Horizontal Movement of Kindergarten Children During a Primary School Fire Evacuation Drill

Egress strategies and designing effective evacuation routes is pivotal to building fire safety and comprise a significant part of building codes. With the rise of performance-based design of fire safety applications, analysing occupant behaviour and their movement during fire events has been experiencing a great demand in the recent decade and prompted research on formulating human behaviour in fire science. The focus of the research was to evaluate movement characteristics of occupants along the egress routes, including vertical and horizontal components. As a result, analytical models have been adopted to allow fire engineers to estimate the flow of building occupants [1] and overall evacuation time, which are necessary for evaluation of Required Safe Egress Time (RSET) for performance-based analysis.

The early research on mass motion in buildings outlines the relation of movement with the density of building occupants and their speed, clothing, and human anthropological characteristics [2, 3]. Later works continued by collecting data on occupants’ movement have been based on unannounced, semi announced and announced evacuation drills in various occupancies. As a result, evacuation parameters better formulation and essentials for occupant evacuations such as pre-evacuation decision-making time [4, 5], downward or upward [6‐8] speed of occupant over stairs, movement speed in horizontal sections such as corridors and movement characteristic near exit doors [9] have been studied and helped to develop analytical and computational evacuation models which are summarized in [10].

Buildings encompass various population groups that may not share the same characteristics as populations dominated by adults, as considered in the previous works. The elderly, people with disabilities and children are among the groups with physical and mental abilities different from adults in processing emergency cues and evacuating the buildings [11]. Previous research on elderly people highlighted a different evacuation behaviour affected by floor layouts, instruction and assistance by others, and occupants’ characteristics. Research on disabled people found a movement speed of much less than the ones reported for adults [12], and it may vary depending on the types of assistance provided to disabled and elderly people during evacuation. Familiarity with evacuation plans and moving as a group found to be common behaviour among disabled and elderly people [13]. In another evacuation drill conducted at a university, the authors found differences between female and male disabled individuals in terms of evacuation movement times, and familiarity with the building greatly affected the overall evacuation time [14].

Like elderly and disabled people, children are among the vulnerable groups during fire emergencies, and the research community does not well understand their evacuation abilities and behaviours due to limited data available in the literature. In recent years, attempts to learn the evacuation behaviour of children with different age ranges have been prompted. Hamilton et al. conducted a series of evacuations in schools in Ireland to evaluate the pre-evacuation decision-making time, vertical and horizontal movement of children aged 4–12 years [15, 16]. They highlighted the age dependency and effect of handrails on school children’s speed. The study also found that pupils use the entire width of exit doors, and their maximum flow is higher than those reported for adults, movement speed-age dependency. Evacuation drills in schools with children at pre-school age showed that familiarity with escape routes, children’s age, and environmental conditions, including instructions by staff determine factors in the evacuation behaviours of children [17, 18]. In a series of evacuation drills, the movement of children aged 1–15 years old in schools in Denmark was studied along stairs and exit doors [19‐21] in which a significant discrepancy between available theories for adults’ evacuation movement and children data obtained by evacuation drills was reported. Ono et al. conducted a series of evacuation drills in a Brazilian school with pupils aged 6–14 years and noted the effect of instructions by school staff as an important factor to determine children and adolescent evacuation speed [22‐24]. Similar conclusions were reported by others with evacuation drills in China [25], in Spain [26], and in Russia [27, 28] that led to the same findings.

Although the previous works shed light on the overall movement behaviour of children in the school evacuation, more research should be carried out to address children’s evacuation behaviour. In essence, the behaviour should be analysed under various scenarios, building layout, and egress routes design to compile a database that allows validation of models, developing development simulation methods, and analytical theories specific to children. Moreover, the current data is also geographically limited to certain parts of the world that may not represent children’s populations of different backgrounds, cultures, and demography. Therefore, more data on children’s evacuation must be collected to address these shortcomings in the available data. To fill these gaps, this study (1) expands the emperical database by analysing two semi-announced drills with 135 kindergarten children in Muscat, Oman - an under-represented population and (2) systematically examines how exit-route components (corridor and two door width: 75 cm vs. 105 cm) affect speed, flow, and congestion patterns. Our findings offer child-specific correlations and challenge current egress-width prescriptions (e.g. NFPA 101's minimum clear width of 810 mm), supporting the development of tailored evacuation models and design recommendations for early-education facilities. The above gaps were the motivation of current work to expand available data on children’s evacuation and further scrutinise the evacuation parameters with respect to the change of escape route components layout. Current work evaluates kindergarten children’s movement on horizontal components, including exit doors of various widths through semi-announced evacuation drills in a kindergarten in Muscat, Oman.

The semi-announced evacuation drills were conducted with over 135 kindergarten participants aged 4 to 7 years old. The evacuation times recorded for both EV-1 and EV-2 are consistent with the ranges reported in the literature for school buildings with 2 to 4 storeys. Cuesta et al. reported evacuation times ranging from 70 to 114 s for 70 to 162 children aged 6 to 16 years old [26]. Ono et al. noted evacuation times of 267 to 368 s for a population of 393 to 906 children aged 6 to 14 years old [22]. A study among children in a Finnish 3-storey school found an evacuation time of 228 s for 249 participants [30]. Kholshevnikov et al. reported an evacuation time of 630 s for 220 children aged 7 to 17 [27]. Najmanová et al. recorded evacuation times for children aged 3 to 6 in a kindergarten, with times ranging from 81 to 186 s [31].

The 68-second delay calculated for the arrival of children at the exit door following the fire alarm cue falls within the range reported in previous studies. Najmanová and Ronchi documented a range of 57–164 s for the first participant’s arrival at the exit door in a school building [31]. Lárusdóttir reported a pre-evacuation time ranging from 10 to 545 s for children aged 3–6 years, with an average of 114 s [21]. Similarly, Cuesta et al. observed pre-evacuation times of 4–166 s for children aged 6–16 years based on five evacuation drills conducted in a Spanish school [32]. Although the definition of pre-evacuation time and the delay time noted here are not identical, the wide range highlights that delay times can vary significantly depending on numerous factors, including the type of warning systems, instructions provided by staff, the layout of the building, and the ages of the children involved.

Children’s speed and flow through the corridor revealed a clear discrepancy between the correlations fitted to the current data and those reported by [18]. Similar differences are observed for flows near the exit door in Figs. 6 and 7, where the correlations deviate beyond the confidence bounds of the data scatter and fitted trends. Several factors may explain the evident differences between the two studies. The data in Najmanová and Ronchi’s research are based on a mix of evacuation types, including unannounced, semi-announced, and announced drills. These variations could influence the behaviour of children and staff during evacuation, particularly their movement speeds. Furthermore, in Najmanová and Ronchi’s study [18], the movement of children in the corridors was categorised based on their state, such as walking or running. In contrast, no such categorisation was applied in this study, as most pupils were observed walking. This distinction likely resulted in slower speeds for comparable densities. Moreover, the evacuation drills in this study were conducted without staff assistance, whereas in some evacuations studied by Najmanová and Ronchi, staff provided instructions and support to participants. Such assistance could significantly affect the speed and flow of movement through corridors and doorways.

The width of the exit door significantly influences the distribution of speed and flow both near and through the exit. Wider exit doors allow higher movement speeds and increased flow rates across various densities. Existing empirical formulas, such as those presented by the Society of Fire Protection Engineers Handbook, typically assume a linear relationship between door width, after deducting a boundary width, and flow through the exit door. Although some researchers, such as Gwynne et al. [9], have criticised these models and recommended incorporating additional door parameters, such as design features, into flow calculations, the method remains widely used in fire engineering practice. The relationship between exit width and participant flow is evident in the present study. By comparing the exit width ratio between EV-1 and EV-2, a value of 0.66 was obtained. Similarly, averaging the flow rates at a density of 3.07 persons/m² for EV-1 and 3.09 persons/m² for EV-2, and then calculating the flow ratio yielded a value of 0.60. The ratio of these two values yields approximately 1, confirming a strong correlation between exit width and flow for the children participating in these drills.

From a practical perspective, the findings suggest that minimum exit door widths prescribed by codes may not always ensure optimal evacuation efficiency for kindergarten and primary school settings. For instance, while NFPA 101 [33] recommends a minimum clear door width of 810 mm, the results of this study indicate that narrower exits could lead to congestion, particularly in scenarios involving young children. In the UK, Approved Document B (ADB) sets a minimum width of 750 mm for a maximum number of 60 people and increases to 850 mm for a maximum number of people of 110 or an alternative of 5 mm per person [34]. These observations highlight the importance of considering occupancy-specific evacuation needs when setting regulatory standards. The exit door width attempted in these drills corresponds to the minimums mandated by fire safety codes in different countries. Considering the evacuation times in this study, the evacuation times with minimum width noted by the codes may surpass the ideal evacuation times of less than 3 min, at least for the context of kindergarten and primary schools. This shows the minimum width may not be applicable for different occupancies, and it may yield different performances, with the potential to cause complete congestion at the exit doors.

When using the current results for comparison or modelling purposes, several limitations of the study should be considered. The evacuation drills may not be fully comparable to actual events involving fire emergencies in buildings. Therefore, caution is advised when comparing the presented results with real-life incidents. Additionally, the uncertainties associated with the repeatability of evacuation drills remain unknown, as only a single evacuation was conducted in this study. However, conducting evacuation drills in a single school for data collection is not uncommon, and similar approaches have been used in previous studies [25, 26], which are generally considered to provide an acceptable level of accuracy within the research community. Factors such as staff engagement, the frequency of drills in a building, and the methods of warning can introduce variability and should be accounted for when comparing results across studies. This study made every effort to ensure that children and staff were unaware of the evacuation drills beforehand. However, the possibility that some participants may have been alerted by pre-evacuation activities or management cannot be ruled out. This could have influenced their perception of the emergency cues and movement speeds. It was assumed that all children participating in this study were healthy and physically capable of movement over vertical and horizontal egress components. Nonetheless, it is possible that some participants may have had undisclosed disabilities or conditions that could have affected their movement, which was not revealed to the research team.

Children represent a vulnerable population during fire evacuations, with behaviours and movement patterns that differ significantly from those of adults. This study investigated the evacuation behaviours of kindergarten children (ages 4–7) through semi-announced drills in a primary school. By analysing movement speed, flow, and behaviour in horizontal components, including corridors and exit doors of varying widths, the study highlighted several unique characteristics of young children during evacuations.

Key findings reveal that children do not adhere to the privacy zones typically maintained by adults and exhibit strong group dynamics, such as waiting for friends or moving in clusters. These behaviours often lead to congestion and slower evacuation times, particularly in high-density scenarios or at narrow exit doors. The study also underscores the critical influence of exit door width on flow rates and evacuation efficiency, aligning with prior research while emphasizing the need to tailor evacuation models to account for behavioural factors specific to children.

The results have practical implications for fire safety engineering and building design. Regulatory minimums for exit widths, such as those prescribed by NFPA 101, may not always suffice for optimal evacuation performance in schools, particularly for younger children. Design standards should consider both physical and behavioural factors, ensuring egress routes accommodate the unique needs of child occupants.

This study contributes valuable data to the limited body of knowledge on children’s evacuation behaviours but also highlights the need for further research. Future studies should investigate a wider range of scenarios, including unannounced drills, diverse building layouts, and different cultural contexts. Incorporating staff-assisted evacuations and exploring the impact of environmental and psychological factors on children’s behaviour can provide deeper insights to refine evacuation models and improve fire safety strategies.