Risk Assessment of Pedestrian Evacuation under the Influence of Fire Products

Traditional methods are using FDS and Pathfinder for numerical simulation of fire evacuation and do not consider the impact of fire products on pedestrians, which will lead to erroneous evacuation results. In order to explore the impact of fire products on pedestrian evacuation when a fire occurs in a building, the pedestrian evacuation risk assessment results are closer to the actual situation. By establishing a full-scale model based on the pedestrian evacuation speed coefficient, a numerical simulation of the evacuation of pedestrians under the influence of fire products is carried out, and the influence of spray intensity and smoke exhaust rate on pedestrian evacuation is analysed at the same time. The results show that, compared with traditional methods, the fire evacuation model based on the evacuation speed coefficient can better reflect the evacuation effect of the real fire scene. At the same time, adding spray and smoke exhaust devices can give pedestrians more time to escape.


Introduction
With the development of computer technology and the indepth study of human behavior and psychology, researchers gradually began to widely use computers to directly simulate the movement state of people in buildings and established a large number of evacuation models. e evacuation efficiency of a fire is affected by many factors, not only to consider the differences between individuals in the evacuation process, such as cognitive level [1], physiological conditions, mobility, psychological pressure, and other factors [2] impact on the speed of pedestrian evacuation. In addition, the structural details of the building such as evacuation exit conditions [3,4], emergency lighting system [5], escalators and stairs [6,7], etc. will also affect the evacuation of personnel to varying degrees. In addition, there are many related in-depth and meticulous researches on different groups and locations of fires. rough the use of computer simulations, a variety of models have been established to study the laws and characteristics of personnel evacuation behavior, making this research direction continue to mature. research on the evacuation of personnel during the fire started earlier in foreign countries. As early as 1909, the United States published a research report on personnel evacuation behavior [8]. In the early 1930s, the National Bureau of Standards conducted evacuation simulation studies in normal environments and fire drill environments. Japanese scientist Togawa, British scientist Melink, Booth, etc., through video data and related fire records, based on a large number of observations, summed up the empirical calculation formula for time in the evacuation process [9]. " e behavior of people in Fires" written by the British scientist Wood through a large number of fire case investigations is regarded as a representative work in the field of human fire behavior [10]. In the theoretical study of personnel evacuation, Henderson first proposed the macro model of personnel evacuation [11]. In this article, the movement behavior of pedestrians is similar to the flow of gas or fluid. German traffic flow expert Helbing and his colleagues based on Lewin's concept of social power and improved it on the basis of Henderson's fluid dynamic equations and jointly constructed a social power pedestrian traffic flow model in 1995. Subjective expectations, the interrelationships between people, the interaction between people and the environment, the psychological reactions of personnel, and other forces are quantified, which is described by the concept of social force. In 1986, U. Frisch, B. Hasslacher, and Y. Pomeau collaborated and proposed a lattice gas model. e state variables such as space, time, and speed and position of the model are discrete, so the speed of calculation is increased. Von Neumann et al. proposed the cellular automaton in 1948. Since the time, space, and state in the model are discrete variables, this highly discrete calculation can perform high-speed simulation operations for high-performance computers. Moreover, it can easily reproduce different complex phenomena in fire or self-organization, mutual attraction, and chaos in the process of dynamic evolution of fire and has strong ability to simulate various physical systems and natural phenomena.
In recent years, with the continuous occurrence of fires, domestic research on the evacuation simulation model of people in the process of fire has been increasingly emphasized, and domestic units at all levels have begun to gradually carry out research on all aspects of fires. Among them, Northeastern University Chen and others have comprehensively studied the problems of personnel evacuation in fires, including the general evacuation behavior characteristics and disaster avoidance strategies of people during fires [12][13][14] and established personnel emergency in major accidents. e evacuation model [15], the simulation model of human evacuation behavior after a fire in a large public building [16], and the simulation model of personnel evacuation based on virtual reality technology [17]. Workers such as Founder and Lu of the City University of Hong Kong established a space grid evacuation model [18,19]. Wuhan University Founder and others have developed the evacuation simulation software for high-rise buildings [20]. After establishing the State Key Laboratory of Fire Science, the University of Science and Technology of China gradually began to organize related research: Song et al. quantified and optimized friction and repulsion by combining existing social force models based on the classic cellular automaton model e calculation rules of force improve the efficiency of calculation by optimizing continuous model and discrete model and improve the quantitative description of repulsive force and friction force, which greatly improves the calculation speed and saves its accuracy and improves the cell Automaton model [21].
On the basis of the existing cellular automaton traffic flow model and pedestrian flow model, Yang proposed a basic model for simulating the escape of people in a fire [22]. By introducing the concept of "total hazard map," Good simulation of the intelligence of personnel and the role of fire scenes makes the model more reasonable and convenient in determining the escape route of personnel. In addition, he also developed a microscopic discrete evacuation model of personnel behavior based on cellular automata theory [23], so that the individual characteristics of each evacuee can be simulated. Zheng et al. simulated the dynamics of pedestrian evacuation based on cellular automata-based underground flood diffusion [24]. Choi and Chi studied the optimal route selection model for fire evacuation based on hazard prediction data [25].
Sun and Wang conducted relevant research on the simulation of the evacuation preparation time and exit width on the evacuation of people [26]. In addition, they and other scientists conducted a major fire in urban high-rise buildings under the leadership of academician Fan Weicheng, the project expert team leader and research on key basic issues of prevention and control. Secondly, there are also many related researches on individual differences and related phenomena in the process of fire, such as psychological factors [27], cognitive level [1], group phenomena [28], the effect of indicators on personnel evacuation during the fire [29] and research on other aspects. At present, the main scenes of the main personnel evacuation research are: research on the school gymnasium [30], teaching building [31,32], etc. In addition, there are also many researches on the impact factors of the fire process on the fire environment on people, including smoke regulation [33], smoke toxicity [34], and many other aspects.
According to relevant statistics of the American Fire Protection Association, deaths in fires are mainly caused by poisoning and suffocation, with a general proportion of 80% from poisonous smoke, 13% from high temperature, and 7% from other causes [35]. erefore, when assessing the risk of building fires, it is very necessary to consider the evacuation of pedestrians under the influence of fire products. Many scholars have conducted research on this. Sun et al. [36] conducted numerical simulation studies on the spread and development of fire and the evacuation of pedestrians for teaching buildings; Song et al. [37] conducted fire control operations and pedestrians safety in the terminal building; Quantitative simulations were made; Chen [38] conducted a fire dynamic evolution process and numerical simulation of pedestrians evacuation of the subway station and gave indicators for judging the safe evacuation of pedestrians; Liu et al. [39] carried out pedestrians under fire in high-rise buildings Evacuation research; Li and Zhang [40] conducted a fire simulation and safety evacuation study in a connected dormitory building using the fire simulation software FDS and the pedestrians evacuation software Pathfinder.
Most of the previous researches used the fire dynamic simulation software FDS and the personnel evacuation software Pathfinder together, but did not connect the fire products with the evacuation. When conducting the evacuation research under fire conditions, the fire products did not have any impact on the pedestrians in the evacuation, which would lead to false evacuation conclusions. In order to reduce the error caused by traditional method, make the evacuation result closer to the real scene. erefore, the author intends to use FDS and Pathfinder to conduct simulation research on evacuation of people under the influence of fire products and provide certain guidance for pedestrian evacuation risk assessment of building fires based on the simulation results.

Theoretical Framework
As shown in the figures, Figure 1 is the traditional method. By setting three types of detectors at the exit of the fire dynamic simulation software FDS, the shortest time for the three types of sensors to reach the specified critical hazard value is regarded as Available Safe Egress Time (ASET), in the pedestrian evacuation software Pathfinder; V c represents the pedestrians to evacuate at a fixed walking speed. When the pedestrians evacuate to the exit, the evacuation time of the pedestrian is regarded as the Required Safe Egress Time (RSET). Although comparing ASET and RSET to assess whether the pedestrian is safe to escape, this method does not consider the impact of fire products on pedestrian. Pedestrians always escape at a fixed speed V c during the evacuation process, which is inconsistent with the actual situation. Figure 2 shows the proposed method in this paper. As shown in the figure, different from the previous model establishment, in this paper, BIM software Revit is used to establish the research model, which can be directly imported into FDS and Pathfinder, and the authenticity and restoration degree of the model is much higher than previous modeling methods. In FDS, CO volume fraction, CO 2 volume fraction, O 2 volume fraction, and temperature are used as ASET, and the ASET is refined. e refined ASETcan get the injury status of the person. Use visibility, CO volume fraction, and temperature as indicators that affect the evacuation of pedestrians. By converting the three into Evacuation Speed Coefficient and importing the Speed Modifier that comes with Pathfinder, the pedestrian evacuation speed is V t at this time, which means that during the evacuation process, pedestrians' speed will change with the change of time. It can be considered that the fire products have an impact on the evacuation of people. By getting RSET in Pathfinder and comparing it with the refined ASET, we can get the four states of pedestrians (No injury, Minor injured, Seriously injured, and Dead).

Composition and Influence of Fire Products.
e fire products are mainly divided into fire smoke and hightemperature hot plume. ese products will reduce the physiological movement ability of pedestrians, affect the choice of escape routes, and even affect the nervous system of pedestrians, causing pedestrians to make wrong judgments during the evacuation process. e hazards of fire products are as follows: (i) e shading of smoke: Combustible materials will produce a large number of solid particles and other gas-phase products and associated droplets during the combustion decomposition process, which are collectively referred to as flue gas. e accumulation of smoke will result in reduced visibility, which will affect pedestrians' field of vision and negatively affect the evacuation of pedestrians. (ii) e toxicity of smoke: Once pedestrians in the fire inhale toxic and harmful fumes, some mechanisms and functions of body organs will be affected. If inhaled for a long time, they may even cause serious injury or death. e most significant effects of fire smoke on the human body are CO and CO 2 . e former weakens the binding of hemoglobin to O 2 , resulting in insufficient O 2 content in the blood and insufficient support; the latter affects the respiratory rate and depth of the human body, causing pedestrians to inhale toxic and harmful gases faster. In the fire scene, the rapid consumption of oxygen during the fire will also cause suffocation and other situations, which will hinder the escape. e effects of different volume fractions of CO, CO 2 , and O 2 on the human body are shown in Tables 1-3. (iii) High-temperature hot plume. High-temperature thermal plumes mainly affect the circulatory system and nervous system of the human body. When the human body is exposed to high temperature for a long time, it will cause the skin blood vessels to dilate, and a lot of sweating will condense the blood, resulting in a rapid heartbeat, increased blood pressure, abnormal co-ordination and accuracy of human movements, resulting in reduced brain reaction speed and attention. Under normal circumstances, the tolerance of pedestrians to different temperatures is shown in Table 4.

Fire Product Impact Function.
Zhu [41] believes that the speed of pedestrian evacuation in a fire is mainly affected by visibility, CO, and flue gas temperature, and can be divided into the following three aspects: (i) Visibility: e visibility of the fire, which is mainly caused by the smoke volume fraction, affects the field of vision of the pedestrians during the evacuation process, thereby affecting the escape speed of the pedestrians. (ii) CO: e impact on the body organs of pedestrians is mainly reflected in the toxicity of anesthesia and stimulation of toxicity. Long-term inhalation may cause serious injury or death. Discrete Dynamics in Nature and Society (iii) Temperature: Affect the physical function of pedestrians, make pedestrians feel hot, dizzy and dizzy, and produce nervousness.
Zhu [41] introduced the concept of equivalent speed coefficient, quantified the impact of the 3 products on the human body, and introduced visibility coefficient f(v), CO influence

CO 2 volume fraction (%)
Human reaction symptoms 0.55 e human body will not produce any symptoms within 6 h 1-2 Cause discomfort 5 Feeling unable to gasp, causing poisoning within 30 minutes 7-10 Lose consciousness within minutes to death Shortness of breath, headache, dizziness, slow movements 10-12 Nausea and vomiting, inability to move or even paralysis Less than 6 Within 6-8 min, pedestrians will die Discrete Dynamics in Nature and Society coefficient f(c), and temperature influence coefficient f(θ). e product of these 3 coefficients represents the total effect of fire products on the rate of human evacuation [41].
(1)  [42] summarized data on the speed of pedestrians walking in smoke based on 10 different empirical formulas and provided a representation of walking. e relationship between the speed and visibility functions, the visibility impact coefficient and visibility are as follows: (2) Theorem 2. VIS represents the visibility of pedestrians during evacuation, m. Figure 3 shows the influence of visibility on the coefficient of evacuation. As shown in Figure 3, the curve varies linearly between 0.6 m < VIS < 3 m. When VIS > 3 m or VIS < 0.6 m, the visibility coefficient is invariable, always 1 or 0.2. [41] fitted the formula of the effect of CO on human body speed measured by different volume fractions of Milke [43] as a formula. And the function of CO volume fraction, the expression is as follows:

Theorem 3. p is the volume fraction of CO, %, t is the exposure time, min.
Zhu [41] thinks that when the co volume fraction of smoke is less than 0.1%, the evacuation speed to the human body can be neglected. When the volume fraction of the smoke reaches 0.25%, it can cause serious harm to the pedestrians, resulting in the evacuation speed down to 0 at this time, and the pedestrians lose their power of action.

Temperature Influence Function.
Zhu [41] believes that the effects of high-temperature plumes on human body can be divided into 3 stages: (i) When the temperature θ < 30°C is perceived by the human body, it can be considered that the temperature has no effect on the evacuation speed of the person, at this time f(θ) is always equal to 1. (ii) When the temperature sensed by the human body is 30°C < θ < 60°C, during this time period, the human body senses the temperature rise, stimulating the pedestrians to escape from the fire quickly; at this time, the temperature has a positive effect on the speed of evacuation of the pedestrians. (iii) When the temperature sensed by the human body is θ > 120°C, the evacuation speed of the pedestrians is gradually reduced by the temperature until the pedestrians lose their ability to move. At this time, the temperature has a negative effect on the evacuation of the pedestrians.
e expression is as follows:  Figure 4 represents the influence of temperature on the evacuation coefficient of pedestrians. It can be seen from the Figure 4 that when the temperature θ < 30°C, the temperature influence coefficient f(θ) is always 1; when the temperature θ � 60°C, the temperature influence coefficient f(θ) max � 3.333; when the temperature θ > 120°C, the temperature influence coefficient f(θ) is always 0.

Consider the Numerical Simulation of Evacuation under
Fire Products. FDS is a simulation software specially designed to create fire scenarios and view the evolution process of fire dynamics. It allows users to view the results of smoke spreading and fire spreading during the process of fire. Using the software, a building fire model can be obtained, as shown in Figure 5. e model is a full-scale fire model on the first floor of a building, with a total length of 84.0302 m, a total length of 20 m, and a height of 4 m. e fire source is located in a room. Figure 6 shows a schematic diagram of 23 regions in the fire model. 23 sensors divide the corridor into 23 zones. Each region has temperature sensors, CO sensors, visibility sensors, and sprinkler sensors, so as to output the result of fire products in each region. At the same time, the CO sensor, CO 2 sensor, temperature sensor, and O 2 sensor are 4 kinds of sensors, which are mainly used for the calculation of safe evacuation time. e height is 1.7 m., and the specific fire parameters are shown in Table 5.
Pathfinder is the evacuation software based on evacuation dynamics. By setting the corresponding parameters (walking speed, shoulder width, outlet selection, etc.) for each escaping pedestrians in the group, and setting up the escape mode of pedestrians, virtual evacuation of pedestrian movement is simulated. In the Pathfinder 2019, a function of regional pedestrian speed modifier is added, as shown in Figure 7, where the initial value is 1. e initial evacuation speed of a representative, that is, the speed coefficient, corresponds to different values at different times. As time goes on, the fire products begin to affect the pedestrians, and the values continue to decrease; that is, the speed coefficient decreases, and the speed of the pedestrians starts to slow down until the speed is changed to 0. It is used to establish evacuation model under the condition of fire product impact. e fire model in the FDS can be directly imported into the Pathfinder. According to the location of the 23 detection regions in the FDS, 23 evacuation regions are manually divided into Pathfinder. Due to the fact that there is a certain delay in the fire control in the real fire scenario, in order to be closer to the actual evacuation situation, there are 5 kinds of response behaviors for different rooms. As shown in Figure 8, the response delay time of different behavior pedestrians is shown in Table 6.
ere are 6 kinds of working conditions for evacuation. ere are different evacuations in different conditions, and each condition is divided into 2 scenarios, namely, Scenario 1, without considering the effects of fire products; Scenario 2, taking into account the influence of fire products on the parameters of working conditions. e parameter settings of the 6 working conditions are shown in Table 7.
Assuming existence 0.1 ≤ p < 0.25, when t � t 0 . At that time p � p 0 ; when t � t i+j . At that time p � p i+j . e relationship between time and volume fraction is shown in Table 8. Suppose the initial time is t i . When t � t i , the person has been exposed to the volume fraction p i for 0. For the three kinds of fire product influence coefficients, the visibility influence coefficient and the temperature influence coefficient can be directly imported into Pathfinder, and for the CO influence coefficient, there are two variables, volume fraction and duration, which cannot be directly imported into Pathfinder, and the eorem 4 needs to be modified.
Get the following formula: In this fire simulation calculation process, FDS records data about every 0.5 s, the simulation time is 500 s, and a total of 1001 data points; it can be considered that at the moment, the person has been exposed to the volume fraction for 0.5 s, then the formula (5) change into When p < 0.1 or p > 0.25, there are only two states 1 or 0. At this time, p has no relationship with the exposure time t. It can be considered that the conversion method is applicable to the entire function expression. Figure 9 shows a comparison of speed coefficient curves between region 1 and region 22. As shown in Figure 9, the temperature has a positive effect on evacuation in the early stage of the fire, which dilutes the negative effects of visibility and CO on the human body. As the fire proceeds, CO reaches the dangerous point first, and the co speed coefficient returns to zero, leading to the total speed coefficient returning to zero.     Table 9, we can see that the evacuation time under the condition of considering the impact of fire products is significantly larger than the evacuation time without considering the impact of fire products. In the 2 scenarios, the maximum relative error is 60.69% when the evacuation is completed, and the smallest relative error is 39.59%. For the     Discrete Dynamics in Nature and Society safe evacuation of pedestrians depends on 2 characteristic time: one is the time needed for the fire to develop to the safety threat to the pedestrians, that is, Available Safe Egress Time (ASET), and the second is the time needed to evacuate to the safe region, that is, Required Safe Egress Time, RSET, so as to ensure that the pedestrians can evacuate safely. We must ask RSET < ASET because of the great difference between the evacuation results under the 2 scenarios. e author will re-quantify the target ASET of safe evacuation. For ASET's judgment, CO 2 , O 2 , CO, and temperature are used to judge whether the pedestrians are safe evacuation targets. ese indicators can have a certain degree of influence on human health, and for 3 kinds of ASET, combined with the influence degree of fire smoke and temperature on human body, the index of ASET is quantified in Tables 10 and 11. Take the time when the 4 indicators reach the prescribed value, respectively, and take the shortest time as the corresponding available safe evacuation time ASET. According to the 4 sensors set at the doorway in the model, 4 indexes meet the prescribed value corresponding to the time Table 12. erefore, it can be considered that ASET1 � 129.53 s, ASET2 � 261.52 s, and ASET3 are not found in the study. erefore, it is not worth considering ASET3. It is worth noting that the risk index adopted by the speed coefficient is the same as that adopted by the ASET, but this may lead to different results. It may appear that when the ASET2 arrives, the pedestrians still have the power of action. e use of a single measuring point instead of the data of a single region may cause some errors for larger regions, but this is acceptable, and the optimization of errors is not considered in this paper. e evacuation time under 6 working conditions is compared with ASET1 and ASET2, and the degree of injury in 2 scenarios is studied, taking working conditions 1 (200 pedestrians) and working conditions 6 (700 pedestrians) as an example. Figure 10(a) shows the evacuation of 700 pedestrians in 2 scenarios. It can be seen that when we get to ASET1 � 129.53 s, there are 379 pedestrians left in Scenario 1 and 550 pedestrians in Scenario 2. At that time, the remaining number is in the condition of slight injury. As the evacuation proceeds, the pedestrians in Scenario 1 have been evacuated before ASET2, and the pedestrians have not In Scenario 2, when the number of evacuees was 700, 656 pedestrians completed evacuation at 301 s, and 44 pedestrians did not escape because of the regional speed coefficient of 0. e human body bears no more than 1 minutes, or the human body is particularly harmful, and immediately dies in a short time   rough the analysis of the above 2 results, the evacuation time and the degree of injury are quite different in the 2 scenarios. In Case 6, for example, when the evacuation number is 700, the pedestrians in Scenario 1 begin to evacuate completely without serious injury or loss of action. Even 44 of them stayed in the fire because of their loss of action. is is truer to the actual situation, because in the actual fire scenario, when the number of evacuees is large, it is easier for pedestrians to become unconscious due to heavy injuries. erefore, it can be considered that the evacuation results under fire conditions are closer to the real situation than the traditional simulation results. A fire scenario can only correspond to an evacuation model, and the fire products have a direct impact on the evacuation of pedestrians.

Influence of Spray Intensity on Speed Coefficient.
Zhao and Qian [44] research shows that different spray intensities can inhibit the temperature change in the fire field; Khoat et al. [45] analysed the fire characteristics after spray fire extinguishment in the cabin fire. During the fire extinguishing process, it is believed that the presence of spray has a significant effect on reducing the temperature of the smoke layer. Yang et al. [46] think that the spray intensity can prevent the visibility of the no fire source region from drastically decreasing, Wang et al.'s [47] numerical simulation showed that the CO volume fraction of the spray coverage region is higher than that of the no spray condition; Sun [48]used the fire simulation software FDS and fire test and found that within a certain range, the greater the spray intensity, the smaller the CO volume fraction in the chamber. It is found that within a certain range, the larger the spray intensity, the smaller the CO in the chamber. All the above studies prove that the spray intensity has an impact on the 3 kinds of analytical products. In order to explore the influence of spray intensity on the speed coefficient, 3 working conditions are set up in this paper. On the basis of the original model, a spray sensor is added at each 3.5 m height in the region, and a total of 23 sprinkler sensors are set up. e specific spray intensity settings are shown in Table 13.
A total of 4 representative regions were selected, of which 15 were located outside the fire room, adjacent to the fire region, 23, 8, and 1 gradually away from the fire region. e effect of spray intensity on the speed coefficient is shown in Figure 11. From Figure 11, it can be seen that the spray intensity has little effect on the speed coefficient, but with the increase of spray intensity, the time of curve zero return is postponed. at is, the time lag of pedestrian speed coefficient is 0.    For region 15, because the region is close to the fire region, in the middle and late fires, the greater the spray intensity, the greater the speed coefficient, while for region 1 and region 8, in the middle of the fire, the greater the spray intensity, the speed coefficient. On the contrary, the speed coefficient is smaller. is is because, in this period of time, the most important indicator of spray influence is temperature. e higher the spraying intensity, the higher the temperature in the region, but the temperature always has a positive effect on the human body, stimulating pedestrians to accelerate away from the region.

Influence of Smoke Rate on Speed Coefficient.
Wang et al. [49] research shows that mechanical exhaust can effectively reduce the highest temperature and the highest volume fraction of CO in the fire field and delay the beginning time of the smoke visibility. Zhu and Li [50] carried out the small size fire smoke simulation test and found that the exhaust efficiency has a certain effect on the temperature of the hot smoke layer. In order to explore the influence of smoke rate on the speed coefficient, 4 kinds of working conditions are set up in this paper. Due to the large number of detection regions, 6 regions were taken as comparison objects, and smoke exhaust ports were set up in the height of 22 regions, including region 22, region 18, region 14, region 10, region 16, and region 2. e length and width of smoke outlets were all 0.5 m. e smoke emission rate corresponding to different conditions is Table 14. As can be seen from Figure 12, smoke exhaust has a significant effect on the speed coefficient. In the middle of the fire, the speed coefficient decreases with the increase of exhaust gas rate. is is because the temperature has a positive effect on the human body at that time. e higher the smoke exhaust rate and the lower the temperature, the smaller the stimulating effect on the human body. e time lag of speed coefficient returning to zero is delayed, and the effect of smoke exhaust on speed coefficient is much better than that of spray.

Hospital Building Information Modeling.
e Second Hospital of Medical University is called the Second Affiliated Hospital of Harbin Medical University. Its outpatients building has 7 floors, including 6 floors above ground and 1 floor underground. According to the on-site inspection, the − 1st to 4th floors are mostly outpatients' rooms, the 5th to 6th floors are mainly ward regions, and the distribution of the 7-story buildings is shown in Table 15 Figure 23 shows the display effect of a certain part of the hospital in the model. Figure 24 shows the actual display effect of this part of the hospital.

Analysis of Evacuation Dynamic Parameters.
By combining data from multiple papers, in this study, the number of evacuees was planned to be 2640. In this study, by          Table 16 shows the reaction time of different pedestrians in the fire, and Table 17 shows the evacuation speed of different pedestrians. rough onsite visits to the outpatients building of the Second Hospital of Medical University, combined with the actual functions of each floor of the outpatients building, different staff proportions were set for each floor. e specific staff type ratio distribution is shown in Table 18 below.

Results
Cheng et al. [51] conducted a risk assessment of pedestrian evacuation in subway stations under large-scale sports activities. Chen et al. [52] synthesized the preevacuation behavior data of the residents obtained through the investigation and discussed the influencing factors of the evacuation time. Choochart and ipyopas [53] used software Pathfinder to study passenger evacuation of Airbus A330-300 aircraft. de Lama et al. [54] used a virtual environment to study human behavior in fire. Velásquez et al. [55] studied the gait body mass index of building risk assessment through graph theory. is article will combine the research results of various scholars and fire products to conduct the evacuation risk assessment of building fire pedestrians.
As shown in the Figure 25, the person has no behavior in the first 10 s, and the person starts to move after 10 s. is is because in all types of pedestrians, the reaction time of healthy men is the shortest, and healthy men begin to escape 10 s after the fire alarm. It can be seen from the curve that the evacuation curve of the pedestrians in the two scenarios is almost the same in the first 185 s or so, because the fire    1  20  10  20  10  10  10  10  10  1  180  90  180  90  135  90  45  90  2  60  30  60  30  45  30  15  30  3  60  30  60  30  45  30  15  30  4  120  40  40  40  40  80  20  20  5  125  25  50  100  10  0  10  0  6  125  25  50  100  10  0   products in Scenario 2 did not have a great impact on the pedestrians during this period, but in the 185 s. After that, the impact of fire products is getting bigger and bigger, the slope of the curve in Scenario 2 is getting smaller and smaller, and the slope represents the efficiency of pedestrians evacuation per unit time. With the evacuation time, the evacuation time of Scenario 1 is 792.3 s. In the Scenario 2, the evacuation time of the pedestrians was 1946 s (of which 15 pedestrians failed to evacuate), and the difference between the evacuation results was 145.61%.     Table 19 were obtained. Combine the data in the Table and the evacuation situation of each exit to determine the degree of injury and the type of accident. As can be seen from Figures 27-34, the pedestrians at the seven exits 1, 3, 4, 5, 6, 7, and 8 were not injured during the evacuation process. Before the minor injury, the pedestrians had fled the fire. rough calculation and analysis, the 7 exits had a total of 2539 evacuees, accounting for 96.17% of the total number.
In Exit 2's evacuation situation, 61 pedestrians had fled the fire before ASET1 � 92 s, and the remaining 25 pedestrians began to be injured. At this time, the remaining 25 pedestrians entered a state of minor injury. With the evacuation time, at 190 s, Twenty-five pedestrians with minor injuries escaped from the fire site; the total time from entering the minor injury state to completely escaping from the fire site was 98 s, accounting for 51.58% of the total evacuation time of the exit. is is because the fire products have a great impact on the evacuation speed of pedestrians; as a result, pedestrians are slow to move. For the same distance, compared with not considering fire products, pedestrians will spend more time to escape under the influence of fire products Figure 35 shows the location of region 12 in pathfinder.
rough the search for data, the reason for the failure of pedestrians evacuation is that the temperature of the regions 1-12 where the pedestrians is located exceeds 120°C, and  Discrete Dynamics in Nature and Society during the development of the fire, the temperature reached the lethal temperature, and the CO did not reach the lethal volume fraction. So this time, the study takes temperature as the research object. As shown in the Figure 36, the fitting curve of the evacuation situation of the regions 1-12 and the temperature change situation. At the 182 s, 15 pedestrians in the region 1-12 because the temperature of the region reached 129.1°C and the evacuation coefficient became 0. at is to say, the 15 pedestrians lost their mobility and entered a comatose state. Without considering any rescue, when t � 186 s, according to the data in Table 11, it can be considered that the temperature reached the lethal state, and these 15 pedestrians went from a severely injured coma to a dead state. According to the information in Table 19, it can be judged that the fire accident is a major accident.

Conclusions
(1) Compared with traditional research methods, the evacuation time of the evacuation model under the influence of fire products is longer, and the speed of the evacuation will change with time until the evacuation fails. is method is closer to the evacuation result of pedestrians in a real fire sceneIt can be close to the evacuation result of people in the real fire scene. (2) e spray intensity has little effect on the speed coefficient. In the early and mid-term of the fire, the smaller the spray intensity, the greater the irritation to the human body. is is because the presence of spray leads to a slower temperature rise, which will have a positive effect on evacuation. In the later stage of the fire, the greater the spray intensity, the weaker the influence of the fire products, the longer it takes for the speed coefficient to return to zero, and the more time the personnel will have to escape from the fire. (3) e smoke exhaust rate has a significant effect on the speed coefficient. is is because the presence of smoke can bring out the fire smoke and high-temperature heat plume, reducing the impact of fire products in the early stage. In the later stage of the fire, the higher the smoke exhaust rate, the more evacuation time is available for personnel, and the effect of smoke exhaust on evacuation is far better than spraying.

Data Availability
e data used to support the findings of this study are available from the corresponding author upon request.

Conflicts of Interest
e authors declare that they have no conflicts of interest.

18
Discrete Dynamics in Nature and Society