Szczegóły publikacji
Opis bibliograficzny
Bayesian networks and Support Vector Classifier in damage risk assessment of RC prefabricated building structures in mining areas / Janusz RUSEK, Krzysztof Tajduś, Karol FIREK, Adrian JĘDRZEJCZYK // W: 2020 5th International conference on Smart and sustainable technologies (SpliTech) [Dokument elektroniczny] : September 23-26, 2020, Split, Croatia : [virtual 2020]. — Wersja do Windows. — Dane tekstowe. — [Croatia] : IEEE, 2020. — ISBN: 978-1-7281-7363-4; e-ISBN: 978-953-290-105-4. — S. 1-8. — Wymagania systemowe: Adobe Reader. — Bibliogr. s. 7-8, Abstr. — Publikacja dostępna online od: 2020-11-04
Autorzy (4)
- AGHRusek Janusz
- Tajduś Krzysztof
- AGHFirek Karol
- AGHJędrzejczyk Adrian
Słowa kluczowe
Dane bibliometryczne
| ID BaDAP | 130284 |
|---|---|
| Data dodania do BaDAP | 2020-09-24 |
| Tekst źródłowy | URL |
| DOI | 10.23919/SpliTech49282.2020.9243718 |
| Rok publikacji | 2020 |
| Typ publikacji | materiały konferencyjne (aut.) |
| Otwarty dostęp |  |
| Wydawca | Institute of Electrical and Electronics Engineers (IEEE) |
Abstract
The development of methods for capturing and collecting information with regard to the issues of Structural Health Monitoring and Building Maintenance Management allows for more accurate description of multifactorial processes in the whole life cycle of building structures. This includes the problem of damage risk assessment, which is a complex issue, consisting of a number of factors resulting from the applied structural, technological and material solutions, quality of maintenance or variability of originally assumed operating conditions. Additionally, this problem is getting significantly more complicated when it is necessary to take into account external influences from changing environmental impacts. This is the case with existing building developments located in areas subjected to damaging effects of underground mining in the form of continuous ground deformation and mining tremors. In the long run, the external environmental effects resulting from the decommissioning of the mines should also be taken into account. These are qualitatively similar to continuous deformations of the terrain, but as a rule, they significantly cover much larger area which was not considered to be affected by negative mining impacts during mining operations. From the social and economic points of view, not only the prediction of the extent and intensity of damage, but also the diagnosis of its causes are important. The Bayesian inference framework offers a possibility of creating a universal model for assessing the risk of damage to existing building structures. However, in the case of complex processes, this structure is not known and its search is a basic methodological problem. In the conducted research presented in this article, the Bayesian network structure was extracted using selected score-based learning methods. Currently, the adopted approach is the subject of intensive research, especially in the field of medical science, biology and genetics. Therefore, implementation of such methodology is an innovation in the interdisciplinary field of civil engineering and mining. However, due to the fact that this methodology is still being developed, it needs to be verified. For this purpose, the SVC (Support Vector Classifier) method was used. The effectiveness of the adopted methodology was finally confirmed. This was done by comparing the results obtained for the selected Bayesian network and Support Vector Machine classifier. This formed the basis for formulating conclusions on the applied methodology and on the influence of specific factors on the modelled damage risk process in the group of existing RC prefabricated buildings structures.
