Heritage buildings reflect the varied communities’ knowledge, beliefs, and customs and provide a sense of continuity and identity to an area by giving it character, meaning, and quality (Historic England 2008).
During the last 20 years, BIM (Building Information Modelling) has been used in the AEC industry and is today widely used internationally (Antonopoulou & Bryan, 2017).
However, in recent years, a new field has emerged within BIM, called HBIM (Historic Building Information Modelling), focusing on the conservation, management and preventive measures of heritage and historical buildings.
According to Historic England (2017), both BIM and HBIM offer a framework for better data management in a multi-disciplinary and collaborative environment.
The world is currently bombarded with data, which is steadily increasing due to rapidly expanding technology (Bilal et al., 2016). As a result, in the last few years, research in using Big Data analytics as a decision-making support system has increased drastically as the amount of data created by individuals continues to increase exponentially.
Thus this research paper aims to exploit the use of IoT and open meteorology data to report and analyse the status of different facade elements in an HBIM or digital twin model for an enhanced decision-making platform.
Development of Framework
The research paper started out by carrying out a bibliometric analysis to analyse the current research in the field. A total of 162 research papers were carefully selected, reviewed and ran through four different analyses. The results of the final analysis showed that most research papers focus on the representation of what exists and seek to develop a workflow from existing to digital twin, rather than focusing on conservation and prevention.
The proposed framework links the geo-localisation of a building (fixed input) and input data (meteorological, manual or sensor) to the four analytic levels to create three different outputs, based on the level of detail needed or available.
The framework derives from the problem statement that heritage building facades tend to surfer and deteriorate a considerable amount as the years pass. Those damages and degradation can, in some cases, amount to a hefty cost in refurbishment, which, unfortunately, not everyone can cover the expense. This leads to the overall deterioration of heritage buildings and hence the loss of the affluent human’s cultural history.
Thus, the intent is to create a visual representation of zones that are prone to damages in the present and future and to act as a report on response needed for refurbishment, conservation and prevention through the help of deep neural networks.
During the first part of the conception of the framework, the research looked at the different types of data both available and required related to heritage building facades. three categories emerged, open data, manual data, and sensor data, of which each category required different equipment and technologies.
Succeeding the categorisation of data, and in order to make sense of those data, each data set must be connected to the four analytical stages: descriptive, diagnostic, predictive and prescriptive. There, the data can be identified and analysed to predict future trends and act as a decision-making support on what to do, based on the results from the four stages.
The results of the analysis can then be expressed in three levels of development depending on the level of detail needed: level one, level two, and level three. The planar segmentation acts as a simple 2d representation, the geometrical segmentation focuses on the overall building area and the survey segmentation focuses on individual building parts.
As the framework is applied to more existing buildings, a deep neural network is then created, where data can then be automatically recognised to allow for predictions and prescriptions of future needs.
Future Research
This research aimed at developing a basic framework using Big Data as a driving tool to enhance the conservation and prevention process in heritage building. In order to fully develop a deep neural network, a further study would need to be conducted with hundreds of case studies. Furthermore, past a digital case study, having a physical case study would allow for sensor data to be collected to act as a live health monitor.
Although some current HBIM literature touch on the basics of factors affecting heritage buildings, conducting research on defining all the factors, such as moisture percentage, or flood, would be useful to grasp the use and power of Big Data fully. Likewise, it would be intriguing to search for the possibility to link the digitalisation of old documents with data analytics to explore and classify historical data.
Lastly, further research could also be conducted on the digital representation of material decay on heritage buildings and how those decay can be linked to sensors and HBIM models.