispad Kabinet” for the Groningen field - June 2018
71
2024, CS1
2024, CS2
2024, CS3
2024
Seismic Risk Assessment for Production Scenario “Basispad Kabinet” for the Groningen field - June 2018
72
2026, CS1
2026, CS2
2026, CS3
2026
Seismic Risk Assessment for Production Scenario “Basispad Kabinet” for the Groningen field - June 2018
73
Figure 5-6
Graphs showing the number of buildings exceeding a given average annual collapse rate for CS1 (top left), CS2 (top right), and CS3 (bottom left), for the “Basispad
Kabinet” production scenario for the assessment periods years 2020, 2022, 2024 and 2026. The named structural systems denote the top-five ranked according
to the number of buildings with a collapse rate of at least 10
-5
/year. Breakdown of the structural systems (bottom right) contributing to LPR over the 10
-5
/year
threshold for the assessment periods years 2020, 2022, 2024 and 2026 for the “Basispad Kabinet” scenario.
5.2
Maps of Buildings compared to the Meijdam-Norm Risk Levels
The maps of figure 5.7 show all buildings exceeding mean LPR>10
-5
/year for the years 2020, 2022, 2024 and 2026.
Different colours represent different dominant building typologies. For the purpose of this risk assessment, the
Groningen building stock has not been adjusted for the ongoing strengthening operations.
Seismic Risk Assessment for Production Scenario “Basispad Kabinet” for the Groningen field - June 2018
76
2022
Seismic Risk Assessment for Production Scenario “Basispad Kabinet” for the Groningen field - June 2018
77
2024
Seismic Risk Assessment for Production Scenario “Basispad Kabinet” for the Groningen field - June 2018
78
2026
Figure 5.7
Maps of all buildings exceeding mean LPR>10
-5
/year for the years 2020, 2022, 2024 and 2026. Different colours indicate different building typologies.
2020
2022
2024
2025
Figure 5.7
Map indicating individual building with Local Personal Risk exceeding 10
-5
/year for the years 2020, 2022, 2024
and 2026 and production scenario “Basispad Kabinet”.
Seismic Risk Assessment for Production Scenario “Basispad Kabinet” for the Groningen field - June 2018
80
5.3
Structural Upgrading Program
The probabilistic assessment of the number of buildings that do not meet the Meijdam Norm does not immediately
translate into an estimate of the structural strengthening scope. There are three main reasons why the scope of the
structural upgrading plan will in general be larger than the probabilistic assessment of the number of buildings that
do not meet the Meijdam norm.
Efficiency of identifying buildings with LPR >10
-5
has not yet been proven.
The Hazard and Risk Assessment is a probabilistic assessment and does not directly identify each individual
building that needs to be included in the structural upgrading plan. Through an inspection program these
buildings will have to be identified. A risk-based inspection program will be able to identify these buildings with
reasonable efficiency.
Remaining uncertainty in hazard and risk assessment.
Significant progress has been made towards assessing the risk from Groningen earthquakes. However,
uncertainty remains in the estimate of the number of buildings that do not meet the norm based on mean LPR
> 10
-5
/year. Especially building inspections can help reduce this uncertainty.
Differences between the hazard and risk assessment and NEN-NPR building code.
Ultimately the structural upgrading scope will be based on the NEN-NPR building code. Improvement of the
Hazard and Risk Assessment Updating and calibration of the building code with the latest technical insight from
laboratory experiments and modelling are likely to reduce the difference.
The probabilistic estimate of the number of buildings, where the Meijdam-Norm Safety Level is exceeded, does
therefore not directly translate into an estimate of the structural strengthening scope. However, the Hazard and Risk
Assessment provides a useful tool for prioritisation of building inspections.
Ultimately the structural upgrading scope
will be based on the assessment of individual buildings based on the NEN-NPR building code.
Seismic Risk Assessment for Production Scenario “Basispad Kabinet” for the Groningen field - June 2018
81
5.4
References
1.
Seismic risk assessment for a selection of seismic risk production scenarios for the Groningen field - Addendum
to: Induced Seismicity in Groningen Assessment of Hazard, Building Damage and Risk (November 2017), Jan
van Elk, Assaf Mar-Or, Leendert Geurtsen, Per Valvatne, Eddy Kuperus and Dirk Doornhof, March 2018.
Seismic Risk Assessment for Production Scenario “Basispad Kabinet” for the Groningen field - June 2018
82
Appendix A – Abbreviations
EZK
Ministry of Economic Affairs and Climate Policy
GTS
Gasunie Transport Services BV
GY
Gas-year (12-months period following 1
st
October). This was introduced for practical reasons. The gas-
year starts with the 6 coldest months of the year avoiding a winter period to be split over two one-year
time periods, such as a calendar year.
H-gas
High Calorific Gas (Gas from most gas field has a higher calorific content than gas from the Groningen gas
field)
HRA
Hazard and Risk Assessment
L-Gas
Low Calorific Gas (Groningen gas had due to the nitrogen content a lower calorific content than gas from
many other gas fields)
LPR
Local Personal Risk
MC
Monte Carlo
N
2
Nitrogen
NAM
Nederlandse Aardolie Maatschappij BV
NFA
No Further Activity
UGS
Underground Gas Storage
A more complete list if abbreviations can be found in “Induced Seismicity in Groningen, Assessment of Hazard,
Building Damage and Risk – November 2017, NAM (Jan van Elk and Dirk Doornhof), November 2017” available from
www.nam.nl
.
Seismic Risk Assessment for Production Scenario “Basispad Kabinet” for the Groningen field - June 2018
83
Appendix
B
–
Verwachtingenbrief
aanvulling
Winningsplan Groningenveld 2016
For convenience, the Expectation Letter has been included in the report in this Appendix.
Seismic Risk Assessment for Production Scenario “Basispad Kabinet” for the Groningen field - June 2018
84
Seismic Risk Assessment for Production Scenario “Basispad Kabinet” for the Groningen field - June 2018
85
Seismic Risk Assessment for Production Scenario “Basispad Kabinet” for the Groningen field - June 2018
86
Seismic Risk Assessment for Production Scenario “Basispad Kabinet” for the Groningen field - June 2018
87
Seismic Risk Assessment for Production Scenario “Basispad Kabinet” for the Groningen field - June 2018
88
Seismic Risk Assessment for Production Scenario “Basispad Kabinet” for the Groningen field - June 2018
89
Appendix C – Implementation of the discrete M
max
Distribution in the Probabilistic Seismic Hazard and
Risk Analysis
The sensitivity of the Probabilistic Seismic Hazard and Risk Analysis to epistemic uncertainties identified on the logic
tree (Fig. C.1) is shown in Figure C.2. Four key factors have been identified: the seismological model, ground motion
model (GMM), building fragility model, and the consequence model. The extent of each grey bar denotes the average
value of the risk metric for the subset of the logic-tree where the given factor is constrained to the lower branch
(lower limit) and then the upper branch (upper limit). Results are shown for 2018-2022 under the 24 bcm/year
production scenario for a single mean local personal risk (LPR) metric, computed as the mean over all populated
buildings and all probability-weighted logic tree branches. Alternative assessment periods and production scenarios
yield similar results for the relative sensitivities.
The Hazard and Risk Assessment of November 2017 (Ref. 1) indicated that the distribution of epistemic uncertainty
identified for the maximum possible earthquake magnitude, M
max
, is a significant contributor to epistemic
uncertainty in the Probabilistic Seismic Hazard and Risk Analysis.
Figure C.1
Summary of the logic-tree used to characterise epistemic uncertainties within the Probabilistic Seismic Hazard
and Risk Analysis. The complete logic tree contains 216 branches which comprises the full factorial
combination of these 5 factors and their associated 2—4 levels. Each logic tree branch was analysed by an
independent Monte Carlo simulation of the unique combination of sub-models. Mean hazard and risk metrics
are derived as the probability-weighted combination of all Monte Carlo results across the complete logic tree.
Seismic Risk Assessment for Production Scenario “Basispad Kabinet” for the Groningen field - June 2018
90
Figure C.2
Tornado plot to indicate the sensitivity of the Probabilistic Seismic Risk Analysis to the identified epistemic
uncertainties. The extent of each bar denotes the values of the risk metric under sequential re-weightings of
the logic-tree where the each factor in turn is constrained to the lower branch (lower limit) and then the upper
branch (upper limit). Results are shown for 2018-2022 under the 24 bcm/year production scenario for logic
tree mean local personal risk for all populated buildings.
The distribution was established by a panel of experts following a workshop at Schiphol Airport, The Netherlands
(Ref. 2). The resulting cumulative distribution function (CDF) for M
max
was represented by eight values (Fig. C.3).
Figure C.3
Table of eight values capturing the cumulative distribution function (CDF) for M
max
taken from “Report on
Mmax Expert Workshop – 8 – 10 March 2016” (Ref. 2, page 9 of the Report from the Expert Panel on Maximum
Magnitude Estimates for Probabilistic Seismic Hazard and Risk Modelling in Groningen Gas Field 25 April 2016).
For practical reasons, this distribution was captured by three branches in the logic tree (Fig. C.4). Including a seven-
branch representation of the uncertainty in M
max
, would have been impractical from a computer run-time
perspective. The M
max
values and associated probabilities on the three branches of the logic tree were chosen to
exactly match the zero, first, second, third, fourth and fifth moments of the M
max
distribution, as established by the
M
max
expert panel.
Seismic Risk Assessment for Production Scenario “Basispad Kabinet” for the Groningen field - June 2018
91
Figure C.4
Logic tree for M
max
as used in the Hazard and Risk Assessment of November 2017, taken from reference 3 pg.
117.
To assess the influence of representing epistemic uncertainty in M
max
by three levels in the logic tree, a single risk
analysis was conducted using the 7-level M
max
distribution established by the M
max
expert panel. This single analysis
was based on the reference production scenario of 24 Bcm/year also used in the Induced Seismicity in Groningen
Assessment of Hazard, Building Damage and Risk of November 2017 (Ref. 1). The hazard maps for these two hazard
assessments with a representation of the uncertainty in the M
max
by 3- and 7- point discrete distributions respectively
are shown in Figure C.5.
Three Levels
Seven Levels
Figure C.5
Hazard maps for 3-point (left) and 7-point (right) discrete probability M
max
distributions under the 24 Bcm/year
production scenario and a 0.21% annual probability of exceedance.
Comparison of these figures shows that 7-point discrete M
max
distribution yields systematically lower Peak Ground
Acceleration (PGA) values (Table C.1). The influence of this representation on LPR for buildings in the Groningen area
is shown in Figure C.5. Consistent with the hazard results, the influence on risk also results in a lower estimate of
buildings exceeding the life-safety norm.
Largest PGA on the Hazard Map
Three Levels
Seven Levels
Three Branches
Seven Branches
Seismic Risk Assessment for Production Scenario “Basispad Kabinet” for the Groningen field - June 2018
92
Including Schildmeer area
(as reported in reference 3).
0.207 g
0.196 g
Excluding Schildmeer area
0.185 g
0.174 g
Table C.1
The largest PGA on the hazard maps (0.21% annual exceedance probability) for the 3-point and 7-point discrete
M
max
distributions under the 24 Bcm/year production scenario.
Three Levels
Seven Levels
Figure C.5
LPR distributions under the 24 Bcm/year production scenario for the 3-point (left) and 7-point (right) discrete
M
max
distributions.
The implementation of a 3-point instead of a 7-point discrete M
max
probability distribution in the logic-tree approach
has not led to under-estimation of Probabilistic Seismic Hazard or Risks metrics. Instead, this the 3-point discrete
M
max
probability distribution is shown to be conservative relative to the 7-point discrete probability distribution
established by the M
max
expert panel.
C.1 References
1.
Induced Seismicity in Groningen, Assessment of Hazard, Building Damage and Risk – November 2017, NAM
(Jan van Elk and Dirk Doornhof), November 2017.
2.
Report on M
max
Expert Workshop – 8 – 10 March 2016, World Trade Centre, Schiphol Airport, The Netherlands,
Independent Expert Panel, July 2016.
Three Branches
Seven Branches
Seismic Risk Assessment for Production Scenario “Basispad Kabinet” for the Groningen field - June 2018
93
Dostları ilə paylaş: |