An open source platform addressing structural stability risk assessment in historical centres

Renzo Carlucci
Charilaos Maniatakis
Philip Fayad
Andrea Di Iorio
Nikolaos Schetakis
Costantine Spyrakos
Haris Saroglou
Alessio Di Iorio
Alexandros Paraskeuas
Nikos Papadopoulos

The management of stability risk assessment in historical centres in a project that aims to design and develop an open platform where blocks of buildings and large structures may be represented in damage maps before the event occurs, addressing damage forecasts for seismic movements impacting on the structural stability of the CH. The structural stability is calculated using a Matlab code developed at the National Technical University of Athens, translated to an open source programming language (Python)
using the software tool GeoPandas to enable operations that would normally require spatial database software (PostGIS). Furthermore, the use of open source technologies like QGIS desktop software will provide open and free access to view and contribute in the future.

Introduction

STABLE (STructural stABiLity risk assEssment, 2018-2022, Horizon2020, MSCA - Marie Sklodowska Curie Actions - RISE - Research and Innovation Staff Exchange - grant agreement nº. 823966) has the goal of developing a digital platform to forecast earthquake damage to Historical City Centers in Europe combining different techniques and methodologies: structural stability models, earthquake simulation tools combined with geotechnical data, remote sensing and in situ monitoring technologies.

The STABLE team is composed of seven different partners from Italy (Sapienza University, Rome; University of Tuscia, Viterbo), Greece (National Technical University and Geosystem Hellas from Athens and FORTH from Rethymno) and Cyprus (Frederick University and System Space Solution Ltd), with specific expertise, coordinated together by ALMA Sistemi Srl (Italian SME), and with a plan of collaborations and staff exchanges in order to allow interdisciplinary research.

Some emblematic case studies of buildings from different countries were selected as test cases: a few blocks in the city centre of Rieti, Italy; city blocks in Nafplion and the Lavrion area, Greece; a few buildings in the central region of Strovolos, a district close to Nicosia, Cyprus; and, recently, some buildings in the historic centre of Rethymno in Crete, Greece.

The goal of the project is to define a risk modelling system through the integration of different data, from SAR satellite images to geological and geotechnical information, in order to create deformation soil maps. Additionally, we use free and open source solutions, as far as possible, to achieve all the above

The risk map generation

The study aims to investigate the vulnerability of the architectonic heritage and to combine the results with hazard scenarios, as a first fundamental step in the seismic risk mitigation process. The two key elements of a vulnerability analysis are the capacity (strength and deformation) of a structure and the seismic demand. From these elements it is possible to predict expected vulnerability functions of possible damage to the structure as an effect of the seismic input. Corresponding fragility curves will be determined, expressing the probability of a structure, belonging to a certain class, reaching or exceeding a particular damage grade given a deterministic estimate of the spectral displacement caused by an expected hazard (e.g. Kappos et al. 2008).
The expected risk is determined according to the following expression:
Risk = f (Hazard, Vulnerability)
where:
Hazard: seismicity of the region and accepted probability of exceedance of the destructive action
Vulnerability: capacity of the structure as described below.

The function is dependent on the various methods for the assessment of the structure’s vulnerability and the hazard of the soil. The type of method chosen depends mainly on the objective of the assessment, but also on the availability of data and technology. In general, a vulnerability function is a relationship which defines the expected damage on a structure or a class of structures as a function of the ground motion (Figure 1). The two key elements of a vulnerability analysis are the capacity of the structure and the seismic demand. In order to estimate the damage (D), the ability of the structure to resist constraints (capacity of the structure) must be compared with the constraints on the structure due to the earthquake ground motion (seismic demand).

To express the seismic demand, until very recently, the “intensity” was used nearly exclusively. This is a descriptive parameter of an earthquake based on observations of the effects of the earthquake on the environment. It has the advantage that historical data on earthquakes are available. However, information on the real ground movement is lost, and empirical relationships between intensity and peak ground acceleration vary a lot. Some methods use peak ground acceleration as the parameter defining the earthquake. Nevertheless, in that case, not only the information on the duration of the earthquake is lost, but also the information on the frequency content. Thus, a better parameter is the spectral acceleration (Sa) or the spectral displacement (Sd).

The estimation of damage probability will be performed by means of a simplified mechanical methodology (Lagomarsino and Giovinazzi 2006). This method provides results of satisfactory accuracy for relatively simple constructions, while the assessment of structural capacity for more complex constructions, such as monuments, requires more elaborated methodologies (Spyrakos et al. 2015). Initially, the required input parameters will be collected, including:

  • Structural data (typology, construction period, use of the building, post-earthquake damage data, detection of different construction phases, etc.)
  • Geometrical data (total height, number of storeys, perimeter, footprint area, average bearing wall thickness of masonry structures, plan view of at least the ground floor, etc.)
  • Satellite data (coordinates of each building, number of building blocks, number of buildings in each building block)
  • Vulnerability data (estimation of vulnerability by rapid assessment methods related to the evaluation of conservation of the structure and materials used)

The Structural Model Processor

This program calculates the risk map of the building stroke of an area for a given seismic hazard by applying a simplified mechanical method. The risk is defined as the probability of a certain level of damage being exceeded. Four levels of damage are examined, namely P1, P2, P3 and P4. The level of damage Pi increases with increasing i = 1 to 4. The probability takes values from zero (totally unlikely) to one (certainty).

All the calculation process has initially been done in MATLAB, a programming platform designed to allow expressions of computational mathematics, which for our experimental process means facilitating the use of mathematical complex functions. With this, it has been possible to analyse data, develop algorithms and create models and applications, with the final aim of defining the capacity curve and the exact displacements related to specific damage states, which may be derived through sophisticated analysis of a selective building case-study.

The code from MATLAB (wrote by C. Maniatakis) has been translated to an open source programming language (Python) using the software tool GeoPandas (translation by N. Schetakis). GeoPandas enables the user to easily do operations in Python that would normally require spatial database software. PostGIS is used for the webGIS representation of the final risk map.
By translating the code to Python we further ensure the open source capability.
The code requires as inputs both Structural data and seismic spectral data.

The structural data

The structural data related to the building to be examined are included in a Geopackage file containing information on the geometry, average height (Z) and different typologies of the building, stored in the parameters Type 1, Type 2, Type 3, Type 4, Type 5, Type 6, Use, WT.
GeoPackage is an open, standards-based, platform-independent, portable compact format for transferring geospatial information.
The following tables better explain the meaning of each parameter:

Typo 1 - general classification

Typologies

Building types

Type 1

Unreinforced Masonry

Rubble stone

1

Adobe (earth bricks)

2

Simple stone

3

Massive stone

4

U Masonry (old bricks)

5

U Masonry – R.C. floors

6

Reinforced/confined masonry

Reinforced/confined masonry

7

Reinforced Concrete

Concrete Moment Frame

8

Concrete Shear Walls

9

Dual System

10

Typo 2 - type of horizontal structure (diaphragm/ slab/ vault)

Typologies

Type 2

Unreinforced and/or reinforced Masonry

wood slabs

1

masonry vaults

2

composite steel and masonry slabs

3

reinforced concrete slabs

4

Typo 3 - presence or not of a soft first story in reinforced concrete structure.

Typologies

Type 3

Reinforced Concrete

soft first storey - pilotis

1

regular along height

2

Typo 4 - presence or not of infill walls in reinforced concrete structure.

Typologies

Type 4

Reinforced Concrete

infill walls

1

no infill walls

2

Typo 5 - ductility class.

Ductility class

Type 5

without ductility class

1

low ductility class

2

medium ductility class

3

high ductility class

4

Typo 6 - irregularity along height for every structure related to a reinforced concrete structure.

Typologies

Type 6

Reinforced Concrete

irregular

1

regular

2

Use - main use of the structure.

Typologies

Use

Reinforced Concrete

residential

1

public

2

WT - The average bearing wall thickness for masonry structures.

If the WT field is unavailable, a value of WT = 0.60 is assigned by the code, assuming an average thickness of 60 cm for bearing walls.

The Seismic hazard data

The seismic hazard data are included in a Geopackage file that includes different seismic acceleration spectra as a grid of points within the area of interest. The side of the grid is proposed to be 5.0 m x 5.0 m. Each point corresponds to an acceleration spectrum which resulted from a seismic hazard analysis, considering the seismic sources and the local soil effect (e.g. Tselentis et al. 2010). In case of active faults at a short distance from the buildings, a more detailed approach would include the examination of the particular phenomena that take place in the nearby field of active faults, the so-called near-fault area (Spyrakos et al. 2008; Maniatakis and Spyrakos 2012).

The Geopackage file includes information on the Geometry and the Spectrum type defined for each point and included in separate “Spectrum” TXT files, where the “name” is the corresponding value included in the relevant field of acceleration shapefile referred to geohazard. For example, if four areas with different seismic hazards are defined, the TXT files 1, 2, 3 and 4 should also be provided. The content of the TXT files is formatted into two columns, with the first including period values and the second spectral acceleration values in g units, without any text, as shown below:

0.01

0.3241

0.02

0.3243

4.0

0.0203

The acceleration values should be provided in the range of 0 ~ 4.0 seconds.

The final output

The code may be executed, provided that the above-mentioned input data are available in the same folder with the code and four figures in TIFF format; namely P1, P2, P3 and P4.

Every figure depicts with colours the value of probability Pi for the given seismic scenario. In Figure 2, indicative results of the code are presented for probabilities P1 and P2. The results regard a seismic scenario that assumes peak ground acceleration equal to a gR =0.30 g on rigid bedrock.

The GIS platform

All the data derived will be stored in the STABLE GIS platform. More specifically, and concerning the processing of the Structural Model processor outputs, automated workflows are used to convert the Georeferenced Fragility curves to actual classified Structural Vulnerability features in an open format.

The basic subject of analysis is a polygon feature covering each building of study. For the total sum of buildings, a classification process should be implemented in order to export the final Structural Vulnerability map. Essentially, each building/feature is classified according to its assigned fragility value from the previous Structural Model processor step.

As illustrated in Figure 3, the automated process through the SML operator takes as input building polygons, deriving from the Structural Model analysis, and extracts different classes of these buildings according to their vulnerability level.

The case studies

In the first phase of the project, the STABLE platform will contain within it selected sample cases. The cases are historical centres located in Cyprus, Greece and Italy.
Nafplio and the historic centre of Rethymno (case study recently included in the research) in Greece, Strovolos in Cyprus and Rieti in Italy (Figure 4). The aim is to start from individual buildings (monuments, historic buildings, etc.) and then hypothesise a future extension to the entire fabric that constitutes the historic centre.

To study how the structure of the building reacts in the presence of a seismic event of lesser or greater intensity, it is necessary to start from a careful study of mere basic maps, such as cadastral map, statistical data (age of the buildings, H maximum, number of floors), geological studies, aerial photos and historical maps. This will become part of a database of buildings that will allow you to select, investigate and observe all the information relating to the structure.

Details on the latest case study in Rethymno

The study area of the research is focused on the old town of Rethymno, on the island of Crete. The old town of Rethymno is one of the best-preserved towns of the Renaissance. Lying in the heart of modern Rethymno, it combines the oriental features of the Turkish period with Renaissance-style Venetian architecture. The plan of the old town of Rethymno has remained almost completely unchanged since the XVI century and it definitely needs a plan for the preservation of its architectural and historical evidence.
The main aim of our approach will be to assess the feasibility of Persistent Scattered Method (PSI) in estimating deformations and displacements in the broader area of the Old Town.
For this study, more than 100 S1B images were downloaded from the Copernicus Sentinel 1- Hub, concerning the monitoring period from January 2015 to June 2018. The Sentinel-1 mission comprises a constellation of two polar-orbiting satellites, operating day and night, and performing C-band synthetic aperture radar imaging, enabling them to acquire imagery regardless of the weather.
Afterwards, a combination of open source software such as the Sentinel Application Platform (SNAP), the Stanford Method for Persistent Scatterers (StaMPS), and the Statistical Cost Network Flow Algorithm for Phase Unwrapping (SNAPHU) package were installed in a Linux environment in order to carry out the PSI processing. The PSI requires at least 20 SAR images (Mora et al. 2013) to perform the analysis in C-band and measure surface deformation over a time period of months or years, removing atmospheric, topographical, and signal noise effects (Figure 5). In PSI, only the coherent pixels with stable phase or amplitude are used in the processing pipeline (Mancini et al. 2021). In this concept, the PSI exploits interferograms with a single master scene. In our case, 107 images in total were incorporated into the study (tab. 1):



Table 1. Analysed Tracks

Period (yyyy-mm-dd)

Master Science Acquisition date (yyyy-mm-dd)

Pass

Image Number

2015-01-05 /
2018-06-30

2017-23-02

Ascending

107



At the final step, the interferometric processing of S1 SLC images will generate a geolocated velocity map over the study area. This output will offer a ground deformation map which will provide us with useful knowledge about deformation phenomena occurring in the study area as a further contribution to geohazard analysis.
The methodology presented here will be applied also to the site of the Fortezza, a Venetian fort built in the XVI century on a rocky hill known as Paleochora. It is a site of particular interest since its walls (total length: 1.370 m) are well preserved, but at the same time, they are exposed to extreme climate events which endanger their integrity.

Actual achievements

The project is based on the application of the simplified structural stability model to the whole area of the city centre under analysis. The simulation of different earthquake scenarios, also considering the geological information available, and the response of the structures involved in the scenario will provide detailed damage maps of the area at building level, enabling preparedness for the seismic events. Public authorities, urban planners and cultural heritage responsible will have the capability to address preventive maintenance and consolidation of the structures that will be most damaged in future earthquakes.
This approach will have the twofold objective to drastically reduce the reconstruction and restoration costs as well as reduce the death toll associated with seismic activity.
The project has been running for almost one year and a half, mostly spent on the definition of the overall system, and will last until October 2023.

Acknowledgment

This project has received funding from the European Union’s Horizon 2020 research and innovation programme under the Marie Sklodowska-Curie grant agreement No 823966.

References

  • Del Soldato M., Solari L., Poggi F., Raspini F., Tomás R., Fanti R. and Casagli N. 2019. “Landslide- Induced Damage Probability Estimation Coupling InSAR and Field Survey Data by Fragility Curves.” in Remote Sensing, 11: 1486, doi:10.3390/rs11121486.
  • Fárová, K., Jelének, j., Kopacková-Strnadová, V., and Kycl, P. 2019. “Comparing DInSAR and PSI Techniques Employed to Sentinel-1 Data to Monitor Highway Stability: A Case Study of a Massive Dobkovicky Landslide, Czech Republic”, in Remote Sensing, 11: 2670. doi:10.3390/ rs11222670.
  • Hanssen, Ramon F. 2001. Radar Interferometry: Data Interpretation and Error Analysis, Kluwer Academic, ISBN 9780792369455.
  • Lagomarsino, S., and Giovinazzi, S. 2006. “Macroseismic and mechanical models for the vulnerability and damage assessment of current buildings.” in Bulletin of Earthquake Engineering, 4 (4): 415-443.
  • Kappos, A. J., Panagopoulos, G., and Penelis, G. G. 2008. “Development of a seismic damage and loss scenario for contemporary and historical buildings in Thessaloniki, Greece.” in Soil Dynamics and Earthquake Engineering, 28 (10-11): 836-850.
  • Kottke A.R. and Rathje E.M. 2008. “Technical manual for Strata, PEER Report 2008/10”, in Pacific Earthquake Engineering Research Center College of Engineering, University of California, Berkeley.
  • Mancini, F., Grassi, F. and Cenni, N. 2021. “A Workflow Based on SNAP–StaMPS Open-Source Tools and GNSS Data for PSI-Based Ground Deformation Using Dual-Orbit Sentinel-1 Data: Accuracy Assessment with Error Propagation Analysis”, in Remote Sensing, 13: 753. https:// doi.org/10.3390/rs13040753
  • Maniatakis, C. A., and Spyrakos, C. C. 2012. “A new methodology to determine elastic displacement spectra in the near-fault region.” in Soil Dynamics and Earthquake Engineering, 35: 41-58.
  • Massonnet, D. and Feigl, K. L. 1998. “Radar interferometry and its application to changes in the earth’s surface”, in Rev. Geophys. 36 (4): 441–500, Bibcode: 1998 RvGeo. 36.441M, doi: 10.1029/97RG03139.
  • Mora, O., Mallorqui, J.J. and Broquetas, A. 2003. “Linear and Nonlinear Terrain Deformation Maps from a Reduced Set of Interferometric SAR Images. IEEE Trans. Geosci.” in Remote Sensing, 41: 2243–2253.
  • Nakamura Y. 1989. “A method for dynamic characteristics estimation of subsurface using microtremor on the ground surface.” in Quarterly Report of Railway Technical Research Institute (RTRI), 30 (1): 25-33.
  • Spyrakos, C. C., Maniatakis, C. A., Kiriakopoulos, P., Francioso, A., and Taflampas, I. M. 2015. “Performance of a post-Byzantine triple-domed basilica under near and far fault seismic loads: Analysis and intervention.” in Handbook of Research on Seismic Assessment and Rehabilitation of Historic Structures, IGI Global: 831-867
  • Spyrakos, C. C., Maniatakis, C. A., and Taflambas, J. 2008. “Evaluation of near-source seismic records based on damage potential parameters: Case study: Greece.” in Soil Dynamics and Earthquake Engineering, 28 (9): 738-753.
  • Tselentis, G. A., Danciu, L., and Sokos, E. 2010. “Probabilistic seismic hazard assessment in Greece–Part 2: Acceleration response spectra and elastic input energy spectra.” in Natural Hazards and Earth System Sciences, 10 (1): 41-49.