Samuel Roeslin | publications

2023

NHESS

Development of a seismic loss prediction model for residential buildings using machine learning – Ōtautahi, Christchurch, New Zealand

Roeslin, S., Ma, Q., Chigullapally, P., Wicker, J., and Wotherspoon, L.

Natural Hazards and Earth System Sciences 2023

HTML

2021

PhD Thesis

Roeslin, S. (2021). Predicting Seismic Damage and Loss for Residential Buildings using Data Science. University of Auckland.

Bib HTML PDF

@thesis{Roeslin2021,
  abbr = {PhD Thesis},
  author = {Roeslin, Samuel},
  city = {Auckland, New Zealand},
  institution = {University of Auckland},
  month = sep,
  title = {Predicting Seismic Damage and Loss for Residential Buildings using Data Science},
  bibtex_show = {true},
  html = {https://hdl.handle.net/2292/57074},
  pdf = {Roeslin_Samuel_Doctoral_Thesis-Compressed.pdf},
  year = {2021},
  selected = {false}
}

2020

BNZSEE

The September 19th, 2017 Puebla, Mexico earthquake

Roeslin, Samuel, Juárez-Garcia, Hugón, Elwood, Kenneth, Dhakal, Rajesh, and Gómez-Bernal, Alonso

Bulletin of the New Zealand Society for Earthquake Engineering Sep 2020

HTML
17WCEE

Feature Engineering for a Seismic Loss Prediction Model Using Machine Learning, Christchurch Experience

Roeslin, Samuel, Ma, Quincy, Chigullapally, Pavan, Wicker, Joerg, and Wotherspoon, Liam

Proceeding of the 17th World Conference on Earthquake Engineering, 17WCEE Sep 2020

Abs PDF

The city of Christchurch, New Zealand experienced four major earthquakes (MW > 5.9) and multiple aftershocks between 4 September 2010 and 23 December 2011. This series of earthquakes, commonly known as the Canterbury Earthquake Sequence (CES), induced over NZ$40 billion in total economic losses. Liquefaction alone led to building damage in 51,000 of the 140,000 residential buildings, with around 15,000 houses left unpractical to repair. Widespread damage to residential buildings highlighted the need for improved seismic prediction tools and to better understand factors influencing damage. Fortunately, due to New Zealand unique insurance setting, up to 80% of the losses were insured. Over the entire CES, insurers received more than 650,000 claims. This research project employs multi-disciplinary empirical data gathered during and prior to the CES to develop a seismic loss prediction model for residential buildings in Christchurch using machine learning. The intent is to develop a procedure for developing insights from post-earthquake data that is subjected to continuous updating, to enable identification of critical parameters affecting losses, and to apply such a model to establish priority building stock for risk mitigation measures. The following paper describes the complex data preparation process required for the application of machine learning techniques. The paper covers the production of a merged dataset with information from the Earthquake Commission (EQC) claim database, building characteristics from RiskScape, seismic demand interpolated from GeoNet strong motion records, liquefaction occurrence from the New Zealand Geotechnical Database (NZGD) and soil conditions from Land Resource Information Systems (LRIS).
ECML PKDD 2019

Data Integration for the Development of a Seismic Loss Prediction Model for Residential Buildings in New Zealand

Roeslin, Samuel, Ma, Quincy, Wicker, Joerg, and Wotherspoon, Liam

Proceeding of the 17th World Conference on Earthquake Engineering, 17WCEE Sep 2020

Abs HTML

In 2010–2011, New Zealand experienced the most damagingearthquakes in its history. It led to extensive damage to Christchurchbuildings, infrastructure and its surroundings; affecting commercial andresidential buildings. The direct economic losses represented 20% of NewZealand’s GDP in 2011. Owing to New Zealand’s particular insurancestructure, the insurance sector contributed to over 80% of losses for atotal of more than NZ31 billion. Amongst this, over NZ11 billion ofthe losses arose from residential building claims and were covered eitherpartially or entirely from the NZ government backed Earthquake Com-mission (EQC) cover insurance scheme. In the process of resolving theclaims, EQC collected detailed financial loss data, post-event observa-tions and building characteristics for each of the approximately 434,000claims lodged following the Canterbury Earthquake sequence (CES).Added to this, the active NZ earthquake engineering community treatedthe event as a large scale outdoor experiment and collected extensive dataon the ground shaking levels, soil conditions, and liquefaction occurrencethroughout wider Christchurch. This paper discusses the necessary datapreparation process preceding the development of a machine learningseismic loss model. The process draws heavily upon using GeographicInformation System (GIS) techniques to aggregate relevant informationfrom multiple databases interpolating data between categories and con-verting data between continuous and categorical forms. Subsequently,the database is processed, and a residential seismic loss prediction modelis developed using machine learning. The aim is to develop a ‘grey-box’model enabling human interpretability of the decision steps.
Eq Spectra

A machine learning damage prediction model for the 2017 Puebla-Morelos, Mexico, earthquake

Roeslin, Samuel, Ma, Quincy, Juárez-Garcia, Hugon, Gómez-Bernal, Alonso, Wicker, Joerg, and Wotherspoon, Liam

Earthquake Spectra Jul 2020

Abs HTML

The 2017 Puebla, Mexico, earthquake event led to significant damage in many build-ings in Mexico City. In the months following the earthquake, civil engineering stu-dents conducted detailed building assessments throughout the city. They collectedbuilding damage information and structural characteristics for 340 buildings in theMexico City urban area, with an emphasis on the Roma and Condesa neighborhoodswhere they assessed 237 buildings. These neighborhoods are of particular interestdue to the availability of seismic records captured by nearby recording stations, andpreexisting information from when the neighborhoods were affected by the 1985Michoaca ́n earthquake. This article presents a case study on developing a damageprediction model using machine learning. It details a framework suitable for workingwith future post-earthquake observation data. Four algorithms able to perform clas-sification tasks were trialed. Random forest, the best performing algorithm, achievesmore than 65% prediction accuracy. The study of the feature importance for the random forest shows that the building location, seismic demand, and building height arethe parameters that influence the model output the most.
JOBE

Displacement-based design of tall earthquake-resistant diagrid systems

Terán-Gilmore, Amador, Roeslin, Samuel, Tapia-Hernández, Edgar, and Cuadros-Hipólito, Edgar

Journal of Building Engineering Nov 2020

Abs HTML

In recent years, the system known as diagrid (diagonal grid) has been used in several tall buildings around the world. Despite the large advantages that the innovative structural system represents in terms of sustainability, its use has not been sufficiently studied for high seismicity zones. This paper develops a displacement-based methodology for the conception and preliminary performance-based seismic design of perimeter diagrid systems. The methodology applies to tall buildings, whose dynamic response is significantly influenced by global flexural drift deformations. The methodology is exemplified by the preliminary design of a 24-story building located in the soft soil condition of Mexico City. The paper provides some detailed background behind the performance-based design and highlights the benefit of simple, reasonable, and safe structural estimations in terms of the monetary and environmental cost. It is concluded first, that the proposed methodology yields diagonal grid systems that adequately satisfy predefined deterministic performance levels; and second, that a perimeter diagonal grid system is an attractive structural option for tall buildings located in high seismicity zones.

2018

Frontiers

Damage Assessment on Buildings Following the 19th September 2017 Puebla, Mexico Earthquake

Roeslin, Samuel, Ma, Quincy T. M., and García, Hugón Juárez

Frontiers in Built Environment Dec 2018

Abs HTML

Understanding seismic risk enables efficient resource allocation in the effort to increase the resilience of our cities and communities. Field reconnaissance and data collection following disasters document the damaging effects of earthquakes to enable lessons and wisdom to accumulate from past events. An important aim of such field data analysis is establishing a better understanding of building performance such as causes of building failures. These lessons provide essential basis to improve our design codes, develop regulations and policies, to increase building resilience by addressing the infrastructure vulnerability. Currently, to make use of the datasets from around the world, significant effort is required to decode the data which often have unique local and regional context and bias. The struggle begins at data collection where there is a lack of consistent methodology and definitions that can adequately cover the regional nuance. This manuscript proposes a new paper-based tool which addresses the need for a global yet detailed universal methodology for building damage assessment post-earthquakes. The new form is based on the GEM taxonomy v2.0 and the European Macroseismic Scale EMS-98. The recent Mexican earthquake from the 19 September 2017 led to significant building damage in the capital Mexico City and the state of Morelos. A team from New Zealand assessed damage throughout the capital and tested the new paper form in Calle La Morena. The street case study presents a novel visualization of the damage data and buildings characteristics which highlights the correlation between the damage and the building features. It is hoped that this kind of illustration will lead to better comprehension of the damage drivers.
NZSEE 2018

The September 19th, 2017 Puebla, Mexico Earthquake – Preliminary report

Roeslin, Samuel, Elwood, Kenneth J., Juarez-Garcia, Hugon, Gomez-Bernal, Alonso, and Dhakal, Rajesh P.

Proceedings of the 2018 New Zealand Society for Earthquake Engineering (NZSEE) Conference Dec 2018

Abs PDF

The magnitude 7.1 Puebla, Mexico earthquake of September 19th, 2017, centred 55 km south of the city of Puebla, led to significant and extensive damage in central Mexico, especially in Mexico City. One month after the event, a team of four researchers from New Zealand travelled to the affected area with the aim of enabling access to observational data that may be relevant to New Zealand’s own disaster mitigation effort. The NZ team was guided by local teams from the Universidad Autónoma Metropolitana-Azcapotzalco (UAM), and they collaborated with the American Concrete Institute (ACI) Disaster Reconnaissance team and the Colegio de Ingenieros Civiles de Mexico (CICM) – Mexican Society of Civil Engineers. The NZ team actively took part in building assessments while conducting passive observations. Preliminary observations have revealed a concentration of building damage in the transition and soft soil zones (i.e. between hard and deep soils). Some areas within Mexico City suffered widespread and extensive damage, Coapa, Jardines de Coyoacan, Culhuacan and Narvarte neighbourhoods. In general, there were 38 collapsed and more than 3000 damaged buildings. Observations reinforce the significance of local site effects, the need of informed, targeted approach in wide-scale building assessments, and the potential role for building instrumentation. The mission provided opportunities to record the performance of various retrofit and repair techniques through examples that were installed following the 1985 Michoacán Earthquake. The team was jointly resourced by the NZ Society for Earthquake Engineering, the University of Auckland, the University of Canterbury, and Universidad Autonoma Metropolitana-Azcapotzalco.
11NCEE

A New Approach for the Seismic Loss Estimation of Existing Buildings

Roeslin, Samuel, and Ma, Quincy T.

Proceedings of the 11th National Conference in Earthquake Engineering, Earthquake Engineering Research Institute (EERI) Dec 2018

Abs HTML

Past seismic events have shown that economic losses following an earthquake disaster could significantly exceed any previous loss expectation. Modern engineering design methodologies define the seismic risk based on life safety criteria but do not explicitly specify the essential criteria leading to economic losses. This research attempts to identify the preponderant parameters affecting financial losses due to building damage following an earthquake event. Existing preevent data collection frameworks are reviewed. A new approach based on the Pareto principle is introduced. Financial and structural survey data from the Christchurch earthquakes are analyzed with the purpose of identifying the principal sources of economic losses in buildings. As a result of the desktop study, a new post-earthquake assessment form based on the GEM building taxonomy is proposed. Remarks following its application in Mexico City following the 2017 Puebla earthquake are presented.

2015

Master Thesis

Roeslin, S. (2015). Use of Diagonal Grids (diagrams) as an Earthquake Resistant System for Tall Buildings Having a Circular Plan. Regensburg University of Applied Sciences.
PDF

2014

Bachelor Thesis

Roeslin, S. (2014). Bewertung der Auswirkungen von Erschütterungereignissen infolge des Betriebs von Geothermieanlagen am Beispiel eines Gebäudes mit Rissschäden. Karlsruhe University of Applied Sciences.
PDF