NIST will develop a model to predict evacuation decision-making during fires through a better understanding and quantification of the risk perceived by individuals during an emergency (individual-based risk perception). By FY16, NIST will deliver a framework for this model. By FY19, NIST will deliver a verified and validated model that improves the accuracy of the evacuation models used in performance-based design of buildings.
Objective: By 20161, to develop the framework for a model that will predict the evacuation decision-making of occupants in a building fire. This model will improve the accuracy of egress modeling tools used in performance-based design (PBD) of buildings.
What is the new technical idea? To develop a model to predict when an individual decides to evacuate based upon his/her quantified risk perception level during the pre-evacuation period (Figure 1). Current models primarily ignore the prediction of evacuation decisions/actions, which can result in significant discrepancies between existing egress models and reality. The main focus in FY14-FY16 is to develop a framework for a quantitative Risk Perception Model for use by egress modeling tools. Similar to the FED toxicity model that predicts "time to incapacitation", the Risk Perception Model will account for the major environmental and individual factors that increase or decrease risk perception in order to predict "time to protective action" (i.e., the evacuation decision) in a building fire emergency.
Performance-based methods not only maximize building design flexibility, but also minimize the cost of installed fire protection2. The human- (or RSET) side of performance-based methods (e.g., evacuation calculations) is the focus here. From FY08 to FY13, EL collected data on people movement in stairwells during evacuation drills. The purpose of this effort was to update and validate the already limited and outdated movement data currently used by egress models and improve the accuracy of these models used for performance-based design.
However, the movement period is only one aspect of building evacuation. The pre-evacuation period consists of 1) a pre-alarm phase, 2) a risk perception phase, which ends when an evacuation decision is made, and 3) a protective action phase. Although ignored by most evacuation models, disaster studies have shown that, during pre-evacuation, individuals engage in an information-sorting and decision-making process that ends when personal risk is perceived, and in turn, the individual decides to evacuate. The pre-evacuation period, especially the risk perception phase, can be significantly longer than the movement period in a building evacuation. By simulating the risk perception phase, the Risk Perception Model will improve the accuracy of evacuation calculations used in performance-based design. This project is directly aligned with the Strategic Roadmap for Reducing the Risk of Fire in Buildings and Communities under the "Improved Egress" approach3.
What is the research plan? The foundation for this phase of the project was developed during Phase 1 (FY11-FY13)4. More specifically, our work on the Evacuation Decision Model5 quantified evacuation decision-making in the 2001 WTC disaster, the Egress Estimator6 (the movement model for stairs and elevators), and Kuligowski's7 conceptual model of the pre-evacuation period of the 2001 WTC disaster.
During Phase 2 (FY14-FY16), this project will be separated into three research tasks with the completion of each task being critical to meeting the overall project objective.
Task 1 will develop the Risk Perception Model framework. In FY14, the first step will involve identifying all of the environmental and individual-based factors that influence risk perception. This will be accomplished through a review of previous literature on risk perception (e.g., Kuligowski dissertation), building fire and disaster studies, non-disaster human decision-making models, and other literature on emergency decision-making. The next step is to develop a table of parameters outlining the data available on each risk perception factor (e.g., its influence on risk perception). This step will help to identify which perception factors have no available data. The last step is to prioritize these factors based on findings from actual emergency situations and to identify the "significant" factors to include in the model's framework. In FY15, EL staff will identify a mathematical function for each significant factor (e.g., step function and linear), which has data available. When no data is available, a function will be proposed in FY16 based on a theoretical understanding. Verification and validation will be necessary in Phase 3 of this project. Also in FY16, the FY15 and FY16 developed functions for each factor will be combined (e.g., additive, multiplicative, and mixed) to predict the level of risk perception, which in turn, predicts the point in time when the evacuation decision is made at some identified risk perception threshold.
Task 2 will collect data to support the risk perception framework being developed in Task 1. In FY14 (based upon work which began in FY13 under Phase 1 of this project), data related to human behavior during fires will be collected, via survey and interviews, and analyzed to determine how mobility impaired individuals evacuate buildings during emergencies and the role their personal risk perception played in their decisions/actions. In FY15, data will be gathered via focus groups with emergency responders and people with disabilities on evacuation procedures for elevator usage for people with disabilities (developed from the survey and interview data collected in FY14).
Task 3 will be technology transfer. Realizing impact from the measurement science produced in Phase 1 requires technology transfer to standards, regulations, and codes committees. More specifically, the following sub-tasks will be completed as part of Task 3 in FY14-16: 1) In FY14, based on insights gained on the people movement on stairs during Phase 1, EL will work with NFPA 101 and the International Building Code to develop codes and standards proposals that will increase evacuation efficiency of occupants on stairs in tall buildings, 2) In FY14 and 15, EL will make recommendations on emergency communication and building codes and standards8 based upon qualitative information from survivors, as well as data on injuries and fatalities associated with the Joplin, MO tornado investigation, and 3) In FY 14-16, EL will continue to assist in the development of standards on toxicity and fire threat to the environment (e.g., ISO TC929).
Task 4 will be to further progress on models/projects developed in Phase 1. In FY14, efforts will be made to continue publications on the verification and validation of egress models (in collaboration with Lund University). Also in FY14, EL will complete a Beta test version of the NIST online Egress User's Group (modeled after the FDS User's Group) to promote appropriate use of egress data and development/evaluation of egress models. Last, based on efforts initiated in FY13, EL will develop a future roadmap for egress and human behavior in fire projects at NIST in FY14.
1 The project is in year 1 of phase 2. The long term vision is to enable EL's Risk Perception Model and best practices to be used in models for performance-based design of emergency systems in buildings.
2 PBD has the potential to deliver over $7B in cost efficiencies (ABCB 2000).
3 Reducing the Risk of Fire in Buildings and Communities: A Strategic Roadmap to Guide and Prioritize Research, NIST Special Publication 1130, April 2012.
4 Phase 1 of this project was titled: Safety of Building Occupants.
5 Reneke, P.A., "Evacuation Decision Model." Natl. Inst. Stand. Technol., Internal Report 7914, 2013.
6 Egress Estimator is the only freely available software tool that calculates the time required to evacuate a building by stairs (using SFPE Handbook movement on stairs equations), elevators (using Klote's ELVAC model calculations), or a combination of both. The tool considers occupant loading and the capacity and number of stairs and elevators to be used for evacuation in a building.
7 Kuligowski, E.D., Terror Defeated: Occupant Sensemaking, Decision-making and Protective Action in The 2001 World Trade Center Disaster, PhD Thesis, University of Colorado, 2011.
8 Specific standards and codes will be listed once recommendations are expressed publicaly.
9 TC92: Fire Safety, Subcommittee 3: Fire threat to people and environment