Building fire safety monitoring system based on solar power supply

Abstract

The application discloses a building fire safety monitoring system based on solar power supply and relates to the technical field of fire safety monitoring.The system comprises a radiation environment modeling module, a risk area identification module, an energy prediction module, a front-end steady-state regulation module and an identification optimization and closed-loop regulation module.The radiation environment modeling module constructs an infrared radiation and airflow coupling observation baseline, collects ventilation intensity, temperature gradient and reflection structure parameters, generates a radiation space-time distribution model and a focusing track mapping; the risk area identification module performs spectrum difference and turbulence analysis based on the focusing track mapping.The application identifies high-risk focusing areas and predicts energy behavior by constructing an infrared radiation environment model, introduces phase conjugate cancellation and time grid sampling, combines feedback regulation and consistency voting mechanism, realizes steady-state closed-loop control of infrared detection, significantly improves robustness and identification accuracy in complex environments, and is suitable for high-rise buildings, underground buildings and energy-limited places.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of fire safety monitoring, in particular to a building fire safety monitoring system based on solar power supply. BACKGROUND

[0002] The building fire safety monitoring system based on solar power supply is an intelligent safety protection system that combines solar photovoltaic power generation technology with building fire safety monitoring system. The system converts light energy into electrical energy through solar panels installed on the outside or roof of the building, and stores the energy in an energy storage device to provide continuous and stable power support for fire monitoring equipment, thereby breaking the dependence on commercial power and achieving self-sufficient energy supply. In terms of function, the system usually integrates fire detection, smoke monitoring, temperature and humidity detection, flammable gas identification, and video monitoring, etc. various sensing devices, and uses wireless communication network to transmit monitoring data to the control center or cloud platform in real time for intelligent analysis and early warning. When abnormal conditions are detected, the system can automatically trigger sound and light alarms, send remote notifications, or link to the sprinkler device to perform emergency response. The system not only improves the real-time and reliability of building fire safety, but also has the advantages of energy saving, environmental protection, flexible installation, and low operation and maintenance cost, and is especially suitable for remote areas, old buildings, and places with limited energy supply conditions.

[0003] The prior art has the following disadvantages:

[0004] In the prior art, during the operation of the building fire monitoring system, an infrared flame detector is usually used to collect and determine flame radiation signals in real time to achieve rapid identification and alarm of the fire source. However, in a building environment where the ventilation system operates frequently and the airflow direction changes dramatically, the air temperature gradient and flow path will fluctuate dynamically over time, especially in the early stage of a fire, the high-temperature airflow forms complex reflection and refraction effects along the surface of the wall, glass curtain wall or metal components, resulting in energy superposition and transient focusing phenomenon of infrared radiation signals in the local space. When the infrared flame detector is in such a high-energy radiation reflection area, its sensing unit may receive energy input that exceeds the normal recognition range, resulting in a detection saturation false image, which causes the output signal to remain at a stable high level for a long time. Such abnormal signals will be misjudged by the fire source recognition algorithm as data anomalies and trigger the self-inhibition mechanism, causing the system to temporarily stop the flame recognition process, resulting in delayed alarm and failure to confirm the real fire in time.

[0005] The above information disclosed in the background section is only used to enhance the understanding of the background of the present disclosure, and therefore it can include information that does not constitute prior art known to those of ordinary skill in the art. SUMMARY

[0006] The application aims to provide a building fire safety monitoring system powered by solar energy to solve the problems in the background art.

[0007] To achieve the above-mentioned purpose, the application provides the following technical solution: a building fire safety monitoring system powered by solar energy, comprising a radiation environment modeling module, a risk area identification module, an energy prediction module, a front-end steady-state regulation module, and an identification optimization and closed-loop regulation module:

[0008] The radiation environment modeling module constructs a coupling observation baseline of infrared radiation and airflow, collects ventilation intensity, temperature gradient, and reflection structure parameters, constructs a radiation spatiotemporal distribution model, and generates an energy focusing trajectory mapping.

[0009] The risk area identification module, based on the focusing trajectory mapping, performs spectral difference and turbulence analysis, identifies radiation hotspots, extracts saturated trigger sequences, and generates a corresponding risk spatiotemporal window set.

[0010] The energy prediction module, based on the risk spatiotemporal window set, constructs a causal residual playback chain, replays historical radiation sequences, predicts the energy peak position and duration in the next cycle, and forms a prediction parameter set required for sampling scheduling.

[0011] The front-end steady-state regulation module, based on the prediction parameter set, implements phase conjugate radiation cancellation and reversible time grid sampling, dynamically adjusts the integration time, suppresses energy peak superposition, and forms a stable input of the detection signal.

[0012] The identification optimization and closed-loop regulation module uses the stable input signal to perform path variation and energy return regulation, performs cross-channel voting by injecting reflection bias, optimizes the identification weight and alarm threshold in real time, and completes the closed-loop steady-state adaptive regulation of the detection system.

[0013] Preferably, the energy focusing trajectory mapping generation step is as follows:

[0014] Establish a coupling observation baseline of infrared radiation and airflow, continuously collect ventilation intensity, temperature gradient, and reflection surface geometric parameters.

[0015] Based on the collected data, construct a radiation spatiotemporal distribution model, and map the ventilation intensity into an airflow vector field, combine the temperature gradient and three-dimensional building structure to form an infrared radiation flux field.

[0016] Identify high-frequency infrared focusing areas in the flux field, extract energy transmission links and calculate the degree of energy density aggregation.

[0017] Generate a focusing trajectory mapping based on the energy transmission links, depict the spatiotemporal migration path of infrared energy under airflow disturbance conditions, and form a structured analysis input.

[0018] Preferably, the risk space-time window set generation step is as follows:

[0019] Based on the energy focusing trajectory mapping, select the continuous time section with high energy density, perform polarization spectrum difference calculation to identify non-thermal source type energy anomaly points;

[0020] Call the three-dimensional topological relationship of the building structure in the corresponding area, combine the airflow vector direction, flow rate change rate and temperature difference field to construct the turbulence disturbance configuration atlas and determine the high-risk reflection area;

[0021] Extract the infrared output curve of the hot spot area, identify the rising segment that continuously exceeds the steady-state threshold to form a saturation trigger sequence;

[0022] According to the time interval and spatial position contained in the saturation trigger sequence, a multi-dimensional mapping relationship is established, a risk space-time window set is constructed, and an evolution trend index is generated.

[0023] Preferably, when performing polarization spectrum difference calculation, infrared spectrum data under different polarization directions are collected by combining multi-angle waveband filter group and linear polarizer group to extract interference fringe structure formed by multiple reflections and distinguish non-thermal source type high-energy focusing area.

[0024] Preferably, the prediction parameter set generation step is as follows:

[0025] According to the risk space-time window set, extract the infrared radiation energy sequence and environmental parameters, construct the energy evolution trajectory under the unified sampling period, and extract the high similarity historical sequence;

[0026] Based on time causality, construct residual playback chain, embed time offset and energy difference residual data, and form historical behavior reproduction path;

[0027] Map the prediction path to the building coordinate system, output the prediction time position of the energy peak value and the saturation probability;

[0028] Construct the prediction parameter set and input the collection process to realize sampling scheduling and integral time adjustment.

[0029] Preferably, the prediction parameter set includes prediction time position pair, peak value amplitude interval, risk level, disturbance response delay, historical playback path number and sampling beat adjustment suggestion value, which is used to guide the front-end sampling beat dynamic sorting and integral time window real-time adjustment.

[0030] Preferably, the stable input formation process of the detection signal is as follows:

[0031] According to the prediction parameter set, construct the phase conjugate radiation wavefront structure which is spatially reversed, time reversed and energy amplitude matched with the infrared input path, realize wavefront cancellation control;

[0032] By combining the predicted beat and the residual intensity, a reversible time grid sampling structure with independent integral beats is constructed to achieve the misalignment control of the integral period and the energy peak.

[0033] Normalization and weighted correction are performed on multi-channel input signals to achieve energy compensation and lateral attenuation of redundant paths;

[0034] Based on the periodic characteristics in the predicted parameters, a dynamic allocation strategy for the integral capacity is implemented to construct a non-uniform time integral model;

[0035] The regulated input sequence is compared with the predicted parameters to construct an energy residual model and complete the next cycle of regulation configuration.

[0036] Preferably, the following steps are taken to dynamically adjust the identification weights and alarm thresholds using a stable input signal, performing path variational calculation and energy return control, injecting reflection bias and conducting cross-channel voting:

[0037] Based on the infrared detection output after steady-state regulation, Hamiltonian variational path calculation is performed to obtain the optimal energy migration trajectory from the disturbance source to the sampling point.

[0038] Energy back-tracing traps are set up according to the optimal path to guide high-energy inputs that deviate from the path into the auxiliary path and limit their energy bandwidth.

[0039] Dynamically inject reflection bias values ​​to adjust the phase of reflected waves between channels and achieve output consistency;

[0040] Based on the channel stability and residual evaluation results, the identification weight and alarm judgment threshold are dynamically adjusted;

[0041] The current cycle output is checked for error. If the deviation exceeds the limit, a new round of path and turnaround adjustment is triggered.

[0042] The technical effects and advantages provided by the present invention in the above technical solution are as follows:

[0043] This invention, based on environmental modeling, accurately depicts the spatial focusing trajectory of infrared radiation under complex airflow disturbances. Combined with high-risk area identification and energy behavior prediction, it enables proactive prediction and intervention in potential interference areas. At the sampling front end, a phase conjugate radiation cancellation mechanism and a reversible time grid sampling structure are introduced to effectively reduce the transient superposition of energy peaks and ensure signal input stability. The system also performs path reconstruction and energy foldback adjustment based on dynamic feedback of detection results. Through a cross-channel consensus voting mechanism, it optimizes identification weights and alarm thresholds, constructing a closed-loop steady-state adaptive control capability for the entire infrared sensing process. This solution significantly improves the robustness and identification accuracy of infrared detection in fire monitoring, possesses excellent environmental adaptability, and is particularly suitable for high-rise buildings, underground spaces, and areas with limited energy supply, promoting the intelligent transformation of fire monitoring from passive response to prediction-driven, adaptive control. Attached Figure Description

[0044] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments recorded in this invention. For those skilled in the art, other drawings can be obtained based on these drawings.

[0045] Figure 1 This is a schematic diagram of the building fire safety monitoring system based on solar power according to the present invention. Detailed Implementation

[0046] Exemplary embodiments will now be described more fully with reference to the accompanying drawings. However, these exemplary embodiments can be implemented in many forms and should not be construed as limited to the examples set forth herein; rather, they are provided so that the description of this disclosure will be more complete and fully convey the concept of the exemplary embodiments to those skilled in the art.

[0047] This invention provides, for example Figure 1 The building fire safety monitoring system based on solar power shown includes a radiation environment modeling module, a risk area identification module, an energy prediction module, a front-end steady-state control module, and an identification optimization and closed-loop adjustment module.

[0048] The radiation environment modeling module establishes a coupled observation baseline for infrared radiation and airflow, continuously collects ventilation intensity, temperature gradient and reflector geometric parameters, constructs a radiation spatiotemporal distribution model, and generates an energy focusing trajectory map for analyzing radiation convergence behavior caused by airflow disturbance.

[0049] This step establishes a coupled observation baseline for infrared radiation and airflow disturbance, creating a mapping relationship between infrared energy distribution and building interior environmental parameters. This provides data support and a model foundation for subsequent identification of anomalous energy focusing phenomena and energy behavior prediction. The specific steps are as follows:

[0050] Typical ventilation areas within buildings were selected as observation targets. Based on the building's structural layout, air vent distribution, lighting orientation, and wall materials, multiple infrared acquisition nodes and environmental sensing devices were installed. The infrared acquisition nodes employed wide-band detectors covering the mid-to-long-wave infrared frequency band. During installation, the angle between the field of view and the reflecting surface was considered to ensure that radiation behavior from different angles could be detected. The environmental sensing devices were used to simultaneously collect data on ventilation intensity, air temperature gradient, and surface temperature changes of surrounding objects within the area. Ventilation intensity was measured in real-time by high-sensitivity miniature anemometers. The temperature gradient was measured using multi-point distributed thermocouples deployed at different heights and wall corners to form vertical and horizontal thermal profiles. The geometric parameters of the reflecting surface were obtained using a structural scanner, constructing a three-dimensional building reflection model that includes features such as spatial location, material properties, and surface roughness.

[0051] Multi-source data were time-synchronized and spatially aligned to construct a spatiotemporal distribution model of infrared radiation under airflow disturbance. Specifically, ventilation intensity changes were mapped to an airflow vector field, and this was used as the driving force to introduce flow direction, velocity gradient, and turbulence parameters into the existing 3D building structure model. Initial boundary conditions for the radiation transfer path were constructed by combining this with the temperature gradient distribution. Infrared energy response data were then used to reconstruct the temporal trajectory of reflection and refraction within the 3D structure. By comparing the intensity distribution curves of infrared radiation at different locations in space under different ventilation conditions, an infrared radiation flux field covering the entire observation space was constructed. This flux field not only reflects the concentrated or sparse state of infrared energy at different time points but also clarifies the dynamic modulation effect of ventilation disturbance on the infrared transmission path, thus providing structural input for subsequent energy focusing behavior modeling.

[0052] Based on the constructed infrared radiation flux field, the radiation focusing areas most significantly affected by ventilation within the building were further identified. By tracking high-frequency intensity abrupt changes in infrared energy across time and space, the spatial locations where abnormal infrared energy convergence occurs due to airflow disturbances were determined. Combined with reflective surface geometry information, structural combinations exhibiting multiple reflections and focusing within the radiation reflection path were extracted. Furthermore, a complete transmission link of infrared energy from the source point to the detection point was established, and the energy transfer efficiency changes along each link were calculated to quantify the degree of energy density concentration. For locations exhibiting high-frequency focusing behavior, a temporally continuous energy change sequence was constructed to analyze the persistence, regularity, and predictability of the focusing behavior. This process not only reveals the physical mechanism of radiation convergence but also clarifies the dominant role of building structures in the distribution of radiation energy, laying the foundation for focusing trajectory modeling.

[0053] Based on the identified focusing areas and their energy transmission links, a focusing trajectory map is generated to characterize the spatiotemporal migration path of infrared energy under airflow disturbance conditions. This trajectory map, with time as the horizontal axis, spatial location as the vertical axis, and energy density as the weighting parameter, depicts the complete path structure of infrared radiation from the initial reflecting surface through multiple propagations to the final sensing surface over a continuous time period. Furthermore, the trajectory map is integrated with changes in building internal thermal parameters to achieve dynamic response and identification of radiation intensity. The statistical characteristics of the trajectory map can be used to determine whether a particular structural combination exhibits a high-risk focusing tendency under specific ventilation conditions, and the evolution trend of this tendency. This result not only possesses a high-precision description capability of radiation convergence phenomena but also provides a continuous and quantitative input basis for subsequent dynamic reflection hotspot identification and causal playback analysis.

[0054] The risk area identification module, based on energy focusing trajectory mapping, performs polarization spectral difference calculation and turbulence configuration analysis to identify radiation hotspots in local space, extract detector saturation trigger sequences caused by radiation enhancement, and generate a risk spatiotemporal window set corresponding to the sequence.

[0055] Based on the previously constructed infrared radiation focusing trajectory mapping results, the radiation hotspot areas inside the building are precisely identified and dynamically classified, and key sequences that trigger detector saturation responses are extracted to construct a risk spatiotemporal window set for subsequent prediction and intervention. The specific steps are as follows:

[0056] Based on the infrared energy focusing trajectory mapping formed in the previous steps, a continuous time segment with significantly increased energy density is selected as the starting point for analysis, and the polarization characteristic distribution of all infrared signals within the corresponding spatial region is obtained. Within this region, a multi-channel spectral scanning array is constructed by combining linear polarizers and multi-angle band filters to achieve synchronous acquisition of infrared bands under multiple polarization directions at the same time point. By comparing the spectral differences between adjacent bands and polarization directions, the multi-level interference fringe structure formed on the highly reflective surface is extracted, thereby identifying non-thermal energy anomalies caused by multiple reflections and surface enhancement. This process can distinguish the physical properties of locally high-energy focusing regions indirectly excited by airflow disturbances from signals from real flames or heat sources, providing high-precision input for subsequent determination of the cause of energy enhancement.

[0057] The identified spatial location and time point are used as the analysis window. Within this window, the 3D topological relationships of the corresponding region in the building structure model are invoked to calculate the instantaneous distribution characteristics of airflow vector direction, velocity change rate, and temperature difference field. Based on these parameters, a turbulent disturbance configuration map of the region under the current time conditions is constructed. This configuration map is combined with infrared radiation flux data for overlay analysis to determine whether specific airflow disturbances trigger enhanced energy focusing behavior, with particular attention to velocity reversal, shear force concentration, and persistent vortex residence in the surface normal direction. By verifying path correspondence and overlap with the dynamic focusing path in the energy trajectory mapping, high-risk reflection regions with long-term airflow disturbance triggering mechanisms are further identified, and their evolution path trends in the overall space are determined.

[0058] Within the identified high-risk focusing areas, hotspot regions overlapping with the sensing surface of the infrared acquisition device or intersecting with its projection path are extracted. The energy rise, peak hold, and decay processes in each region are analyzed segmentally over time. The original voltage output curve of the infrared signal is recalibrated for energy density, identifying rising segments that continuously exceed the steady-state threshold range, forming initial characteristic sequences representing potential triggers for optical saturation. Further, by combining changes in environmental parameters within the time window, such as airflow disturbance frequency, wind speed amplitude variation range, and wall temperature response curves, multidimensional clustering analysis is performed on these saturation characteristic segments to screen out energy-enhanced response trajectories with high repeatability and interference exclusion, ultimately forming a set of trigger sequences that induce detector saturation response. Each sequence contains a complete time start, peak time point, duration, energy fluctuation range, and associated reflective structure number and airflow characteristic vector.

[0059] Based on the established set of saturated trigger sequences, a risk spatiotemporal window set is constructed. Specifically, the time intervals and spatial projection locations within each saturated trigger sequence are mapped in a multidimensional way to form a three-dimensional matrix structure. Combined with the rate of change of environmental disturbance variables, dynamic weight calculations are performed on each unit to identify sensitive time periods and spatial locations where risks occur. On this basis, similar risk units are merged according to the rules of temporal continuity and spatial proximity, delineating a set of spatiotemporal windows with independent risk expression capabilities. Each risk window includes not only the risk level, start time, duration, and central energy location, but also its frequency of occurrence and evolution trend indicators in historical observation sequences. This set serves as the core input for subsequent causal inference and energy prediction processes, possessing the ability to pre-express high-energy interference areas within buildings, laying a data foundation for achieving stable infrared detection output.

[0060] The energy prediction module constructs a causal residual playback chain based on the risk spatiotemporal window set, executes the replay process of the known radiation sequence, predicts the temporal and spatial location of the infrared radiation energy peak in the next sampling period, outputs the detector saturation probability and energy duration intensity, and forms a set of prediction parameters for sampling scheduling.

[0061] Based on the risk spatiotemporal window set obtained from prior identification, causal analysis and dynamic prediction of infrared radiation energy behavior are performed. By constructing a replay link with temporal memory capabilities, the precise location of future infrared energy peaks is achieved, forming a predictive basis for sampling scheduling. The specific steps are as follows:

[0062] For the established set of risk spatiotemporal windows, the infrared radiation energy sequence, airflow disturbance parameters, temperature gradient changes, reflective structure numbers, and spatial locations corresponding to each window are extracted in chronological order. To ensure the continuity of the analysis and spatial resolution, all input data undergo uniform sampling period calibration to reconstruct energy evolution trajectories at equal time intervals. Based on this, a reference window is established and used as a comparison object to compare energy change patterns in other windows. By calculating the energy rise rate, peak duration, decay curvature, and disturbance response delay, historical sequences with high similarity are extracted as a candidate sequence set. Each sequence in this set records the infrared radiation response characteristics under similar environmental disturbance conditions, providing source data for subsequent replay chain construction.

[0063] Within the candidate sequence set, a residual replay chain is constructed based on temporal causality. Starting from a reference window, this chain sequentially connects multiple historically observed highly similar sequences, and performs residual quantization on the time offset, energy amplitude differences, and perturbation amplitude errors between connected nodes. The residual data is embedded as link weights in the link structure to measure the stability and prediction accuracy of the causal chain. In this chain, each node not only retains the original energy sequence but also includes its background information under specific perturbation conditions. By traversing this residual replay chain, radiative behavior under various similar historical scenarios can be reproduced, and logical deductions of energy evolution trends in future sampling periods can be achieved. In particular, for windows where a complete energy curve has not yet been observed at the current time, interpolation reconstruction is performed using similar paths on the replay chain to generate prediction results with temporal shift capabilities.

[0064] By utilizing the predicted paths generated during the playback process, the time nodes and spatial locations of potential infrared energy peaks within the next sampling period are identified. To this end, each playback path is mapped to the current building structure coordinate system, and spatial matching is performed in conjunction with the boundary conditions of current environmental parameters. By calculating the projection position of the predicted energy peak onto the sensing surface, it is determined whether the area falls within the sensitive area of ​​the acquisition device. If it is within the coverage area, the predicted time corresponding to the peak energy is further analyzed, and the sampling capability is time-aligned with the device's current sampling cycle. Based on this, the energy value range, prediction time error limit, and spatial location uncertainty index of the predicted point are output. Simultaneously, the historical misjudgment rate distribution is overlaid to form a probability model indicating potential detector saturation. By weighted aggregation of multiple predicted points, a high-risk distribution map that may trigger a saturation response within a continuous time period is obtained.

[0065] Based on the prediction results, a set of prediction parameters required for sampling scheduling is constructed. This parameter set includes the predicted time-location pair, peak amplitude range, corresponding risk level, disturbance response delay, historical replay path number, and suggested sampling cycle adjustment value for each potential energy peak point in the next cycle. Furthermore, this parameter set is input into the front-end acquisition process management architecture to achieve dynamic sorting of sampling priorities and real-time adjustment of the integration time window, thereby realizing predictive data acquisition and feedforward anti-interference control. This sampling scheduling mechanism can proactively avoid the saturation risk caused by high-energy focal points, while increasing the proportion of effective sampling data, providing a higher-quality input basis for subsequent energy compensation and alarm judgment. Through this method, the overall detection process no longer relies on static threshold judgment, but instead achieves steady-state control and adaptive intervention based on dynamic prediction.

[0066] The front-end steady-state control module introduces phase conjugate radiation cancellation control to the front end of the detector based on the prediction parameter set, and configures a reversible time grid sampling structure to adjust the integration time node in real time according to the prediction beat, thereby weakening the superposition of energy peaks at the prediction position and forming a stable control over the input of the detection signal.

[0067] Based on the constructed set of prediction parameters, the radiation response path and time sampling structure of the infrared detector front end are intervened and controlled to weaken the cumulative effect of the predicted energy peak on the sensor surface, thus ensuring the stability and distinguishability of the signal input. The specific steps are as follows:

[0068] Based on the spatial location, time node, and amplitude range of the future energy peak provided by the predicted parameter set, an infrared input path with corresponding direction and time sequence is selected. Before the infrared wavefront reaches the detector, a radiation wavefront structure is constructed that is spatially reversed, temporally reversed, and matches the original radiation path in terms of energy amplitude. This structure is realized through an electrically controlled reflective array with nonlinear optical path compensation capability. The array elements are driven according to the phase structure of the predicted wavefront, causing them to emit interference waves with phase conjugation characteristics. The interference wave and the soon-to-arrive original infrared radiation wave undergo wavefront cancellation in free space in front of the sensor, weakening the concentrated superposition effect of energy on the sensing surface. To ensure the stability and real-time performance of this phase conjugation interference process, the wavefront evolution speed and wavelength matching capability of the array are controlled by the time error limit in the preceding predicted path, achieving sub-millisecond-level radiation compensation synchronization.

[0069] After completing spatial wavefront modulation, a reversible time-grid sampling structure is constructed to address residual energy focusing that may still exist in some non-principal axis directions or short-time paths, based on the temporal distribution curves of each prediction point provided by the prediction parameter set. This structure consists of multiple sets of electro-optic switches and high-speed integrating charge-coupled devices (CCDs). Each sampling unit has independent integration on-off cycles at different time points. By nonlinearly interleaving the energy fluctuation rhythm of future high-energy regions with the sampling cycle, the peak energy is ensured to fall within the window of the integration period when it is sensed, preventing high-intensity energy from entering the integration peak period. The control logic of this structure originates from the integration time adjustment suggestions given in the previous steps and is synchronously updated based on the residual intensity curve after wavefront cancellation, ensuring that the energy flux of each sampling cycle stably falls within the dynamic range of the sensor.

[0070] Building upon wavefront cancellation and time sampling adjustment, the input signal equalization among multiple infrared acquisition channels was further analyzed and optimized. The energy density received by each channel within the current period was normalized, and the response amplitudes of each channel were automatically weighted and corrected based on the energy deviation distribution model in the prediction path. If a channel fails to fully perform wavefront cancellation due to its spatial location, other channels provide energy compensation corresponding to the predicted redundant path, thus laterally weakening single-point high-energy fluctuations through a distributed redundant sampling strategy. This process not only improves the tolerance of the entire front-end sampling structure to local abnormal energy inputs but also provides a multi-path comparison basis for signal reconstruction in subsequent data processing stages.

[0071] Based on multi-channel energy equalization, and according to the trend of the duration and repetition period of peak energy in the predicted parameter set, the dynamic allocation logic of the integral capacity in the front-end charge scheduling structure is initiated. This logic slowly transfers some of the integral capacity required for short-term high-energy segments to long-period background energy segments, enabling the front-end acquisition unit to have greater charge redundancy before the peak arrives, thereby further reducing the saturation response caused by integral overflow. This scheduling strategy dynamically adjusts the parameter curves based on the periodic risk level and historical trigger frequency in the predicted path, constructs a non-uniform time integral model, and realizes a peak-avoidance and valley-filling energy processing mechanism, effectively improving the stability of the overall infrared sensing capability under complex interference backgrounds.

[0072] The infrared input signal, influenced by spatial wavefront interferometry, time sampling rearrangement, channel response equalization, and charge capacity scheduling, is fused to generate a steady-state controlled infrared radiation input sequence. This sequence is then compared item by item with the original predicted parameters to construct an energy response residual model for evaluating the control effect. Before the start of the next sampling cycle, the wavefront adjustment structure and time sampling cycle are reconfigured based on the maximum deviation and cumulative distortion rate of the residual model in the current cycle. This achieves cross-cycle dynamic feedback optimization, ensuring that the infrared input signal remains within the detector's linear response range and has sufficient dynamic margin to cope with uncertain behavior under high-frequency energy disturbances. This provides stable, reliable, and unsaturated basic data support for final data discrimination and alarm decision-making.

[0073] The identification optimization and closed-loop adjustment module, based on the detection output after stable control, initiates the Hamilton variational path calculation and energy return trap collaborative mechanism. By dynamically injecting reflection bias values, it achieves consistent voting of cross-channel output results. Based on the voting results, it dynamically adjusts the identification weight and alarm judgment threshold to complete the closed-loop steady-state adaptive adjustment of the infrared detection process.

[0074] For the infrared detection output results after steady-state regulation, a path evolution and energy management mechanism is further introduced to dynamically adjust the output judgment parameters, achieving closed-loop steady-state adaptive regulation of the entire infrared detection process. The specific steps are as follows:

[0075] Based on the infrared output data obtained after wavefront cancellation, time sampling rearrangement, and input signal stabilization in the previous cycle, a path calculation process based on the Hamiltonian variational principle is initiated, targeting the energy variation trend of each channel sampling point in the stable signal sequence. Specifically, by combining the spatial topological location of each channel sensing surface, the physical geometric parameters of the reflection structure, and the historical evolution curve of energy flux, a set of multiple possible energy transmission paths is established. The optimal path trajectory from the initial disturbance source to each channel sampling point is calculated by constructing an objective function that minimizes energy dissipation. This path not only considers maximizing energy transmission efficiency but also dynamically assigns weights to wavefront interference, reflection attenuation, and sampling occlusion factors present on the path, forming a spatial migration model with real propagation constraints. This model determines the most likely migration trajectory of energy from the source point to the sensing point in the current cycle, laying the spatial path foundation for subsequent reflection compensation and cross-channel fusion.

[0076] Based on the obtained optimal energy path trajectory, the deployment logic of the energy return trap is executed to recover, buffer, and adjust energy inputs that deviate from the optimal path trajectory but still possess signal value. This trap structure constructs a reverse guidance zone for incident infrared waves around the sensing area using a variable reflective film array, redirecting high-energy radiation that might cause signal drift back to the main path channel along a secondary path. The trap activation process is controlled by the residual field between the energy prediction model of the previous cycle and the sampling data of the current cycle. When an abnormal energy surge occurs in a sensing area but does not trigger saturation, this energy component is introduced into the return path, and its energy bandwidth is limited by an additional filter bank to prevent misjudgment or threshold shift when it is reprojected onto the sensing surface. This step expands the system's tolerance for non-ideal energy inputs without affecting normal sensing energy acquisition.

[0077] After the energy path trajectory and backtracking trap mechanism are completed, to address the potential output fluctuation differences across multiple sampling channels for the same physical event, a reflection bias value is dynamically injected to improve the consistency of cross-channel results. Specifically, a light-controlled medium layer for phase delay adjustment is embedded in each sensing channel. A unified energy reference curve is constructed based on the sampling results of other channels, and the actual output energy of the current channel is compared with this reference curve. When a significant deviation exists, the phase delay characteristics of the light-controlled medium are adjusted to create a slight misalignment in the time or spatial dimension of the reflected wave received by that channel, offsetting or compensating for the accumulated error within the current period, thereby adjusting the output results of each channel to the same energy response range. The injection of this reflection bias value does not change the original energy path and sensing structure; rather, it is fine-tuned through medium control before output formation, ensuring high cross-channel consistency in the final data used for judgment.

[0078] After obtaining consistent output results, the identification weights and alarm thresholds in the infrared detection process are dynamically adjusted based on the stability of the sampling channel outputs and the residual evaluation results of the energy prediction model. The identification weight adjustment is based on the historical hit rate of the channel and the energy signal-to-noise ratio change in the current cycle, prioritizing the identification contribution of channels exhibiting high consistency and high resolution across multiple cycles. The alarm threshold is dynamically adjusted based on the prediction error and actual misjudgment frequency within consecutive cycles. For environmental conditions with significant interference frequency backgrounds, the threshold is raised to suppress false triggering by non-fire sources; simultaneously, for situations with high-confidence focusing trends in the prediction path, the threshold is appropriately lowered to improve the system's sensitivity to early fire sources. These dynamically adjusted parameters are reconfigured at the end of each cycle and input into the detection decision process for the next cycle.

[0079] The four processes—path optimization, energy return, channel consistency correction, and parameter reconstruction—are unified and converged to generate an integrated infrared detection output value after closed-loop stable control in the current cycle. This integrated value is verified for error range through an internal cross-validation mechanism. If all channel outputs fall within the tolerance deviation band, it is considered to be in steady-state effective. If the deviation rate of any channel exceeds the preset tolerance, it is fed back to the path evolution and return adjustment steps, initiating a new round of correction. Through this closed-loop feedback mechanism, the infrared detection process no longer relies on fixed thresholds or static rules, but rather on an adaptive structure where time-varying input, predicted output, and spatial structure jointly determine the system behavior. This enables continuous and stable operation under complex thermal disturbances and building reflection interference environments, ensuring that alarm decisions have sufficient accuracy, real-time performance, and anti-interference capabilities.

[0080] This invention, based on environmental modeling, accurately depicts the spatial focusing trajectory of infrared radiation under complex airflow disturbances. Combined with high-risk area identification and energy behavior prediction, it enables proactive prediction and intervention in potential interference areas. At the sampling front end, a phase conjugate radiation cancellation mechanism and a reversible time grid sampling structure are introduced to effectively reduce the transient superposition of energy peaks and ensure signal input stability. The system also performs path reconstruction and energy foldback adjustment based on dynamic feedback of detection results. Through a cross-channel consensus voting mechanism, it optimizes identification weights and alarm thresholds, constructing a closed-loop steady-state adaptive control capability for the entire infrared sensing process. This solution significantly improves the robustness and identification accuracy of infrared detection in fire monitoring, possesses excellent environmental adaptability, and is particularly suitable for high-rise buildings, underground spaces, and areas with limited energy supply, promoting the intelligent transformation of fire monitoring from passive response to prediction-driven, adaptive control.

[0081] The foregoing has only described certain exemplary embodiments of the present invention by way of illustration. Undoubtedly, those skilled in the art can modify the described embodiments in various ways without departing from the spirit and scope of the present invention. Therefore, the foregoing drawings and descriptions are illustrative in nature and should not be construed as limiting the scope of protection of the claims of the present invention.

Claims

1. A building fire safety monitoring system based on solar power, characterized in that, It includes a radiation environment modeling module, a risk area identification module, an energy prediction module, a front-end steady-state control module, and an identification optimization and closed-loop adjustment module: The radiation environment modeling module constructs a coupled observation baseline for infrared radiation and airflow, collects ventilation intensity, temperature gradient, and reflection structure parameters, builds a spatiotemporal distribution model of radiation, and generates an energy focusing trajectory mapping. The steps for generating the energy-focused trajectory map are as follows: Establish a coupled observation baseline for infrared radiation and airflow, and continuously collect ventilation intensity, temperature gradient, and reflector geometric parameters; A radiation spatiotemporal distribution model is constructed based on the collected data, and the ventilation intensity is mapped into an airflow vector field. Combined with the temperature gradient and three-dimensional building structure, an infrared radiation flux field is formed. In the flux field, identify the high-frequency infrared focusing region, extract the energy transmission link, and calculate the degree of energy density concentration; Based on the energy transmission link, a focusing trajectory mapping is generated to characterize the spatiotemporal migration path of infrared energy under airflow disturbance conditions and form a structured analysis input; The risk area identification module, based on focused trajectory mapping, performs spectral difference and turbulence analysis to identify radiation hotspots, extract saturation trigger sequences, and generate corresponding risk spatiotemporal window sets. The energy prediction module constructs a causal residual playback chain based on the risk spatiotemporal window set, replays historical radiation sequences, predicts the energy peak position and duration of the next cycle, and forms the prediction parameter set required for sampling scheduling. The steps for generating the prediction parameter set are as follows: Infrared radiation energy sequences and environmental parameters are extracted based on risk spatiotemporal window sets, energy evolution trajectories under a unified sampling period are constructed, and highly similar historical sequences are extracted. Based on temporal causality, a residual playback chain is constructed, embedding time offset and energy difference residual data to form a path for reproducing historical behavior; The predicted path is mapped to the building coordinate system, and the predicted time location and saturation probability of the energy peak are output. Construct a set of prediction parameters and input them into the data acquisition process to achieve sampling scheduling and integration time adjustment; The prediction parameter set includes prediction time location pairs, peak amplitude range, risk level, disturbance response delay, historical replay path number, and sampling beat adjustment suggestion value, which are used to guide the dynamic sorting of front-end sampling beats and real-time adjustment of the integration time window; The front-end steady-state control module, based on the predicted parameter set, implements phase conjugate radiation cancellation and reversible time grid sampling, dynamically adjusts the integration time, suppresses the superposition of energy peaks, and forms a stable input of the detection signal; The identification optimization and closed-loop adjustment module utilizes a stable input signal to perform path variation and energy return control. By injecting reflection bias, it performs cross-channel voting, optimizes identification weights and alarm thresholds in real time, and completes the closed-loop steady-state adaptive adjustment of the detection system.

2. The building fire safety monitoring system based on solar power according to claim 1, characterized in that, The steps for generating the risk spatiotemporal window set are as follows: Based on energy focusing trajectory mapping, continuous time segments with increasing energy density are selected, and polarization spectral difference calculation is performed to identify non-thermal energy anomalies. The three-dimensional topological relationship of the building structure in the corresponding area is invoked, and a turbulent disturbance configuration map is constructed by combining the airflow vector direction, the rate of change of flow velocity and the temperature difference field, and high-risk reflection areas are identified. Extract the infrared output curve of the hot spot area and identify the rising segments that continuously exceed the steady-state threshold to form a saturation trigger sequence; A multidimensional mapping relationship is established based on the time intervals and spatial locations contained in the saturation trigger sequence, a risk spatiotemporal window set is constructed, and an evolution trend indicator is generated.

3. The building fire safety monitoring system based on solar power according to claim 2, characterized in that, When performing polarization spectral difference calculations, infrared spectral data under different polarization directions are collected by combining multi-angle band filters and linear polarizers to extract the interference fringe structure formed by multiple reflections and distinguish non-thermal source high-energy focusing regions.

4. The building fire safety monitoring system based on solar power according to claim 1, characterized in that, The process of forming a stable input of the detection signal is as follows: Based on the predicted parameter set, a phase conjugate radiation wavefront structure that is spatially reversed, temporally reversed, and energy amplitude matched with the infrared input path is constructed to achieve wavefront cancellation control; By combining the predicted beat and the residual intensity, a reversible time grid sampling structure with independent integral beats is constructed to achieve the misalignment control of the integral period and the energy peak. Normalization and weighted correction are performed on multi-channel input signals to achieve energy compensation and lateral attenuation of redundant paths; Based on the periodic characteristics in the predicted parameters, a dynamic allocation strategy for the integral capacity is implemented to construct a non-uniform time integral model; The regulated input sequence is compared with the predicted parameters to construct an energy residual model and complete the next cycle of regulation configuration.

5. The building fire safety monitoring system based on solar power according to claim 4, characterized in that, Using a stable input signal, path variational calculation and energy return control are performed, reflection bias is injected, and cross-channel voting is conducted. The steps for dynamically adjusting the identification weight and alarm threshold are as follows: Based on the infrared detection output after steady-state regulation, Hamiltonian variational path calculation is performed to obtain the optimal energy migration trajectory from the disturbance source to the sampling point. Energy back-tracing traps are set up according to the optimal path to guide high-energy inputs that deviate from the path into the auxiliary path and limit their energy bandwidth. Dynamically inject reflection bias values ​​to adjust the phase of reflected waves between channels and achieve output consistency; Based on the channel stability and residual evaluation results, the identification weight and alarm judgment threshold are dynamically adjusted; The current cycle output is checked for error. If the deviation exceeds the limit, a new round of path and turnaround adjustment is triggered.

Images