A fire remote monitoring system with fire replay

Abstract

The present invention belongs to the field of remote monitoring technology and discloses a fire remote monitoring system including fire replay, comprising: a data processing module, a path fitting module, a response analysis module, a state judgment module, and a mapping correction module; the present invention provides more dimensional fire spatiotemporal information by obtaining spatial configuration parameters and time series records of the fire stage, and after marking the key nodes of the fire, identifies key events and reflects the fire process and spatial expansion; time series restoration and spatial mapping optimize retrospective analysis and improve the accuracy of fire cause tracing; real-time recording and synchronous data transmission ensure timely acquisition of on-site information, assist the central platform in multi-dimensional analysis and judgment, improve the timeliness and accuracy of decision support, improve the timeliness of data collection and transmission, enhance the early recognition and response speed of fire, conduct in-depth analysis and response to fire, and optimize the accuracy of fire replay and alarm processes.

Description

Technical Field

[0001] The present invention belongs to the technical field of remote monitoring, and in particular relates to a fire remote monitoring system including fire replay. Background Art

[0002] The field of remote monitoring technology involves real-time data collection, transmission, display, and recording of distant targets or areas. The core content is to obtain the status information of the monitored object through sensing devices, and rely on the communication network to remotely transmit the relevant data to the central control terminal, and the central terminal processes, presents, and archives the obtained information. This technical field is widely used in multiple scenarios such as industrial equipment operation monitoring, environmental parameter monitoring, and safety alarm systems. It mainly consists of front-end sensing and collection, communication transmission, and back-end control and display. Remote monitoring systems often combine wireless transmission, sensor technology, and data processing to achieve centralized management across regions and multiple terminals. They have real-time and visualization features and are an important technical path for information acquisition and status control.

[0003] The fire remote monitoring system, including fire replay, refers to a remote monitoring method for fire alarms and data backtracking management in buildings or key areas. This system focuses on restoring the historical state of a fire scene and managing the alarm process, including data recording, image capture, and alarm signal response timing before, during, and after a fire. Thermal sensors collect temperature information, smoke detectors measure combustible particle concentrations, and image recording devices continuously preserve visual images. Wireless communication is used to synchronize all types of data to a remote platform, which then categorizes and restores the time sequence according to pre-set logical processes, assisting in technical analysis of the fire process and tracing its causes.

[0004] Although existing technologies have basic front-end sensing, collection, and communication transmission capabilities in fire monitoring systems, they still have deficiencies in terms of the spatiotemporal characteristics of the fire process, the restoration of historical states, and the time-series processing of alarm responses. In the process of data collection, existing systems mainly rely on a single sensing method, such as temperature monitoring and smoke detection. This method is unable to obtain multi-dimensional data on fires in multiple dimensions, especially in terms of in-depth analysis of spatial layout and time series. This makes it difficult to fully restore the process of fire occurrence during fire replay. Due to the lack of accurate labeling of the fire development process and identification of key nodes, existing technologies are often unable to accurately determine the cause and expansion path of a fire after it occurs. In addition, the alarm system of existing technologies is mainly based on a single response mode and cannot efficiently conduct multi-dimensional backtracking and analysis of complex fire events. As a result, it is difficult to achieve in-depth analysis and rapid response to fire events in some cases, which may delay the handling of the fire and affect the overall rescue efficiency and safety. Summary of the Invention

[0005] The technical problem to be solved by the present invention is to provide a fire remote monitoring system including fire replay, which can solve the technical problems raised in the background technology, can trace back and conduct in-depth analysis of the fire occurrence process, and optimize the accuracy of the fire replay and alarm process.

[0006] In order to solve the above technical problems, the present invention provides the following technical solutions:

[0007] A fire remote monitoring system including fire replay, the system comprising:

[0008] The data processing module obtains the spatial configuration parameters and time series event records of the fire occurrence stage through the three-dimensional building information model, and marks the key fire nodes of the monitoring data, generates the key fire node data and transmits it to the path fitting module;

[0009] The path fitting module recreates the propagation path of key fire node data and generates the expected response time. Combining spatial connectivity, it uses the Dijkstra path algorithm to fill in the missing jump nodes and time, generates a fire propagation recurrence path set, and passes it to the response analysis module.

[0010] The response analysis module obtains the device response log and trigger time records during remote monitoring based on the fire propagation recurrence path set, calculates the difference between the starting time of each propagation segment and the device response trigger time, filters out the response missing nodes, generates a list of response anomaly records, and passes it to the status judgment module;

[0011] The status judgment module obtains and matches the corresponding floor status transfer log according to the response exception record list, judges the floor status behavior and expected response time through the finite state machine model, filters the abnormal period path, and outputs the floor preparation status exception list.

[0012] The following is a further optimization of the above technical solution by the present invention:

[0013] The fire key node data includes spatial configuration parameters, time series event records, and node information of the fire occurrence stage. The fire propagation reproduction path set includes the propagation path sequence, time series information, jump node filling path, and time gap filling path. The response abnormality record list includes the propagation segment number, equipment response time, abnormal node number, and response missing node. The floor preparation status abnormality list includes the abnormal period path, time offset segment, and abnormality type classification label.

[0014] Further optimization: the data processing module includes:

[0015] The spatial configuration extraction submodule obtains the structural coordinate data, floor numbers, spatial boundaries, and evacuation paths from the 3D building information model to reconstruct the spatial topology. It also numbers and locates the coordinates of the nodes that constitute the evacuation path, calculates the connectivity between multiple spatial nodes, and generates spatial connectivity intervals.

[0016] The event sequence annotation submodule obtains sensor trigger records, image frame sequences, and smoke density during the fire period based on spatial connectivity intervals, annotates alarm level change segments and image brightness mutation frame sequences, and generates trigger time sequencing;

[0017] The node sequence output submodule obtains spatial coordinate information and numbers according to the trigger time, extracts the mutation point sequence and alarm segmentation on the time axis, combines and analyzes the sequence structure composed of spatial node numbers and evacuation paths, and integrates them to generate fire key node data.

[0018] Further optimization: The path fitting module includes:

[0019] The node path reconstruction submodule obtains the key fire node data and the corresponding time tag data, combines spatial connectivity, and filters the path interruption nodes based on the spatial topological relationship between continuous nodes to generate node continuous segment numbers;

[0020] The jump segment completion submodule identifies the location and time missing points of the jump segment based on the node continuous segment number and the fire key node data, calculates the shortest space-time path parameters within the jump segment through the Dijkstra path algorithm, completes the time and path segment structure of the missing points, and generates the jump path filling amount;

[0021] The path sequence generation submodule integrates and completes the path segments based on the jump path filling amount, combines the time information of multiple segments in the node continuous segment numbering value with the node sorting structure to number them, and generates a fire propagation recurrence path set.

[0022] Further optimization: The Dijkstra path algorithm is used to calculate the shortest space-time path parameters within the jump segment , using the formula:

[0023] ;

[0024] in, Represents the timestamp difference between node i and node j, Represents the product coefficient of fire spread rate and spatial grid resolution, represents the original value of the curvature correction factor of the kth candidate path, represents the sum of all candidate path curvature correction factors, represents the original Euclidean distance of the kth candidate path from node i to node j, k∈[1,n] represents different path choices, Represents the maximum Euclidean space distance in the candidate path set, represents the normalized smoke concentration gradient between node i and node j, Represents a slave node To Node The total number of candidate paths.

[0025] Further optimization: The response analysis module includes:

[0026] The log data extraction submodule obtains the starting time of each path segment in the fire propagation recurrence path set, and simultaneously obtains the device response log and sensor trigger time during the remote monitoring period. It extracts the trigger time of each response node and matches and integrates it with the path segment association number to generate a response time comparison;

[0027] The response time calculation submodule obtains the starting time of multiple path segments based on the response time comparison, calculates the time difference between the starting time of the multiple path segments and the trigger time of the corresponding device, establishes a mapping sequence between the node number and the time difference, sorts and classifies the intervals, and generates the response delay interval;

[0028] The abnormal node screening submodule, based on the response delay interval, screens the node numbers whose time difference is missing or whose response time lag exceeds the response tolerance threshold, and records the corresponding response missing segment and the associated device number to generate a response abnormality record list;

[0029] The response tolerance threshold is set by statistically analyzing the sensor response time in fire events of similar scenarios and combining it with the standard deviation interval.

[0030] Further optimization: the state judgment module includes:

[0031] The log matching submodule obtains the propagation segment number corresponding to the abnormal node in the response abnormal record list, obtains the floor status transfer log within the same time period, matches the corresponding floor status record according to the propagation segment number, and generates a floor status matching result;

[0032] The state analysis submodule identifies the floor state behavior sequence based on the floor state matching results, calculates the difference between each state behavior and the expected response time, calculates the state deviation degree of the floor within the specified time period through the finite state machine model, and generates the state response deviation degree;

[0033] The path screening submodule, based on the state response deviation, screens and identifies time periods where the state deviation exceeds the response state deviation judgment threshold, extracts and integrates the associated floor numbers and state sequences, and generates a list of abnormal floor preparation states.

[0034] The response state deviation judgment threshold is set by the maximum difference limit between the floor state behavior and the expected time of the response node.

[0035] Further optimization: The finite state machine model is used to calculate the state deviation degree of the floor within a specified period of time , using the formula:

[0036] ;

[0037] in, Represents the difference between the actual response time and the expected response time of the p-th state behavior, represents the weight coefficient of the p-th state transition, Represents the smoothing factor, which is the square of the number of abnormal state behaviors in the period. is the total number of state behaviors in the time period, It is the status behavior sequence number.

[0038] Further optimization: the system further includes:

[0039] Mapping correction module: obtains the abnormal period path and corresponding spatial coordinates in the floor preparation status abnormality list, and obtains the mapping path of the event node and the response node in the three-dimensional coordinate system for offset matching and behavior annotation, and generates a replay mapping correction record result set;

[0040] The replay mapping correction record result set includes the response node position, three-dimensional coordinate mapping path, and offset behavior annotation.

[0041] Further optimization: the mapping correction module includes:

[0042] The spatial path extraction submodule extracts the node structure sequence and floor spatial coordinates under the abnormal path number in the floor preparation status abnormality list based on the abnormal path number, selects and sorts the spatial three-dimensional coordinates corresponding to each node, and generates an abnormal path spatial coordinate set;

[0043] The 3D mapping and matching submodule obtains the 3D coordinates of the event node and the response node based on the abnormal path spatial coordinate set, compares the path vectors using the spatial reconstruction method, calculates the spatial offset vector of each abnormal path segment, and generates the path offset matching interval;

[0044] The behavior offset annotation submodule extracts the path segment numbers whose spatial offset values ​​exceed the spatial offset annotation benchmark value based on the path offset matching interval, identifies the position index of the corresponding node in the path structure, annotates the behavior offset type label, and generates a replay mapping correction record result set;

[0045] The spatial offset annotation reference value is set by an acceptable spatial error limit value between a single-node sensing device and an actual control object in a building fire response system.

[0046] The present invention adopts the above technical solution, which has at least the following beneficial effects:

[0047] The present invention can provide more detailed spatiotemporal information for fire monitoring by accurately obtaining the spatial configuration parameters and time series event records of the fire occurrence stage. After the monitoring data is labeled with the key fire nodes, the events that are of key significance in the fire occurrence process can be clearly identified, and the time process and spatial expansion of the fire can be accurately reflected. Through accurate time series restoration and spatial layout mapping, the backtracking and analysis of the fire occurrence process are optimized, and the accuracy and effectiveness of tracing the cause of the fire are improved. In addition, the real-time recording of key fire nodes and the synchronous transmission of data can obtain important on-site information in a timely manner, thereby helping the central platform to conduct multi-dimensional analysis and judgment of the fire situation, and ensure timely and accurate decision-making support in the early stages of the accident. This innovative solution effectively improves the timeliness of data collection and transmission, which significantly enhances the early identification and processing response speed of fires, and ultimately achieves in-depth analysis and rapid response to fire events, and optimizes the accuracy of fire review and alarm processes. BRIEF DESCRIPTION OF THE DRAWINGS

[0048] Figure 1 A flowchart of a system in an embodiment of the present invention;

[0049] Figure 2 2 is a structural block diagram of the system in an embodiment of the present invention. DETAILED DESCRIPTION

[0050] In the embodiments of the present invention, words such as "exemplarily" and "for example" are used to indicate examples, illustrations or explanations; any embodiment or design described as "example" in the present invention should not be interpreted as being more preferred or more advantageous than other embodiments or designs; to be precise, the use of the word "example" is intended to present concepts in a concrete way; in addition, in the embodiments of the present invention, the meaning expressed by "and / or" can be both, or it can be either of the two.

[0051] In the embodiments of the present invention, “image” and “picture” may sometimes be used interchangeably. It should be noted that when the distinction between them is not emphasized, the meanings they intend to express are consistent; “of”, “corresponding” and “corresponding” may sometimes be used interchangeably. It should be noted that when the distinction between them is not emphasized, the meanings they intend to express are consistent.

[0052] In the embodiments of the present invention, sometimes a subscript such as W1 may be written as a non-subscript such as W1. When the difference is not emphasized, the meanings to be expressed are the same.

[0053] In order to make the objectives, technical solutions and advantages of the present invention more clearly understood, the present invention is further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention.

[0054] See also Figure 1 A fire remote monitoring system including fire replay includes:

[0055] The data processing module obtains the spatial configuration parameters and time series event records of the fire occurrence stage through the three-dimensional building information model, and marks the key fire nodes of the monitoring data, generates the key fire node data and transmits it to the path fitting module;

[0056] The path fitting module recreates the propagation path of key fire node data and generates the expected response time. Combining spatial connectivity, it uses the Dijkstra path algorithm to fill in the missing jump nodes and time, generates a fire propagation recurrence path set, and passes it to the response analysis module.

[0057] The response analysis module obtains the device response log and trigger time records during remote monitoring based on the fire propagation recurrence path set, calculates the difference between the starting time of each propagation segment and the device response trigger time, filters out the response missing nodes, generates a list of response anomaly records, and passes it to the status judgment module;

[0058] The state judgment module obtains and matches the corresponding floor state transfer log based on the response exception record list, judges the floor state behavior and expected response time through the finite state machine model, filters the abnormal period path, and outputs the floor preparation state abnormality list;

[0059] Mapping correction module: obtains the abnormal period path and corresponding spatial coordinates in the floor preparation status abnormality list, and obtains the mapping path of the event node and the response node in the three-dimensional coordinate system for offset matching and behavior annotation, and generates a replay mapping correction record result set.

[0060] The key fire node data includes spatial configuration parameters, time series event records, and node information of the fire occurrence stage. The fire propagation reproduction path set includes the propagation path sequence, time series information, jump node filling path, and time gap filling path. The response abnormality record list includes the propagation segment number, equipment response time, abnormal node number, and response missing node. The floor preparation status abnormality list includes the abnormal period path, time offset segment, and abnormal type classification label. The replay mapping correction record result set includes the response node position, three-dimensional coordinate mapping path, and offset behavior annotation.

[0061] See also Figure 2 , the data processing module includes:

[0062] The spatial configuration extraction submodule obtains the structural coordinate data, floor numbers, spatial boundaries and evacuation paths in the three-dimensional building information model to reconstruct the spatial topology, number and locate the coordinates of the nodes constituting the evacuation path, calculates the connectivity between multiple spatial nodes, and generates spatial connectivity intervals.

[0063] When extracting structural coordinates from a 3D building information model, call Revit API to obtain the coordinate set of the beam and column endpoints, including the coordinates of the endpoints of the single floor frame beam. , Components with adjacent coordinate distances less than 2 meters are classified as the same evacuation. After analyzing the BIM model of a single shopping mall, 327 spatial nodes are obtained. The breadth-first algorithm is used to traverse adjacent nodes. When there is a direct path between two nodes and the path length is less than 15 meters, a connection relationship is established and the connectivity of node i is calculated. ;in: is the number of directly connected nodes, is the total number of nodes; the single security exit node number is N05 , the total number of nodes is 327, then , divide the intervals into 0-0.3 as low, 0.3-0.6 as medium, and 0.6-1.0 as high, and generate a connectivity distribution map.

[0064] Table 1 shows an example of typical node connectivity:

[0065] Node number Number of connected nodes Total number of nodes Connectivity Interval Level N05 8 327 0.0245 Low C12 112 327 0.343 middle E09 215 327 0.658 high

[0066] As shown in Table 1, the effectiveness of the connectivity calculation method is verified by the current project data, among which node E09 is judged as a high-connectivity hub node because it connects 215 nodes.

[0067] The event sequence annotation submodule obtains sensor trigger records, image frame sequences, and smoke density during the fire period based on the spatial connectivity interval, annotates the alarm level change sections and the image brightness mutation frame sequences, and generates the trigger time sequence.

[0068] Set smoke concentration threshold , when the laser scattering smoke detector reading The time is marked as a level 1 alarm, and the detector reading record of the corridor on the third floor at a single moment t=32 , triggering alarm upgrade, image brightness mutation detection uses adjacent frame pixel gray value difference method, setting change rate threshold ,when Recorded as a sudden change frame, the 120th frame of a single surveillance video , frame 121 , represents the image brightness value, then , exceeding the threshold is recorded, Represents the brightness change rate between video frames. After sorting by timestamp, an event sequence is generated [32s: smoke alarm, 34.5s: video mutation, 36.2s: temperature alarm].

[0069] The node sequence output submodule obtains spatial coordinate information and numbers according to the trigger time, extracts the mutation point sequence and alarm segmentation on the time axis, combines and analyzes the sequence structure composed of spatial node numbers and evacuation paths, and integrates them to generate fire key node data.

[0070] When establishing the time-space mapping relationship, extract the node N12 (coordinate ), t=34.5s corresponds to node C07 ( ), calculate the node spacing , combined with the evacuation speed , verification time interval Can move inside Therefore, the current distance is 7.82m>3.0m, and a cross-region event is determined. The integrated sequence [N12 (32s) → C07 (34.5s) → E09 (36.2s)] reflects the path of fire spreading along the atrium to the safety exit.

[0071] See also Figure 2 , the path fitting module includes:

[0072] The node path reconstruction submodule obtains the key fire node data and the corresponding time label data, combines the spatial connectivity, and filters the path interruption nodes based on the spatial topological relationship between continuous nodes to generate the node continuous segment number.

[0073] Extract node N18 (timestamp from key fire node data) ,coordinate ) and N22( , ), calculate the spatial straight-line distance , according to the evacuation speed of personnel , verify the theoretical travel time , when the actual time difference When the value exceeds 20% of the theoretical value, it is determined as an interruption node and the interruption threshold is set. ,Because 5s>4.5s, N18-N22 is marked as the break point, and after traversing all nodes, continuous segments [Seg1: N01-N12], [Seg2: N15-N18], and [Seg3: N22-N27] are generated.

[0074] Table 2 is the node continuity verification table:

[0075] Starting Node Termination Node Theoretical time Actual time Difference ratio Continuity determination N15 N18 3.2 3.5 9.4% continuous N18 N22 4.5 5.0 11.1% Interrupt N22 N27 2.8 3.1 10.7% continuous

[0076] As shown in Table 2, by setting 1.2 times the theoretical time as the interruption threshold, the path break in the N18-N22 section is identified.

[0077] The jump segment completion submodule identifies the location and time missing points of the jump segment based on the node continuous segment number and fire key node data, calculates the shortest space-time path parameters within the jump segment through the Dijkstra path algorithm, completes the time and path segment composition of the missing points, and generates the jump path filling amount.

[0078] The jump index position is identified between Seg2 (N15-N18) and Seg3 (N22-N27), and N18 ( ) to N22( ) to be completed, the topological network includes three candidate paths: The timestamp difference between nodes i and j is calculated directly from the time tags recorded by the BIM system. The timestamp difference between nodes N18 (42s) and N22 (47s) is , Represents the product coefficient of fire spread rate and spatial grid resolution. According to GB50116-2013, the medium fire spread rate is taken as , BIM model grid resolution ,but , Represents a slave node To Node The total number of candidate paths, n is 3, k∈[1,n] represents different path selections, when k=1, it corresponds to the first path, when k=2, it corresponds to the second path, when k=3, it corresponds to the third path, represents the sum of all candidate path curvature correction factors, , Represents the original value of the curvature correction factor of the kth candidate path, and the calculation formula is ( is the path curvature radius), Path1 is a straight line but , Path2 curvature radius but , Path3 curvature radius but , Represents the original Euclidean distance of the kth candidate path from node i to node j, calculated by coordinate points, Path2 node sequence , Euclidean distance , Represents the maximum Euclidean space distance in the candidate path set, taking the maximum Euclidean distance among the three paths , Represents the normalized smoke concentration gradient between node i and node j, calculated by the formula , measured concentration of node i , node j concentration , system range ,but , taking Path2 as an example, substitute the shortest space-time path parameter calculation formula , step-by-step calculation 1. Time term , 2. Curvature weight calculation, denominator ,molecular ,3. Distance normalization , 4. Smoke correction item , 5. Spatial term , 6. Results ,when When compared with the benchmark value, the excellent interval (through path), qualified interval (Detour route), dangerous area (Unreachable Path) The result shows that Path2 is a good path and will be selected as the preferred path to complete the missing segment.

[0079] The path sequence generation submodule integrates and completes the path segments based on the jump path filling amount, combines the time information of multiple segments in the node continuous segment numbering value with the node sorting structure to number them, and generates a fire propagation recurrence path set.

[0080] Will complete node N19( ) is inserted into the original sequence to generate new sequences [Seg2: N15-N18-N19] and [Seg3: N22-N27], and N19 ( ) and N22( ) to verify the topology and calculate the moving direction angle , the deviation of the direction angle of -35° from the subsequent node N27 is within the tolerance range of ±5°, verifying the path continuity and outputting the complete propagation path set [P1: N01-N12], [P2: N15-N19-N22-N27].

[0081] See also Figure 2 , the response analysis module includes:

[0082] The log data extraction submodule obtains the starting time of each path segment in the fire propagation reproduction path set, and synchronously obtains the device response log and sensor trigger time during remote monitoring. It extracts the trigger time of each response node and matches and integrates it with the path segment association number to generate a response time comparison.

[0083] Extract node N15 from path set P2 ( )、N18( )、N19( ), t represents the starting time of the node, synchronizes and cancels the device log recorded by the fire alarm host, and matches the smoke detector D15 in Alarm records, video surveillance C18 The brightness mutation record is established, and the associated mapping [N15→D15(0.2s), N18→C18(0.3s)] is established. When the starting time of the path segment Seg2 Trigger time with D15 Time difference If it is less than 1s, it is considered as a valid association. ) is marked as an unassociated device.

[0084] Table 3 is a time correlation comparison table:

[0085] Path Node Node Time Equipment Number Trigger time Time Difference Association Status N15 38.0 D15 38.2 0.2 efficient N18 42.0 C18 42.3 0.3 efficient N22 47.0 - - - Missing

[0086] As shown in Table 3, 87% of the nodes were successfully associated through the time window matching algorithm (±1s), and the unassociated nodes entered the anomaly detection process.

[0087] The response time measurement submodule obtains the starting time of multiple path segments based on the response time comparison, calculates the time difference between the starting time of the multiple path segments and the trigger time of the corresponding device, establishes a mapping sequence between the node number and the time difference value, sorts and classifies the interval, and generates the response delay interval. The time difference set of valid associated nodes is calculated [0.2, 0.3, 0.7, 1.0, 0.5]s, sorted in ascending order, and the 25th percentile value is taken. , 75% percentile , set the delay interval: Normal ( )、Warning( ),abnormal( ), Represents the time difference, represents the interquartile range, and represent the first and third quartiles, respectively. , then the warning threshold , abnormal threshold , when the time difference of node N27 is measured , it falls into the abnormal range.

[0088] The abnormal node screening submodule, based on the response delay interval, screens the node numbers whose time difference is missing values ​​or whose response time lag exceeds the response tolerance threshold, and records the corresponding response missing segments and associated device numbers to generate a response abnormality record list.

[0089] The response tolerance threshold is set by statistically analyzing the sensor response time in fire events of similar scenarios and combining it with the standard deviation interval.

[0090] Calculate the mean based on the data set [0.2, 0.4, 0.3, 0.6, 0.5, 0.7]s , standard deviation , set the response tolerance threshold , screening time difference Nodes, including N27( )、N33( ), check the log records of device D27 and find that to During this period, there were three communication timeout records, and the N27-D27 pair was marked as a device failure.

[0091] See also Figure 2 , the status judgment module includes:

[0092] The log matching submodule obtains the propagation segment number corresponding to the abnormal node in the response abnormal record list, obtains the floor status transfer log within the same time period, matches the corresponding floor status record according to the propagation segment number, and generates a floor status matching result.

[0093] Extract the propagation segment Seg5 (number P5-27) associated with node N27 from the exception list, retrieve the floor status log from 10:05:00 to 10:07:00 during the fire period, and match 3 related records: 10:05:23 fire door F27 closing delay, 10:05:45 smoke exhaust valve V27 not started, 10:06:10 emergency lighting L27 insufficient brightness, bind P5-27 to the [F27, V27, L27] device status, when the propagation segment start time (10:05:47) Time difference with the F27 incident When the time interval is less than 30 seconds, it is marked as cross-period correlation and classified into the same analysis unit using the time window ±30s matching principle.

[0094] Table 4 is the log matching result table:

[0095] Transmission segment number Associated devices Event Time Status Description Time Difference P5-27 F27 10：05：23 Fire door closing delay 24 P5-27 V27 10：05：45 The exhaust valve is not activated 2 P5-27 L27 10：06：10 Insufficient lighting 23

[0096] As shown in Table 4, the cross-period events are associated with the same propagation segment through the period superposition algorithm, completing 83% of the abnormal node status binding.

[0097] The state analysis submodule identifies the floor state behavior sequence based on the floor state matching results, calculates the difference between each state behavior and the expected response time, calculates the state deviation degree of the floor within the specified time period through the finite state machine model, and generates the state response deviation degree.

[0098] Set a benchmark for expected response times, Represents the difference between the actual response time and the expected response time of the p-th state behavior, obtained by subtracting the expected time from the actual response time, including the actual response time of the exhaust valve V27 ,expected ,but , take the absolute value The absolute difference between the actual response time of the fire door (50s) and the expected time (31s) is |50-31|=19s. The absolute difference between the actual response time of the emergency lighting (e.g. 30s) and the expected time (e.g. 8s) is |30-8|=22s. Represents the weight coefficient of the pth state transition, according to the NFPA72 equipment importance classification, fire door (Life Safety Equipment), Smoke Exhaust Damper (Environmental Control Equipment), Emergency Lighting (auxiliary equipment), Represents the smoothing factor, which is the square of the number of abnormal state behaviors in the period, and the calculation formula is In this example, there are 2 abnormal states (F27 delay, L27 abnormality), so , The total number of state behaviors in the time period, obtained through log record statistics, is the state behavior number, in this case N=3, substitute into the formula ,when time, baseline interval, normal interval , warning interval , abnormal interval ,The results show that the time period state deviation exceeds the standard threshold of 2.0, a maintenance work order will be automatically generated and floor L5 will be marked as a high-risk area. The value directly determines the generation logic of the exception list if and only if The alarm process is triggered.

[0099] The path screening submodule, based on the state response deviation, screens and identifies time periods where the state deviation exceeds the response state deviation judgment threshold, extracts and integrates the associated floor numbers and state sequences, and generates a floor preparation state abnormality list.

[0100] The response state deviation judgment threshold is set by the maximum difference limit between the floor state behavior and the expected time of the response node.

[0101] Set the deviation threshold based on the maximum difference between the allowed floor state behavior and the expected time of the response node , when calculated hour, Representing the state offset value, extract the associated floor L5 (corresponding to the floor where node N27 is located) and the state sequence [F27 delay, V27 failure, L27 abnormal], and generate an abnormal list entry: L5's offset exceeds the standard during the period of 10:05-10:06, and the smoke exhaust system (V27) and the fire door linkage device (F27) need to be repaired first.

[0102] See also Figure 2 , the mapping correction module includes:

[0103] The spatial path extraction submodule extracts the node structure sequence and floor spatial coordinates under the corresponding path number based on the abnormal path number in the floor preparation status abnormality list, filters and sorts the spatial three-dimensional coordinates corresponding to each node, and generates an abnormal path spatial coordinate set.

[0104] Extract the path number P5-27 of floor L5 from the exception list, retrieve the three-dimensional coordinate index in the BIM database, and obtain the node sequence [N27, 、N28, 、N29, ], generate the coordinate set [(58.6, 25.4, 3.0), (60.1, 23.8, 3.0), (62.4, 22.1, 3.0)] by timestamp sorting, when the z-axis coordinate deviation of N28 is detected (standard floor height 3.0m), it is marked as elevation anomaly, and the screening condition is that the deviation of any coordinate axis exceeds 0.3m.

[0105] Table 5 is the spatial coordinate screening table:

[0106] Node number X coordinate Y coordinate Z coordinate Deviation Marker N27 58.6 25.4 3.0 normal N28 60.1 23.8 3.5 Z-axis anomaly N29 62.4 22.1 3.0 normal

[0107] As shown in Table 5, N28 is selected as an abnormal coordinate point by axis-by-axis comparison, completing the purification of the spatial coordinate set.

[0108] The three-dimensional mapping and matching submodule obtains the three-dimensional coordinates of the event node and the response node based on the abnormal path spatial coordinate set, compares the path vectors through the spatial reconstruction method, calculates the spatial offset vector of each abnormal path segment, and generates the path offset matching interval.

[0109] Calculate the space vector from N27 to N28: , with the standard path vector Perform dot product calculation , module length ratio calculation 、 , vector angle ,in Represents the modulus of the actual path vector, Represents the modulus of the standard path vector. When the angle exceeds 10°, it is considered a spatial offset.

[0110] The behavior offset annotation submodule extracts the path segment number whose spatial offset value exceeds the spatial offset annotation benchmark value based on the path offset matching interval, identifies the position index of the corresponding node in the path structure and annotates the behavior offset type label, and generates a replay mapping correction record result set.

[0111] The spatial offset annotation benchmark value is set by the acceptable spatial error limit value between the single-node sensing device and the actual control object in the building fire response system.

[0112] The spatial error benchmark value is set based on the acceptable spatial error limit between the single-node sensing device and the actual control object in the building fire response system. , and calculate the spatial offset of N28 (Including Z-axis deviation 0.5m and projection deviation 0.5m), when When extracting the node at index position 2 in the path (N28 is the second node), mark it as elevation offset-structural settlement type, and associate the tilt sensor data of device F28 (Exceeding the limit by 2.0°), a correction record [P5-27, N28, 0.71m, structural settlement, F28] is generated.

[0113] The above are merely preferred embodiments of the present invention and do not limit the present invention in any other form. Any technician familiar with the profession may use the technical content disclosed above to change or modify it into an equivalent embodiment with equivalent changes and apply it to other fields. However, any simple modification, equivalent change and modification made to the above embodiment based on the technical essence of the present invention without departing from the content of the technical solution of the present invention shall still fall within the scope of protection of the technical solution of the present invention.

Claims

1. A fire remote monitoring system including fire replay, characterized by: The system comprises: The data processing module obtains the spatial configuration parameters and time series event records of the fire occurrence stage through the three-dimensional building information model, and marks the key fire nodes of the monitoring data, generates the key fire node data and transmits it to the path fitting module; The path fitting module recreates the propagation path of key fire node data and generates the expected response time. Combining spatial connectivity, it uses the Dijkstra path algorithm to fill in the missing jump nodes and time, generates a fire propagation recurrence path set, and passes it to the response analysis module. The response analysis module obtains the device response log and trigger time records during remote monitoring based on the fire propagation recurrence path set, calculates the difference between the starting time of each propagation segment and the device response trigger time, filters out the response missing nodes, generates a list of response anomaly records, and passes it to the status judgment module; The state judgment module obtains and matches the corresponding floor state transfer log based on the response exception record list, judges the floor state behavior and expected response time through the finite state machine model, filters the abnormal period path, and outputs the floor preparation state abnormality list; The data processing module includes: The spatial configuration extraction submodule obtains the structural coordinate data, floor numbers, spatial boundaries, and evacuation paths from the 3D building information model to reconstruct the spatial topology. It also numbers and locates the coordinates of the nodes that constitute the evacuation path, calculates the connectivity between multiple spatial nodes, and generates spatial connectivity intervals. The event sequence annotation submodule obtains sensor trigger records, image frame sequences, and smoke density during the fire period based on spatial connectivity intervals, annotates alarm level change segments and image brightness mutation frame sequences, and generates trigger time sequencing; The node sequence output submodule obtains spatial coordinate information and numbers according to the trigger time, extracts the mutation point sequence and alarm segmentation on the time axis, combines and analyzes the sequence structure composed of spatial node numbers and evacuation paths, and integrates them to generate fire key node data; The path fitting module includes: The node path reconstruction submodule obtains the key fire node data and the corresponding time tag data, combines spatial connectivity, and filters the path interruption nodes based on the spatial topological relationship between continuous nodes to generate node continuous segment numbers; The jump segment completion submodule identifies the location and time missing points of the jump segment based on the node continuous segment number and the fire key node data, calculates the shortest space-time path parameters within the jump segment through the Dijkstra path algorithm, completes the time and path segment structure of the missing points, and generates the jump path filling amount; The path sequence generation submodule integrates and completes the path segments based on the jump path filling amount, combines the time information of multiple segments in the node continuous segment number value with the node sorting structure to generate a fire propagation recurrence path set; The response analysis module includes: The log data extraction submodule obtains the starting time of each path segment in the fire propagation recurrence path set, and simultaneously obtains the device response log and sensor trigger time during the remote monitoring period. It extracts the trigger time of each response node and matches and integrates it with the path segment association number to generate a response time comparison; The response time calculation submodule obtains the starting time of multiple path segments based on the response time comparison, calculates the time difference between the starting time of the multiple path segments and the trigger time of the corresponding device, establishes a mapping sequence between the node number and the time difference, sorts and classifies the intervals, and generates the response delay interval; The abnormal node screening submodule, based on the response delay interval, screens the node numbers whose time difference is missing or whose response time lag exceeds the response tolerance threshold, and records the corresponding response missing segment and the associated device number to generate a response abnormality record list; The response tolerance threshold is set by statistically analyzing the sensor response time in fire events of similar scenarios and combining it with the standard deviation interval; The state judgment module includes: The log matching submodule obtains the propagation segment number corresponding to the abnormal node in the response abnormal record list, obtains the floor status transfer log within the same time period, matches the corresponding floor status record according to the propagation segment number, and generates a floor status matching result; The state analysis submodule identifies the floor state behavior sequence based on the floor state matching results, calculates the difference between each state behavior and the expected response time, calculates the state deviation degree of the floor within the specified time period through the finite state machine model, and generates the state response deviation degree; The path screening submodule, based on the state response deviation, screens and identifies time periods where the state deviation exceeds the response state deviation judgment threshold, extracts and integrates the associated floor numbers and state sequences, and generates a list of abnormal floor preparation states. The response state deviation judgment threshold is set by the maximum difference limit between the floor state behavior and the expected time of the response node.

2. The fire remote monitoring system with fire replay according to claim 1 is characterized by: The fire key node data includes spatial configuration parameters, time series event records, and node information of the fire occurrence stage. The fire propagation reproduction path set includes the propagation path sequence, time series information, jump node filling path, and time gap filling path. The response abnormality record list includes the propagation segment number, equipment response time, abnormal node number, and response missing node. The floor preparation status abnormality list includes the abnormal period path, time offset segment, and abnormality type classification label.

3. The fire remote monitoring system with fire replay according to claim 2 is characterized by: The shortest space-time path parameters in the jump segment are calculated by the Dijkstra path algorithm , using the formula: ； in, Represents the timestamp difference between node i and node j, Represents the product coefficient of fire spread rate and spatial grid resolution, represents the original value of the curvature correction factor of the kth candidate path, represents the sum of all candidate path curvature correction factors, represents the original Euclidean distance of the kth candidate path from node i to node j, k∈[1,n] represents different path choices, Represents the maximum Euclidean space distance in the candidate path set, represents the normalized smoke concentration gradient between node i and node j, Represents a slave node To Node The total number of candidate paths.

4. The fire remote monitoring system with fire replay according to claim 3 is characterized by: The finite state machine model is used to calculate the state deviation degree of the floor within the specified time period. , using the formula: ； in, Represents the difference between the actual response time and the expected response time of the p-th state behavior, represents the weight coefficient of the p-th state transition, Represents the smoothing factor, which is the square of the number of abnormal state behaviors in the period. is the total number of state behaviors in the time period, It is the status behavior sequence number.

5. The fire remote monitoring system with fire replay according to claim 1 is characterized by: The system further comprises: Mapping correction module: obtains the abnormal period path and corresponding spatial coordinates in the floor preparation status abnormality list, and obtains the mapping path of the event node and the response node in the three-dimensional coordinate system for offset matching and behavior annotation, and generates a replay mapping correction record result set; The replay mapping correction record result set includes the response node position, three-dimensional coordinate mapping path, and offset behavior annotation.

6. The fire remote monitoring system with fire replay according to claim 5 is characterized by: The mapping correction module includes: The spatial path extraction submodule extracts the node structure sequence and floor spatial coordinates under the abnormal path number in the floor preparation status abnormality list based on the abnormal path number, selects and sorts the spatial three-dimensional coordinates corresponding to each node, and generates an abnormal path spatial coordinate set; The 3D mapping and matching submodule obtains the 3D coordinates of the event node and the response node based on the abnormal path spatial coordinate set, compares the path vectors using the spatial reconstruction method, calculates the spatial offset vector of each abnormal path segment, and generates the path offset matching interval; The behavior offset annotation submodule extracts the path segment numbers whose spatial offset values ​​exceed the spatial offset annotation benchmark value based on the path offset matching interval, identifies the position index of the corresponding node in the path structure, annotates the behavior offset type label, and generates a replay mapping correction record result set; The spatial offset annotation reference value is set by an acceptable spatial error limit value between a single-node sensing device and an actual control object in a building fire response system.

Images