A concealed fire optical fiber early warning method and system for an external wall fire-retardant insulation layer

Abstract

This invention discloses a fiber optic early warning method and system for concealed fires in exterior wall flame-retardant insulation layers, relating to the field of fiber optic sensing and early warning technology. It establishes a discrete graph structure reflecting the connectivity of the building interface and assigning length weights by projecting a heat-generating physical component onto the outer interface of the exterior wall flame-retardant insulation layer. The maximum temperature transition value of the temperature ranking sequence of basic temperature measurement points is extracted, and a set of thermal anomaly segments is generated by combining segmentation effectiveness and spatial continuity. A graph connectivity search algorithm is executed within the discrete graph structure, stripping temperature measurement points reached by external heat structures as interpretable parts, and obtaining effective uninterpreted areas through set difference operations. Finally, the mileage center value is calculated based on the fiber optic deployment mileage of temperature measurement points within the uninterpreted areas, and the alarm is located. This invention achieves mathematical decoupling between photovoltaic attached thermal shadows and the actual concealed fire heat core, effectively overcoming false alarms and missed alarms under multi-source heat aliasing, and solving the problem of spatial positioning distortion caused by complex facade wiring.

Description

Technical Field

[0001] This invention relates to the field of fiber optic sensing and early warning technology, specifically to a fiber optic early warning method and system for concealed fires in exterior wall flame-retardant insulation layers. Background Technology

[0002] With the widespread application of building-integrated photovoltaics (BIPV) technology, photovoltaic curtain walls or externally mounted photovoltaic modules are becoming increasingly common in building exterior wall projects. In composite facade systems composed of photovoltaic modules, module backsheets, junction boxes, DC connectors, concealed cable trays, module frames, mounting frames, rear ventilation cavities, and the underlying fire-retardant insulation layer, the spatial structure is complex and interwoven. Due to the dense electrical wiring and its constant exposure to the outdoor environment, the fire-retardant insulation layer of the exterior wall poses an extremely high fire risk. Such concealed fires typically lurk within the insulation layer, in the seams of the insulation panels, at the edges of holes, in areas where cables penetrate the wall, or between the insulation layer and the protective layer. In the early stages, they manifest as localized smoldering, low-intensity heating, or the formation of isolated localized heat cores, making it extremely difficult to detect open flames and smoke from outside the building. To detect early abnormal temperature rises, distributed temperature-sensing optical fibers are typically laid on the outer interface of the fire-retardant insulation layer in practice to achieve large-scale and long-distance linear continuous temperature monitoring.

[0003] In the actual operation of photovoltaic building composite facades, heat-generating components inevitably generate a large amount of external heat, including backsheet heating during normal photovoltaic module operation, hot spot effects caused by partial shading, junction box heating, poor DC connector contact, and heat accumulation in hidden cable trays. The heat generated by these external devices is conducted along the normal direction to the vicinity of the outer interface of the exterior wall's fire-retardant insulation layer, forming an externally attached heat shadow. Existing fiber optic fire alarm solutions mainly employ fixed-temperature alarms, differential-temperature alarms, or single-rate-of-temperature alarms. Their judgment logic highly relies on absolute temperature thresholds or time-dimensional temperature gradients, completely detached from the complex three-dimensional spatial topology of the building interior. When the fiber optic cable extracts an abnormal temperature signal from the basic temperature measurement point, existing early warning systems can only mechanically output a surface result indicating a localized temperature exceeding the limit, unable to spatially trace the heat transfer path based on the connectivity of the thermal anomaly at the building interface. Because it is difficult to decouple and separate the heat conducted by external photovoltaics from the actual hidden fire heat core inside the fire-retardant insulation layer of the exterior wall at the algorithm level, existing systems are prone to false fire alarms when photovoltaic equipment is operating under high load, or may mask actual early smoldering fires under the background of photovoltaic heating, resulting in missed alarms. In addition, given the complex wiring and routing of three-dimensional facades, traditional one-dimensional coordinate extraction methods can lead to severe distortion of the location points for hidden fire alarms, making it impossible to provide fire-fighting linkage with high-confidence heat-triggered electrical alarm signals and accurate location data. Summary of the Invention

[0004] To address the shortcomings of existing technologies, this invention proposes a fiber optic early warning method and system for concealed fires in exterior wall flame-retardant insulation layers. This method solves the problem that in photovoltaic building composite facade scenarios, the external attached heat shadow generated by the operation of photovoltaic equipment overlaps with the actual concealed fire heat core inside the exterior wall flame-retardant insulation layer. Existing single temperature threshold alarm technologies cannot trace heat sources and mathematically decouple them based on the heat transfer connectivity of the building's three-dimensional topological interface. This leads to fire early warning systems being prone to false alarms and missed alarms, and also causes alarm location distortion when wiring is folded back in complex facades.

[0005] To achieve the above objectives, the present invention provides the following technical solution: A fiber optic early warning method for concealed fires in exterior wall fire-retardant insulation layers includes: The heat-generating solid component is projected along the normal direction to the outer interface of the fire-retardant insulation layer of the exterior wall, forming an external heat structure projection set, and establishing a discrete graph structure that reflects the connectivity of the building interface and assigns length weights. The system acquires basic temperature measurement points from distributed temperature sensing fiber optic data acquisition and constructs a temperature sorting sequence based on these points. It then extracts the temperature transition amounts between adjacent step points in the temperature sorting sequence of the basic temperature measurement points, retrieves the maximum temperature transition amount and its corresponding segmentation position index, and extracts basic temperature measurement points with higher values ​​than the segmentation position index to generate a set of thermal anomaly candidate points. The system performs spatial continuity merging on the thermal anomaly candidate point set and filters out segments with fewer temperature measurement points than the minimum continuous measurement point threshold to obtain the current thermal anomaly segment set. Within the discrete graph structure, a graph connectivity search algorithm is executed on the specific thermal anomaly segments contained in the current thermal anomaly segment set. When the cumulative path length exceeds the preset maximum effective heat transfer span, the program is truncated. The basic temperature measurement point reached by the external thermal structure projection set is taken as the interpretable part. The set difference operation is performed on the specific thermal anomaly segment and the interpretable part. Based on the minimum continuous measurement point number threshold, a second spatial continuity segmentation is performed to obtain the effective uninterpreted area. If the valid unexplained area is determined to be a non-empty set, the center value of the mileage is calculated based on the fiber optic deployment mileage of the basic temperature measurement points within the valid unexplained area. The actual temperature measurement point with the fiber optic deployment mileage closest to the center value of the mileage is selected as the location point for the concealed fire alarm, and a thermally triggered electrical alarm signal is output.

[0006] Furthermore, projecting the heat-generating solid component along the normal direction to the outer interface of the fire-retardant insulation layer of the exterior wall includes: The projection dimension reduction operator is applied to the heat-generating physical component to generate point projection, line projection, and area projection; Point projection is generated by extracting the three-dimensional centroid coordinates of the heating entity and projecting them along the normal direction onto the outer interface of the fire-retardant insulation layer of the exterior wall. Line projection is generated by extracting the boundary contour line or central axis of the heating entity and projecting it along the normal direction. Area projection is generated by extracting the three-dimensional bounding box contact surface contour of the heating entity on the outer interface of the fire-retardant insulation layer of the exterior wall.

[0007] Furthermore, a discrete graph structure reflecting the connectivity of building interfaces and assigned length weights is established, including: Analyze the intersection lines of polygonal patches in a building information model or extract vertices from a 3D model mesh to obtain connection nodes; The basic temperature measurement points, the projection elements in the external heat structure projection set, and the construction connection nodes are used as the node set of the discrete graph structure. When any two nodes in the node set are located on the outer interface of the same continuous insulation layer, the boundary of the same cable trough, or the adjacent interface of the same metal keel, and there is no separation between any two nodes, a connecting edge is generated in the edge set of the discrete graph structure. When there is an air gap between any two nodes and no continuous structural interface, or when there is a building expansion joint partition or a fireproof seal that completely blocks the connection, the isolation rule will not generate a connection edge.

[0008] Furthermore, length weights are assigned, including: By using the globally unique data index of nodes in the discrete graph structure, the original three-dimensional spatial coordinates of the nodes in the building information model or surveying data can be reversed. The actual three-dimensional geodesic line or the cumulative length of the three-dimensional polyline is calculated based on the original three-dimensional spatial coordinates, and the actual three-dimensional geodesic line or the cumulative length of the three-dimensional polyline is assigned as the length weight to the connecting edge.

[0009] Furthermore, the maximum value of the temperature transition satisfies the preset segmentation validity determination conditions, including: The maximum value of the temperature transition is evaluated to be greater than or equal to a preset multiple of the temperature resolution of the fiber optic temperature measurement system, and greater than or equal to a preset multiple of the median of all temperature transitions within the calculation partition where the base temperature measurement point is located. When the calculation partition where the basic temperature measurement point is located is in a constant temperature state, resulting in a median temperature transition of zero, the default value of the preset multiple of the temperature resolution of the fiber optic temperature measurement system is used as the sole criterion for determining the validity of the segmentation.

[0010] Furthermore, the minimum number of consecutive measurement points threshold is obtained through the following steps: Obtain the spatial resolution and sampling interval of the fiber optic temperature measurement system; Calculate the ratio of the spatial resolution of the fiber optic temperature measurement system to the sampling interval of the temperature measurement points. Round up the ratio of the spatial resolution of the fiber optic temperature measurement system to the sampling interval of the temperature measurement points and add an endpoint compensation value to generate a minimum threshold for the number of continuous measurement points.

[0011] Furthermore, within the discrete graph structure, a graph connectivity search algorithm is performed on the specific thermal anomaly segments contained in the current set of thermal anomaly segments. When the accumulated path length exceeds the preset maximum effective heat transfer span, the program is truncated, including: In a discrete graph structure, a virtual super source point is constructed. The virtual super source point is connected to all nodes in the external heat structure projection set with virtual edges of zero weight. The single-step graph connectivity search algorithm is executed with the virtual super source point as the single search starting point. The preset maximum effective heat transfer span is calculated by reading the material attribute fields of the corresponding entity in the building information model and matching them with the system's preset material thermodynamic attenuation mapping table.

[0012] Furthermore, when performing the one-time graph connectivity search algorithm, it includes: When the discrete graph structure is in a disconnected state due to fire prevention blocking, the cumulative path length of the basic temperature measurement point that cannot be reached by the single graph connectivity search algorithm is marked as the set maximum penalty threshold constant or a specific isolation overflow mark. The temperature anomaly in the isolation area caused by fire prevention blocking is characterized as a hidden fire inside the fire-retardant insulation layer of the external wall.

[0013] Furthermore, the center value of the mileage is calculated based on the fiber optic cable deployment mileage of the basic temperature measurement points within the effective unexplained area, including: When the sampling interval of all basic temperature measurement points in the effective unexplained area is constant, the arithmetic mean of the fiber optic deployment mileage of all basic temperature measurement points is taken as the mileage center value. When there are differences in the sampling interval of the basic temperature measurement points within the effective unexplained area, the actual fiber optic coverage length represented by each basic temperature measurement point is used as the weighting coefficient, and a weighted average algorithm is used to calculate the mileage center value.

[0014] Furthermore, a fiber optic early warning system for concealed fires in exterior wall flame-retardant insulation layers is proposed to realize the fiber optic early warning method for concealed fires in exterior wall flame-retardant insulation layers as described above, including: The projection mapping module is used to project the heat-generating solid components onto the outer interface of the fire-retardant insulation layer of the exterior wall along the normal direction, forming an external heat structure projection set, and establishing a discrete graph structure that reflects the connectivity of the building interface and assigns length weights. The anomaly extraction module is used to acquire the basic temperature measurement points of the distributed temperature sensing fiber optic acquisition, construct a temperature sorting sequence based on the basic temperature measurement points, extract the temperature transition amount between adjacent step points in the temperature sorting sequence of the basic temperature measurement points, retrieve the maximum value of the temperature transition amount and its corresponding segmentation position index, when the maximum value of the temperature transition amount meets the preset segmentation validity judgment condition, extract the basic temperature measurement points higher than the segmentation position index to generate a set of thermal anomaly candidate points, perform spatial continuity merging on the thermal anomaly candidate point set, and filter out the segments with the number of temperature measurement points less than the minimum continuous measurement point number threshold to obtain the current thermal anomaly segment set; The search decoupling module is used to perform a graph connectivity search algorithm on the specific thermal anomaly segments contained in the current set of thermal anomaly segments within the discrete graph structure. When the cumulative path length exceeds the preset maximum effective heat transfer span, the program is truncated. The basic temperature measurement point reached by the external thermal structure projection set is taken as the interpretable part. The set difference operation is performed on the specific thermal anomaly segment and the interpretable part. The spatial continuity is divided twice according to the minimum continuous measurement point number threshold to obtain the effective uninterpreted area. The location alarm module is used to calculate the center value of the fiber optic cable deployment based on the fiber optic cable mileage of the basic temperature measurement points within the valid unexplained area if the valid unexplained area is determined to be a non-empty set. The actual temperature measurement point with the fiber optic cable deployment mileage closest to the center value of the mileage is selected as the location point for the hidden fire alarm, and a thermally triggered electrical alarm signal is output.

[0015] Compared with existing technologies, it has the following advantages:

[0016] This solution proposes a fiber optic early warning method and system for concealed fires in exterior wall flame-retardant insulation layers. By establishing a discrete graph structure reflecting the connectivity of building interfaces and combining it with a projection dimensionality reduction operator, external heat-generating structures on complex facades are mapped to the same heat transfer reference surface, breaking through the technical bottleneck of traditional fiber optic temperature measurement which relies solely on absolute temperature or temperature rise rate for isolated discrimination. After extracting the temperature step signal from the basic temperature measurement points, a graph connectivity search algorithm is used to propagate along the solid heat transfer topology network. Set difference operations are performed on the acquired thermal anomaly segments and the interpretable parts reachable from the topology network. This mechanism can decouple and separate the attached heat shadow formed by the normal operation of external photovoltaic equipment from the actual concealed fire heat core inside the flame-retardant insulation layer under complex facade thermodynamic backgrounds. By filtering out temperature anomalies supported by legitimate external heat sources, the system overcomes the shortcomings of conventional early warning systems, which are prone to false alarms and missed alarms in multi-source heat overlap environments, significantly improving the confidence level of early concealed fire identification.

[0017] To address the alarm coordinate offset problem caused by spatial back-and-forth cabling, this solution abandons the conventional three-dimensional Euclidean geometric center calculation logic. Instead, relying on the effective uninterpreted area obtained after decoupling, it directly calculates the mileage center value on the one-dimensional topological dimension of the fiber optic deployment and maps it to the nearest objective temperature measurement node. This mileage addressing mechanism avoids the spatial coordinate averaging and suspension errors caused by the back-and-forth bending of cables on complex building surfaces, ensuring that the final output alarm coordinates are strictly anchored to the actual fiber optic nodes. By combining the minimum consecutive measurement point number threshold derived from spatial resolution and sampling interval, a secondary spatial continuity segmentation is performed, further filtering out spatially isolated discrete background noise. This ensures that the heat-triggered electrical alarm signal sent to the downstream fire protection system not only has accurate traceability in terms of fire source attributes but also meets the precise requirements for on-site engineering verification in terms of spatial addressing. Attached Figure Description

[0018] Figure 1 This is a schematic diagram of the method flow of the present invention.

[0019] Figure 2 This is a schematic diagram of the system framework of the present invention. Detailed Implementation

[0020] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0021] Please see Figure 1 This application provides a fiber optic early warning method for concealed fires in exterior wall flame-retardant insulation layers; The method specifically includes the following steps: Step 1: Establish a unified coordinate system based on the exterior wall facade and acquire basic physical data to construct the interface attachment path data for external heat source radiation. Use a two-dimensional or three-dimensional exterior wall facade coordinate system to acquire distributed heat-sensing fiber optic deployment data, extract all basic temperature measurement points on the distributed heat-sensing fibers, and obtain the spatial coordinates of these basic temperature measurement points, including horizontal and vertical axis coordinates. Acquire the sampling interval and spatial resolution of the fiber optic temperature measurement host. Acquire spatial positioning data for external heat sources such as the edge of the photovoltaic module backsheet, junction box, DC connector, hidden cable tray, module frame, and installation keel. Simultaneously, acquire spatial boundary data for the outer interface of the insulation layer, the edge of the fireproof isolation strip, the joints of the insulation board, expansion joints, and the fireproof sealing area.

[0022] Specifically, the coordinate system establishment and distributed temperature-sensing fiber optic data acquisition obtain multi-dimensional position and temperature characteristic information through sensor hardware and spatial reference plane settings. The sampling interval represents the actual cable length between two adjacent physical sampling points of the temperature measurement host, and the spatial resolution represents the size of the smallest spatial thermal event that the temperature-sensing system can accurately identify and independently locate. Unifying the spatial dimension of multi-source heterogeneous data can correct the spatial misalignment between the one-dimensional linear mileage data of the fiber optic cable and the building facade's volumetric structure, providing reliable positional anchors for constructing a thermodynamic topological network.

[0023] It should be noted that the spatial location data of the external heat source of the photovoltaic system and various spatial boundary data are not simulated data generated internally by the system. In actual engineering implementation, the above-mentioned spatial boundary data are obtained by importing data from the building information model, or by analyzing computer-aided design as-built drawings, or by actual measurement and input through on-site laser mapping equipment. Clearly defining the physical source interface of the data ensures the accurate mapping and feasibility of the early warning system in complex building environments.

[0024] The various heat-generating physical components acquired from the photovoltaic side are projected along the normal direction of the target heat transfer interface onto the outer interface of the fire-retardant insulation layer of the exterior wall, forming a unified two-dimensional mapping object. All two-dimensional mapping objects are combined to form an external thermal structure projection set. When performing the projection dimensionality reduction operator, point projection is generated by extracting the three-dimensional centroid coordinates of the heat-generating component and projecting them onto the target plane; line projection is generated by extracting the component boundary contour line or central axis and projecting it; and area projection is generated by extracting the three-dimensional bounding box contact surface contour of the component on the target plane.

[0025] Specifically, the external thermal structure projection set refers to the effective heat transfer geometric footprint left on the heat-conducting interface by photovoltaic heating three-dimensional components suspended or attached to the outer side of the exterior wall. Point projection, line projection, and area projection correspond to the geometric dimensionality reduction equivalent representation of heating components of different volumes and shapes on the same two-dimensional heat transfer main interface. By reducing the dimensionality of the complex and scattered three-dimensional facade equipment heat sources to the outer interface of the insulation layer coplanar with the fiber optic deployment location, the dimensionality of the subsequent graph connectivity search in computer matrix operations can be greatly reduced, while simultaneously reproducing the real physical state of heat adhesion and diffusion on the surface of the solid structure.

[0026] It should be noted that when executing the projection logic for heat-generating components, if the building facade has local curved surfaces or non-coplanar angled structures, the projection rule follows the principle of projecting the normal to the nearest heat transfer surface. That is, the normal of the interface of the nearest effective insulation layer to the heat-generating component is used as the projection direction. This ensures that the obtained projection of the external heat source structure can reflect the primary contact area of ​​heat intrusion into the insulation system, thereby avoiding the coordinate distortion caused by conventional vertical coordinate system projection on complex building facades.

[0027] A network of interface attachment propagation paths is established based on the connectivity of building interfaces, and this network is then abstracted into a discrete graph structure that can be searched using computer algorithms. The discrete graph structure consists of a set of nodes and a set of edges. Nodes fall into three categories: the first category consists of all basic temperature measurement points; the second category consists of all projected elements in the external thermal structure projection set; and the third category consists of construction connection nodes, which are obtained and imported by parsing the intersection lines of polygonal faces in the building information model or extracting vertices from the 3D model mesh. Between any two nodes in the node set, a connection edge is generated in the edge set only if they are located on the same continuous insulation layer outer interface, the same cable tray boundary, or the same adjacent interface of the same metal keel, and there is no partition between them. If there is an air gap between two nodes and no continuous construction interface, a building expansion joint partition, or a fireproof seal completely blocking the connection, an isolation rule is forcibly triggered to prevent the generation of a connection edge, and a length weight is assigned to each successfully generated connection edge.

[0028] Specifically, the interface attachment propagation path network is established based on the connectivity of building structure interfaces to simulate the potential trajectory of heat propagation along the outer interface of the flame-retardant insulation layer and adjacent metal structural attachments. The isolation rule refers to the constraint condition in the algorithm that actively cuts off node connectivity based on the barrier principle in fire engineering. By mapping the thermal barrier and heat transfer capacity of the exterior facade's fire-resistant structure in a discrete data space, and by calculating the cutting off of illegal heat transfer paths through the algorithm, misjudgments caused by heat transfer paths being blocked by firewalls despite close spatial geometric straight-line distances are effectively avoided.

[0029] It should be noted that, to ensure the subsequent algorithms possess realistic thermodynamic calculation value, each connecting edge generated in the edge set must be assigned a specific length weight. When quantifying edge weights, the system uses the globally unique data index of the graph structure node to reverse-address its original 3D spatial coordinates in the Building Information Model or surveying data. Based on this, it calculates the actual cumulative length of the 3D geodesic or 3D polyline and uses this as the explicit length weight. This parameter compensates for the perspective shortening effect that occurs when reducing 3D to 2D, ensuring that the discrete graph structure possesses a realistic spatial resistance scale.

[0030] Using the nodes corresponding to all projected elements as source nodes, a shortest path search is performed in the discrete graph structure, incorporating the length weights of each connecting edge. Within the basic logical framework, the shortest interface attachment path from the temperature measurement point to the entire external thermal structure is determined, expressed as: In the formula, L is the shortest interface attachment path, which is the shortest heat transfer span of a single temperature measuring point when facing the common threat of multiple heat sources on the facade. It is used to quantify the heat transfer resistance of the temperature measuring point affected by the heat source of the photovoltaic system. The single-source attachment path length is the sum of the side lengths of the continuous physical interfaces that the heat source must cross to spread from a single projected element along the network surface to the temperature measurement point, reflecting the spread cost of a single path. n is the total number of projected elements in the external heat structure projection set, used to define the data range for comparison operations.

[0031] If a temperature measurement point in a discrete graph structure has no connected edges with any of the projected elements, then the cumulative value of the path length corresponding to that temperature measurement point is marked as infinity.

[0032] Specifically, the above operations replace traditional Euclidean geometric straight-line distances with graph-theoretic topological measured distances, achieving mathematical decoupling of thermal paths in environments with physical cascading of dual heat sources. While calculating the shortest path is a fundamental technique in data processing, it is transformed here into a core benchmark for thermodynamic propagation deduction, providing baseline data for subsequently identifying the root cause of thermal anomalies.

[0033] It should be noted that, as a preferred underlying algorithm implementation to improve system computing power, to avoid the overhead of calculating each node according to the aforementioned mathematical boundary definitions, the system constructs a virtual super source point in the discrete graph structure. This super source point is connected to all nodes in the external heat structure projection set by virtual edges with zero weight. Thus, only a single search algorithm is needed to simultaneously obtain the shortest interface attachment path from all basic temperature measurement points to the nearest heat projection, significantly improving the real-time computing efficiency under massive nodes.

[0034] It should be further explained that, due to the fire barrier causing the discrete graph structure to be disconnected, the algorithm marks the cumulative path length of unreachable temperature measurement points as infinity. In the computer system, this infinity represents a set maximum penalty threshold constant or a specific isolation overflow flag. This means that the temperature measurement point cannot receive attached heat transfer from the corresponding photovoltaic device thermodynamically and topologically, allowing the system to directly characterize the temperature anomaly occurring within the isolation area caused by the fire barrier as a hidden fire inside the fire-retardant insulation layer of the external wall.

[0035] Step Two: Obtain the current temperature sequence collected by the distributed temperature-sensing fiber optic temperature measurement host to extract the current thermal anomaly segment. Based on the building's exterior facade orientation, floor level, or photovoltaic module array distribution characteristics, the overall monitoring area is divided into several calculation partitions. For any calculation partition, extract the actual temperature acquisition value of each basic temperature measurement point within that partition at the current moment, and construct a key-value pair data structure by linking the data index of the basic temperature measurement point with the corresponding actual temperature acquisition value. Perform an ascending sort operation on the actual temperature acquisition value field in the key-value pair data structure in the numerical domain to construct a temperature sorting sequence.

[0036] Specifically, dividing the computational space into zones is intended to isolate the environmental base temperature differences caused by varying facade sunlight conditions, floor ventilation, and insulation structures. Binding data indexes to temperature values ​​creates a key-value pair data structure, maintaining the correlation between spatial and temperature attributes during subsequent sorting. Performing ascending sort removes spatial connectivity limitations of temperature measurement points, and identifying potential high-temperature anomaly clusters based on thermodynamic numerical step distribution reduces baseline drift interference caused by large-scale environmental temperature fluctuations, providing a stable data background for anomaly signal extraction.

[0037] It should be noted that the boundaries of the computational partitions are not randomly generated. In practice, the spatial boundaries of the computational partitions are directly generated by parsing the firebreak boundaries, continuous insulation zone boundaries, or photovoltaic module seams from the Building Information Model (BIM). The data import interface ensures that data samples within the same computational partition are under similar heat exchange conditions, avoiding logical discontinuities caused by subjective manual demarcation.

[0038] By iterating through the above temperature sorting sequence and subtracting each item, the temperature transition between adjacent step points is obtained. Its expression is: In the formula, The temperature transition is the temperature difference between two adjacent positions in the sorted sequence, used to quantify the step gradient within the temperature dataset. The temperature value at the i-th position is the temperature data at the i-th position after the temperature is sorted in ascending order within the partition. This is the temperature data of the next higher order.

[0039] The system retrieves the maximum temperature transition value in the sorted sequence and extracts the segmentation position index corresponding to this maximum value. The system then performs a segmentation validity determination, evaluating whether the maximum temperature transition value simultaneously satisfies a preset multiple of the fiber optic temperature measurement system's temperature resolution and a preset multiple of the median of all temperature transition values ​​within the segment. If the determination conditions are met, the system backmaps the temperature measurement points corresponding to the segmentation position index higher than the pre-built key-value pair mapping relationship to the fiber optic deployment coordinate system, generating a set of candidate thermal anomaly points.

[0040] Specifically, searching for the maximum temperature jump aims to identify data gaps between normal temperature backgrounds and abnormal heat sources. A preset multiple of temperature resolution and a preset multiple of the median jump are used as dual verification indicators to determine whether the numerical gap simultaneously exceeds the inherent measurement resolution limit of the sensor hardware and the normal thermal fluctuation level of the current environment. This dual verification mechanism filters out false alarms caused by forced data segmentation when there are no obvious fire precursors, improving the reliability of high-temperature group screening.

[0041] It should be noted that the preset multiple in the above judgment process is preferably three times, and this preset multiple is only used to verify whether the data segmentation has a reliable signal-to-noise ratio, and is not used as a constant temperature threshold to directly trigger a fire alarm. When the calculation zone where the basic temperature measurement point is located is in a constant temperature state, resulting in a median temperature transition of zero, the system defaults to using the preset multiple of the temperature resolution of the fiber optic temperature measurement system as the sole criterion for judging the segmentation validity when performing double verification, ensuring the robustness of the algorithm under extremely low temperature difference fluctuation boundaries.

[0042] Based on the fiber optic deployment sequence, a spatial continuity merging operation is performed on the set of candidate thermal anomaly points. When candidate points are located in the same fiber optic channel and are adjacent or spaced no more than one sampling interval along the fiber optic deployment mileage, they are merged into a single continuous candidate thermal anomaly segment. Subsequently, a spatial filtering operation is performed to determine whether the total number of continuous temperature measurement points contained in each candidate thermal anomaly segment meets the minimum continuous measurement point threshold. The calculation formula is as follows: In the formula, The minimum number of consecutive measurement points threshold is the minimum number of temperature measurement points required to determine the validity of a continuous segment. It is used to filter out isolated spatial points that are not reliably located by the system. Spatial resolution of a fiber optic temperature measurement system represents the size of the smallest spatial thermal event that the temperature sensing system can accurately identify and independently locate. The sampling interval for temperature measurement points represents the actual cable length between two adjacent sampling points on the temperature measurement host. The system identifies candidate segments containing a number of consecutive temperature measurement points greater than or equal to the minimum threshold for the number of consecutive measurement points as the current set of thermal anomaly segments.

[0043] Specifically, the spatial continuity merging operation restores the high-level signal in the one-dimensional data domain to a heated object with spatial scale. The formula for the minimum continuous measurement point number threshold is designed based on the discrete point segment mapping principle. By rounding up the ratio of spatial resolution to sampling interval and adding endpoint compensation values, the spatial length span is accurately converted into a discrete temperature measurement point number index. This filtering threshold is used to intercept and eliminate signal artifacts or single-point electromagnetic noise with a spatial span lower than the inherent sensing limit of optical fiber, ensuring that the abnormal segments input to the subsequent causal interpretation model have spatial connectivity and reasonable size.

[0044] Step 3: Traverse the extracted set of current thermal anomaly segments, mapping each specific thermal anomaly segment to a pre-constructed discrete graph structure. Based on the interface attachment path graph, interpret the external attachment paths of the thermal anomaly segments to extract valid uninterpreted areas, and perform classification and labeling operations on the node states in the graph structure memory. The system identifies basic temperature measurement point nodes that match the spatial location of a specific thermal anomaly segment and updates their state attributes to thermal anomaly temperature measurement nodes; it identifies other basic temperature measurement point nodes not included in a specific thermal anomaly segment and maintains or updates their state attributes to non-thermal anomaly temperature measurement nodes; the state of the constructed connection nodes in the original discrete graph structure remains unchanged.

[0045] Specifically, mapping anomalous sections to a discrete graph structure can build a data interaction bridge between one-dimensional fiber optic temperature signals and the three-dimensional topological structure of building facades. Performing node state classification and labeling operations aims to establish boundary barriers for subsequent graph connectivity search algorithms. Non-thermal anomaly temperature measurement nodes act as termination boundaries for heat transfer searches in the mathematical model, representing spaces that have not yet been infiltrated by heat or where heat has physically attenuated. This limits the disordered expansion of the algorithm's search path, ensuring that computing power is focused on high-probability heat transfer channels.

[0046] It should be noted that the above node state marking operation is implemented by appending Boolean state labels or categorical enumeration values ​​to the node objects in the discrete graph structure. This memory-level attribute update method achieves dynamic injection of temperature field data without destroying the topological connectivity of the original discrete graph structure, avoiding the computational cost caused by frequently rebuilding the topological matrix as the time series changes.

[0047] Using a virtual super-source node in the discrete graph structure as the single starting point, a graph connectivity search algorithm is executed in the updated discrete graph structure. The search path extends stepwise along the established connecting edges in the graph structure, allowing crossings of constructed connecting nodes and thermal anomaly measurement nodes, and forcibly terminating the search branch when it reaches a non-thermal anomaly measurement node or encounters a break without a connecting edge. The system summarizes all thermal anomaly measurement nodes that can be continuously reached along the topology network from the virtual super-source node and groups them into the interpretable part of the specific thermal anomaly segment.

[0048] Specifically, the graph connectivity search algorithm simulates the diffusion process of external heat adhering to and spreading along the solid interface of a building's exterior wall. Extracting interpretable components aims to identify and aggregate high-temperature anomalies that can be directly traced back to the heat-generating components of the photovoltaic system in topological space. This mechanism uses topological connectivity instead of spatial straight-line distance to allocate heat source attribution, reducing the probability of misjudgments of fires caused by the overlapping projections of normal heat load from the photovoltaic system and internal fire conditions in three-dimensional space.

[0049] It should be noted that during graph connectivity search, to prevent excessive extrapolation that violates thermodynamic laws on pure structural channels lacking fiber optic temperature measurement point constraints, the system sets cumulative length weight constraints on paths continuously passing through structural connection nodes. When two thermal anomaly temperature measurement nodes are only connected by structural connection nodes, and the cumulative path length between them exceeds the preset maximum effective heat transfer span, the system truncates the search branch. The maximum effective heat transfer span is calculated by reading the material property fields of the corresponding entities in the building information model and matching them with the system's preset material thermodynamic attenuation mapping table, and is used to characterize the limit distance of external heat conduction along unmonitored channels.

[0050] Perform a set difference operation between the specific thermal anomaly segment and the interpretable portion obtained from the search to obtain the uninterpreted region. The expression is: In the formula, The unexplained region is a set of temperature measurement points in a specific thermal anomaly segment that cannot be continuously covered by the external thermal structure projection nodes along the interface network. It is used to characterize anomaly segments that have lost the explanation support of external heat sources. For a specific thermal anomaly section, it is a set of spatially continuous high-temperature anomaly measurement points extracted from the current fiber optic temperature sequence. The interpretable part is the set of temperature measurement points in the abnormal section that can be continuously reached from the external heating structure via a topological path.

[0051] After identifying the unexplained area, a second spatial continuity segmentation is performed based on the fiber optic cable deployment mileage, decoupling it into at least one independent continuous sub-segment. For any continuous sub-segment, the system counts the total number of temperature measurement points it contains and determines whether this total is greater than or equal to a predetermined minimum threshold for the number of continuous measurement points. If the condition is met, the continuous sub-segment is retained and classified as a valid unexplained area; if the unexplained area is an empty set or all the segmented continuous sub-segments do not meet the condition, the fire tracing for that specific thermal anomaly segment is terminated.

[0052] Specifically, the set difference operation is performed to mathematically decouple the dual heat sources aliased in the fiber optic temperature field, stripping away the attached heat generated by the operation of external photovoltaic equipment. Spatial continuity quadratic segmentation is used to deconstruct the spatially fragmented segments resulting from the removal of intermediate nodes during the set difference operation. The filtered continuous sub-segments are characterized as valid unexplained areas, identifying independent heat cores lacking legitimate external heat source support, thus providing high-confidence independent evidence for qualitatively determining hidden fires inside the exterior wall fire-retardant insulation layer.

[0053] It should be further explained that the minimum continuous measurement point threshold mentioned above directly calls the parameter value established in step two based on the optical fiber spatial resolution and sampling interval calculation, thus unifying the global spatial filtering scale. When the uninterpreted area is an empty set or is filtered out due to fragmentation, the system classifies the current specific thermal anomaly segment as an externally attached thermal shadow caused by normal heating of photovoltaic modules or electrical overheating, thereby filtering out false alarm signals that trigger non-building fire alarm systems.

[0054] Step 4: Obtain the valid unexplained area extracted in the previous steps, generate a concealed fire alarm event based on the unexplained area, and output a heat-triggered electrical alarm signal. Determine whether the valid unexplained area is an empty set. If it is, the system classifies the current thermal anomaly as an environmental thermal fluctuation that can be explained by externally attached thermal shadows, and maintains the heat-triggered electrical alarm signal at a non-alarm level. If the valid unexplained area is a non-empty set, the system identifies the area as a candidate heat core for a concealed fire inside the exterior wall flame-retardant insulation layer. Extract all continuous temperature measurement points and their corresponding fiber optic cable mileage within the valid unexplained area, and calculate the mileage center value of the candidate heat core based on the fiber optic cable mileage sequence. Traverse all specific temperature measurement points within the valid unexplained area, and select the actual temperature measurement point whose fiber optic cable mileage is closest to the mileage center value as the concealed fire alarm location point. The expression for solving the concealed fire alarm location point is: In the formula, The concealed fire alarm location point is a fiber optic sensing node that represents the location of a concealed fire and is used to guide on-site investigation. The specific temperature measurement points contained within the effective unexplained area are individual local temperature measurement units that constitute the current independent heat core; The fiber optic mileage corresponding to a specific temperature measurement point is the distance coordinate of that temperature measurement point on the one-dimensional fiber optic topology. The mileage center value of this effective uninterpreted region is the midpoint of the distribution of this continuously heated segment on the one-dimensional fiber topology. (Symbol) This indicates the calculation of the absolute difference between two one-dimensional mileage values.

[0055] Specifically, determining whether the valid unexplained area is an empty set aims to perform the final identification of the fire cause. It calculates the mileage center value and performs a proximity mapping operation to the nearest actual temperature measurement point, anchoring continuous temperature anomaly segments to objective fiber optic nodes. This one-dimensional mileage addressing mechanism replaces the traditional three-dimensional Euclidean space geometric center solution, effectively avoiding the suspended positioning misalignment caused by three-dimensional coordinate averaging when fiber optic cables are laid in complex back-and-forth cabling on building facades. This ensures that the alarm positioning coordinates correspond to the actual laid fiber optic nodes, improving the efficiency of on-site verification by fire control center personnel based on coordinates. After establishing the positioning point, the system simultaneously extracts the highest characteristic temperature or maximum relative temperature rise within the valid unexplained area, and instantiates a concealed fire alarm event object in memory, along with the alarm positioning point and its corresponding three-dimensional spatial coordinates, fiber optic channel number, the span of the start and end temperature measurement points of the valid unexplained area, and temperature array data. Subsequently, this concealed fire alarm event object is sent to an external fire alarm controller or building safety management platform, and the thermal triggering electrical alarm signal output to the fire protection system is set to an effective trigger level. By encapsulating core state scalars such as the highest characteristic temperature, the system has made a leap from algorithmic data analysis to physical security control, providing direct data criteria for downstream fire protection systems to perform hierarchical linkage.

[0056] It should be noted that when calculating the aforementioned mileage center value, the system adaptively switches the calculation model based on the sampling interval distribution characteristics of the fiber optic temperature measurement host. When the sampling interval of all basic temperature measurement points within the effective uninterpreted area is constant, the arithmetic mean of the fiber optic deployment mileage of all basic temperature measurement points is used as the mileage center value. When there are differences in the sampling interval of the basic temperature measurement points within the effective uninterpreted area, the actual fiber optic coverage length represented by each basic temperature measurement point is used as a weighting coefficient, and a weighted average algorithm is used to calculate the mileage center value. This adaptive calculation model corrects for the spatial positioning offset caused by uneven cabling density.

[0057] It should be further clarified that when the system outputs concealed fire alarm events, it is only responsible for outputting the state parameters of the independent thermal core, and does not interfere with or replace the inherent fire linkage classification logic of the downstream fire alarm controller. Downstream fire-fighting equipment, based on its built-in mandatory fire protection standards, and in conjunction with parameters such as the received highest characteristic temperature and spatial span, independently determines and outputs warning commands or fire alarm commands. This interface boundary design ensures the versatility and platform compatibility of this warning method as an independent sensing algorithm module, avoiding limitations on its practical engineering application scope due to binding to specific fire alarm levels.

[0058] Furthermore, refer to Figure 2 As shown, a fiber optic early warning system for concealed fires in exterior wall fire-retardant insulation layers is proposed to realize the fiber optic early warning method for concealed fires in exterior wall fire-retardant insulation layers as described above, including: The projection mapping module is used to project the heat-generating solid components onto the outer interface of the fire-retardant insulation layer of the exterior wall along the normal direction, forming an external heat structure projection set, and establishing a discrete graph structure that reflects the connectivity of the building interface and assigns length weights. The anomaly extraction module is used to acquire the basic temperature measurement points of the distributed temperature sensing fiber optic acquisition, construct a temperature sorting sequence based on the basic temperature measurement points, extract the temperature transition amount between adjacent step points in the temperature sorting sequence of the basic temperature measurement points, retrieve the maximum value of the temperature transition amount and its corresponding segmentation position index, when the maximum value of the temperature transition amount meets the preset segmentation validity judgment condition, extract the basic temperature measurement points higher than the segmentation position index to generate a set of thermal anomaly candidate points, perform spatial continuity merging on the thermal anomaly candidate point set, and filter out the segments with the number of temperature measurement points less than the minimum continuous measurement point number threshold to obtain the current thermal anomaly segment set; The search decoupling module is used to perform a graph connectivity search algorithm on the specific thermal anomaly segments contained in the current set of thermal anomaly segments within the discrete graph structure. When the cumulative path length exceeds the preset maximum effective heat transfer span, the program is truncated. The basic temperature measurement point reached by the external thermal structure projection set is taken as the interpretable part. The set difference operation is performed on the specific thermal anomaly segment and the interpretable part. The spatial continuity is divided twice according to the minimum continuous measurement point number threshold to obtain the effective uninterpreted area. The location alarm module is used to calculate the center value of the fiber optic cable deployment based on the fiber optic cable mileage of the basic temperature measurement points within the valid unexplained area if the valid unexplained area is determined to be a non-empty set. The actual temperature measurement point with the fiber optic cable deployment mileage closest to the center value of the mileage is selected as the location point for the hidden fire alarm, and a thermally triggered electrical alarm signal is output.

[0059] The above embodiments are only used to illustrate the technical methods of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical methods of the present invention without departing from the spirit and scope of the technical methods of the present invention.

Claims

1. A fiber optic early warning method for concealed fires in exterior wall flame-retardant insulation layers, characterized in that, include: The heat-generating solid component is projected along the normal direction to the outer interface of the fire-retardant insulation layer of the exterior wall, forming an external heat structure projection set, and establishing a discrete graph structure that reflects the connectivity of the building interface and assigns length weights. The system acquires basic temperature measurement points from distributed temperature sensing fiber optic data acquisition and constructs a temperature sorting sequence based on these points. It then extracts the temperature transition amounts between adjacent step points in the temperature sorting sequence of the basic temperature measurement points, retrieves the maximum temperature transition amount and its corresponding segmentation position index, and extracts basic temperature measurement points with higher values ​​than the segmentation position index to generate a set of thermal anomaly candidate points. The system performs spatial continuity merging on the thermal anomaly candidate point set and filters out segments with fewer temperature measurement points than the minimum continuous measurement point threshold to obtain the current thermal anomaly segment set. The preset segmentation validity determination condition is as follows: the maximum value of the evaluated temperature transition is simultaneously greater than or equal to a preset multiple of the temperature resolution of the fiber optic temperature measurement system, and greater than or equal to a preset multiple of the median of all temperature transitions in the calculation partition where the base temperature measurement point is located. When the calculation partition where the base temperature measurement point is located is in a constant temperature state, resulting in the median of temperature transitions being zero, the preset multiple of the temperature resolution of the fiber optic temperature measurement system is used as the only segmentation validity determination condition by default. Within the discrete graph structure, a graph connectivity search algorithm is executed on the specific thermal anomaly segments contained in the current thermal anomaly segment set. When the cumulative path length exceeds the preset maximum effective heat transfer span, the program is truncated. The basic temperature measurement point reached by the external thermal structure projection set is taken as the interpretable part. The set difference operation is performed on the specific thermal anomaly segment and the interpretable part. Based on the minimum continuous measurement point number threshold, a second spatial continuity segmentation is performed to obtain the effective uninterpreted area. The program truncation refers to terminating the continued extension of the corresponding search branch when the cumulative path length exceeds the preset maximum effective heat transfer span. If the valid unexplained area is determined to be a non-empty set, the center value of the mileage is calculated based on the fiber optic deployment mileage of the basic temperature measurement points within the valid unexplained area. The actual temperature measurement point with the fiber optic deployment mileage closest to the center value of the mileage is selected as the location point for the concealed fire alarm, and a thermally triggered electrical alarm signal is output.

2. The fiber optic early warning method for concealed fires in an exterior wall flame-retardant insulation layer according to claim 1, characterized in that, Projecting a heat-generating solid component along the normal direction to the outer interface of the fire-retardant insulation layer of the exterior wall, including: The projection dimension reduction operator is applied to the heat-generating physical component to generate point projection, line projection, and area projection; Point projection is generated by extracting the three-dimensional centroid coordinates of the heating entity and projecting them along the normal direction onto the outer interface of the fire-retardant insulation layer of the exterior wall. Line projection is generated by extracting the boundary contour line or central axis of the heating entity and projecting it along the normal direction. Area projection is generated by extracting the three-dimensional bounding box contact surface contour of the heating entity on the outer interface of the fire-retardant insulation layer of the exterior wall.

3. The fiber optic early warning method for concealed fires in an exterior wall flame-retardant insulation layer according to claim 1, characterized in that, Establish a discrete graph structure that reflects the connectivity of building interfaces and assigns length weights, including: Analyze the intersection lines of polygonal patches in a building information model or extract vertices from a 3D model mesh to obtain connection nodes; The basic temperature measurement points, the projection elements in the external heat structure projection set, and the construction connection nodes are used as the node set of the discrete graph structure. When any two nodes in the node set are located on the outer interface of the same continuous insulation layer, the boundary of the same cable trough, or the adjacent interface of the same metal keel, and there is no separation between any two nodes, a connecting edge is generated in the edge set of the discrete graph structure. When there is an air gap between any two nodes and no continuous structural interface, or when there is a building expansion joint partition or a fireproof seal that completely blocks the connection, the isolation rule will not generate a connection edge.

4. The fiber optic early warning method for concealed fires in an exterior wall flame-retardant insulation layer according to claim 3, characterized in that, Assigning length weights includes: By using the globally unique data index of nodes in the discrete graph structure, the original three-dimensional spatial coordinates of the nodes in the building information model or surveying data can be reversed. The actual three-dimensional geodesic line or the cumulative length of the three-dimensional polyline is calculated based on the original three-dimensional spatial coordinates, and the actual three-dimensional geodesic line or the cumulative length of the three-dimensional polyline is assigned as the length weight to the connecting edge.

5. The fiber optic early warning method for concealed fires in an exterior wall flame-retardant insulation layer according to claim 1, characterized in that, The minimum number of consecutive measurement points threshold is obtained through the following steps: Obtain the spatial resolution and sampling interval of the fiber optic temperature measurement system; Calculate the ratio of the spatial resolution of the fiber optic temperature measurement system to the sampling interval of the temperature measurement points. Round up the ratio of the spatial resolution of the fiber optic temperature measurement system to the sampling interval of the temperature measurement points and add an endpoint compensation value to generate a minimum threshold for the number of continuous measurement points.

6. The fiber optic early warning method for concealed fires in an exterior wall flame-retardant insulation layer according to claim 1, characterized in that, Within the discrete graph structure, a graph connectivity search algorithm is performed on the specific thermal anomaly segments contained in the current set of thermal anomaly segments. If the accumulated path length exceeds the preset maximum effective heat transfer span, the program is truncated, including: In a discrete graph structure, a virtual super source point is constructed. The virtual super source point is connected to all nodes in the external heat structure projection set with virtual edges of zero weight. The single-step graph connectivity search algorithm is executed with the virtual super source point as the single search starting point. The preset maximum effective heat transfer span is calculated by reading the material attribute fields of the corresponding entity in the building information model and matching them with the system's preset material thermodynamic attenuation mapping table.

7. A fiber optic early warning method for concealed fires in an exterior wall flame-retardant insulation layer according to claim 6, characterized in that, When performing a one-step graph connectivity search algorithm, the following are included: When the discrete graph structure is in a disconnected state due to fire prevention blocking, the cumulative path length of the basic temperature measurement point that cannot be reached by the single graph connectivity search algorithm is marked as the set maximum penalty threshold constant or a specific isolation overflow mark. The temperature anomaly in the isolation area caused by fire prevention blocking is characterized as a hidden fire inside the fire-retardant insulation layer of the external wall.

8. The fiber optic early warning method for concealed fires in an exterior wall flame-retardant insulation layer according to claim 1, characterized in that, The center value of the mileage is calculated based on the fiber optic cable deployment mileage of the basic temperature measurement points within the valid unexplained area, including: When the sampling interval of all basic temperature measurement points in the effective unexplained area is constant, the arithmetic mean of the fiber optic deployment mileage of all basic temperature measurement points is taken as the mileage center value. When there are differences in the sampling interval of the basic temperature measurement points within the effective unexplained area, the actual fiber optic coverage length represented by each basic temperature measurement point is used as the weighting coefficient, and a weighted average algorithm is used to calculate the mileage center value.

9. A fiber optic early warning system for concealed fires in an exterior wall flame-retardant insulation layer, used to implement the fiber optic early warning method for concealed fires in an exterior wall flame-retardant insulation layer as described in any one of claims 1-8, characterized in that, include: The projection mapping module is used to project the heat-generating solid components onto the outer interface of the fire-retardant insulation layer of the exterior wall along the normal direction, forming an external heat structure projection set, and establishing a discrete graph structure that reflects the connectivity of the building interface and assigns length weights. The anomaly extraction module is used to acquire the basic temperature measurement points of the distributed temperature sensing fiber optic acquisition, construct a temperature sorting sequence based on the basic temperature measurement points, extract the temperature transition amount between adjacent step points in the temperature sorting sequence of the basic temperature measurement points, retrieve the maximum value of the temperature transition amount and its corresponding segmentation position index, when the maximum value of the temperature transition amount meets the preset segmentation validity judgment condition, extract the basic temperature measurement points higher than the segmentation position index to generate a set of thermal anomaly candidate points, perform spatial continuity merging on the thermal anomaly candidate point set, and filter out the segments with the number of temperature measurement points less than the minimum continuous measurement point number threshold to obtain the current thermal anomaly segment set; The preset segmentation validity determination condition is as follows: the maximum value of the evaluated temperature transition is simultaneously greater than or equal to a preset multiple of the temperature resolution of the fiber optic temperature measurement system, and greater than or equal to a preset multiple of the median of all temperature transitions in the calculation partition where the base temperature measurement point is located. When the calculation partition where the base temperature measurement point is located is in a constant temperature state, resulting in the median of temperature transitions being zero, the preset multiple of the temperature resolution of the fiber optic temperature measurement system is used as the only segmentation validity determination condition by default. The search decoupling module is used to perform a graph connectivity search algorithm on the specific thermal anomaly segments contained in the current set of thermal anomaly segments within the discrete graph structure. When the cumulative path length exceeds the preset maximum effective heat transfer span, the program is truncated. The basic temperature measurement point reached by the external thermal structure projection set is taken as the interpretable part. The set difference operation is performed on the specific thermal anomaly segment and the interpretable part. The spatial continuity is divided twice according to the minimum continuous measurement point number threshold to obtain the effective uninterpreted area. The program truncation refers to terminating the continued extension of the corresponding search branch when the cumulative path length exceeds the preset maximum effective heat transfer span. The location alarm module is used to calculate the center value of the fiber optic cable deployment based on the fiber optic cable mileage of the basic temperature measurement points within the valid unexplained area if the valid unexplained area is determined to be a non-empty set. The actual temperature measurement point with the fiber optic cable deployment mileage closest to the center value of the mileage is selected as the location point for the hidden fire alarm, and a thermally triggered electrical alarm signal is output.

Images