Venue building evacuation route planning and fire safety management methods based on digital operation and maintenance

Abstract

The present invention discloses a venue building evacuation route planning and fire safety management method based on digital operation and maintenance, which relates to the field of intelligent fire management technology. The method includes: constructing a digital twin model of the venue building based on BIM and the Internet of Things, using the digital twin model in combination with fire emergency event instance data to construct a fire monitoring model; collecting real-time multi-source fire safety data through the venue building's firefighting equipment for fire detection and early warning, determining fire scene dynamic parameters to update the fire monitoring model, and obtaining fire dynamic characteristics in different venue building areas; obtaining key factors affecting evacuation efficiency, combining an improved rat swarm algorithm to plan evacuation routes, and visually displaying the optimal evacuation routes for different venue building areas in a preset manner. The present invention dynamically and in real time performs evacuation route planning and comprehensive monitoring and management of venue fire safety based on the development trend of the fire, thereby reducing losses in emergency situations such as fires.

Description

Technical Field

[0001] The present invention relates to the field of smart fire management technology, and more specifically, to a venue building evacuation route planning and fire safety management method based on digital operation and maintenance. Background Art

[0002] Evacuation route planning and fire safety management in venue architectural design are critical components for ensuring personnel safety. With the acceleration of urbanization and the increase in large public buildings, the rapid and safe evacuation of personnel in emergency situations has become a crucial issue in architectural design. Evacuation route design should adhere to the following principles: The shortest path principle ensures that personnel can reach safety areas by the shortest possible route; the two-way evacuation principle ensures that two-way evacuation passages are designed to avoid congestion caused by a single route; the barrier-free principle ensures that evacuation passages are free of obstructions for rapid passage; and the clear signage principle ensures that evacuation signs are clearly visible to ensure rapid identification in an emergency.

[0003] The configuration of fire protection facilities within venue buildings is the foundation of fire safety management. These primarily include automatic sprinkler systems, fire alarm systems, fire hydrants and fire extinguishers, and smoke exhaust systems. Intelligent fire management integrates sensors, monitoring equipment, and data analysis technologies to achieve comprehensive monitoring and management of venue fire safety. This enables real-time monitoring of fire risks, automatically activates emergency response plans, and optimizes fire resource allocation through big data analysis. Evacuation route planning and fire safety management within venue building design are crucial components for ensuring personnel safety. Scientific evacuation route design, advanced simulation technology, and intelligent management systems can effectively improve venue safety and reduce losses in emergencies such as fires. Therefore, using digital twin operations and maintenance to implement intelligent evacuation route planning and fire safety management within venue buildings is a pressing issue. Summary of the Invention

[0004] In order to solve the above technical problems, the present invention proposes a venue building evacuation route planning and fire safety management method based on digital operation and maintenance, which can dynamically plan evacuation routes and comprehensively monitor and manage venue fire safety in real time according to the development trend of the fire, thereby reducing losses in emergency situations such as fire.

[0005] The present invention provides a venue building evacuation route planning and fire safety management method based on digital operation and maintenance, comprising the following steps:

[0006] Build a digital twin model of the venue building based on BIM and the Internet of Things, and use the digital twin model combined with fire emergency event instance data to build a fire monitoring model;

[0007] Fire detection and early warning are carried out by collecting real-time multi-source fire safety data through the fire safety equipment of the venue building. The fire scene dynamic parameters are determined based on the multi-source fire safety data and twin data to obtain the fire dynamic characteristics of different venue building areas;

[0008] Based on the dynamic characteristics of fires in different venue building areas, the key factors affecting evacuation efficiency are obtained. Based on these key factors and combined with the improved rat swarm algorithm, evacuation path planning is performed to obtain the optimal evacuation path for different venue building areas.

[0009] The optimal evacuation routes of the different venue building areas are visualized in a preset manner, and fire safety equipment is used for route navigation.

[0010] In this solution, a digital twin model of the venue building is constructed based on BIM and the Internet of Things. The digital twin model is combined with fire emergency event instance data to build a fire monitoring model. Specifically:

[0011] Obtaining construction design data of the target venue building, using a BIM platform to construct a structural geometric digital model of the venue building based on the construction design data, and splicing and combining the structural geometric digital models to obtain a three-dimensional model of the venue building;

[0012] Assign the location and attributes of firefighting equipment in the venue building to the three-dimensional model, aggregate real-time sensory data of firefighting safety equipment in the venue building through the Internet of Things to generate multi-source fire safety data, and use a data interface to integrate the multi-source fire safety data into the digital twin model of the venue building, thereby achieving a connection between the physical space and the twin space and generating a digital twin model of the venue building;

[0013] By matching, aligning, and digitally mapping multi-source fire safety data with the digital twin model, fire emergency event instance data that meets similarity standards is retrieved based on the structural characteristics of the venue building. The pre-processed fire emergency event instance data is then used to simulate and verify the digital twin model.

[0014] The digital twin model of the venue building is divided into grid blocks. The local structural features in each grid block are used to screen the pre-processed fire emergency events. The multi-source fire safety data perceived by fire safety equipment in the screened fire emergency event instance data is imported into the digital twin model to obtain twin data.

[0015] Preset fire spread verification indicators and equipment status verification indicators, compare the twin data obtained through simulation with the instance data, adjust the twin model through comparison deviation, define fire risk indicators through multi-source fire safety data, build a monitoring threshold database corresponding to the fire risk indicators, and generate a fire monitoring model.

[0016] In this solution, real-time multi-source fire safety data is collected by the venue building’s fire safety equipment for fire detection and early warning. The dynamic parameters of the fire scene are determined based on the multi-source fire safety data and twin data. Specifically,

[0017] Collect multi-source fire safety data of the venue building and import it into the fire monitoring model to monitor the environmental status and equipment operation status of the venue building in real time, use the multi-source fire safety data to construct a fire safety data vector matrix, compare the fire safety data vector matrix with the monitoring threshold vector matrix, and obtain a residual vector matrix;

[0018] Fire detection and early warning are performed through the positive and negative values of each residual vector in the residual vector matrix to obtain twin data of the fire monitoring model, and the twin data is used to fuse data with multi-source fire safety data to obtain multi-source fusion data;

[0019] Fire dynamic data sequences are extracted from fire emergency event instance data to construct data samples. The data samples are divided into training sets and validation sets and imported into the AlexNet network. Feature extraction is performed through convolutional layers of different sizes, and activation functions are used for activation calculation. The pooling layer uses downsampling to reduce feature dimensions and achieve further feature extraction.

[0020] The reduced-dimensional features are imported into the fully connected layer. The fire classification results of the training samples are obtained through three fully connected layers and then verified using the validation set. The current network parameters are retained when the preset performance standards are met.

[0021] The last fully connected layer of the trained AlexNet network was removed, and the real-time multi-source fire safety data after the early warning was used as input. The corresponding one-dimensional feature vector was obtained through the fully connected layer, and the fire scene dynamic parameters were constructed using the one-dimensional feature vector.

[0022] In this solution, the fire dynamic characteristics of different venue building areas are obtained, specifically:

[0023] The fire scene dynamic parameters of different venue building areas are obtained based on the digital twin model of the venue building after grid processing. The fire scene dynamic parameters of different venue building areas are spatially modeled by the graph neural network. The association between the building structure similarity and channel connection setting areas of different venue building areas is used to obtain the spatial dependency information of the fire scene dynamic parameters based on the association between different venue building areas using the graph neural network;

[0024] In the graph convolution layer of the graph neural network, the state of each venue building area is updated through the information aggregation mechanism, and the spatial structure embedding matrix of the fire scene dynamic parameters is constructed to obtain a spatial embedding vector sequence;

[0025] The spatial embedding vector sequence output by the graph neural network is input into the LSTM network. The relationship between the key features in the spatial embedding vector sequence is obtained through the forget gate and memory cell state update, and the fire dynamic characteristics of the dynamic parameters of the fire scene in different venue building areas that change with time and space are output.

[0026] In this scheme, the key factors affecting evacuation efficiency are obtained based on the dynamic characteristics of fire in different venue building areas, specifically:

[0027] Correlation analysis is used to obtain factors that can affect evacuation efficiency from fire emergency event instance data. The obtained factors are then subjected to feature screening using the mRMR algorithm to obtain a set of factors that affect evacuation efficiency.

[0028] Taking each influencing factor in the factor set as the principal component direction in turn for principal component projection, obtaining the characteristic scatter point distribution corresponding to each influencing factor, and obtaining the characteristic parameters of each influencing factor according to the fire dynamic characteristics of different venue building areas;

[0029] Conduct principal component analysis on the characteristic parameters of each influencing factor in different venue building areas, obtain the principal component influencing factors as the principal component direction for principal component projection, and obtain the characteristic scatter point distribution of different venue building areas;

[0030] The similarity between the characteristic scatter point distribution of different venue building areas and the characteristic scatter point distribution corresponding to each influencing factor is calculated, and the influencing factors that meet the preset similarity standards are selected as the key factors of each venue building area.

[0031] In this plan, the venue building area evaluation is carried out based on the key factors mentioned above, specifically:

[0032] According to the key factors, the average parameter values corresponding to the personnel tolerance limits during evacuation of different key factors are read from the fire emergency event instance data to generate a benchmark evaluation matrix, and a risk assessment model for the venue building area is constructed through learning based on the benchmark evaluation matrix;

[0033] Obtain the characteristic parameters of each key factor in different venue building areas and import them into the risk assessment model. Obtain the Euclidean distance between the characteristic parameters of the key factors and the parameter vectors in the benchmark assessment matrix. Use comparative weighted calculation to obtain the risk values of different venue building areas.

[0034] Different venue building areas are graded according to the risk values, and the grading results are used to mark the digital twin model after grid processing. When the risk grade is greater than the preset risk level, it is deemed inaccessible.

[0035] In this scheme, the improved rat swarm algorithm is used for evacuation path planning, specifically:

[0036] Fire safety equipment collects real-time multi-source fire safety data to obtain the distribution of trapped people in buildings and venues. Soft clustering is performed by setting cluster centers based on the safety exits of the buildings and venues to match the trapped people with the nearest safety exits.

[0037] Based on the distribution of trapped people and the matching safe exits, the starting and target points of the paths are generated for the trapped people. The improved rat swarm algorithm is used for evacuation path planning. The evacuation paths are randomly obtained through the risk distribution of different venue building areas to initialize the rat swarm population.

[0038] Using elite reverse learning to process the rat population to expand the diversity of the population, defining an objective function with minimum evacuation time as the goal, mapping the objective function to the fitness function, and calculating the fitness value;

[0039] The mouse with the minimum fitness value is selected as the local optimal solution. The position of the mouse is updated according to the two stages of chasing prey and attacking prey in the mouse swarm algorithm. The fitness value is recalculated and the current and global optimal solutions are updated.

[0040] The position of the individual mouse is updated using the Levy flight strategy, and the fitness values before and after the update are compared. If the fitness value after the update is smaller, the updated individual mouse is taken as the optimal solution, otherwise the original optimal solution is retained. After the iterative optimization is completed, the optimal evacuation path for different trapped people is obtained based on the optimal solution.

[0041] In this solution, the optimal evacuation routes for the different venue building areas are visualized in a preset manner, and fire safety equipment is used for route navigation, specifically:

[0042] Use the 3D model of the venue building to visually mark the optimal evacuation paths for trapped people in different venue building areas. When the risk distribution in the venue building changes, the updated optimal evacuation path will be replaced and displayed;

[0043] Generate indication information based on the current optimal evacuation path, convert the indication information into corresponding modal data based on the properties of fire safety equipment in different venue building areas, and use different fire safety equipment to provide path guidance for trapped people.

[0044] Compared with the prior art, the present invention has the following beneficial effects:

[0045] The present invention dynamically adjusts the evacuation route according to the real-time fire spread situation to avoid congestion and dangerous areas, and uses an improved rat swarm algorithm to plan the evacuation route to find the optimal evacuation route. In the evacuation route planning, it automatically avoids dangerous areas such as fire, smoke, and collapsed structures to ensure that people can evacuate quickly and safely. BRIEF DESCRIPTION OF THE DRAWINGS

[0046] In order to more clearly illustrate the technical solutions in the embodiments or exemplary embodiments of the present invention, the following briefly introduces the drawings required for use in the embodiments or exemplary descriptions. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained according to these drawings without paying any creative work.

[0047] Figure 1 A flow chart showing the venue building evacuation route planning and fire safety management method based on digital operation and maintenance is shown;

[0048] Figure 2 A flow chart for determining fire scene dynamic parameters based on multi-source fire safety data and twin data is shown;

[0049] Figure 3 A flowchart of evacuation path planning using the improved rat swarm algorithm is shown;

[0050] Figure 4 A block diagram of the venue building evacuation route planning and fire safety management system based on digital operation and maintenance is shown. DETAILED DESCRIPTION

[0051] In order to more clearly understand the above-mentioned objects, features and advantages of the present invention, the present invention is further described in detail below in conjunction with the accompanying drawings and specific embodiments. It should be noted that, in the absence of conflict, the embodiments of the present application and the features therein can be combined with each other.

[0052] In the following description, many specific details are set forth to facilitate a full understanding of the present invention. However, the present invention may also be implemented in other ways different from those described herein. Therefore, the scope of protection of the present invention is not limited to the specific embodiments disclosed below.

[0053] Figure 1 A flow chart of the venue building evacuation route planning and fire safety management method based on digital operation and maintenance is shown.

[0054] like Figure 1 As shown, the first embodiment of the present invention provides a venue building evacuation route planning and fire safety management method based on digital operation and maintenance, including:

[0055] S102, constructing a digital twin model of the venue building based on BIM and the Internet of Things, and using the digital twin model in combination with fire emergency event instance data to construct a fire monitoring model;

[0056] S104: Collect real-time multi-source fire safety data through the venue building's fire safety equipment to perform fire detection and early warning, determine fire scene dynamic parameters based on the multi-source fire safety data and twin data, and obtain fire dynamic characteristics in different venue building areas;

[0057] S106, obtaining key factors affecting evacuation efficiency based on the dynamic characteristics of fires in different venue building areas, performing evacuation route planning based on the key factors and an improved rat swarm algorithm to obtain optimal evacuation routes for different venue building areas;

[0058] S108: Visually display the optimal evacuation routes of the different venue building areas in a preset manner, and use fire safety equipment to perform route navigation.

[0059] It should be noted that the construction design data of the target venue building is obtained, and a structural geometric digital model of the venue building is constructed according to the construction design data using the BIM platform, including floor layout, evacuation passages, safety exits, etc. The structural geometric digital models are spliced and combined to obtain a three-dimensional model of the venue building; the location and attributes of the fire-fighting equipment in the venue building are assigned to the three-dimensional model, and the real-time perception data of the fire safety equipment in the venue building is aggregated through the Internet of Things to generate multi-source fire safety data. The fire safety equipment includes automatic sprinkler fire extinguishing systems, monitoring systems, fire alarm systems, emergency lighting systems, smoke exhaust systems, etc. The multi-source fire safety data is integrated into the digital twin model of the venue building using a data interface to achieve the connection between the physical space and the twin space and generate a digital twin model of the venue building; the temperature, humidity, smoke concentration and other environmental parameters in the venue are monitored in real time to promptly detect safety hazards such as fire, and by monitoring the status of fire-fighting equipment, lighting systems, elevators and other equipment, their normal operation in emergency situations is ensured; the multi-source fire safety data is matched, aligned and digitally mapped with the digital twin model.

[0060] According to the structural characteristics of the venue building, fire emergency event instance data that meets the similarity standard is retrieved, and the pre-processed fire emergency event instance data is used to perform simulation verification of the digital twin model; the digital twin model of the venue building is divided into grid blocks, and the local structural features are used in each grid block to screen the pre-processed fire emergency events, and the multi-source fire safety data perceived by the fire safety equipment in the screened fire emergency event instance data is imported into the digital twin model to obtain twin data; fire spread verification indicators and equipment status verification indicators are preset, and the fire spread verification indicators include fire spread speed, temperature distribution, smoke concentration distribution, etc. The equipment status verification indicators include equipment start-up time, operating status, failure rate, etc.; the twin data obtained through simulation are compared with the instance data, the twin model is adjusted through comparison deviation, the fire risk indicators are defined through multi-source fire safety data, a monitoring threshold database corresponding to the fire risk indicators is constructed, and a fire monitoring model is generated.

[0061] Figure 2 A flow chart for determining fire scene dynamic parameters based on multi-source fire safety data and twin data is shown.

[0062] According to an embodiment of the present invention, real-time multi-source fire safety data is collected by the fire safety equipment of the venue building for fire detection and early warning. The dynamic parameters of the fire scene are determined based on the multi-source fire safety data and twin data, specifically:

[0063] S202: Collect multi-source fire safety data of the venue building and import it into the fire monitoring model to monitor the environmental status and equipment operation status of the venue building in real time; use the multi-source fire safety data to construct a fire safety data vector matrix; compare the fire safety data vector matrix with the monitoring threshold vector matrix to obtain a residual vector matrix;

[0064] S204: Perform fire detection and early warning based on the positive and negative values of each residual vector in the residual vector matrix to obtain twin data of the fire monitoring model, and use the twin data to fuse with multi-source fire safety data to obtain multi-source fused data;

[0065] S206, extracting fire dynamic data sequences from fire emergency event instance data to construct data samples, dividing the data samples into training sets and validation sets, and importing them into the AlexNet network. Feature extraction is performed through convolutional layers of different sizes, and activation calculation is performed using activation functions. The feature dimension is reduced by downsampling using a pooling layer to achieve further feature extraction.

[0066] S208, importing the reduced-dimensional features into the fully connected layer, obtaining the fire classification results of the training samples through three fully connected layers, and verifying the fire classification results using the validation set, retaining the current network parameters when the preset performance standards are met;

[0067] S210: Remove the last fully connected layer of the trained AlexNet network, use the real-time multi-source fire safety data after the warning as input, obtain the corresponding one-dimensional feature vector through the fully connected layer, and use the one-dimensional feature vector to construct fire scene dynamic parameters.

[0068] It should be noted that fire risk indicators, such as temperature thresholds, smoke concentration thresholds, and gas concentration thresholds, are defined based on multi-source fire safety data. The fire safety data vector matrix is compared with the monitoring threshold vector matrix, and warning rules are set based on this comparison. For example, if the temperature exceeds the threshold and the smoke concentration continues to rise, and the temperature residual vector and smoke concentration residual vector are determined to be positive, a fire warning is triggered, and firefighting equipment is activated for fire extinguishing and smoke exhaust. Feature extraction is performed on the real-time multi-source fire safety data after the fire warning based on the AlexNet network. The original AlexNet network has eight weighted layers, including five convolutional layers and three fully connected layers. The first, second, and fifth convolutional layers each contain a maximum pooling layer, and the last layer of the three fully connected layers is a softmax function. Feature extraction is performed through convolutional layers with convolutional kernels of different sizes, activated using the ReLU activation function, and input into the pooling layer. The pooling operation reduces the feature dimensionality for further feature extraction, and the fully connected layer outputs a one-dimensional feature vector.

[0069] The fire dynamic parameters of different venue building areas are obtained based on the digital twin model of the venue building after grid processing. The fire dynamic parameters of different venue building areas are spatially modeled in the graph neural network. The association between areas is set by the building structure similarity and channel connection of different venue building areas. The spatial dependency information of the fire dynamic parameters is obtained based on the association between different venue building areas using the graph neural network; the state of each venue building area is updated through the information aggregation mechanism in the graph convolution layer of the graph neural network, and the spatial structure embedding matrix of the fire dynamic parameters is constructed to obtain a spatial embedding vector sequence; the spatial embedding vector sequence output by the graph neural network is input into the LSTM network. The LSTM generates a hidden state for each time step. The hidden state contains the information summary of the previous input sequence. The relationship between the key features in the spatial embedding vector sequence is obtained through the forgetting gate and memory cell state update, and the fire dynamic characteristics of the fire dynamic parameters of different venue building areas that change with time and space are output.

[0070] It should be noted that the Pearson correlation coefficient was used to obtain factors that can affect evacuation efficiency from fire emergency event instance data. The obtained factors were then subjected to feature screening using the mRMR algorithm to obtain a set of factors affecting evacuation efficiency, including fire scene temperature, carbon monoxide gas concentration, smoke concentration, corridor visibility, number of personnel, etc. Each influencing factor in the factor set was sequentially used as the principal component direction for principal component projection to obtain the characteristic scatter point distribution corresponding to each influencing factor. The characteristic parameters of each influencing factor were obtained based on the fire dynamic characteristics of different venue building areas. Principal component analysis was performed on the characteristic parameters of each influencing factor in different venue building areas, and principal component projection was performed using the principal component influencing factors as the principal component direction to obtain the characteristic scatter point distribution of different venue building areas. The characteristic scatter point distribution of different venue building areas was calculated for similarity with the characteristic scatter point distribution corresponding to each influencing factor, and the influencing factors that met the preset similarity criteria were selected as the key factors of each venue building area.

[0071] Based on the key factors, the average parameter values corresponding to the tolerance limits of personnel during evacuation corresponding to different key factors are read in the fire emergency event instance data to generate a benchmark assessment matrix. For example, the human body injury temperature of 60 degrees Celsius in building fires, the extreme dimming degree of smoke concentration on personnel visibility, and the carbon monoxide concentration that causes personnel poisoning are obtained. Based on the benchmark assessment matrix, a risk assessment model for venue building areas is constructed through learning; the characteristic parameters of each key factor in different venue building areas are obtained and imported into the risk assessment model, the Euclidean distance between the characteristic parameters of the key factors and the parameter vectors in the benchmark assessment matrix is obtained, and the risk values of different venue building areas are obtained using comparative weighted calculation. , expressed as:

[0072]

[0073] in, represents the benchmark evaluation matrix, represents the matrix transpose, represents the Euclidean distance operation, Characteristic parameters that represent key factors.

[0074] Different venue building areas are graded according to the risk values, and the grading results are used to mark the digital twin model after grid processing. When the risk grade is greater than the preset risk level, it is deemed inaccessible.

[0075] Figure 3 A flowchart of evacuation path planning using the improved rat swarm algorithm is shown.

[0076] According to an embodiment of the present invention, an improved rat swarm algorithm is used to plan evacuation routes, specifically:

[0077] S302: Using fire safety equipment to collect real-time multi-source fire safety data, the distribution of trapped people in the building is obtained. Cluster centers are set according to the safety exits of the building to perform soft clustering, and the nearest safety exit is matched for the trapped people.

[0078] S304: Generate path starting points and destination points for the trapped people based on their distribution and matching safe exits, use an improved rat swarm algorithm to plan evacuation paths, and initialize rat swarm populations by randomly obtaining evacuation paths based on the risk distribution of different venue building areas.

[0079] S306, using elite reverse learning to process the rat population to expand the diversity of the population, defining an objective function with minimum evacuation time as the goal, mapping the objective function to a fitness function, and calculating a fitness value;

[0080] S308, selecting the mouse individual corresponding to the minimum fitness value as the local optimal solution, updating the position of the mouse individual according to the two stages of chasing prey and attacking prey in the mouse swarm algorithm, recalculating the fitness value and updating the current and global optimal solutions;

[0081] S310, using the Levy flight strategy to update the position of the individual mouse, compare the fitness values before and after the update, if the fitness value after the update is smaller, then the updated individual mouse is taken as the optimal solution, otherwise the original optimal solution is maintained, after the iterative optimization is completed, the optimal evacuation path for different trapped people is obtained according to the optimal solution.

[0082] It should be noted that soft clustering is performed by setting cluster centers based on the building's emergency exits. Trapped individuals are matched to the nearest emergency exit based on occupant distribution and exit capacity. If the density of an emergency exit is too high or the evacuation route is too congested, individuals are intelligently assigned to other qualified emergency exits to balance the load across exits. The building's spatial information is abstracted into a mathematical network model consisting of grid blocks and a collection of evacuation channels. The rat swarm algorithm can help find the optimal evacuation path for evacuation routes within the building, ensuring rapid and safe evacuation. During the search process, rats gravitate toward the optimal path (leader). Elite reverse learning is used to initialize the population diversity during the global search process of the population enhancement algorithm, improving convergence speed. Random evacuation paths are used to initialize the rat swarm, with each rat representing a possible evacuation path. The fitness function is calculated by defining an objective function based on the minimum evacuation time. Preferably, the objective function can be expressed as a weighted combination of path length, congestion level, and safety factor. A preset number of mice are selected as the elite population. The reverse solution is calculated based on the dynamic boundary and updated. The fitness value is calculated and compared before and after the update. If the reverse solution has a smaller fitness, the reverse solution is used as the initial population. In the mouse swarm algorithm, the positions of individual mice are updated in the two stages of chasing prey and attacking prey. The Lévy flight is used to update the positions of individual mice to improve the algorithm's optimization ability and ability to escape local extremes. The Lévy flight strategy is expressed as: ,in represents the rat swarm after the Levy flight strategy is updated, Represents the rat swarm before Levi's flight strategy update. represents the random step size, Represents a random search path. Compare the fitness values of the mouse before and after the update. If the updated fitness value is smaller, use the updated mouse position as the optimal solution.

[0083] It should be noted that the optimal evacuation paths for trapped personnel in different areas of the venue building are visually marked using the three-dimensional model of the venue building. When the risk distribution in the venue building changes, the updated optimal evacuation path is replaced and displayed. When marking the evacuation path, different colors or arrows are used to indicate the path direction and priority. Dangerous areas such as fire, smoke, and collapsed structures are identified in the model to remind people to avoid them. Instruction information is generated based on the current optimal evacuation path. The instruction information is converted into corresponding modal data based on the properties of the fire safety equipment in different venue building areas. Different fire safety equipment is used to provide path instructions and navigation for trapped personnel. For example, the evacuation path is sent to the trapped person's mobile device and path navigation is performed through the mobile device. The evacuation path is converted into voice instruction information and lighting instruction information, and people are guided to evacuate along the evacuation path through voice broadcast and lighting. Visual display and path navigation of evacuation paths can significantly improve evacuation efficiency and safety.

[0084] Figure 4 A block diagram of the venue building evacuation route planning and fire safety management system based on digital operation and maintenance is shown.

[0085] The second embodiment of the present invention provides a venue building evacuation route planning and fire safety management system 4 based on digital operation and maintenance, which includes a fire safety data acquisition unit 401, a fire monitoring unit 402, an evacuation route planning unit 403, an evacuation route visualization unit 404 and a fire safety equipment management unit 405;

[0086] The fire safety data acquisition unit collects real-time multi-source fire safety data through the fire safety equipment of the venue building and performs data preprocessing;

[0087] The fire monitoring unit builds a digital twin model of the venue building based on BIM and the Internet of Things, and uses the digital twin model combined with fire emergency event instance data to build a fire monitoring model for fire detection and early warning;

[0088] The evacuation path planning unit determines the fire scene dynamic parameters based on the multi-source fire safety data and the twin data, obtains the fire dynamic characteristics of different venue building areas, obtains the key factors affecting the evacuation efficiency based on the fire dynamic characteristics of different venue building areas, and performs evacuation path planning based on the key factors in combination with the improved rat swarm algorithm to obtain the optimal evacuation path for different venue building areas;

[0089] The evacuation path visualization unit visualizes the optimal evacuation paths of the different venue building areas in a preset manner;

[0090] The fire safety equipment management unit is responsible for managing the status of fire safety equipment in the venue building, and generating path navigation instructions of different modes according to the properties of different fire safety equipment for path navigation.

[0091] The third embodiment of the present invention provides a computer-readable storage medium, which includes a venue building evacuation route planning and fire safety management method program based on digital operation and maintenance. When the venue building evacuation route planning and fire safety management method program based on digital operation and maintenance is executed by a processor, it implements the steps of the venue building evacuation route planning and fire safety management method based on digital operation and maintenance.

[0092] In the several embodiments provided in this application, it should be understood that the disclosed methods can be implemented in other ways. The embodiments described above are merely illustrative. For example, the division of the units is merely a logical function division. In actual implementation, there may be other division methods, such as: multiple units or components can be combined, or can be integrated into another system, or some features can be ignored or not executed. In addition, the coupling, direct coupling, or communication connection between the components shown or discussed can be through some interfaces, indirect coupling or communication connection of devices or units, which can be electrical, mechanical or other forms.

[0093] Those skilled in the art will appreciate that all or part of the steps of the above-mentioned method embodiments may be implemented by hardware associated with program instructions, and the aforementioned program may be stored in a computer-readable storage medium. When the program is executed, the program executes the steps of the above-mentioned method embodiments. The aforementioned storage medium includes various media that can store program codes, such as mobile storage devices, read-only memories (ROMs), random access memories (RAMs), magnetic disks, or optical disks.

[0094] Alternatively, if the integrated units described above are implemented as software functional units and sold or used as standalone products, they can also be stored on a computer-readable storage medium. Based on this understanding, the technical solutions of the embodiments of the present invention, or the portion that contributes to the prior art, can be embodied in the form of a software product. This computer software product, stored on a storage medium, includes instructions for enabling a computer device (such as a personal computer, server, or network device) to execute all or part of the methods described in various embodiments of the present invention. The aforementioned storage media include various media capable of storing program code, such as removable storage devices, ROM, RAM, magnetic disks, or optical disks.

[0095] The above description is only a specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any technician familiar with this technical field can easily think of changes or replacements within the technical scope disclosed by the present invention, which should be covered by the scope of protection of the present invention.

Claims

1. A venue building evacuation route planning and fire safety management method based on digital operation and maintenance, characterized in that: The following steps are involved: Build a digital twin model of the venue building based on BIM and the Internet of Things, and use the digital twin model combined with fire emergency event instance data to build a fire monitoring model; The digital twin model of the venue building is divided into grid blocks. The local structural features in each grid block are used to screen the pre-processed fire emergency events. The multi-source fire safety data perceived by fire safety equipment in the screened fire emergency event instance data is imported into the digital twin model to obtain twin data. Preset fire spread verification indicators and equipment status verification indicators, compare the twin data obtained through simulation with the instance data, adjust the twin model based on the comparison deviation, define fire risk indicators based on multi-source fire safety data, build a monitoring threshold database corresponding to the fire risk indicators, and generate a fire monitoring model; Fire detection and early warning are carried out by collecting real-time multi-source fire safety data through the fire safety equipment of the venue building. The fire scene dynamic parameters are determined based on the multi-source fire safety data and twin data to obtain the fire dynamic characteristics of different venue building areas; Based on the dynamic characteristics of fires in different venue building areas, the key factors affecting evacuation efficiency are obtained. Evacuation path planning is carried out based on these key factors and combined with an improved rat swarm algorithm to obtain the optimal evacuation path for different venue building areas. The rat swarm algorithm uses elite reverse learning and Levy flight strategy to optimize the population; Visually display the optimal evacuation routes of the different venue building areas in a preset manner, and use fire safety equipment to navigate the routes; Fire safety equipment in the venue building collects real-time multi-source fire safety data for fire detection and early warning. Fire scene dynamic parameters are determined based on the multi-source fire safety data and twin data, specifically: Collect multi-source fire safety data of the venue building and import it into the fire monitoring model to monitor the environmental status and equipment operation status of the venue building in real time, use the multi-source fire safety data to construct a fire safety data vector matrix, compare the fire safety data vector matrix with the monitoring threshold vector matrix, and obtain a residual vector matrix; Fire detection and early warning are performed through the positive and negative values of each residual vector in the residual vector matrix to obtain twin data of the fire monitoring model, and the twin data is used to fuse data with multi-source fire safety data to obtain multi-source fusion data; Fire dynamic data sequences are extracted from fire emergency event instance data to construct data samples. The data samples are divided into training sets and validation sets and imported into the AlexNet network. Feature extraction is performed through convolutional layers of different sizes, and activation functions are used for activation calculation. The pooling layer uses downsampling to reduce feature dimensions and achieve further feature extraction. The reduced features are imported into the fully connected layer. The fire classification results of the training samples are obtained through three layers of fully connected layers and then verified using the validation set. The current network parameters are retained when the preset performance standards are met. The last fully connected layer of the trained AlexNet network is removed, and the real-time multi-source fire safety data after the early warning is used as input. The corresponding one-dimensional feature vector is obtained through the fully connected layer, and the fire scene dynamic parameters are constructed using the one-dimensional feature vector; Based on the dynamic characteristics of fire in different venue building areas, the key factors affecting evacuation efficiency are obtained, specifically: Correlation analysis is used to obtain factors that can affect evacuation efficiency from fire emergency event instance data. The obtained factors are then subjected to feature screening using the mRMR algorithm to obtain a set of factors that affect evacuation efficiency. Taking each influencing factor in the factor set as the principal component direction in turn for principal component projection, obtaining the characteristic scatter point distribution corresponding to each influencing factor, and obtaining the characteristic parameters of each influencing factor according to the fire dynamic characteristics of different venue building areas; Conduct principal component analysis on the characteristic parameters of each influencing factor in different venue building areas, obtain the principal component influencing factors as the principal component direction for principal component projection, and obtain the characteristic scatter point distribution of different venue building areas; Calculate the similarity between the characteristic scatter point distribution of different venue building areas and the characteristic scatter point distribution corresponding to each influencing factor, and select the influencing factors that meet the preset similarity standards as the key factors of each venue building area; The venue building area evaluation is conducted based on the key factors, specifically: According to the key factors, the average parameter values corresponding to the personnel tolerance limits during evacuation of different key factors are read from the fire emergency event instance data to generate a benchmark evaluation matrix, and a risk assessment model for the venue building area is constructed through learning based on the benchmark evaluation matrix; Obtain the characteristic parameters of each key factor in different venue building areas and import them into the risk assessment model. Obtain the Euclidean distance between the characteristic parameters of the key factors and the parameter vectors in the benchmark assessment matrix. Use comparative weighted calculation to obtain the risk values of different venue building areas. , expressed as: ， in, represents the benchmark evaluation matrix, represents the matrix transpose, represents the Euclidean distance operation, Characteristic parameters representing key factors; Different venue building areas are graded according to the risk values, and the grading results are used to mark the digital twin model after grid processing. When the risk grade is greater than the preset risk level, it is deemed inaccessible.

2. The venue building evacuation route planning and fire safety management method based on digital operation and maintenance according to claim 1 is characterized in that: A digital twin model of the venue building is constructed based on BIM and the Internet of Things, specifically: Obtaining construction design data of the target venue building, using a BIM platform to construct a structural geometric digital model of the venue building based on the construction design data, and splicing and combining the structural geometric digital models to obtain a three-dimensional model of the venue building; Assign the location and attributes of firefighting equipment in the venue building to the three-dimensional model, aggregate real-time sensory data of firefighting safety equipment in the venue building through the Internet of Things to generate multi-source fire safety data, and use a data interface to integrate the multi-source fire safety data into the digital twin model of the venue building, thereby achieving a connection between the physical space and the twin space and generating a digital twin model of the venue building; By matching, aligning and digitally mapping multi-source fire safety data with the digital twin model, fire emergency event instance data that meets the similarity standards is retrieved based on the structural characteristics of the venue building, and the pre-processed fire emergency event instance data is used to simulate and verify the digital twin model.

3. The venue building evacuation route planning and fire safety management method based on digital operation and maintenance according to claim 1 is characterized in that: Obtain the fire dynamic characteristics of different venue building areas, specifically: The fire scene dynamic parameters of different venue building areas are obtained based on the digital twin model of the venue building after grid processing. The fire scene dynamic parameters of different venue building areas are spatially modeled by the graph neural network. The association between the building structure similarity and channel connection setting areas of different venue building areas is used to obtain the spatial dependency information of the fire scene dynamic parameters based on the association between different venue building areas using the graph neural network; In the graph convolution layer of the graph neural network, the state of each venue building area is updated through the information aggregation mechanism, and the spatial structure embedding matrix of the fire scene dynamic parameters is constructed to obtain a spatial embedding vector sequence; The spatial embedding vector sequence output by the graph neural network is input into the LSTM network. The relationship between the key features in the spatial embedding vector sequence is obtained through the forget gate and memory cell state update, and the fire dynamic characteristics of the dynamic parameters of the fire scene in different venue building areas that change with time and space are output.

4. The venue building evacuation route planning and fire safety management method based on digital operation and maintenance according to claim 1 is characterized in that: The improved rat swarm algorithm is used for evacuation path planning, specifically: Fire safety equipment collects real-time multi-source fire safety data to obtain the distribution of trapped people in buildings and venues. Soft clustering is performed by setting cluster centers based on the safety exits of the buildings and venues to match the trapped people with the nearest safety exits. Based on the distribution of trapped people and the matching safe exits, the starting and target points of the paths are generated for the trapped people. The improved rat swarm algorithm is used for evacuation path planning. The evacuation paths are randomly obtained through the risk distribution of different venue building areas to initialize the rat swarm population. Using elite reverse learning to process the rat population to expand the diversity of the population, defining an objective function with minimum evacuation time as the goal, mapping the objective function to the fitness function, and calculating the fitness value; The mouse with the minimum fitness value is selected as the local optimal solution. The position of the mouse is updated according to the two stages of chasing prey and attacking prey in the mouse swarm algorithm. The fitness value is recalculated and the current and global optimal solutions are updated. The position of the individual mouse is updated using the Levy flight strategy, and the fitness values before and after the update are compared. If the fitness value after the update is smaller, the updated individual mouse is taken as the optimal solution, otherwise the original optimal solution is retained. After the iterative optimization is completed, the optimal evacuation path for different trapped people is obtained based on the optimal solution.

5. The venue building evacuation route planning and fire safety management method based on digital operation and maintenance according to claim 1 is characterized in that: The optimal evacuation routes for the different venue building areas are visualized in a preset manner, and fire safety equipment is used for route navigation, specifically: Use the 3D model of the venue building to visually mark the optimal evacuation paths for trapped people in different venue building areas. When the risk distribution in the venue building changes, the updated optimal evacuation path will be replaced and displayed; Generate indication information based on the current optimal evacuation path, convert the indication information into corresponding modal data based on the properties of fire safety equipment in different venue building areas, and use different fire safety equipment to provide path guidance for trapped people.

Images