|Ph.D Student||Rozenfeld Ophir|
|Subject||Lean Scheduling for Safety:|
Development of a Time Dependent Risk Level Model
|Department||Department of Civil and Environmental Engineering||Supervisors||Professor Yehiel Rosenfeld|
|Professor Rafael Sacks|
|Full Thesis text|
This work describes an idea of lean safety-conscious scheduling and process planning, implemented by a knowledge based model. An innovative model for assessing safety risks in construction was developed, based on the fundamental assumption that risk level is time-dependent. Construction is the leading industry in causing fatal work accidents in almost every country around the world. The budget spent by construction companies on accident prevention is mostly insufficient and inefficiently disbursed. Apart from its potentially disastrous human impacts, inadequate safety disturbs production flow. This research proposes a conceptual risk-level forecasting model and puts it into practice, by developing a practical method for predicting the safety risk level at a construction site during the project and for each individual worker. Having the ability to assess the expected risk level could avoid the reactive costs caused by poor safety performance, and help to distribute the proactive costs in the most efficient way.
The construction industry differs from others in many aspects, among which and above all, are the dynamic nature of the construction site, the large number of subcontractors involved in the process, and the high exposure to environmental conditions. For these reasons, an appropriate method for risk analysis in construction should refer to 4D information (the changing 3D physical environment over time) which is being updated throughout the construction process. The proposed model draws on three sources of information; the first is independent data concerning job safety analysis of construction activities, which was gathered by an extensive survey in Israeli construction sites; the second is the site and building 3D model in which regions and their elevations are detailed; the third is the project schedule from which real time information about locations of workers and risks is derived.
Two important implementations are enhanced by using this model. The first is 'risk leveling', a preventive method for reducing the risk level generated by two different independent activities, of which one poses the danger that the other's workers are exposed to. Rescheduling one of the activities can eliminate the exposure, and therefore, reduce the risk. The second is managing safety prevention in an efficient way by using the risk level prediction to create a priority list of preventive actions.