EP2758949A1

EP2758949A1 - Evacuation system for planning escape routes in a building, method, and computer program

Info

Publication number: EP2758949A1
Application number: EP12759154.3A
Authority: EP
Inventors: Rajesh Reddy NANDYALA
Original assignee: Robert Bosch GmbH
Current assignee: Robert Bosch GmbH
Priority date: 2011-09-20
Filing date: 2012-09-17
Publication date: 2014-07-30
Anticipated expiration: 2032-09-17
Also published as: DE102011083024A1; EP2758949B1; WO2013041479A1

Abstract

Escape routes in buildings are usually defined during the planning of the building and marked by means of signs during the construction of the building. The escape routes are optionally modified again during operation of the building, in particular if usage changes are made to sub-areas of the building. In addition to the static indication possibilities for marking escape routes by means of signs, dynamic signs of escape routes have also become known, which dynamic signs are activated only when a hazardous situation arises. The invention relates to an evacuation system (1) for planning escape routes in a building, comprising a model module (2) that provides a model of the building, an event module (3) that provides at least one escape event in the building, a person distribution module (4) that provides a relevant person distribution in the building, an escape route module (5) for planning the escape routes on the basis of the model, the at least one escape event and the relevant person distribution as input variables, and a person prediction device (9) that estimates the relevant person distribution in the building using a person distribution model.

Description

description

title

Evacuation systems for the planning of escape routes in a building,

Procedure as well as computer program

State of the art

The invention relates to an evacuation system for planning escape routes in a building with a model module providing a model of the building with an event module providing at least one escape event in the building with a person distribution module providing and distributing a relevant person distribution in the building one

Escape route module for planning the escape routes on the basis of the model, the at least one escape event and the relevant person distribution as input variables.

Usually, escape routes in buildings are defined during building planning and signposted during the construction of the building. If necessary, in the operation of the building, especially in changes in use of

Parts of the building, the escape routes adapted again. The

Escape routes are usually marked with static signs or other indications, so that escape route seekers can easily find the escape routes in case of danger and follow them.

In addition to the static information about the escape route description also dynamic signage of escape routes have become known, which are only activated when a danger occurs.

The document DE 10 2008 042 391 A1, which forms probably the closest prior art, relates to a fire safety device with a Input module, which is used to receive fire data in one

Secured area is formed with an evaluation module, which is designed for processing the fire data and to form a processing result, and with an output module which for activating and / or controlling security actions based on the processing result of

Evaluation module is formed, wherein the evaluation module a

Prediction unit to predict the fire course based on the fire data as a result of processing. As a security action, it is proposed, inter alia, to carry out an optimization of escape routes. Furthermore, the use of surveillance cameras, intrusion sensors,

Access sensors and other sensors that provide information on existing persons in the building and their whereabouts, to detect current accumulations of people and, for example, to detect panic in front of escape doors.

Disclosure of the invention

In the context of the invention, an evacuation system for the planning of

Escape routes in a building with the features of claim 1, a

Method for planning escape routes in a building with the features of claim 9 and a computer program with the features of

Claim 10 proposed. Preferred or advantageous embodiments of the invention will become apparent from the dependent claims, the following description and the accompanying drawings.

The evacuation system according to the invention is suitable and / or designed for planning escape routes in a building. The building may be a private, industrial or public building. For example, the building can be considered a private home, an office building, a department store, a

Production building, a library, an administration building or a

Warehouse trained. The evacuation system serves as

Starting to provide and / or update the escape routes in the building. An escape route is understood to mean a route within the building from a starting point to an end point. The starting point is

in particular as a current residence of persons trained. The endpoint is particularly useful as a savepoint, such as an exit, or as an intermediate point, e.g. a collection point or a distribution point in the building formed. In particular, the escape routes also include evacuation paths in the building.

The evacuation system includes a model module that provides a model of the building. The model of the building may be used as an architectural model, such as a 2D floor plan or be designed as a 3D model.

In particular, the model can be represented within the framework of the building data modeling (BIM). Alternatively or additionally, the model may be formed as a graph having a plurality of nodes and edges, the edges connecting the nodes. Such graphs can be advantageously used in the mathematical treatment of path problems.

Furthermore, the evacuation system comprises an event module which provides at least one escape event in the building. In particular, that is

Event module designed as an input module, which is signal-technically connectable or connected to a sensor system in the building. In which

Escape event is in particular a fire or fire message. Such a fire or fire message can be generated for example by an automatic fire alarm, a manual fire alarm or by an oral message. Optionally, the fire or

Fire alarm are routed via a fire control panel, which verifies the fire or fire alarm and passes it on to the event module. In addition to fire or fire alarm are also other messages, such as.

Persons threats, earthquakes, smoke or gas reports as

Escape events conceivable. Generally speaking, an escape event is an event that motivates people in the building to leave the building as quickly as possible.

The evacuation system comprises a person distribution module, which provides a relevant person distribution in the building. Under one

Person distribution is understood in particular a list in which Whereabouts in the building, in particular rooms and corridors or - mathematically - nodes and optionally complementary edges, a number of people is assigned to the whereabouts. The relevant

Person distribution is the distribution of persons, which should be considered when planning escape routes.

Furthermore, the evacuation system comprises an escape route module, which is used to plan the escape routes on the basis of the model, the at least one

Escape event and the relevant person distribution is designed as input variables. The escape route module thus implements an algorithm that the

Escape routes in the model plans so that the persons can be brought as fast as possible, safely and / or effectively before the escape event in safety. Such algorithms, in particular based on graphene, are well known. For example, an algorithm of the shortest paths of Edsger Dijkstra from the year 1959 is known. Ford and Fulkerson have one

Maximum Dynamic Flow Algorithm published in 1956, published in the Canadian Journal of Mathematics, Volume 8, pp.399-404.

In the context of the invention, a person prediction device

suggested that the relevant person distribution in the building below

Estimates the use of a person distribution model. Particularly preferably, the person distribution model for the building is comprehensive, so that at least 80%, preferably at least 90% and in particular at least 98% of all rooms are to be planned for the escape routes, in the

Person distribution model are included. In particular, this is based

Person distribution model on a-priori knowledge about the distribution of persons or contains this at least, which was used to form the person distribution model. The advantage of the invention is to be seen in particular in that the relevant

Personnel distribution in the case of an escape event is not currently recorded, ie measured, but must be made available through the person distribution model. In particular, sensor systems in buildings are usually not suitable for accommodating an up-to-date distribution of persons and, moreover, usually do not completely cover the building, so that a current

Personnel distribution not detected or measured with sufficient certainty can be. In contrast, the use of a

Personal prediction facility to estimate a relevant person distribution with a high level of security and to make it the basis for planning the escape routes.

With the input of the relevant person distribution, it is the

Evacuation system possible, the escape routes also depending on

Planning the escape route capacities in advance. So have to

For example, for sub-areas in the building in which there are no or only a few people, the escape routes do not have large capacities. In contrast, escape routes, which come from areas with a high number of people, have a very high capacity to avoid congestion or panic on the escape routes. The personal prediction device particularly preferably forms part of the

Persons distribution module. However, it should be noted that the distribution of the evacuation system into different modules is preferably functional or organizational. For example, the entire evacuation system can be designed as a data processing system, in particular as a computer system, so that the

different modules are parts of a program. It is also possible that the various modules in the evacuation system are designed so that their functions mix with each other. In a preferred embodiment of the invention is the

Personal predictive device for modeling a time-dependent

Person distribution trained. This aspect of the invention assumes that the distribution of persons in the building follows certain time-dependent rules. For example, the distribution of persons can vary depending on the time of day

People in the morning, e.g. start at 8:00 and until evening, e.g. Stop at 18:00. Outside these working hours, the office building will be largely empty. Even within working hours, the person distribution will be e.g. change over the day depending on shared breaks. A daytime-dependent person distribution can systematically this conditional

Represent changes. In addition, it is advantageous to have a weekday-dependent Modeling the distribution of persons, since in many buildings on the working days will be a different distribution of persons than on the weekend days. Particularly advantageous may also be a holiday dependent

Person distribution are modeled because in many buildings, for example. on a holiday like Christmas, a different person distribution will prevail than on a normal working day. Also a month dependent

Modeling of person distribution can be beneficial when in the

Building e.g. a seasonal operation, such as a nursery or a department store for camping items, is present. A correspondingly time-dependent modeling of the person distribution can take into account some or even all of these aspects and thus always provide a relevant person distribution at the time of "planning the escape routes".

In one possible embodiment, the person distribution model is statically predetermined. For example, it is possible that when planning a

Building, e.g. Office building, presence data of the persons in the building are estimated as plan data and these plan data form the basis for the person distribution model. This embodiment is easy to implement. However, such a static

Person distribution model does not respond to a dynamic usage behavior.

Therefore, it is particularly preferred that the person distribution model is created using previous presence data of the persons in the building or in subareas thereof. On the one hand, an immediate distribution of persons for the areas for which these data are available can be derived from the earlier presence data. On the other hand, on one

Person distribution in partial areas of the building, of which no previous presence data are present, e.g. the flows of persons in these areas are taken into account. Furthermore, through the

Use of previous attendance data of the persons also changes of the person distributions are considered, so that the

Person distribution model is dynamically adapted to the usage behavior of the building. In addition to the distribution of persons in the building, flows of people can also be modeled so that e.g. popular or often frequented

Paths in the building can be identified. In a particularly preferred embodiment of the invention, the evacuation system comprises a sensor system having a plurality of sensors distributed in the building, wherein the sensors are configured to occupy a number of persons in the building or in parts of the building as the

To record attendance data. Possible sensors are e.g.

Motion sensors, surveillance cameras, burglary sensors,

Access sensors, access control devices, etc. With the mentioned

Sensors recorded presence data form the input or the

Database for creating or updating the

Persons distribution model.

More preferably, the person distribution model is modeled or trained over time using machine learning techniques. If the learning of the person distribution model over a longer period, such. over several months or years, the person distribution model can estimate the relevant person distribution with a very high security. Modeling can be done, for example, by cluster analysis, data mining and, in particular, by the use of Gaussian mixture models (GMM).

In a preferred embodiment of the invention that sets

Person distribution module to a second function by current

Presence data from the sensor system with the relevant

Personality distribution of the person distribution model compares. The

Person distribution module is thus as a kind of sensor for monitoring the

Person distribution in the building.

A first possible, advantageous use of the comparison can be implemented as part of a review, whether the planned escape routes are actually used. Should turn out in the comparison that the planned

Escape routes or in particular a planned escape route is not used, it may be assumed that this planned escape route is blocked. In this case, the escape route module may suggest an alternative to the potentially blocked escape route, so that a safe evacuation of the people in the building can take place. Alternatively or additionally, it can be advantageously checked whether there is a panic or a significant concentration of persons in the building. For a person's concentration, the relevant

Personal distribution adapted and the escape routes to be updated. Similarly, panic can be counteracted by the generation of further escape routes of panic.

Another object of the invention relates to a method for planning escape routes in a building, preferably using the

Evacuation system according to one of the preceding claims or as described above, wherein based on a model of a building, at least one escape event in the building and a relevant person distribution in the building as input variables at least one escape route is planned in the building as a starting point. In the context of the invention, it is proposed that the relevant person distribution in the building be estimated using a person distribution model.

The planned escape routes in the building thus constitute an output of the process or the evacuation system. The escape routes can subsequently be implemented automatically or with the interposition of a fire control center, by providing dynamic escape route indicators, such as, eg. LED arrows, LCD displays, illuminated signs, etc. are controlled so that the persons in the building, the planned escape route is displayed.

Another object of the invention relates to a computer program with program code means having the features of claim 10.

Further features, advantages and effects of the invention will become apparent from the following description of preferred embodiments of

Invention. 1 shows a schematic representation as an overall overview of an evacuation system as an exemplary embodiment of the invention; FIG. 2 is a block diagram of the model module of the evacuation system in FIG. 1;

FIG. 3 is a block diagram of the event module of the evacuation system in FIG. 1;

FIG. 4 is a block diagram of the person distribution module of FIG

Evacuation system in Figure 1;

FIG. 5 is a flowchart illustrating the operation of the

Escape route module in Figure 1;

Figure 6 a, b a model of a building to be evacuated in different

representations;

Figure 7 a, b, c partial graph of the model in Figure 6b;

FIG. 8 shows a further flow chart for illustrating the mode of operation of the escape route module in FIG. 1.

FIG. 1 shows a schematic block diagram

Evacuation system 1 for planning escape routes in a building as an embodiment of the invention. The representation in FIG. 1 shows in particular an overview of the functional modules and optionally

supplementary modules. The overview is to be understood as an organizational overview or function overview and does not limit the evacuation system 1 to the assignment of functions to individual modules.

The evacuation system 1 is used to create or update

Escape routes F in a building. The evacuation system 1 can

For example, be designed as a program in a data processing device, in particular in a computer. The building is - abstractly considered - a plurality of distributed rooms, which are connected by corridors. Thus, the evacuation system 1 can be used for a variety of different types of buildings. The evacuation system 1 comprises a model module 2, an event module 3, a person distribution module 4 and an escape route module 5. The escape route module 5 is used to plan the escape routes F as an output variable and gets from the other modules 2, 3, 4 different input variables as the basis for the

Calculation of escape routes. Thus, the escape route module 5 receives from the

Model module 2 a model M of the building. The model M of the building may be formed as a floor plan, a 3D model or another representation. Optionally, the model module 2 can also provide a graph of the building as a model. Details on this will be explained below in connection with FIG.

The event module 3 transfers to the escape route module 5 escape events A, ie information or data of events in the building, which can trigger an escape. The type of escape event A is described below in

In connection with the figure 3 described in detail.

The passenger distribution module 4 transfers to the escape route module 5 a relevant person distribution P, ie information and data about the

Distribution of people in the building. Details of that

Person distribution module 4 are disclosed in connection with the figure 4.

A in the figure 1 of the model module 2 to the event module 3 or

Personnel distribution module 4 extending arrow indicates that the model M, in particular the graph or partial information or data about it are also passed on to the event module 3 and the person distribution module 4, so that e.g. the escape events A a position in the building or the relevant person distribution P can be determined based on the model M.

On the basis of the model M, of the at least one escape event A and the relevant person distribution P, at least one escape route F is generated by the escape route module 5 and - as shown by way of example in FIG. 1 - transferred to a signaling system or a fire control center 6. The

Fire station or the signaling system 6 set the escape routes F in the

Build the buildings so that they become clear to the people in the building. For example, depending on the escape routes F markings and / or signaling devices, in particular emergency exit signaling devices, so switched that the at least one escape route F is displayed in the building. Furthermore, FIG. 1 shows a sensor system 7 with a plurality of sensors which are distributed in the building. The sensors are, for example, as automatic or manual fire detectors, surveillance cameras,

Intruder alarm, access control systems, motion detectors, etc. trained. On the one hand, the sensor system 7 serves to provide an input to the event module 3, for example by an alarm signal from an automatic

Fire alarm is passed to the event module 3 and is forwarded by this as escape event F to the escape route module 5. On the other hand, the sensor system 7 is optionally connected by signal technology to the passenger distribution module 4, this function being explained in conjunction with FIG.

2 shows a schematic block diagram of the model module 2, wherein a floor plan model is shown within the model module 2 on the left side and a graph of the same part of the building on the right side. The floor plan model is shown to scale, with M1 a first

Room, M2 a second room and M3, for example, an input area are shown. Between the spaces M1 ... M3 are different passes of the D1 ... D5 as breakthroughs, e.g. Doors, drawn. M4 denotes a safety output from the illustrated subregion.

On the right side, the same subarea is shown, but this time as a graph, where the nodes M1 ... M4 correspond to the spaces on the left side and the edges D1 ... D5 correspond to the passageways with the same designations. Each node M1... M4 has two parameters, which in

curly brackets are shown, wherein the first parameter designates the initial number of persons in the room and the second parameter the maximum person capacity of the room. The initial one

Number of persons in the room is provided by the person distribution module 4 or estimated as a relevant person distribution P. The edges D1 ... D5 also have two parameters, namely the passage time along the passage and the maximum flow capacity along the passage per Time unit. For example, a long gear has a longer transit time than a short gear or a wider gear has a higher flow capacity than a narrow gear. Instead of providing the relevant person distribution P in the model module 2, the model module 2 can provide the graph without the initial number of persons, the number of persons then being in the

Escape route module 5 are used according to the specifications of the person distribution module 4.

FIG. 3 shows the event module 3 in a block diagram. The event module 3 takes over from the sensor system 7 sensor messages, in particular alarm messages, and assigns them in one

Allocation device 10 to the position in the building. Thus, for example, in the case of an alarm message from an automatic fire alarm, an initial position of a fire or fire in the building or in the model M is determined. This real event R is transferred as escape event A to the escape route module 5. Optionally, the or the

Real events R passed to a simulation device 8, which starting from the real event R a fire or a development

Smoke development in the building estimates and outputs a simulation result S as escape event A. To implement the simulation device 8, for example, the software: Fire Dynamics Simulator (FDS) can be used, which was originally developed by K. B. McGrattan at NIST (US). In one possible exemplary embodiment, it is also conceivable that a multiplicity of simulation results for the building are already precalculated in the simulation device 8, so that the occurrence of a real event R can be resorted to without great calculation times and thus very quickly to the simulation result S.

FIG. 4 shows a block diagram of the person distribution module 4. The person distribution module 4 uses as input variables data from the model module 2, in particular the model M or the graph G as well as inputs from the sensor system 7. As an output variable, the passenger distribution module 4 provides a relevant person distribution P to the escape route module 5. The person distribution module 4 comprises a person prediction device 9 which stores the relevant person distribution P in the building

Estimating use of a person distribution model PM. The person distribution model PM is created by a model device 1 1, wherein the sensor data from the sensor system 7 for modeling

be used. Many of the sensors used in the sensor system 7 can serve as a secondary function for detecting persons in subregions, such as e.g. Spaces, be used. So it is possible, for example, with a

Security camera counting people in a surveillance area.

It is also possible to record a number of passing persons via an access control system. Also other sensors, in particular motion sensors, such as e.g. Burglary sensors can be used to detect persons in the subareas. These presence data are used to make a realistic and thus very reliable

Person distribution model PM to create. The person distribution model PM maps the distribution of persons in the building depending on time and place.

Through the person distribution model PM, it is possible to estimate the number of persons per room, depending on the time of day and / or the day of the week. The estimated number of persons or the relevant person distribution P is used to supplement the graph, optionally in the model module 2 or in the escape route module 5. The number of persons is used in particular to predict an evacuation time from a room and / or in time to potential problems in comparison with the

Results from the event module 3, in particular with the

Simulation results of the spread of fire and smoke, to point out, or plan the escape routes differently. Optionally, in addition, it is possible that the person distribution model PM

Information about the normal or usual directions of movement of persons has. For example, it is possible to determine which route people from a certain area prefer to use. These preferred ways can preferably be used in the escape route planning, since then as escape routes are chosen paths that normally use the people from the room and therefore also know. In a very simple implementation, the over the sensors of the

Sensor system 7 averaged detected presence data and then represent the person distribution model PM. However, it is also possible to use a self-learning system, which receives the presence data as input and creates a model from it.

In a possible addition of the embodiment receives the

Person prediction 9 additional information, especially current presence data from the sensor system 7. This current

Presence data can be used for verification and, if necessary, correction of the estimated, relevant person distribution P. It is also possible to check whether the planned escape routes are being used or possibly - if not used - to be classified as impassable by comparing the current presence data, so that the

Escape route module 5 proposes alternative escape routes. By comparison, it is also possible, e.g. Identify person accumulations or panic behavior, so that 5 different or complementary escape routes can be planned by the escape route module.

FIG. 5 illustrates in a flow chart a possible embodiment of the method of operation of the escape route module 5.

In a first step 100, the building is optionally taken over by the model module 5 as a model model M or the model M is converted by the escape route module 5 into a graph. The graph G has a plurality N of nodes n and a plurality E of edges e. Each node N has as a parameter a maximum capacity corresponding to the number of persons staying in the space allocated to the node. As a second parameter, each node has the current number of persons in this room. Each edge e has as parameter a maximum throughput capacity per unit time and as a second parameter the time unit which is necessary to traverse the edge of a person.

In FIGS. 6 a, b, the transition from a model M to a graph is compared. In the figure 6a is the floor plan of a building on three Floors I, II, III shown. Above each side

Stairways are the three floors I, II, III interconnected. To facilitate the assignment are in the plan in the figure 6a the

Designations of nodes shown in Figure 6b. Of particular note is the two collective outputs N21 and N22, which are additionally marked Exit # 1 and Exit # 2 in the picture on the right.

According to FIG. 5, in step 200, the graph in FIG. 7b is subdivided into partial regions, ie smaller graphs <G1, G2, Gk>, each of the small graphs <G1, G2, Gk> having at least one escape route to an output , In the present example, the graph has been divided into three sections G1, G2 and G3, as shown in Figs. 7a, b, c. The division is also shown by sketched lines in FIG. 7b.

In FIG. 7 a, it can be seen that on the ground floor I as partial area G1, starting from the staircases N17 and N19, there is an immediate edge to the collecting outlets N21 and N22.

FIG. 7 b shows the partial area G 2 for the second floor II, wherein it can be seen that, starting from the staircases N 13 and N 15, the paths to the exits N 21 and N 22 extend over the staircases N 17 and N 19 of the ground floor I.

In the figure 7 c, the partial area G3 for the third floor III is shown, wherein, starting from the staircase on the third floor N5 and N8, the escape route via the vestibule N10 and the staircase N13 of the second floor II and the staircase N17 of the ground floor I goes to the output N21. In the same way, a possible path from the staircase N8 in the third floor III on the vestibule N12, the staircase N 15 of the second floor II and the staircase N 19 of the ground floor I to the exit N22. Thus, the three sections G1, G2 and G3 each use the same

Collective outputs N21 and N22 and share a majority of the nodes or the edges on the way to the outputs N21 and N22.

According to FIG. 5, in a step 300, considering the relevant person distribution P and the escape route events A for each Sub-areas G1, G2 and G3 planned escape routes to the collective exits N21 and N22. At this point it should be underlined that no planning for

Escape routes in the entire building G is performed, as this planning would require a significantly higher calculation effort. Instead, the planning is carried out in parallel and in particular independently of each other for the subregions G1, G2, G3. Thus, the evacuation system 1 is suitable for very branched buildings, since these only - divided according to the principle of divida et impera - into subareas and the subareas must be treated independently. The required computing power thus increases linearly with the complexity of the building and not with the linearity, as would be the case if the escape routes were calculated for the building as a whole.

In a step 400, based on the escape events A of a

Priority device 12 (Figure 1) distributed a priority for the subregions G1, G2 and G3. It is a consideration of this embodiment that fire very often breaks out only in a partial area of a building, so that it is first necessary to evacuate this subarea. Subsequently, only the escape routes for the sub-area are output and forwarded to the signaling system or the fire center 6, which have the highest priority. Thus, escape routes are initially output only for persons from the sub-area G1, G2, G3 with the highest priority.

In a next step 500 it is checked whether there are still free capacities for evacuation at the collective outputs N21 and N22 or at the nodes and edges previously shared by the subregions G1, G2, G3. For example, there may be very few people present in the highest priority subarea, so despite the

Evacuation efforts the stairwells will not reach their load limit in terms of capacity. In this case, it is possible to start concurrent with the evacuation of a second subarea with a lower priority.

The planning of the escape routes and the setting of the priorities for the subareas G1, G2, G3 is updated in an iterative process, on the one hand new escape events A - in particular with regard to simulated Smoke propagation results - as well as changes of the relevant ones

Person distribution to be checked.

So it is possible, for example, that certain escape routes are no longer passable by spreading fire or spreading smoke, as explained in the following example:

It is assumed that an alarm case in the third floor III, ie in the sub-area G3. Thus, the third level III is assigned the highest priority and proposed at a time t = 0, the following escape routes from the rooms N1, N6 and N7:

N1 -N3-N2-N5-N10-N13-N17-N21

N1 -N3-N4-N8-N12-N15-N19-N22

N6-N3-N2-N5-N10-N13-N17-N21

N6-N3-N4-N8-N12-N15-N19-N22

N7-N3-N2-N5-N10-N13-N17-N21

N7-N3-N4-N8-N12-N15-N19-N22

From the event module 3, in particular from the simulation device 8, the propagation of the fire or smoke is simulated on the basis of the fire or fire message and, as a result, the following table is output, which indicates the passability or impassability of the nodes in the next time:

Thus, in the period from t = 0s to t = 250s, the aforementioned escape routes are passable, after the time t = 250, at least the node N4 is impassable, so that the escape routes are dynamically restricted as follows: N1 -N3-N2-N5-N10-N13-N17-N21

[N1 -N3-N4-N8-N12-N15-N19-N22] blocked

N6-N3-N2-N5-N10-N13-N17-N21

[N6-N3-N4-N8-N12-N15-N19-N22] locked

N7-N3-N2-N5-N10-N13-N17-N21

[N7-N3-N4-N8-N12-N15-N19-N22] locked

Should no new escape routes be found, e.g. after t <400 s, the evacuation system 1 can issue a message to the fire center 6 that external assistance is needed.

In particular, when the person distribution P has fallen far in the highest priority subarea, due to the

Capacity check on shared nodes and edges N21 and N22, the next higher priority sub-range is preferred during evacuation.

FIG. 8 shows a more detailed flowchart of the method and, subsequently, a possible pseudocode for the method in FIGS. 5 and 8 as an exemplary embodiment:

Step A

First, an escape event, e.g. triggered via the sensor system 7. Step B

The subarea affected by the flight event is determined, whereby a high or the highest priority is assigned in this non-safe subarea. The sub-domain may be expanded based on simulation results from the simulator 8 to ensure that all individuals in potentially unsafe sub-areas are covered for evacuation.

Step C

The escape routes are planned by the escape route module 5 at least for the sub-area with the highest priority and optionally for all sub-areas. When planning the escape routes, the results from the

Simulation device 8 taken into account. Step D

The evacuation of the persons in the highest priority subarea is initiated immediately. For the calculation of the evacuation time, the person distribution model PM is used, for the calculation of the escape routes can optionally be added to the in the person distribution model PM

identified preferred ways of persons are used.

Steps

It is checked if different, affected by the escape event

Subareas each have individual outputs or collective outputs. If the different subregions have individual outputs, the evacuation according to step F is initiated in parallel for the subregions. In the event that the sections to be evacuated share the collective outputs, according to step G, first the evacuation for the sub-area with the highest priority is triggered. The evacuation then takes place via the free, unimpaired escape routes.

Steps H, I, J

During the evacuation, new data are constantly being acquired by the sensor system 7 and the simulation device 8, the graph is updated, in particular with regard to nodes and edges that can be used for an escape route and, if appropriate, the escape routes are dynamically based on the sensor and sensor

Adjusted simulation information and updated the escape route display in the building.

Step K

It is checked if the evacuation was successful. According to step L, in an - at least partially unsuccessful evacuation - an inventory of the trapped persons carried out and requested external help. In a successful evacuation is checked in step M, if there are more people in the building. According to step N becomes

if necessary, determine that the evacuation is complete or, according to step O, evacuate the next higher priority subarea, wherein in the

Loop is returned to the step G. // Beginning of the embodiment //

A) Input:

1) G (N, E): a graph G with a set of nodes N and a set of edges E; Each node n e N has two properties:

Maximum Node Capacity (s): non-negative integer // maximum node capacity Initial Node Occupancy (s): non-negative integer // initial node assignment Each edge e e has two properties:

Maximum Edge Capacity (s): non-negative integer // maximum edge capacity Travel time (s): non-negative integer // travel time

2) S: set of source nodes, S _Ξ N; (Rooms with persons)

3) D: set of destination nodes, D _ = N; (outputs)

4) Sensor Information: Current sensor information of temperature detectors, smoke detectors, motion sensors and surveillance cameras etc.

5) Simulation Information: Simulation of fire with a known origin based on the first fire alarm for the estimation of fire and fire

Smoke distribution.

6) movement model: Gaussian mixed model of passenger movement in the building;

B) Pre-process network: The graph G (N, E) is divided into 'k' small graphs <G1, G2,. , , , Gk> s, so that each subdivided graph has a path or routes to an exit or exits.

C) Output: Dynamic escape routes with plans of fugitives on each route.

D) Method:

For all the 'k' graphs <G1, G2,. , , , Gk> {

Initial Number of Occupants = Get Number of Occupants (Movement Model); Get Evacuation Path (Graph, Gi) // The escape routes are generated in parallel for all graphs.

}

Priority Areas = Get Priorities (Sensor Information);

While (Graph, Gk has high priority) do

{

Initial Number of Occupants in Graph, Gk = Get Number of Occupants (Movement Model);

Evacuate occupants in Graph, Gk

Maximum capacity or not used by Graph, Gk

Estimate the capacities in shared nodes and evacuate occupants from other Graphs, Gj where j! = K.

Fire and Smoke spread = Get affected nodes and edges (Simulation Information);

Generate Dynamic Evacuation Paths = Get Evacuation Path (Graph, Gk)

Blocked occupants = Get External help (Graph, Gk) // People are due to

Fire and smoke are blocking and need external help.

}

E) Function: Get Evacuation Path (Graph, Gk)

While any source node s has evacuee do

{

find route R <nO, n1,. , , nk> with earliest destination arrival time between routes between all (s, d) pairs where s s, d e D, nO = s, nk = d)

flow = min (number of evacuees still at source node s,

Available Edge Capacity (all edges on route R),

Available Node Capacity (all nodes from n1 to nk on route R)

);

for i = 0 to k - 1 do {

tDash = t + Travel_time (e (ni, ni + 1))

Available edge capacity (e (ni, ni + 1), t) reduced by flow;

Available Node Capacity (ni + 1, tDash) reduced by flow;

}

} Output evacuation paths;

// END of the embodiment //

Instead of the described function E) GetEvacuationPath can eg. For example, the following algorithm can be used:

In one step, the source nodes and the collective outputs of the subarea are captured. For the sub-area II in FIG. 7b, it should be assumed by way of example that the nodes N9 and N14 are designed as source nodes, ie as nodes with persons to be evacuated. Nodes N21 and N22 constitute the collective outputs.

In a next step, the shortest paths between all

Source node and all collective outputs calculated. The calculation can be done, for example, using the Dijkstra algorithm described above. For subsection II in FIG. 7b, the following paths are thus determined:

N9-N1 1 -N10-N13-N17-N21

N9-N1 1 -N12-N15-N19-N22

N14-N1 1 -N10-N13-N17-N21

N14-N1 1 -N12-N15-N19-N22

Subsequently, the persons to be evacuated are evacuated as a function of the capacities of the nodes and edges.

For example, the subregions can also be subdivided such that existing fire doors are selected as collective exits.

Claims

claims

1 . An evacuation system (1) for planning escape routes in a building comprising a model module (2) providing a model (M) of the building with an event module (3) providing at least one escape event (A) in the building with a person distribution module (4), which is a relevant

Person distribution (P) in the building provides, with an escape route module (5) for planning the escape routes (F) on the basis of the model (M), the at least one escape event (A) and the relevant person distribution (P) as input variables, characterized by a Personal prediction device (9), which contains the relevant

Person distribution (P) in the building using a

Person distribution model (PM).

2. evacuation system (1) according to claim 1, characterized in that the Personenprädiktionseinrichtung (9) a time-dependent, in particular daytime, a weekday-dependent, a month-dependent and / or a holiday-dependent person distribution (PM) models in the building.

3. evacuation system (1) according to one of claims 1 or 2, characterized

characterized in that the person distribution model (PM) is created using previous presence data of the persons in the building or parts thereof.

4. evacuation system (1) according to claim 3, characterized in that the evacuation system (1) comprises a sensor system (7) with a plurality of sensors distributed in the building, wherein the sensors are formed are to include a number of persons in the building or portions thereof as the presence data.

Evacuation system (1) according to claim 4, characterized in that the sensors motion sensors, burglar sensors,

Access control devices and / or surveillance cameras, etc.

include.

Evacuation system (1) according to one of the preceding claims 4 or 5, characterized in that the person distribution module (4) is adapted to compare current presence data from the sensor system with the person distribution model (PM).

Evacuation system (1) according to claim 6, characterized in that in the comparison, a check is carried out whether the planned escape routes are used.

Evacuation system (1) according to claim 6 or 7, characterized in that in the comparison of a panic or a person concentration is detected.

Method of planning escape routes (F) in a building,

preferably using the evacuation system (1) according to one of the preceding claims, characterized in that based on a model (M) of a building, at least one

Escape event (A) and a relevant person distribution (P) as input variables at least one escape route (F) is planned in the building, the relevant person distribution (P) is estimated in the building using a Personenverteilungsmodells (PM).

0. Computer program with program code means to complete all the steps of

A method according to claim 9, when the program is executed on a computer and / or on the evacuation system (1) of any one of claims 1 to 8.