Search

CN-121999402-A - Urban office building fire personnel turn-back behavior evacuation efficiency analysis system and method

CN121999402ACN 121999402 ACN121999402 ACN 121999402ACN-121999402-A

Abstract

The invention discloses a system and a method for analyzing the evacuation efficiency of a foldback behavior of fire personnel in an urban office building, which belong to the field of emergency safety management of building and particularly comprise the steps of collecting video and environmental parameter data in the fire evacuation process in real time through a video monitoring unit and a sensor unit, sending the video and environmental parameter data to a behavior analysis module for identifying foldback behavior and extracting characteristics after cleaning and fusing through a preprocessing unit, and an evacuation efficiency evaluation module for evaluating the evacuation efficiency by combining the foldback behavior characteristics, the environmental parameters and the evacuation time.

Inventors

  • KANG YONG
  • XIA XIAOXUE
  • CHENG ZHIYUAN
  • MA SHUYE

Assignees

  • 北京石油化工学院

Dates

Publication Date
20260508
Application Date
20240419

Claims (9)

  1. 1. The system for analyzing the evacuation efficiency of the retracing behaviors of fire workers in the urban office building is characterized by comprising a data acquisition and processing module, a behavior analysis module, an evacuation efficiency evaluation module, an intelligent optimization module and an innovation application module; The data acquisition and processing module is used for acquiring real-time data in the evacuation process and processing the original data; The behavior analysis module is used for analyzing the foldback behavior of the personnel and the influence factors thereof in the evacuation process; The evacuation efficiency evaluation module is used for evaluating evacuation efficiency; The intelligent optimization module is used for intelligently analyzing the turn-back behavior in the evacuation process, dynamically adjusting the evacuation scheme and providing optimization suggestions; the innovative application module is used for integrating knowledge and technologies in different fields into the evacuation efficiency analysis system.
  2. 2. The urban office building fire personnel foldback behavior evacuation efficiency analysis system according to claim 1, wherein the data acquisition and processing module comprises a video monitoring unit, a sensor unit, a preprocessing unit and a data storage unit, and the behavior analysis module comprises a foldback behavior identification unit, a behavior feature extraction unit and a behavior influence factor analysis unit; The system comprises a video monitoring unit, a sensor unit, a preprocessing unit, a data storage unit, a behavior feature extraction unit, a behavior influencing factor analysis unit and a data storage unit, wherein the video monitoring unit is used for capturing video data in an evacuation process in real time through a camera installed in an office building, the sensor unit is used for monitoring fire disaster multidimensional information in real time through an integrated sensor, the preprocessing unit is used for cleaning collected original data, removing noise and abnormal values and fusing data from different sources to form fused data, the data storage unit is used for storing the fused data in a database, the foldback behavior recognition unit is used for recognizing foldback behaviors in the evacuation process through video analysis and personnel positioning data and classifying and quantifying the foldback behaviors, the behavior feature extraction unit is used for extracting foldback behavior features, and the behavior influencing factor analysis unit is used for analyzing factors influencing the foldback behaviors.
  3. 3. The urban office building fire personnel turn-back behavior evacuation efficiency analysis system according to claim 2, wherein the evacuation efficiency evaluation module comprises an evacuation time calculation unit, an evacuation efficiency evaluation unit and a scenario simulation and prediction unit, the intelligent optimization module comprises an evacuation strategy optimization unit, a real-time decision support unit and an interactive training unit, and the innovation application module comprises a coordination unit, an early warning unit and a visual display unit; The system comprises an evacuation time calculation unit, an evacuation efficiency evaluation unit, a scenario simulation and prediction unit, an evacuation strategy optimization unit, a real-time decision support unit, an interactive training unit, a pre-warning unit, a visual display unit and an early warning unit, wherein the evacuation time calculation unit is used for precisely calculating evacuation required time by using an advanced personnel positioning technology and a path planning algorithm and predicting an evacuation bottleneck region, the evacuation efficiency evaluation unit is used for combining evacuation time and retrace behavior characteristics to construct an evacuation efficiency evaluation model and analyze evacuation efficiency influence factors, the scenario simulation and prediction unit is used for predicting personnel evacuation conditions under different fire scenes by using a scenario simulation technology, the evacuation strategy optimization unit is used for intelligently generating an optimized evacuation strategy according to evacuation efficiency evaluation results and scenario simulation prediction, the real-time decision support unit is used for integrating real-time data monitoring, behavior analysis, efficiency evaluation and optimization strategies to provide real-time decision support, the interactive training unit is used for simulating fire evacuation scenes by using a virtual reality technology to provide interactive training, the cooperative unit is used for realizing interconnection and intercommunication with an emergency management system, the pre-warning unit is used for constructing an intelligent pre-warning model based on real-time data and historical data, and the visual display unit is used for displaying analysis results by using a visual technology.
  4. 4. The system for analyzing the evacuation efficiency of the turn-back behavior of fire personnel in the urban office building according to claim 3, wherein the specific workflow of each unit corresponding to the data acquisition and processing module, the behavior analysis module, the evacuation efficiency evaluation module, the intelligent optimization module and the innovation application module comprises the following steps: The method comprises the steps that a video monitoring unit in a data acquisition and processing module captures video data in an evacuation process in real time through a camera installed in an office building, a sensor unit monitors fire conditions in real time through a temperature sensor, a smoke sensor, a humidity sensor, a pressure sensor and a biological sensor, the positions, the moving conditions and the psychological states of personnel are recorded, the video data captured by the video monitoring unit and parameter data acquired by the sensor unit are transmitted to a preprocessing unit, the preprocessing unit receives the video data captured by the video monitoring unit and the parameter data acquired by the sensor unit, a machine learning algorithm is used for automatically cleaning and removing abnormal values and noise, data from different sources are fused to form fusion data, and after preprocessing is completed, the fusion data are stored in a data storage unit, and the fusion data are transmitted to a behavior analysis module, an evacuation efficiency evaluation module and an intelligent optimization module; The behavior analysis module analyzes the fusion data, the foldback behavior recognition unit invokes video monitoring data from the data storage unit, recognizes foldback behaviors in the evacuation process by utilizing an image recognition and machine learning algorithm, transmits recognition results to the behavior feature extraction unit, further analyzes the foldback behaviors, extracts foldback behavior features, and sends the foldback behavior feature data to the behavior influence factor analysis unit, wherein the behavior influence factor analysis unit synthesizes the feature data and the environment parameter data of the foldback behaviors, analyzes factors influencing the foldback behaviors of products, and sends the feature data and the environment parameter data of the foldback behaviors to the evacuation efficiency evaluation unit and the scene simulation and prediction unit; An evacuation time calculation unit in the evacuation efficiency evaluation module calculates evacuation required time according to personnel movement data acquired by the sensor unit and fire development data captured by the video monitoring unit, sends evacuation required time results to the evacuation efficiency evaluation unit, and the evacuation efficiency evaluation unit evaluates the evacuation efficiency by combining evacuation time, turn-back behavior characteristics and environmental parameters, and sends evaluation results to the scene simulation and prediction unit and the intelligent optimization module; The evacuation strategy optimizing unit in the intelligent optimizing module intelligently generates an optimized evacuation strategy according to the evaluation result and the prediction result, the real-time decision support unit integrates real-time data monitoring, behavior analysis and efficiency evaluation, provides real-time decision support, sends the real-time decision to the interactive training unit, and the interactive training unit simulates a fire evacuation scene by using a virtual reality technology to provide interactive training for staff; The collaboration unit in the innovation application module performs data sharing and strategy linkage through interconnection and intercommunication with the emergency management system, the early warning unit builds an intelligent early warning model based on real-time data and historical data, and the visual display unit displays analysis results of the modules in an intuitive mode by utilizing a visual technology.
  5. 5. The urban office building fire personnel foldback behavior evacuation efficiency analysis system according to claim 4, wherein the behavior feature extraction unit extracts the features of the foldback behavior by adopting a strategy combining space-time feature extraction and behavior pattern analysis, extracts the space-time features of the foldback behavior from the video sequence through the 3D convolutional neural network, performs clustering, and encodes the clustering result into a vector form.
  6. 6. The urban office building fire personnel turn-back behavior evacuation efficiency analysis system according to claim 5, wherein the scenario simulation and prediction unit adopts a comprehensive simulation and bottleneck analysis strategy, and the specific steps include: A1, inputting historical data and real-time data into a scene simulation and prediction unit, setting different evacuation scenes according to actual conditions, and determining the number of simulated personnel, behavior rules and environment settings; A2, constructing a fluid dynamic model according to data preparation and scene setting, and setting a model initial state; a3, running a simulation algorithm, simulating evacuation processes in different scenes, and recording personnel flow, density and speed in the simulation process, wherein the formula is as follows: Where a represents crowd density, p represents pressure in the evacuation scene, The pressure correction coefficient is indicated as such, Represents the gradient, pi represents the viscosity of the fluid, Indicating the fluid viscosity influencing factor, v indicating the crowd flow velocity, g indicating the gravitational acceleration, A correction factor representing gravitational acceleration, F representing external forces in an evacuation scenario, Represents the external force intensity in the evacuation scene, H represents the density correction term in the simulation process, Correction coefficients representing density correction terms; A4, setting the maximum personnel flow of the outlet area as R max , analyzing the simulation result, identifying the bottleneck area in the evacuation process, comparing the flow, density and speed parameters of different outlet areas, and judging whether the outlet area is the bottleneck area or not, wherein the calculation formula is as follows: Wherein R n represents the flow rate of people in the nth exit area, n represents the number of exit areas, alpha n represents the crowd density in the nth exit area, v n represents the crowd moving speed in the nth exit area, S n represents the area of the nth exit area, f 1 (θ n ,δ n ,T n ) and Representing a combined function, θ n representing a crowd moving direction for an nth exit area, δ n representing an obstacle density for the nth exit area, T n representing crowd moving time for the nth exit area, λ n representing personnel movement capability for the nth exit area, The influence of the external environment on the personnel behaviors when the crowd moves in the nth exit area is represented, gamma represents the influence of age on the personnel behaviors, and t represents the influence of the early and late peaks on the personnel behaviors; A5, if R n ≥R max , the exit area is a bottleneck area, analyzing problems and reasons existing in the bottleneck area, predicting evacuation problems and risks based on simulation results and bottleneck analysis, and outputting prediction results.
  7. 7. An urban office building fire personnel foldback behavior evacuation efficiency analysis method implemented based on the urban office building fire personnel foldback behavior evacuation efficiency analysis system according to any one of claims 1-6, comprising: S1, introducing edge computing equipment at a camera and a sensor, collecting behavior data and psychological state data, building information data and environmental parameter data of personnel in the evacuation process in real time through the camera and the sensor, and preprocessing the collected data to obtain fusion data; S2, training the fusion data by using a deep learning model and an adaptive learning algorithm, extracting the characteristic of the reentry behavior, and analyzing factors influencing the reentry behavior by combining psychology and behavioral knowledge; S3, constructing a simulation system for automatically learning and optimizing an evacuation strategy based on a deep learning and reinforcement learning technology, performing evacuation simulation for multiple times according to the fusion data and the analysis result of the foldback behavior, automatically calculating an evacuation efficiency index, and intelligently generating an evacuation strategy optimization scheme according to an evacuation efficiency evaluation result and a transfer learning technology; s4, verifying the effectiveness of an optimization scheme through simulation, applying the optimized evacuation strategy to an actual office building, organizing staff for evacuation exercise, collecting actual data in the exercise process, analyzing the behavior of the staff in evacuation, and adjusting and optimizing the evacuation strategy according to actual application feedback; S5, constructing an intelligent evacuation indication system according to an evacuation strategy, building a real-time linkage mechanism between the intelligent evacuation indication system and a fire protection system, transmitting fire information to the intelligent evacuation indication system in real time by the fire protection system when a fire disaster occurs, rapidly generating an evacuation indication according to the fire information, transmitting the evacuation indication to personnel in a display screen and voice prompt mode, and updating the evacuation indication in real time by the system according to the fire disaster development condition.
  8. 8. The method for analyzing the evacuation efficiency of the return behavior of fire personnel in an urban office building according to claim 7, wherein the specific step of S2 comprises the steps of: S201, marking the collected fusion data X= { X i ,y j ,z k ,m l } by using region segmentation or key point marking, and dividing the marked fusion data into a training set, a verification set and a test set, wherein X i represents behavior data of personnel in an evacuation process, y j represents psychological state data of the personnel in the evacuation process, z k represents building information data, m l represents environment parameter data, i represents the quantity of the behavior data, j represents the quantity of the psychological state data, k represents the quantity of the building information data, and l represents the quantity of the environment parameter data; S202, constructing a deep learning model architecture, determining the number of network layers, the number of neurons and an activation function, and adding a self-adaptive learning algorithm and an incremental learning mechanism into the deep learning model; S203, training the deep learning model by using a training set, adjusting model parameters by using a back propagation algorithm and an optimizer, and adjusting super parameters and optimizing model performance by using a verification set; s204, after training is completed, automatically extracting the characteristics of the foldback behavior from the input data by using forward propagation of the model; S205, analyzing the characteristic of the reentry behavior by combining psychological and behavioral knowledge, and analyzing the factors influencing the reentry behavior by using a statistical analysis method.
  9. 9. The method for analyzing the evacuation efficiency of the return behavior of fire personnel in the urban office building according to claim 8, wherein the evacuation efficiency index in S3 comprises evacuation time, evacuation distance and evacuation speed.

Description

Urban office building fire personnel turn-back behavior evacuation efficiency analysis system and method Technical Field The invention belongs to the field of building emergency safety management, and particularly relates to a system and a method for analyzing the evacuation efficiency of the foldback behavior of fire personnel in an urban office building. Background Along with the acceleration of the urban process, the number of high-rise office buildings is continuously increased, and accordingly, higher requirements on the safety of the office buildings are brought along, wherein a fire disaster is one of main safety threats facing the office buildings, and when the fire disaster occurs, how to effectively evacuate personnel in the buildings and reduce casualties becomes a problem to be solved urgently. The traditional evacuation methods often depend on broadcast notification, emergency lighting and indication marks, however, the methods have certain limitations in practical application, such as the broadcast notification may be interfered by noise of a fire scene, so that information is not clear, the emergency lighting and the indication marks may fail due to power interruption, and the methods often neglect psychological and behavioral characteristics of personnel in the evacuation process, so that the evacuation efficiency is not high, therefore, a system and a method capable of analyzing the foldback behavior of the personnel in the fire evacuation process and optimizing the evacuation efficiency according to the foldback behavior are developed, and the method has important practical application value. The invention discloses a simulation method, an evacuation early warning method and a system for personnel evacuation process under fire, wherein the evacuation early warning method obtains corresponding samples by detecting fire sources and prompts personnel to evacuate according to evacuation routes corresponding to the samples, the samples are obtained by the simulation method for the personnel evacuation process under the fire, the simulation method comprehensively considers correlation among building fire spread, structural failure and personnel evacuation in the simulation process and simulates the evacuation routes corresponding to the fire sources according to preset fire sources, the evacuation routes adopted in the invention are obtained according to actual simulation conditions, and the risk uncertainty of fire spread, the resistance uncertainty of structural failure and the environmental uncertainty of personnel evacuation are considered, so that the personnel safety under the condition of real-time evacuation efficiency and structural failure uncertainty, especially the safety of fire rescue people can be improved. The prior art has the following problems that 1) although the evacuation early warning method considers the correlation among fire spread, structural failure and personnel evacuation, the personnel returning behavior in the fire is not deeply analyzed, 2) although the evacuation route is obtained according to the actual simulation situation, if the personnel behavior pattern of the returning behavior is not carefully analyzed, the obtained evacuation route is not optimal, 3) only the risk uncertainty of the fire spread and the resistance uncertainty of the structural failure are considered, the uncertainty of the personnel behavior is not fully considered, 4) the evacuation route is formulated depending on the preset fire source and the simulation result, and the flexibility and the adaptability are lacking when the complex and changeable situation in the actual fire is faced. Disclosure of Invention Aiming at the defects of the prior art, the invention provides a system and a method for analyzing the evacuation efficiency of the foldback behavior of fire personnel in urban office buildings, which are characterized in that edge computing equipment is introduced to collect and preprocess multi-mode data in the evacuation process in real time, a deep learning model and a self-adaptive learning algorithm are utilized to extract foldback behavior characteristics and analyze influence factors, a simulation system is constructed based on the deep learning and the reinforcement learning, an evacuation strategy is automatically optimized, an optimization scheme is generated through migration learning, after simulation verification, the simulation system is applied to the actual office buildings, data is collected through evacuation exercise, an intelligent evacuation indication system is developed according to a feedback adjustment strategy, the intelligent evacuation indication system is linked with a fire protection system in real time, and evacuation indications are rapidly generated and updated according to the fire situation, so that the evacuation efficiency is improved. In order to achieve the above purpose, the present invention provides the following technical solutions: the s