Application of building information modeling and damage detection technology in disaster recovery and reconstruction

In the post-disaster recovery and reconstruction phase, building assessment is a very important first step in the process of repairing damaged buildings. In practices, the building assessment still needs building visual inspection and manual analysis which requires a lot of energy and time. Various emerging technologies in the construction sector that can be used to solve problems, for example: Building Information Modeling (BIM), image processing, artificial intelligence. The study aims to review the application of BIM and damage detection technology in postdisaster buildings assessment process. Furthermore, the study focuses more specifically on review of the technology application related to BIM and artificial intelligence for damage detection on crack or concrete spalling in post-disaster recovery and reconstruction. The framework of the automatic integration of damage detection technology and BIM was developed as a way to generate retrofitting designs automatically based on field inspection and building information in post-disaster recovery and reconstruction.


Introduction
According to United Nations Office for Disaster Risk Reduction [1], the number of recorded disasters around the world in 2000-2019 increased sharply doubled compared with the previous twenty years. The reported disasters also caused enormous building damage which resulted in economic losses and social problems. In the post-disaster reconstruction stage, repairing damaged buildings often takes a very long time, so it causes many buildings abandoned for years after the disaster.
The repairing process of damaged buildings is an important part of the recovery and reconstruction phase after a disaster event. In the post-disaster short term recovery, building assessment are a key activity following a disaster [2]. Meanwhile, for the long term, postdisaster reconstruction is a very important stage because it provides long-term development guidance [3]. So, in dealing with buildings affected by disasters, building assessment is the first step that determine the strategy for building repairs. Unfortunately, post-disaster building assessments still often employ conventional methods to carry out visual inspections of buildings that require a lot of effort and time [4]. Moreover, the limited access to resources to disaster affected areas and shortage of qualified people to handle building assessments aggravate the challenges in the post-disaster recovery and reconstruction phase [5].
The enormous potential of the application of technology emerges as a solution to manual process that require great effort and manpower in the building assessment process. Several various emerging technologies in the construction sector have great potential to solve those problems, such as Building Information Modelling (BIM), image processing, artificial intelligence. The data on BIM and automation in visual inspection can be used to support the post-disaster reconstruction process [6]. The automation of these technologies can facilitate reconnaissance teams and civil engineers in determining repair and retrofit strategies of disaster-damaged buildings. Then, it is important to study the application of these technologies in post-disaster building assets in order to identify further developments.
The study aims to review the implementation of BIM and damage detection technology in post-disaster assessment buildings. Furthermore, the study focuses more specifically on review of the technology application related to BIM and image processing for damage detection on crack or concrete spalling in post-disaster recovery and reconstruction stage. This study also develops a framework for the automatic integration of damage detection technology and BIM for future developments of building assessment process in postdisaster reconstruction.

The phase of disaster recovery and reconstruction
The description of the recovery and reconstruction phases in this section focuses on construing the concepts, functions, and stages in relation to building assessments. There are several definitions and observations about how the phase after a disaster occurs. Alexander [8] contended that disasters tend to be a recurring event so as to form a cycle which can be divided into four phases. These phases are mitigation, preparedness, response, and recovery that shown in the Figure 1. Two stages occurred after a disaster is the response and recovery, including reconstruction.
In the figure, recovery describes as a phase after the disaster which is a continuation of the response phase or often referred to as emergency action. Recovery is the process of repairing damage, restoring services and rebuilding facilities after a disaster has struck [8]. Meanwhile, reconstruction is part of recovery that focuses on the process of repair, rebuilding and post-disaster development by building new forms of protection so as not to recreate past vulnerabilities [9]. Furthermore, Quarantelli [10] emphasizes that reconstruction refers to the post-impact rebuilding of physical structures that were destroyed or damaged in a disaster. According to Lindell [11], disaster recovery had a function to identify the physical impact of disasters, short-term recovery, long-term reconstruction, and recovery management to monitor the performance of all three functions. Disaster recovery activities that contain these four functions can be carried out sequentially and simultaneously.
In the process of recovery and reconstruction, the building assessment is the first step to determine the condition of the buildings affected by the disaster. The following step after assessing the building condition is to determine the treatment strategy and planning according to the results of the assessment, then proceed with the reconstruction activities. Brunsdon and Smith [12] developed a conceptual model of the recovery process consisting of five key stages. The five stages consist of impact assessment, restoration proposals, funding arrangements, regulatory processes, and physical construction. The below figure shows the scheme of the initial to final stages in recovery process. It shows that a proper and prompt assessment determines the next steps to the execution of the physical reconstruction. [12].

Fig. 2. Key Stages in Recovery Process
Related to the physical reconstruction process for buildings, damage assessment is the main function that must be carried out immediately so that the building can immediately recover. Furthermore, the recovery of the building function also considers the threat of further disasters so that building retrofitting activities in the reconstruction phase can ensure the strength of the building as the updated conditions.

Building assessment process
Building assessment after disaster carried out in three stages with primary consideration for safety based on the technical aspects of the building structure [2,8,13,14,15]. The first stage aims to quickly identify the level of damage of a building according to visually observed damage. This stage categorizes the building into various potential hazards or safe. The categories can be divided into three or more classifications that describe the strength of the building structure as unsafe, limited entry, or safe.
The secondary stage aims to assess building damage more precisely and quantitatively, and then it can be used to make solution for repairs or retrofit strategy of damaged buildings based on technical and economic aspects. This stage can be divided into two technical solutions, namely general repair based on technical standards according to the type of damage and repair that requires technical in-depth analysis.
For the final stage, there are several versions with different purposes. There are evaluation systems that use the final stage to investigate the engineering in more detail of damaged buildings that require more in-depth analysis or questionable conditions. However, some use this stage to ensure the strengthening of the building considering its long-term use. The building assessment steps are determined through field inspections, process of the damage level evaluation, and the type of reinforcement required. Generally, field inspections are focused on the first two stages to determine the safety of the buildings and to evaluate the damage level of the targeted buildings. Then, the detailed evaluation supports the retrofitting analysis and strategy. For special building condition, the second stage is followed by final stages that do in-depth structural analysis by experts. In the assessment for concrete frame buildings, the degree of damage is related to structural performance. Moehle and Deierlein [16] described building performance based on deformation and structural performance level by FEMA 356. There are three levels of structural performance, i.e.: immediate occupancy, life safety, and collapse prevention. The immediate occupancy level describes the condition of the building experiencing minor cracking and minor spalling of concrete cover on the primary elements. The life safety level describes the condition of the building having extensive damage to beams, joint cracks in primary elements, major cracking and hinge formation in ductile elements. Meanwhile, collapse prevention describes a condition close to collapse with extensive cracking and hinge formation in ductile elements, and extensive concrete spalling on columns and beams.
Referring to Figure 3, the condition prevention of collapse shows that the building is in the unsafe category while the other two conditions can be categorized as safe. Visually, the condition of the building with lateral deflection or partially collapse can be seen from the outdoor inspection. The condition of damage life safety to damage threshold has indoor building damage that requires an indoor inspection to determine the level of building damage. The procedures and forms in the field manual by ATC-20-1 show how to quickly investigate the structural condition of a building. The results of these investigations can determine the level of performance of the building structure which can be followed by structural analysis for more details. Thus, building assessment has clear methods and parameters to determine the condition of buildings after a disaster event.

Application of BIM and damage detection technology for post disaster reconstruction
BIM refers to one or more virtual models of buildings that are built digitally to support the building life cycle in the design, construction, fabrication, operation, maintenance, and modification phases of buildings in the future [17]. BIM provides the capability to conduct a building damage assessment as part of a restoration project in the post-disaster phase [9,18,19]. Sertyesilisik [20] stated that access to pre-event or existing information on buildings for damage assessment and access to initial design requirements can support recovery after a disaster event. On the other hand, the concept of multidimensional BIM, the information contained in the developing BIM has the potential for the creation of data and information management systems are centralized with the system of the other platforms. The below table presents research on the use of BIM in the recovery and reconstruction phases.
The application of damage detection technology in the construction sector has experienced great development using both 2D and 3D image capture methods [26,27]. Up to now, research on damage detection is much more applied to detect damaged buildings in an area using aerial photography than research on damage detection in a building. In fact, detection of damage to a building has many benefits for quickly assessing the condition of the building, especially after a disaster. The research of automatic damage detection application to a building lead to the use of various methods combined with artificial intelligence. The development of research on damage detection techniques is widely applied in the post-disaster phase and integrated with information from BIM. Table 2 shows research topics on the application of automatic damage detection and the integration with building assessment and BIM.  [21] x BIM as an effective tool to facilitate informed decision-making [7] x BIM and Virtual Permitting Framework in post-disaster recovery [22] x Automated approach for structural elements diagnosis based on BIM [23] x Blockchain technology and BIM for smart contract in post-disaster event [24] x BIM-based approach for automating the engineering analyses for postearthquake [25] x Access structural information and design requirements for reconstruction design [18] x x x Table 2. Development of BIM and damage detection technology integration in disaster recovery and reconstruction phase.

Integration of BIM and damage detection
Integrating outdoor building inspection with the BIM [28] BIM and Artificial Intelligence (AI) damage detection; assessment for masonry and concrete buildings [32,33] Compile as-damaged BIM models and a versatile laser scanning emulator [27] BIM and damage detection data exchange and challenges [34] Automated damage detection method Unmanned Aerial Vehicles (UAV)-based building inspection [28] UAV with deep learning on a Generative Adversarial Network (GAN) [35] A wall-climbing Unmanned Aerial System (UAS) [29] Image-based 3D reconstruction using a mobile light detection and ranging (LIDAR) system [31] Terrestrial Laser Scanning (TLS) [27] Automated damage classification An automated approach to recognize concrete crack patterns from images [36] Mask RCNN method for crack detection and segmentation [37] An image-based 3D reconstruction method and 3D crack detection algorithm [26] The researchers found that damage detection technology can be very useful for building assessments, especially if it is supported by BIM. Interestingly, these studies are still not fully integrated in their use in damage assessing a building after a disaster. Building assessment still requires complete integration with automated damage detection with BIMbased. The fully integration on building assessment can be more effective and easier in the disaster recovery and reconstruction phase. 2D and 3D images as input in the field inspection during the building assessment are still not perfectly connected with the as-built model in BIM. The opportunity for process automation is very open to develop BIM applications and direct damage detection systems through images on inspection; so that the building assessment process is sensible to be more efficient and effective.

Development of Post-Disaster Building Assessment Using BIM and Automatic Damage Detection
To expand the development of damage detection in the recovery and reconstruction phases using BIM, a framework was established based on a detailed evaluation of the building assessment process, as shown in Figure 4. The general framework describes a process that integrates BIM, damage detection using images-based analysis, retrofitting procedure, and artificial intelligence supports for damage classification. The integration on the framework has the potential to automatically generate retrofit design and reconstruction costs to facilitate the recovery and reconstruction process. The framework is divided into four parts, there are: building assessment indoor survey, automatic damage detection, automatic repair design process, and automatic QTO-cost for repair/retrofit. The first stage is taking pictures to find out the damage to the structural elements of a building. This image is taken through an indoor inspection with a digital camera which is then collected in a list of building damage. Then, the second stage is to detect the identity of the elements and classify the damage to each element automatically using information from BIM and the building structure damage dataset. This stage begins with detecting damaged parts by comparing the results of the inspection and as-built drawings with a focus on building structural elements. The damage detection results were classified into four categories: crack, spalling, exposed rebar, and other damages. Deep learning can detect damage and predict the type of damage based on datasets from trained damage detection and classification on structural elements of a building. The third part is the automatic repair design to detect the collected damages on inspection stage. The repair design is based on international standard of structure repair procedures and best practices of structure engineer. Structure repair procedures are for structural element damages that can be handled by general repair techniques. Meanwhile, more complex building structural system damages still require in-depth technical analysis which handled by professional structure engineers in engineering evaluation, especially if there are questionable structural damage assessment results. Then, the last part is to calculate the quantity and price for the repair and retrofit strategy. This part generates the quantity take-off and cost calculation for building repair based on bill-of-quantity method and price information at the location of inspected building.
This framework ensures the building assessment and repair process can work within a system. This system can be able to assist engineers in assessing disaster damage more quickly and generate building repair plans in the disaster recovery and reconstruction phase.

Conclusion
Great potential in the post-disaster building assessment process can be further developed through integrating information on BIM and damage detection. Building assessments in the initial recovery and reconstruction phase that focus on assessment based on visual inspection can be developed using various technologies with reference to the procedure of field inspection and structural analysis. Research developments on the integration of BIMbased automatic damage detection can assess buildings more effectively and easily in the disaster recovery and reconstruction process.
The framework of automated retrofitting plan for buildings detailed evaluation based on BIM and damage detection technology was developed as a way to automatically generate retrofitting designs using field inspection data and building information. Integrating BIM, damage detection using images-based analysis, retrofitting procedures, and artificial intelligence support for damage classification can be useful for further developments in post-disaster reconstruction.