A context-based representation and reasoning formalism to support construction safety planning

Construction industry has the highest potential for occupational hazard events among all United States??? industries. One of the major causes is construction employers or employees??? negative attitudes toward safety requirements imposed by different construction safety documents, such as construction safety regulations and project safety plans. Such attitudes towards safety requirements include ignorance, negligence and disobedience and the first two can be dealt with by raising construction project participants??? awareness of safety requirements through better construction safety planning. The huge number of safety requirements from different construction safety documents may hinder project participants from carefully searching through them for identifying applicable safety requirements. In addition, current approaches and tools for raising awareness of safety requirements are not sufficient. For example, traditional keyword-based search through a large number of construction safety specifications can help find safety requirements which contain the search keywords. However, there will be other requirements which are semantically relevant but are not found just because they do not contain these specific keywords. Hence, there is a need for a formalized approach to automating the identification of applicable construction safety requirements. This approach should enable identifying safety requirements more efficiently (i.e. saving time in the identification process) and effectively (i.e. better identify both directly and inferentially relevant requirements), and make it feasible to more easily raise project participants??? awareness of safety requirements. To address the above need, a Construction Safety Documents Management Framework is developed in this research. The developed Framework comprises the following components: (1) a computer interpretable model for representing safety requirements of construction safety documents to enable automated reasoning about them; (2) a semantically-rich model for representing concepts acquired from construction safety documents which describe contextual information to which the imposed requirements apply; and (3) reasoning mechanisms to reason about the above two models for evaluating concepts and safety requirements??? applicability to given project contexts. For the representation of construction safety requirements, the Ordered Hierarchy of Content Object (OHCO) approach is adopted to build the representation model; Extensible Markup Language (XML) is used in this research to implement the OHCO-based model. Ontological modeling, on the other hand, is leveraged to model semantically-rich concepts that describe construction contexts. In addition, the developed reasoning mechanisms utilize the ontological relationships between modeled concepts to automatically evaluate each concept???s applicability. Construction safety requirements??? applicability then can be evaluated by reasoning about the requirements??? applicability conditions and exceptions, which are represented using concepts defined in the concept ontologies. Safety requirements can be classified according to their evaluated applicability. The approach is validated for Job Hazard Analysis documents through computational experiments in multiple representation and reasoning test cases. The validation results show that the developed Framework can successfully evaluate the applicability of safety requirements imposed by these JHA documents and identify safety requirements applicable to given contexts. During the validation, the advantages, limitations and practical implications of the developed Framework are discovered. The potential directions of future research on improving the Framework to reinforce its capability and to remove its limitations are lastly presented