Generative model-based building evacuation simulation for safety design

Stair evacuation plays a crucial role in building evacuation since stairs are generally the only means to evacuate high-rises on fire. To ensure safety stair design, the Life Safety Code suggests using a performance-based design approach, which requires evacuation simulations. Most of existing simulations, however, do not consider the structure of stairs and fatigue of evacuees, and these simulations are not validated by real emergency events or experiments. This paper is on improving the simulation of pedestrian flow in the stairs of high-rises by addressing these issues. A new Cellular Automata simulation model is developed where the simulation map is divided into zones based on the stair structure, and the rule of evacuees’ movement for each zone is appropriately defined to simulate turning behavior. To validate the simulation, a fire drill was held in two high-rise buildings. In this drill, evacuees felt tired after a walk. The simulation results demonstrate that, compared with the simulation without fatigue factor, our simulation can predict the evacuation time more accurately. Building designers can make evacuation plans and strategies based on the new simulation.

Under the new situation, high-speed maglev transportation has gradually become the new development direction of the railway field in the future. Its vehicle operation control system (VOCS for short) as a key subsystem has a great impact on the safety and stability of the entire high-speed maglev train control system. The high-speed maglev VOCS has the characteristics of complex system structure and frequent interaction between its modules. Finding hidden safety hazards and design errors in the system design and development process and ensuring the safety and efficiency of system design and development has become an important task of related projects. On this background, modelling and simulation of typical scenarios of high-speed maglev VOCS are absolutely necessary in the early stage of project development, drawing on the traditional wheel-rail vehicle system simulation verification research results, with the help of advanced computer simulation methods. Using Simulink / Stateflow tools for the parking point stepping functional scenario as an example, the validity and correctness of the high-speed maglev VOCS were verified to ensure the operation safety of the high-speed maglev train through.

Evacuation simulations are among essential analysis to assess the level of life safety in buildings. Currently, a number of evacuation modeling tools are available, each with unique characteristics. But, complexity of human behaviors made most of these tools ignore behavioral data in their models; and among those models which account for human behaviors, rule based modeling and rational decision making approaches are used for decision process. However, to achieve more realistic evacuation calculations, human-like decision making should be integrated to evacuation simulations. In this research we present a novel approach to build multi-agent evacuation simulations that exhibit more realistic simulation behaviors. We propose a model that integrates individual differences, such as personality, age, and gender, with naturalistic decision making, where human decision making in natural environments is investigated. The resulted decision process is applied to BDI agent architecture within a building evacuation system.KeywordsPersonality TraitEvacuation TimingPreference WeightingEmergency EvacuationNaturalistic DecisionThese keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.

Evacuation models represent one of the main tools used to assess building fire safety. Existing evacuation models are based on data from evacuation drills in western and developed countries, while there is a need for data from South Asian and developing countries (in general). This paper presents the results of an evacuation drill carried out in a library of a developing country in South Asia (Pakistan). The data of 85 student evacuees were recorded and analysed. Pre‐evacuation and walking speed data have been reported in this paper which are compared with the available data for library evacuations in other countries. The average pre‐evacuation time of 24.3 s (± standard deviation [SD] = 19.83 s) obtained in this study was shorter than previously published data for similar building types in other parts of the world. Furthermore, unimpeded horizontal speeds (mean = 0.74 m/s, SD = 0.28 m/s) were also smaller than the values reported in the available literature. Stair descending speeds (mean = 0.97 m/s, SD = 0.36 m/s) were within the ranges of stair movement data listed in previous studies. These findings represent a pioneering dataset for future building designs and evacuation simulations in developing countries.

This research studied the fire evacuation simulation and patient assistance who were unable to help themselves out a hospital building. The studied building is a high and large building as Thai law which was 10th floors on 90,842.18 square meters and able to accommodate 4,017 occupants that calculated by the NFPA 101 standard which simulated from 3evacuation cases. The experiment result found that all people can be moved out from the building within 3,048 seconds, 2,931 seconds, and 2,958.3 seconds respectively which didn’t over than the standard as High Building Controlled under Thai Law. The reason that time-limited needed as 1 hour when transferred the people out of the building because enhance the safety of people, especially critically ill patients. Thus, the researchers suggested more channels of evacuation through the elevators by testing for 2 situations. The result of 2situations showed that this method can be decreased the fire evacuation frill time as 2,073 seconds and 1,979.5 seconds respectively.

Modeling and safety strategy of passenger evacuation in a metro station in China

In order to find out the evacuation process of people when fire occurs in a high-rise building, the evacuation process in a ten-storey telecom office building is simulated under different opening conditions by using FDS+Evac software. The temperature field, smoke layer height and toxic gas index are analysed quantificationally. The results prove that about 70% of time is wasted in the stairs because of the stoppage during people evacuation. The exit numbers have little effect on the evacuation time if the number of people is small and they are familiar with the exits. However, adding another stair case in the building is able to reduce 1/3 time of evacuation.

Human behaviour and evacuation time for large underground comprehensive buildings during fire risk process

In the present work, pedestrian behaviour in crowds is assumed to be a coupled phenomena. Therefore, the use of continuousand individual‐models in different regions according to its local density is suggested. (© 2008 WILEY‐VCH Verlag GmbH & Co. KGaA, Weinheim)

Modelling people behaviour during emergencies has become an essential issue in attempting to increase safety aspects in buildings. This paper evaluates people’s choice behaviour for evacuation of tall buildings. A Stated Preference (SP) questionnaire was designed to understand underlying factors behind people behaviour and predict the likelihood of selecting evacuation lifts as opposed to stairs. Various scenarios including six different navigational cases, three levels for the density of people on stairs, three different number of people in the lift lobby and three vertical positions for refuge floors were administrated to 566 participants. A mixed logit model approach was then used to investigate how those factors influence the occupant’s decision-making as well as to capture the heterogeneity of different preferences among people. Traditionally, lifts were not allowed to be used in case of emergency, but the results indicate that people would tend to choose evacuation lifts in situations when they are suggested as the main exit option, and situations when stairs are overcrowded. Thus, if people are navigated by dynamic signs to use evacuation lifts, the percentage of lift users could go approximately from 70% to 80% for refuge floors between 15 and 55, respectively. In contrast, in situations when people have to make a decision between using lifts or stairs to evacuate, stairwells with fewer people as well as overcrowded refuge floors could lead to a decision in favour of stairs. This study represents the first SP experiment combining people decisions, pre-event opinions and beliefs related to evacuation lifts and stairs to understand their route choices for evacuation from tall buildings. The findings of this study can be used in the development of behavioural models for evacuation simulations of tall buildings.

Methodological Approaches for Use Virtual Reality to Develop Emergency Evacuation Simulations for Training, in Emergency Situations

Integrated agent-based simulation and multi-criteria decision making approach for buildings evacuation evaluation

In this study, it was analyzed by a numerical analysis that plan/design considerations for ensuring the spectator safety of large arena audiences in a fire emergency evacuation plan. The latest issue, the 'performance-based design', fire and evacuation plan is important. But nowadays 'Specification-based design' is in common. In evacuation simulation, congestion of exit and aisle is ignored because only evacuation time of large-space is mainly analyzed. In smoke flow,'smoke filling effect' tends to be overrated. From now on, when design a field house, it is needed not 'smoke filling effect' and 'large-space evacuation' analysis, but analyzing 'whole building evacuation time' for ensuring fire evacuation safety of spectator.

With the development of computer processors, vast numerical simulation tools are widely used by fire engineers to determine the spread of fire and smoke. However, the fire modeling practices are often highly time‐consuming and cost‐intensive especially for building geometry information which does limit the further implementation of building fire safety analysis. Although nowadays building information modeling (BIM) has become a buzzword in the Architecture, Engineering, and Construction (AEC) field to facilitate information integration and interoperability, data sharing and exchange are still weak in the traditional interoperability between BIM applications and fire simulation software, since the data schema of them are totally different. In this paper, a quick and accurate approach for information sharing for building fire safety analysis between mainstream BIM applications and widely used fire simulation software has been successfully implemented, based on Industry Foundation Classes (IFC) and Fire Dynamics Simulator (FDS). And, both geometrical building information and sematic information can be shared by this approach through the transformation of coordinate systems, outer database, and IFC file extension. The BIM model restoring fire simulation results can support other performance‐based design of building, such as structural fire‐resistant design and evacuation design. And, all of the analysis results including building fire simulation, structure safety design, and evacuation simulation can be integrated in Autodesk Revit, establishing a framework of IFC‐ and FDS‐based information sharing for building fire safety analysis successfully. A gymnasium has been taken as a case study to illustrate the capability of this framework.

The gold standard for complying Performance Requirements is based on a Quantitative Probabilistic Risk Assessment method. This case study demonstrates the application of this approach to performance based design of a six-storey commercial building with an open stair interconnecting four storeys. Computational Fluid Dynamics based and zone fire as well as evacuation simulations are used to quantify consequences whilst detailed event trees underpinned by statistical data and analysis are utilised to calculate corresponding probabilities. Results are combined in a trade-off analysis tool which calculates the Expected Risk to Life (ERL) based on the trial design features included in each design option. The approach was used to determine a preferred design that achieves an acceptably low ERL and compliance with the Performance Requirements of the Building Code of Australia. The benchmark ERL was set as 1.36 deaths/1000 fires or a probability of death from a fire of 1.36 × 10−3 based on local statistical data. To obtain an optimum fire safety design (Alternative Solution) a layered approach was adopted in which fire safety systems were added until the risk to occupants in the building due to a fire is the same or less than the benchmark ERL. Eventually three sets of trial design were considered and in all cases the calculated ERL were roughly 22% lower than the benchmark. Eventually the trial design with the least number of fire safety systems was recommended as the Alternative Solution. The trade-off analysis shows the sprinklers and wall-wetting sprinklers in the office area resulted in a 20-fold difference in the building wide ERL, each.