Substation transformer remote fire-fighting early warning method and system based on Internet of Things

Abstract

The invention provides a transformer substation transformer remote fire-fighting early warning method and system based on the Internet of Things, and the method comprises the steps: constructing a fire-fighting space model in a three-dimensional manner according to the spatial positions of a fire-fighting device and a transformer in a transformer substation; performing fire risk monitoring on the transformer, and when a fire occurs in the transformer, determining a fire combustion point corresponding to each fire time point and predicting a fire spreading direction based on the fire images of the plurality of fire time points of the transformer; according to each fire combustion point and the fire spreading direction, a fire-fighting strategy is optimized in combination with the fire-fighting space model, and the fire-fighting strategy is used for fire-fighting treatment; using the device state information of the fire-fighting device in the fire-fighting treatment to generate a maintenance early warning of the fire-fighting device; and recording fire-fighting images of a plurality of fire-fighting time points in fire-fighting processing, and generating fire-fighting early warning according to each fire-fighting image and the device state information. Therefore, the accuracy of fire early warning and the efficiency of emergency disposal are remarkably improved.

Description

Technical Field

[0001] This application relates to the field of fire early warning technology, and in particular to a remote fire early warning method and system for substation transformers based on the Internet of Things. Background Technology

[0002] In substation operation, the fire protection system is a crucial protective facility for ensuring the safety of electrical equipment and personnel. Its monitoring and early warning capabilities directly affect the timeliness and effectiveness of fire prevention and control. With the expansion of the power grid and the increasing complexity of the operating environment, dispatch monitoring centers and on-site maintenance departments have placed higher demands on the real-time and accurate monitoring of the fire protection system's operational status.

[0003] While existing substations are equipped with fire alarm systems and can send alarm or fault signals to the dispatch and monitoring center, the transmitted content is mostly general, aggregated signals, lacking granular monitoring of the operational status of core components. For example, the dispatch center cannot know in real time whether the automatic fire alarm system, main transformer sprinkler system, and gas extinguishing system are in automatic mode, nor can it ascertain whether fire hydrants are functioning properly or whether the fire water tank level is sufficient. Furthermore, some substations still use traditional fire management methods, resulting in narrow information coverage, isolated data, and insufficient targeting in fire dispatch and response, hindering the full utilization of information technology in fire prevention and control.

[0004] The core problem resulting from this is that the existing substation fire protection system has shortcomings in information acquisition and transmission, such as insufficient granularity, limited coverage, and isolated data. Dispatch and maintenance personnel have difficulty grasping the operating status of key components in a timely and accurate manner, which restricts the improvement of fire early warning and emergency response capabilities. Summary of the Invention

[0005] The purpose of this application is to address at least one of the aforementioned technical deficiencies, particularly those in the prior art that restrict the improvement of fire early warning and emergency response capabilities.

[0006] Firstly, this application provides a remote fire early warning method for substation transformers based on the Internet of Things, the method comprising:

[0007] Based on the spatial location of fire-fighting equipment and transformers in the substation, a three-dimensional fire-fighting space model is constructed.

[0008] Fire risk monitoring is performed on transformers, and when a fire occurs in a transformer, the fire burning point corresponding to each fire time point and the direction of fire spread are determined based on fire images of the transformer at multiple fire time points.

[0009] Based on the individual fire ignition points and the direction of fire spread, fire-fighting strategies are optimized using a fire-fighting space model. These strategies are then used to handle fire-fighting situations.

[0010] Utilize the device status information of fire protection equipment during fire protection operations to generate early warnings for the maintenance of fire protection equipment;

[0011] Record fire images at multiple fire time points during firefighting operations, and generate fire warnings based on each fire image and device status information.

[0012] In one embodiment, the steps for fire risk monitoring of a transformer include:

[0013] Obtain the normal concentration and real-time concentration of each gas type within the transformer's operating range, as well as the normal temperature and real-time temperature of the transformer.

[0014] Calculate the gas concentration deviation based on the normal and real-time concentrations of each gas type, and calculate the temperature deviation based on the normal and real-time temperatures.

[0015] Calculate transformer anomaly values ​​based on gas concentration deviation and temperature deviation;

[0016] The fire risk of transformers can be predicted by using transformer anomaly values ​​and preset transformer anomaly value thresholds.

[0017] In one embodiment, the steps of determining the fire burning point corresponding to each fire time point and predicting the direction of fire spread based on fire images of multiple fire time points of a transformer fire include:

[0018] Obtain images of the transformer in normal operating condition;

[0019] Compare the fire images at each fire time point with normal working images to extract the fire burning point corresponding to each fire time point;

[0020] The fire burning point corresponding to each fire time point is mapped to the fire space model to determine the direction of fire spread.

[0021] In one embodiment, the step of mapping the fire combustion point corresponding to each fire time point to a fire-fighting space model and determining the direction of fire spread includes:

[0022] Map the fire burning point corresponding to each fire time point to the fire space model to obtain the coordinates of the burning point corresponding to each fire time point;

[0023] Based on the coordinates of the combustion point corresponding to each fire time point, calculate the combustion midpoint corresponding to each fire time point, and calculate the change vector of the combustion midpoint corresponding to adjacent fire time points;

[0024] The average value of the change vectors at each combustion midpoint is taken to obtain the fire spread vector, and the fire spread vector is normalized to obtain the fire spread direction.

[0025] In one embodiment, the fire-fighting device includes the steps of deactivating sprinklers and optimizing the fire-fighting strategy based on each fire point and the direction of fire spread, in conjunction with a fire-fighting space model, including:

[0026] The burning area is determined based on the coordinates of each fire point in the fire-fighting space model;

[0027] Based on the device coordinates of the calling sprinkler in the fire space model and the coordinates of the combustion points adjacent to the calling sprinkler, the spray vector is calculated, and the spray angle is calculated using the spray vector.

[0028] Calculate the cross product of the fire spread direction and the spray vector to obtain the atomization vector, and then perform operations on the spray vector and the atomization vector to obtain the atomization angle;

[0029] The fire-fighting strategy for the combustion zone is optimized by using spray angle and atomization angle.

[0030] In one embodiment, the fire-fighting device includes a call nozzle and a pipe, and the device status information includes the spray volume of the call nozzle and the water consumption of the pipe.

[0031] The steps for generating maintenance early warnings for fire protection devices using device status information during fire suppression include:

[0032] Calculate the average spray volume and the mode of spray volume based on the spray volume of each nozzle, and determine the normal spraying range based on the average spray volume and the mode of spray volume.

[0033] Sprinklers whose spray volume is outside the normal spray range are recorded as abnormal sprinkler calls, and maintenance warnings for abnormal sprinkler calls are generated.

[0034] Calculate the abnormal water consumption value based on the water consumption of the pipeline and the spray volume of each sprinkler head.

[0035] When the abnormal water consumption value exceeds the preset consumption ratio, a pipeline maintenance warning is generated.

[0036] In one embodiment, the step of generating a fire warning based on various fire images and device status information includes:

[0037] The fire intensity change value is determined by using various fire images, and the feasible value of fire treatment is determined based on the device status information;

[0038] Based on the fire intensity change value and the fire control feasibility value, the fire control status is determined, and a fire warning is generated based on the fire control status.

[0039] In one embodiment, the step of determining the fire intensity change value using various fire images includes:

[0040] The fire image at each fire time point is mapped to the fire space model to obtain the remaining burning range at each fire time point;

[0041] Calculate the fire intensity change value based on each remaining burning area.

[0042] In one embodiment, the step of determining the feasible value for fire suppression based on device status information includes:

[0043] Extract the remaining fire-fighting resources at each fire-fighting time point from the device status information;

[0044] Calculate the reduction rate of the burning range based on each remaining burning range, and calculate the consumption rate of fire-fighting resources based on each remaining value of fire-fighting resources.

[0045] The feasible value of fire treatment is calculated based on the reduction ratio of the combustion range and the proportion of fire-fighting resource consumption.

[0046] Secondly, this application provides a remote fire early warning system for substation transformers based on the Internet of Things, including: one or more processors, and a memory;

[0047] The memory stores computer-readable instructions, which, when executed by one or more processors, perform the steps of any of the IoT-based remote fire early warning methods for substation transformers in the above embodiments.

[0048] As can be seen from the above technical solutions, the embodiments of this application have the following advantages:

[0049] This application provides a remote fire early warning method and system for substation transformers based on the Internet of Things (IoT). By using IoT technology to create a 3D model of the spatial location information of fire-fighting devices and transformers within the substation, it achieves precise spatial positioning of fire scenarios. When a fire occurs, it can analyze multi-moment fire images to identify the combustion point and predict the direction of fire spread. Combined with the spatial model, it optimizes fire-fighting strategies, thereby selectively dispatching appropriate fire-fighting systems and devices, improving the timeliness and effectiveness of fire response. Simultaneously, the method collects real-time operational status information of fire-fighting devices during fire suppression and generates fire and maintenance early warnings by combining multi-moment fire images. This solves the problems of insufficient information granularity, limited coverage, and data isolation in existing technologies, enabling dispatch and maintenance personnel to comprehensively, in real-time, and accurately grasp the operational status of key components, significantly improving the accuracy of fire early warning and the efficiency of emergency response. Attached Figure Description

[0050] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0051] Figure 1 A flowchart illustrating the IoT-based remote fire early warning method for substation transformers provided in this application embodiment;

[0052] Figure 2 A schematic diagram of fire-fighting angles provided for embodiments of this application;

[0053] Figure 3 A schematic diagram of the structure of the IoT-based remote fire early warning device for substation transformers provided in this application embodiment;

[0054] Figure 4 This is a schematic diagram of the internal structure of a computer device provided in an embodiment of this application. Detailed Implementation

[0055] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0056] This application provides a remote fire early warning method for substation transformers based on the Internet of Things. The following embodiments illustrate this method by applying it to computer equipment. It is understood that the computer equipment can be various devices with data processing capabilities, including but not limited to a single server, server cluster, personal laptop, desktop computer, etc. Figure 1 As shown, the method includes:

[0057] S101: Based on the spatial location of fire-fighting equipment and transformers in the substation, construct a three-dimensional fire-fighting space model.

[0058] A substation is a comprehensive power facility used for the transformation, distribution, and control of electrical energy. The transformer is the core equipment in a substation that transforms electrical energy into voltage; its spatial location serves as the reference for establishing a three-dimensional coordinate system. Firefighting equipment refers to a collection of devices installed in a substation for fire detection, alarm, and extinguishing, including fire pipes, sprinklers, and deluge valves. Spatial location refers to the parameter information characterizing the position, orientation, and size of a target in three-dimensional space under a unified reference system. Three-dimensional refers to the three-dimensional modeling of an object within the space formed by three orthogonal coordinate axes (X, Y, and Z). A fire protection spatial model is a data model that integrates the geometric positions, structural features, and annotation information of transformers and firefighting equipment in a three-dimensional coordinate system using computer equipment.

[0059] Specifically, the computer equipment first acquires the spatial location information of the transformer, which can be obtained through a laser scanner, total station, BIM data, or high-precision surveying equipment. Then, the computer equipment establishes a three-dimensional coordinate system with the transformer's grounding vertex as the origin. The direction perpendicular to the ground and upward through the origin is defined as the Z-axis, the direction parallel to the ground and to the right through the origin is defined as the Y-axis, and the direction perpendicular to both the Z-axis and Y-axis through the origin is defined as the X-axis. Next, the computer equipment fills the three-dimensional coordinate system with the transformer's geometric shape and dimensions based on the spatial location information, achieving a precise representation of the transformer in three-dimensional space.

[0060] Computer equipment acquires the spatial location information of fire-fighting devices, which can also be provided by on-site surveying equipment, IoT sensor data, or design drawings. The computer equipment maps the location of each fire-fighting device in an established three-dimensional coordinate system and labels fire pipes, sprinklers, and deluge valves, including information such as equipment type, number, functional parameters, and installation height. The computer equipment then integrates the three-dimensional representation data of the transformer and fire-fighting devices to generate a fire-fighting spatial model containing geometric and semantic information for subsequent fire risk analysis and fire-fighting strategy development.

[0061] It is understandable that a fire protection spatial model needs to be constructed based on the spatial location of fire protection devices and transformers in the substation. This is to create a unified three-dimensional coordinate system in computer equipment, accurately correlate core equipment with fire protection devices, and calculate their location relationships, coverage areas, and accessibility. This approach provides high-precision spatial data support for fire monitoring and response, avoiding deficiencies in location accuracy, real-time performance, and information completeness. This allows dispatching and maintenance personnel to intuitively and accurately grasp the equipment layout and the status of fire protection facilities, improving the targeting and efficiency of fire response, and ultimately enhancing the overall protective capability of the substation's fire protection system.

[0062] S102: Conduct fire risk monitoring on transformers, and when a fire occurs on a transformer, determine the fire burning point corresponding to each fire time point and predict the direction of fire spread based on fire images of the transformer at multiple fire time points.

[0063] Fire risk monitoring refers to the continuous detection and assessment of potential fire signs or existing combustion states in transformers using sensor data and image data such as temperature, smoke, and flame spectral / visible light characteristics. Fire images refer to image frames acquired by fixed or rotatable imaging devices during a fire, timestamped, and used to describe the flame morphology and intensity distribution. A fire time point refers to the instant of image acquisition recorded using a unified time reference, including... Multiple discrete moments. The fire ignition point refers to the spatial location representing the strongest flame or highest temperature at a given fire time point. The fire spread direction refers to the main direction vector of the flame's influence range or the movement of the ignition point between adjacent or consecutive fire time points, used to reflect the fire development trend.

[0064] Specifically, the computer equipment receives multi-source data streams from an infrared thermal imager, a visible light camera, and a temperature / smoke sensor, and performs a fusion judgment on temperature abrupt changes, smoke density increases, and flame texture features. When several consecutive frames meet preset trigger conditions (such as a combination of temperature threshold and flame shape threshold), it is determined that "a transformer fire has occurred," and the imaging channel associated with the transformer is locked into a high-frequency sampling mode. At the same time, a unified clock is used to timestamp each frame of the image, forming a sequence of fire images at multiple fire time points.

[0065] Next, the computer equipment performs preprocessing on the fire images at each fire time point, such as denoising, contrast stretching, and white balance correction. Then, it performs flame region segmentation, which may include temperature threshold / color model / flame texture discrimination, and calculates intensity or temperature-weighted centroids on the segmentation results. The centroid locations are then combined with maximum search and morphological thinning to suppress noise, obtaining the coordinates of the fire burning point corresponding to that time point. To improve robustness, the computer equipment can perform trajectory smoothing (such as Kalman filtering) on ​​burning points in adjacent frames and remove transient outliers, outputting... Time-location pair.

[0066] The computer device takes a multi-time-point combustion point sequence as input, calculates the displacement vector of adjacent time points and performs principal direction fitting, such as least squares linear fitting or RANSAC robust estimation, to obtain the main propagation direction and confidence level. When there is large deformation or multiple combustion points, the computer device can simultaneously perform optical flow / deformation field calculations on the outer contour of the flame region, and perform weighted fusion of the combustion point displacement and the leading edge propagation direction of the outer contour to output the propagation direction prediction result at the current moment.

[0067] It is understandable that the reason for determining the fire burning point and predicting the fire spread direction at each time point based on fire images from multiple fire time points during a fire is that multi-time series can provide dynamic change information on the center of flame intensity and the shape boundary. Based on this, computer equipment can obtain stable location of the combustion source and extract the main direction vector that evolves over time, thereby avoiding misjudgments caused by noise or obstruction in single-frame judgments. The position and direction results output by the joint time-space analysis can improve the positioning accuracy and the reliability of trend judgment, shorten the response decision time, and thus improve the substation's handling efficiency and safety assurance capabilities in the early stages of a fire.

[0068] S103: Based on the individual fire ignition points and fire spread direction, optimize the fire-fighting strategy using the fire-fighting space model. The fire-fighting strategy is used for fire-fighting operations.

[0069] Firefighting strategy refers to a set of fire response instructions that can be directly issued, including device identification, action type, and parameters. Firefighting action refers to the execution process of the firefighting strategy on the on-site devices.

[0070] Specifically, the computer equipment receives the fire ignition points and fire spread directions from each fire, maps them to the fire protection space model, generates a predicted impact zone in the model based on the direction vector, such as a strip or cone-shaped area along the direction, and spatially overlays it with sensitive targets such as the transformer body, oil tank / cable trench, and adjacent live equipment to determine the set of targets that need to be prioritized for protection and treatment, as well as their hazard levels.

[0071] The computer equipment retrieves fire-fighting devices located within or near the influence zone from the model, calculates the visibility / accessibility, coverage overlap rate, and action distance of each sprinkler / nozzle relative to the combustion point, and forms a set of candidate actions by combining media adaptation rules (such as priority water spraying for outdoor equipment and optional gas extinguishing for enclosed spaces); at the same time, it performs rapid verification of the hydraulic and gas release parameters of the pipeline network (pipe diameter, pressure, available flow rate, set time window) and eliminates actions that do not meet supply or safety constraints.

[0072] The computer equipment can construct an optimization variable vector containing device start-up and shutdown parameters with the goal of "maximizing the effective coverage of the combustion point and predicted spread path, minimizing the number of activated devices and water / chemical dosage, minimizing response time, and minimizing secondary risks to critical equipment." This vector is then solved under constraints such as supply capacity, interlocking mechanisms, and electrical safety distances to obtain the optimal fire-fighting strategy. When the model is large, a combination algorithm of zoned start-up and shutdown and greedy-local search can be used to obtain a near-optimal solution, with confidence and constraint margin indicators attached to the strategy. The computer equipment can perform consistency verification of the optimal strategy (device interlocking, channel blockage, maintenance isolation status), generate an executable instruction set (including device ID, start time, target flow / duration / pulse mode, etc.), and distribute it via the industrial communication bus. Simultaneously, feedback monitoring is enabled to perform closed-loop comparison of measured values ​​of valve position, pressure, and flow, adjusting the strategy as needed based on the latest combustion point and direction increments.

[0073] It is understandable that optimizing firefighting strategies based on individual fire ignition points and fire spread directions, combined with fire space models, is necessary. This approach allows for the matching and calculation of fire development trends with the spatial layout and performance parameters of firefighting equipment. This enables the precise selection of firefighting equipment that effectively covers the fire source and its spread path, while meeting constraints such as supply capacity and safe distances. Furthermore, it allows for the rational determination of start-up and shutdown sequences and operational parameters. This process not only reduces ineffective spraying and redundant coverage, lowering the risk of secondary damage to equipment in non-affected areas, but also significantly shortens fire response time, improves firefighting efficiency, and enhances the targeted nature of the response and overall protective effect while minimizing resource consumption.

[0074] S104: Utilize the device status information of fire protection devices during fire protection operations to generate maintenance early warnings for fire protection devices.

[0075] Among them, device status information refers to the status quantities and event records reported in real time or polled by fire-fighting devices and their controllers, sensors, and linkage units during fire-fighting operations. This includes, but is not limited to, valve position / open / closed status, valve motor current, pipeline pressure, instantaneous flow rate, pump speed, pump suction / differential pressure, sensor self-tests and fault codes, controller communication link health, power supply voltage / backup battery voltage, execution response delay, and records of manual intervention. Maintenance early warning refers to maintenance reminders generated by computer equipment based on device status information and determined by rules or models.

[0076] Specifically, during fire response, the computer equipment interacts with fire monitoring terminals, sensors, and control units within the substation to acquire real-time status information of these devices. The computer equipment performs data preprocessing on the collected status information, including time synchronization between different data sources, removal of invalid data due to communication anomalies, interpolation compensation for missing data, and normalization of numerical data for subsequent unified analysis. After data processing, the computer equipment establishes a mapping relationship between each status piece of information and its corresponding specific fire-fighting device in the system database, achieving a one-to-one correspondence between data and physical equipment.

[0077] Next, the computer equipment analyzes the processed device status information based on a pre-defined fault diagnosis rule base or machine learning prediction model. When abnormal device operation, performance degradation, or potential failure trends are detected, a maintenance warning is automatically generated. The maintenance warning may include the device identifier of the abnormality, the type of abnormality, the detection time, related operating data, and suggested maintenance measures. This maintenance warning can be pushed to the maintenance personnel's terminal through the dispatch monitoring platform, enabling maintenance personnel to arrange maintenance in a timely manner before the equipment completely fails, thereby improving the continuous availability of the fire protection system and the reliability of fire response.

[0078] It is understandable that the reason for using the status information of fire protection equipment to generate maintenance early warnings in fire fighting is to monitor the health status of the equipment in real time during operation and trigger maintenance prompts in advance when performance degradation or abnormal trends are detected, thereby preventing equipment failure during critical firefighting phases. This not only improves the continuous availability and reliability of the fire protection system but also ensures the stable operation of critical firefighting equipment during fire response, thus improving the overall efficiency and safety of firefighting operations.

[0079] S105: Records fire images at multiple fire time points during fire handling, and generates fire warnings based on each fire image and device status information.

[0080] Among them, fire time points refer to the time nodes marked at preset time intervals or when key events are triggered during fire fighting, used to record the fire status and progress of the response at that moment. Fire images refer to image data acquired by monitoring cameras, thermal imaging equipment, or other image acquisition devices at various fire time points, reflecting the fire range, smoke concentration, sprinkler effect, and changes in the on-site environment. Fire warnings refer to the alert information generated by computer equipment after comprehensively analyzing fire images and equipment status information, used to indicate potential fire spread trends, risks of inadequate response, or potential equipment failure, and include the corresponding level, location, and recommended response measures.

[0081] Specifically, during firefighting operations, the computer equipment establishes a communication connection with image acquisition units deployed within the substation. It automatically records fire images at preset time intervals or when triggered by critical events (such as sudden changes in flame temperature, rapid increases in smoke concentration, or sprinkler activation). These images can originate from high-definition visible light cameras, infrared thermal imaging equipment, or a combination of both, ensuring accurate acquisition of on-site fire information under varying lighting, smoke, or obstruction conditions.

[0082] Subsequently, the computer equipment synchronizes and correlates the fire images at each fire-fighting time point with the device status information collected during the same period, ensuring a one-to-one correspondence between image content and device operating status. Based on this, the image data undergoes preprocessing, including noise filtering, brightness / contrast optimization, and smoke and fire feature enhancement, to improve the accuracy of subsequent fire feature extraction.

[0083] Next, the computer equipment uses fire detection algorithms to analyze fire images, extracting key features such as the flame area, smoke spread range, and sprinkler coverage effect. This is combined with device status information (such as whether the sprinkler pressure is normal, whether the gas extinguishing cylinder pressure is sufficient, and whether the valves are fully open) for a comprehensive assessment. When abnormal fire spread, insufficient fire extinguishing coverage, or malfunctioning fire suppression equipment are detected, a fire warning is generated. This warning includes the risk level, fire location, cause of the anomaly, and recommended response measures. It can be pushed to maintenance personnel in real time through the dispatch and monitoring platform, enabling them to take swift and targeted action, thereby improving the timeliness and effectiveness of fire response.

[0084] It is understandable that recording fire images at multiple time points during firefighting operations and combining these images with equipment status information to generate fire warnings is necessary to create a continuous record of the fire's evolution over time. This allows for the fusion and analysis of on-site visual information and equipment operational data, enabling a comprehensive assessment of fire development trends, firefighting effectiveness, and equipment operational status. This approach can trigger early warnings when fire spreads, fire suppression coverage is insufficient, or equipment malfunctions are detected early. This allows maintenance personnel to adjust firefighting strategies or intervene with equipment in a targeted manner, preventing the fire from escalating and reducing delays in response, thereby improving the accuracy, timeliness, and reliability of the overall fire response.

[0085] In the above embodiments, IoT technology is used to create a 3D model of the spatial location information of fire-fighting devices and transformers within the substation. This enables precise spatial positioning of fire scenarios. When a fire occurs, the method can analyze multi-moment fire images to identify the combustion point and predict the direction of fire spread. Combined with the spatial model, fire-fighting strategies are optimized, allowing for targeted dispatch of appropriate fire-fighting systems and devices, thus improving the timeliness and effectiveness of fire response. Simultaneously, the method collects real-time operational status information of fire-fighting devices during fire suppression and generates fire and maintenance warnings based on multi-moment fire images. This addresses the problems of insufficient information granularity, limited coverage, and data isolation in existing technologies, enabling dispatch and maintenance personnel to comprehensively, in real-time, and accurately grasp the operational status of key components, significantly improving the accuracy of fire warnings and the efficiency of emergency response.

[0086] In one embodiment, the steps for fire risk monitoring of a transformer include:

[0087] Obtain the normal concentration and real-time concentration of each gas type within the transformer's operating range, as well as the normal temperature and real-time temperature of the transformer.

[0088] Calculate the gas concentration deviation based on the normal and real-time concentrations of each gas type, and calculate the temperature deviation based on the normal and real-time temperatures.

[0089] Calculate transformer anomaly values ​​based on gas concentration deviation and temperature deviation;

[0090] The fire risk of transformers can be predicted by using transformer anomaly values ​​and preset transformer anomaly value thresholds.

[0091] Among them, the normal gas concentration refers to the stable concentration benchmark value of various detectable gases within the transformer's operating range, obtained through long-term monitoring or experimental calibration, under normal transformer operation conditions without faults or leaks. The real-time gas concentration refers to the actual concentration value of various gases collected by a gas detection device at a specific moment during transformer operation. The normal temperature refers to the benchmark value of transformer operating temperature obtained through long-term measurement or experimental calibration under stable environmental and load conditions during normal transformer operation. The real-time temperature refers to the actual operating temperature collected in real-time by a temperature sensor during transformer operation. The gas concentration deviation refers to the difference between the real-time gas concentration and the corresponding normal gas concentration. The temperature deviation refers to the difference between the real-time temperature and the normal temperature. The transformer anomaly value refers to the numerical value characterizing the degree of abnormality in the transformer's operating state, obtained through a preset calculation model based on the gas concentration deviation and temperature deviation. The transformer anomaly value threshold refers to the preset numerical limit used in risk prediction to determine whether the transformer has a fire risk.

[0092] Specifically, the computer equipment first monitors the transformer's operating status in real time using a sensor array positioned around the transformer in the substation. The sensor array includes photoluminescent spectral sensors for detecting the concentrations of various combustible and characteristic gases around the transformer, and an infrared thermal imager for acquiring the temperature distribution on the transformer surface. Through these devices, the computer equipment can simultaneously acquire real-time data on both gas and temperature during transformer operation.

[0093] Subsequently, the computer equipment retrieves pre-stored normal operating data from the database to obtain the gas type zl and corresponding normal gas concentration cq(a) within the transformer's operating range, as well as the normal temperature cwd. Combining this with the real-time gas concentration sq(a) and real-time temperature swd collected by sensors, a basis for comparing the real-time and normal operating states is established.

[0094] Next, the computer equipment calculates the gas concentration deviation and temperature deviation based on the normal gas concentration cq(a), normal temperature cwd, real-time gas concentration sq(a), and real-time temperature swd, respectively. Then, it combines these two deviations using a preset mathematical model or weighted algorithm to obtain the transformer anomaly value ycz, which characterizes the degree of abnormality in the transformer's operating state. This anomaly value reflects risk characteristics such as gas leakage and overheating that could lead to fire. The formula for calculating the transformer anomaly value is:

[0095]

[0096] Finally, the computer equipment obtains the pre-set transformer anomaly threshold yyz and compares it with the calculated ycz. When ycz < yyz / 2, the transformer is determined to have no fire risk; when yyz / 2 ≤ ycz < yyz, an early warning message is generated and the transformer operating data is uploaded to the monitoring platform to prompt management personnel to take inspection measures; when ycz > yyz, a warning message is immediately generated, and the fire-fighting device is linked through the control interface to execute fire-fighting measures, thereby effectively intervening before the risk turns into an actual fire and achieving closed-loop control for early detection and rapid response to fires.

[0097] It is understandable that obtaining the normal gas concentration, real-time gas concentration, normal temperature, and real-time temperature is to establish a quantifiable basis for comparing the real-time state with the normal operating state during transformer operation. This allows for accurate reflection of the changing trends in the chemical and thermal states of the transformer by calculating gas concentration deviations and temperature deviations. Furthermore, by calculating transformer anomalies based on these deviations and comparing them with preset thresholds, fire risks can be assessed quantitatively. This method not only enables timely detection of hidden dangers and early intervention in the early stages of fire risk formation but also reduces false alarms and missed alarms, improving the accuracy and reliability of transformer fire monitoring, thereby significantly enhancing the operational safety and fire response efficiency of substations.

[0098] In one embodiment, the steps of determining the fire burning point corresponding to each fire time point and predicting the direction of fire spread based on fire images of multiple fire time points of a transformer fire include:

[0099] Obtain images of the transformer in normal operating condition;

[0100] Compare the fire images at each fire time point with normal working images to extract the fire burning point corresponding to each fire time point;

[0101] The fire burning point corresponding to each fire time point is mapped to the fire space model to determine the direction of fire spread.

[0102] Among them, the normal working image refers to the reference image that reflects the appearance, structural outline and surrounding environment of the transformer when the transformer is running normally and there is no fire.

[0103] Specifically, when monitoring fires in substation transformers, computer equipment first uses high-definition cameras or thermal imaging devices installed around the transformer to collect multi-angle normal working images of the transformer under normal operating conditions and without fire. These images are then processed for noise reduction, distortion correction, and brightness equalization to obtain normal working image data that serves as a benchmark for fire image comparison.

[0104] Subsequently, during the fire, the computer equipment extracts fire images from the monitoring video stream at preset time intervals (i.e., fire time points) and performs image enhancement and background separation processing on these fire images. Next, the computer equipment uses image registration and differential analysis algorithms to compare the fire images at each fire time point with normal working images pixel by pixel, eliminating interference caused by changes in lighting or camera shake, thereby accurately extracting the location of the fire burning point on the transformer surface or surrounding area at that time point.

[0105] After extracting the fire ignition points, the computer equipment maps the coordinates of the fire ignition points corresponding to each fire time point to a pre-constructed fire protection space model. This fire protection space model includes the three-dimensional spatial relationships of the transformer body, surrounding equipment, building structure, and fire protection devices. By comparing the distribution and changing trends of the fire ignition points in the model at different times, the computer equipment can determine the direction of fire spread in space, thus providing a data foundation for subsequent fire risk assessment and fire protection strategy optimization, achieving closed-loop processing from image acquisition to fire propagation analysis.

[0106] It is understandable that acquiring images of transformers in normal operating condition is to establish accurate baseline data without fire, which can be used for comparison of fire images. Comparing the fire image at each fire time point with the normal operating image can effectively eliminate background interference and accurately locate the fire area. Extracting the fire combustion point corresponding to each fire time point can reveal the spatial location changes of the fire source. Mapping the fire combustion point to the fire protection space model and determining the direction of fire spread can intuitively reflect the fire's propagation trend in the actual equipment and environment. This enables precise identification of the fire's location and spread path, providing data support for fire protection strategy formulation and fire-fighting resource scheduling, thereby improving the timeliness and effectiveness of fire response.

[0107] In one embodiment, the step of mapping the fire combustion point corresponding to each fire time point to a fire-fighting space model and determining the direction of fire spread includes:

[0108] Map the fire burning point corresponding to each fire time point to the fire space model to obtain the coordinates of the burning point corresponding to each fire time point;

[0109] Based on the coordinates of the combustion point corresponding to each fire time point, calculate the combustion midpoint corresponding to each fire time point, and calculate the change vector of the combustion midpoint corresponding to adjacent fire time points;

[0110] The average value of the change vectors at each combustion midpoint is taken to obtain the fire spread vector, and the fire spread vector is normalized to obtain the fire spread direction.

[0111] Here, the combustion point coordinates are the spatial location parameters of the fire's combustion point in the fire-fighting spatial model. The combustion midpoint refers to the geometric center of all combustion point coordinates within the same fire time point. The combustion midpoint change vector is a vector calculated from the positional difference between the combustion midpoints of two adjacent fire time points, representing the displacement of the fire center over time. The fire spread vector is a vector formed by averaging the change vectors of all combustion midpoints, used to comprehensively reflect the overall movement trend of the fire.

[0112] Specifically, firstly, based on the relative position of the fire's burning point on the transformer, the fire's burning point is mapped onto the fire-fighting spatial model to establish a correspondence between the fire's location and the actual physical space. After mapping, each fire's burning point is represented using three-dimensional coordinates to ensure the precise location of the fire within the spatial model.

[0113] Subsequently, the number of fire points b at the current fire time point and the three-dimensional coordinates rsi = (Xi, Yi, Zi) of these fire points in the fire-fighting spatial model are obtained, which are the coordinates of each fire point. Based on the number of fire points and their spatial positions, the geometric center of all fire points is calculated to obtain the fire midpoint rzd corresponding to this time point, thus using a single spatial point to intuitively reflect the overall position of the current fire.

[0114] Next, the combustion midpoints rzdj = (Xj, Yj, Zj) corresponding to c different fire time points are obtained, and the combustion midpoints between adjacent time points are vector-calculated to obtain the combustion midpoint change vector bhx, which is used to represent the displacement trend of the overall fire position in the time series.

[0115] The mean value of all combustion midpoint change vectors bhx is calculated to obtain the overall fire spread vector my = (XM, YM, ZM), which comprehensively reflects the average propagation trend of the fire in space. The formula for calculating the fire spread vector is:

[0116]

[0117] Finally, the fire spread vector is normalized to obtain the spread direction vector myf, which is the direction of fire spread. Its components on the X, Y, and Z axes are denoted as Xd, Yd, and Zd, respectively, achieving a precise quantitative expression of the fire spread direction. The formula for calculating the spread direction vector is:

[0118]

[0119] Understandably, mapping the fire's combustion point corresponding to the fire's time point to a fire-fighting spatial model first converts the two-dimensional or image information of the fire scene into spatial coordinate information, thus achieving precise location of the fire in three-dimensional physical space. Next, the combustion midpoint is calculated based on the combustion point coordinates, and further, the change vector of the combustion midpoint at adjacent time points is obtained, quantifying the trend of fire location change over time. Subsequently, averaging all change vectors and normalizing them eliminates the influence of local disturbances and extracts the main direction of the overall fire spread. This not only accurately reconstructs the fire's propagation trajectory in space but also provides a scientific basis for the deployment of fire-fighting resources and the formulation of fire-fighting strategies, thereby improving the efficiency and targeting of fire response.

[0120] In one embodiment, the fire-fighting device includes the steps of deactivating sprinklers and optimizing the fire-fighting strategy based on each fire point and the direction of fire spread, in conjunction with a fire-fighting space model, including:

[0121] The burning area is determined based on the coordinates of each fire point in the fire-fighting space model;

[0122] Based on the device coordinates of the calling sprinkler in the fire space model and the coordinates of the combustion points adjacent to the calling sprinkler, the spray vector is calculated, and the spray angle is calculated using the spray vector.

[0123] Calculate the cross product of the fire spread direction and the spray vector to obtain the atomization vector, and then perform operations on the spray vector and the atomization vector to obtain the atomization angle;

[0124] The fire-fighting strategy for the combustion zone is optimized by using spray angle and atomization angle.

[0125] In this context, "activated sprinkler head" refers to a sprinkler head that is activated and performs water spraying operations during the operation of the fire protection system. "Combustion zone" refers to the spatial area enclosed by the coordinates of multiple combustion points, representing the area affected by the actual fire. "Device coordinates" refers to the three-dimensional coordinate position of fire extinguishing devices such as sprinklers in the fire protection space model, used to determine their spray direction and range of action. "Spray vector" is the direction vector from the sprinkler device coordinates to the coordinates of the nearest combustion point, used to represent the direction in which the sprinkler sprays the extinguishing medium. "Spray angle" is the angle between the spray vector and a specific reference axis, used to determine the spatial attitude of the sprinkler spray. "Atomization vector" is the vertical vector obtained by calculating the cross product of the fire spread direction and the spray vector, used to determine the atomization diffusion direction of the extinguishing medium. "Atomization angle" is the angle between the spray vector and the atomization vector, used to guide the diffusion characteristics of the extinguishing medium in space.

[0126] Specifically, such as Figure 2 As shown, the computer equipment obtains the three-dimensional coordinates rsi = (Xi, Yi, Zi) of each fire combustion point from the fire space model, and calculates the maximum and minimum values ​​of the coordinate set along each axis, i.e., xmx = max(Xi), xmi = min(Xi), ymx = max(Yi), ymi = min(Yi), zmx = max(Zi), zmi = min(Zi), and constructs an axis-aligned bounding box around the combustion point. The bounding box serves as the initial representation of the combustion area rsq. To improve robustness, extreme outliers can be filtered out or percentile truncation (e.g., taking the 5th–95th percentile) can be used before calculating the extreme values ​​to avoid misjudgment of the area caused by a single noise point.

[0127] Subsequently, the computer equipment determines the injection vector between the nozzle and the combustion point. For each nozzle, its coordinates dp = (Xdy, Ydy, Zdy) in the fire space model are obtained, and the coordinates lj = (Xlr, Ylr, Zlr) of the nearest combustion point to the nozzle are selected. This can be done using nearest neighbor search or by selecting the centroids of several points with the smallest distance or intensity based on weights. Then, the injection vector is calculated. To improve temporal continuity and stability, the injection vector can be a weighted average of the injection vectors of each frame within a short time window as the final ps.

[0128] Next, the computer device calculates the injection angle based on the injection vector ps. The formula for calculating the X-axis injection angle α is:

[0129]

[0130] Similarly, the injection angle β with respect to the Y-axis and the injection angle γ with respect to the Z-axis are obtained.

[0131] Simultaneously, the cross product wh = myf × ps is calculated using the propagation direction vector myf = (Xd, Yd, Zd) and the injection vector ps, yielding the atomization vector wh = (Xh, Yh, Zh). Then, the atomization angle is calculated using the vector angle formula. The formula for calculating the atomization angle whj is:

[0132]

[0133] Finally, the computer equipment uses the spray angle and atomization angle to optimize the fire suppression strategy for the combustion zone. Illustratively, based on the spray angle index, it determines whether the current physical orientation of the nozzle can geometrically cover the combustion zone, for example, if the intersection volume of the projection cone and the combustion envelope exceeds a threshold. Based on the atomization angle, it assesses the lateral diffusion effect of the atomization direction on the flame front, and then selects appropriate spray modes (directional spray, fan-shaped spray, or pulse spray) and parameters (pressure, pulse period, activation duration). In cases of multi-nozzle coordination, optimization algorithms (such as multi-objective scheduling for maximizing coverage and minimizing water volume) can be used to determine the start-stop sequence and synchronization strategy of multiple nozzles to achieve optimal coverage of the combustion zone and effective atomization of the extinguishing medium. Finally, the optimized instructions are sent to the field control unit, and closed-loop monitoring is initiated to adjust based on real-time feedback.

[0134] Understandably, determining the combustion zone based on the coordinates of the fire's ignition point in the fire-fighting spatial model is crucial for accurately pinpointing the flame distribution range, thus providing a precise spatial target for subsequent spraying strategies. Next, the spray vector is calculated based on the coordinates of the nozzles and adjacent combustion points, and the spray angle is further calculated to ensure the direction of the sprayed water flow matches the flame position, improving fire suppression coverage efficiency. Subsequently, the cross product of the fire spread direction and the spray vector is calculated to obtain the atomization vector, and the atomization angle is obtained through calculation with the spray vector. This optimizes the diffusion direction and density of water particles, making them more effective at targeting the flame spread path. Finally, combining the spray angle and atomization angle for fire-fighting strategy optimization ensures spraying accuracy while enhancing atomization coverage, achieving efficient suppression of the combustion zone, reducing the risk of fire spread, and improving fire suppression effectiveness.

[0135] In one embodiment, the fire-fighting device includes a call nozzle and a pipe, and the device status information includes the spray volume of the call nozzle and the water consumption of the pipe.

[0136] The steps for generating maintenance early warnings for fire protection devices using device status information during fire suppression include:

[0137] Calculate the average spray volume and the mode of spray volume based on the spray volume of each nozzle, and determine the normal spraying range based on the average spray volume and the mode of spray volume.

[0138] Sprinklers whose spray volume is outside the normal spray range are recorded as abnormal sprinkler calls, and maintenance warnings for abnormal sprinkler calls are generated.

[0139] Calculate the abnormal water consumption value based on the water consumption of the pipeline and the spray volume of each sprinkler head.

[0140] When the abnormal water consumption value exceeds the preset consumption ratio, a pipeline maintenance warning is generated.

[0141] In fire protection systems, pipelines are enclosed channels used to transport extinguishing media (such as water or foam). They are typically made of metal or high-strength composite materials and connect the water source to each sprinkler head, ensuring efficient and stable delivery of the extinguishing media to the designated sprinkler head location during a fire. Spray volume refers to the volume of extinguishing media sprayed by the designated sprinkler head per unit time. The average spray volume is the arithmetic mean of the spray volumes from all designated sprinkler heads. The mode of spray volume is the most frequently occurring value among all spray volume data. The normal spray range is the range of spray volume determined by the average spray volume and the mode of spray volume, representing the reasonable water volume that a sprinkler head can spray under normal operating conditions. An abnormally designated sprinkler head refers to a sprinkler head whose spray volume is outside the normal spray range, potentially indicating a performance malfunction or abnormal spraying. Water consumption is the total volume of extinguishing media transported by the pipeline within a certain time. Abnormal water consumption value is a value calculated based on the difference between the total water consumption of the pipeline and the spray volume of each designated sprinkler head, used to reflect pipeline leaks or abnormal water supply. The preset consumption ratio is a pre-set threshold ratio used to determine whether water resource consumption is abnormal.

[0142] In this embodiment, the number of sprinklers called during the firefighting process is first obtained based on the device status information, and the spray volume psl of each called sprinkler is obtained. Then, the average spray volume pjp is calculated by performing calculations on the spray volume of each sprinkler and the number of sprinklers, and the mode zsp of the spray volume is extracted. Using the relationship between the average spray volume and the mode, the normal spray interval pqj is determined, where the left and right boundaries of the interval are calculated from the deviations of zsp and the average spray volume, respectively. The left interval of the normal spray interval pqj is denoted as Zpqj, Zpqj = zsp - |zsp - pqj|; the right interval of the normal spray interval pqj is denoted as Ypqj, Ypqj = zsp + |zsp - pqj|; pqj = [Zpqj, Ypqj].

[0143] By cyclically determining whether the spray volume of each nozzle is within the specified range, the nozzles can be divided into normal nozzles and abnormal nozzles. The location and status information of abnormal nozzles are recorded and uploaded to maintenance personnel, thereby achieving automatic identification and early warning of nozzle status.

[0144] Based on the sprinkler head status determination, the total water consumption during the firefighting process, xhl, is further obtained from the device status information, and the abnormal water consumption value, yxh, is calculated in conjunction with the spray volume of each sprinkler head. The formula for calculating the abnormal water consumption value is:

[0145]

[0146] By comparing yxh with a preset loss ratio bl, when yxh exceeds the ratio threshold, it can be determined that there is an abnormality such as leakage or blockage in the pipeline, thus generating a pipeline maintenance warning; if it does not exceed the threshold, it is recorded as a normal state. This method can simultaneously detect both local anomalies of sprinklers and overall pipeline anomalies, improving the comprehensiveness and accuracy of fire protection system operation monitoring.

[0147] Ultimately, based on the assessment results of the sprinkler heads and pipes, corresponding early warning information is generated, and abnormal status data is uploaded to the maintenance personnel's terminal via the network, facilitating timely dispatch of personnel for inspection and repair. This not only maintains the optimal performance of the fire suppression system during fire response but also effectively reduces water waste and equipment wear, thereby improving the safety and economy of the fire protection system and achieving dual optimization of fire prevention and resource management.

[0148] In one embodiment, the step of generating a fire warning based on various fire images and device status information includes:

[0149] The fire intensity change value is determined by using various fire images, and the feasible value of fire treatment is determined based on the device status information;

[0150] Based on the fire intensity change value and the fire control feasibility value, the fire control status is determined, and a fire warning is generated based on the fire control status.

[0151] Among them, the fire intensity change value refers to a numerical indicator quantified by comparing and analyzing fire images at different time points, representing the degree of fire development or weakening. The fire handling feasibility value refers to a numerical result that assesses the effectiveness of fire handling measures under current conditions, combining fire situation and equipment status information. The fire control status refers to the result of comprehensively determining whether the fire is in a controllable, uncontrollable, or stable state based on the fire intensity change value and the fire handling feasibility value.

[0152] In this embodiment, fire monitoring cameras deployed around the transformer are first used to collect fire images at different times. Image processing algorithms are then used to extract information such as flame area features, temperature change characteristics, and smoke density, thereby calculating the fire intensity change value. This fire intensity change value quantifies the development trend of the fire; for example, a positive value indicates fire intensification, a negative value indicates fire weakening, and a zero value indicates the fire is basically stable. This method provides intuitive and quantifiable indicators of fire development in real time, offering a data foundation for subsequent control and decision-making.

[0153] Subsequently, based on the device status information, the current fire-fighting feasibility value is obtained. Specifically, the device status information includes key parameters such as fire pump water pressure, remaining water supply, sprinkler head operating status, power supply status, and pipe unobstructedness. Based on this data, the fire-fighting feasibility value is calculated to measure whether the current fire protection system can effectively perform fire-fighting operations under the remaining resources and operating conditions. For example, when the water supply is sufficient, the sprinkler head coverage is complete, and the pump station is operating normally, the fire-fighting feasibility value is high; conversely, when there is equipment failure or insufficient water pressure, the value is low.

[0154] After obtaining the fire intensity change value and the fire-fighting feasibility value, the fire control status is determined. When the fire intensity change value hbh equals -1 (indicating the fire is weakening) and the fire-fighting feasibility value xhb is greater than 0 (indicating that fire-fighting operations can continue effectively under the current conditions), it is determined that the fire can be controlled before fire-fighting resources are exhausted. In this case, the fire control status is marked as "controllable," and the current fire-fighting strategy can continue without significant adjustments.

[0155] Finally, fire early warning information is generated based on the determined fire control status. If the fire is deemed controllable, the warning will instruct personnel to continue current operations and monitor resource consumption. If the fire is deemed uncontrollable or at risk, a high-level warning will be generated immediately, prompting emergency management personnel to take additional supply or evacuation measures. This closed-loop process ensures the continued effectiveness of firefighting operations and allows for timely triggering of emergency responses in situations of resource constraints or abnormal fire development.

[0156] In one embodiment, the step of determining the fire intensity change value using various fire images includes:

[0157] The fire image at each fire time point is mapped to the fire space model to obtain the remaining burning range at each fire time point;

[0158] Calculate the fire intensity change value based on each remaining burning area.

[0159] The remaining combustion range refers to the area that remains in a state of combustion after being mapped onto the fire space model during firefighting operations.

[0160] In this embodiment, fire images at multiple time points xt are first acquired during firefighting operations using surveillance cameras, infrared imaging devices, or other fire sensing terminals. These fire images contain information such as flames, smoke, and the heat distribution of the affected area. Each fire image is mapped to a pre-established fire space model. Through spatial registration, coordinate transformation, and image segmentation techniques, the burning area in the two-dimensional image is accurately located in the three-dimensional space model to obtain the remaining fire burning range at the corresponding time point. This range is recorded as SRF(1), SRF(2), ..., SRF(xt). This ensures the accurate correspondence between the fire source location and the spatial structure, laying a data foundation for subsequent calculations of fire trend changes.

[0161] Next, based on the combustion range SRF(1) to SRF(xt) at the above xt time points, the fire intensity change value hbh is calculated. The formula for calculating the fire intensity change value is:

[0162]

[0163] When hbh = -1, it indicates that the combustion range from SRF(1) to SRF(xt) shows a monotonically decreasing trend, indicating that the fire has been effectively controlled under firefighting measures; conversely, it can be determined that the fire is still spreading or fluctuating. This calculation method can transform visual information into measurable change indicators, improving the accuracy and real-time performance of fire status assessment.

[0164] It is understandable that mapping fire images at each fire time point to a fire space model allows for a precise correspondence between two-dimensional image information and three-dimensional spatial structure. This enables the determination of the true location and extent of the fire source within a building or area, avoiding spatial misjudgments caused by relying solely on planar images. Furthermore, by calculating the fire intensity change value based on the remaining combustion range at each time point, the shrinkage or spread trend of the fire at different times can be quantified, enabling a dynamic assessment of the fire suppression effect. This transforms intuitive visual information into calculable and comparable numerical indicators, improving the accuracy and real-time performance of fire monitoring.

[0165] In one embodiment, the step of determining the feasible value for fire suppression based on device status information includes:

[0166] Extract the remaining fire-fighting resources at each fire-fighting time point from the device status information;

[0167] Calculate the reduction rate of the burning range based on each remaining burning range, and calculate the consumption rate of fire-fighting resources based on each remaining value of fire-fighting resources.

[0168] The feasible value of fire treatment is calculated based on the reduction ratio of the combustion range and the proportion of fire-fighting resource consumption.

[0169] Among them, the remaining value of fire-fighting resources refers to the quantity or proportion of available fire-fighting resources that have not yet been consumed at a certain fire-fighting time point. The fire-fighting resource consumption ratio refers to the loss of water resources under normal use, that is, the conventional loss caused by factors such as evaporation and equipment friction under fault-free conditions, usually based on historical operating data or a percentage threshold set by industry standards. This ratio serves as a benchmark value to distinguish between normal consumption and additional consumption caused by abnormal conditions such as pipe leaks and sprinkler head blockages, ensuring the accuracy of early warning judgments.

[0170] In this embodiment, the remaining fire-fighting resources syz(1) to syz(xt) at each fire-fighting time point are first extracted from the device status information. This can be achieved by real-time monitoring of the fire pump, water tank balance, pipeline pressure, and cumulative water spray volume of each sprinkler head, and storing the above information in correspondence with time tags. In this way, the remaining resource status during the fire-fighting process can be accurately obtained.

[0171] Subsequently, the remaining combustion ranges SRF(1) to SRF(xt) obtained under the same time series are compared and analyzed to calculate the reduction ratio of the combustion range in each time period. This process can quantify the control speed of the fire by comparing the combustion range at each time point with the previous time point and obtaining the percentage reduction of the combustion area. At the same time, the remaining values ​​syz(1) to syz(xt) at each fire-fighting time point are used to calculate the fire-fighting resource consumption ratio, that is, the proportion of the resources used to the initial available resources between adjacent time points, thereby reflecting the resource consumption efficiency of the fire-fighting process. Finally, the combustion range reduction ratio and the fire-fighting resource consumption ratio are correlated to obtain the fire-fighting feasibility value xhb. The formula for calculating the fire-fighting feasibility value is:

[0172]

[0173] Firefighting feasibility values ​​comprehensively reflect the balance between fire control speed and resource utilization efficiency during firefighting. For example, when the reduction in burning area is significantly higher than the resource consumption rate, it indicates a high probability of extinguishing the fire under current remaining resource conditions; conversely, it suggests greater firefighting difficulty, potentially requiring increased resource input or adjustments to firefighting strategies. This provides a quantitative and dynamic feasibility assessment for firefighting decisions, ensuring a more efficient and controllable firefighting process.

[0174] It is understandable that by extracting the remaining value of fire-fighting resources at each fire-fighting time point from the device status information, the consumption of resources during the fire-fighting process can be monitored in real time. By combining the calculation of the reduction ratio of the combustion range for each remaining combustion range and the calculation of the fire-fighting resource consumption ratio based on the remaining value of fire-fighting resources, the speed of fire control and the efficiency of resource use can be quantified simultaneously. By calculating the fire-fighting feasibility value from both, the possibility of completing fire control under the current resource conditions can be comprehensively judged. This provides data support for optimizing fire-fighting strategies, rationally allocating resources, avoiding resource waste, and improving the success rate of fire-fighting.

[0175] The following describes the IoT-based remote fire early warning device for substation transformers provided in the embodiments of this application. The IoT-based remote fire early warning device for substation transformers described below can be referred to in correspondence with the IoT-based remote fire early warning method for substation transformers described above. Figure 3 As shown, this application provides a remote fire early warning device for substation transformers based on the Internet of Things. The device includes:

[0176] Firefighting space model construction module 201 is used to construct a three-dimensional firefighting space model based on the spatial location of fire-fighting devices and transformers in the substation;

[0177] The fire risk monitoring module 202 is used to monitor the fire risk of the transformer and, when the transformer catches fire, to determine the fire burning point corresponding to each fire time point and predict the direction of fire spread based on the fire images of the transformer at multiple fire time points.

[0178] The fire-fighting strategy optimization module 203 is used to optimize the fire-fighting strategy based on each fire burning point and the direction of fire spread, combined with the fire-fighting space model. The fire-fighting strategy is used for fire-fighting treatment.

[0179] The maintenance early warning generation module 204 is used to generate maintenance early warnings for fire protection devices by utilizing the device status information of fire protection devices during fire protection operations.

[0180] The fire warning generation module 205 is used to record fire images at multiple fire time points during fire handling, and generate fire warnings based on each fire image and device status information.

[0181] In one embodiment, the fire risk monitoring module 202 includes:

[0182] The risk monitoring parameter acquisition unit is used to acquire the normal concentration and real-time concentration of each gas type within the transformer's operating range, as well as the normal temperature and real-time temperature of the transformer.

[0183] The deviation calculation unit is used to calculate the gas concentration deviation based on the normal gas concentration and the real-time gas concentration of each gas type, and to calculate the temperature deviation based on the normal temperature and the real-time temperature.

[0184] The transformer anomaly calculation unit is used to calculate transformer anomalies based on gas concentration deviation and temperature deviation.

[0185] The fire risk prediction unit is used to predict the fire risk of transformers by using transformer anomaly values ​​and preset transformer anomaly value thresholds.

[0186] In one embodiment, the fire risk monitoring module 202 includes:

[0187] The normal operating image acquisition unit is used to acquire normal operating images of the transformer.

[0188] The fire burning point extraction unit is used to compare the fire image at each fire time point with the normal working image and extract the fire burning point corresponding to each fire time point.

[0189] The fire spread direction determination unit is used to map the fire burning point corresponding to each fire time point to the fire space model to determine the fire spread direction.

[0190] In one embodiment, the fire spread direction determination unit includes:

[0191] The combustion point coordinate determination sub-unit is used to map the fire combustion point corresponding to each fire time point to the fire space model to obtain the combustion point coordinates corresponding to each fire time point.

[0192] The combustion midpoint change vector calculation subunit is used to calculate the combustion midpoint corresponding to each fire time point based on the combustion point coordinates corresponding to each fire time point, and to calculate the combustion midpoint change vector corresponding to adjacent fire time points;

[0193] The fire spread direction determination sub-unit is used to take the average value of the change vector of each combustion midpoint to obtain the fire spread vector, and then to normalize the fire spread vector to obtain the fire spread direction.

[0194] In one embodiment, the fire-fighting device includes a sprinkler head, and the fire-fighting strategy optimization module 203 includes:

[0195] The combustion zone determination unit is used to determine the combustion zone based on the coordinates of each fire point in the fire-fighting space model.

[0196] The spray angle calculation unit is used to calculate the spray vector based on the device coordinates of the calling nozzle in the fire space model and the coordinates of the combustion point of the adjacent calling nozzle, and to calculate the spray angle using the spray vector.

[0197] The atomization angle calculation unit is used to calculate the cross product of the fire spread direction and the spray vector to obtain the atomization vector, and then perform calculations on the spray vector and the atomization vector to obtain the atomization angle;

[0198] The fire-fighting strategy optimization unit is used to optimize the fire-fighting strategy for the combustion zone by using spray angle and atomization angle.

[0199] In one embodiment, the fire-fighting device includes a sprinkler head and a pipeline, and the device status information includes the spray volume of the sprinkler head and the water consumption of the pipeline; the maintenance early warning generation module 204 includes:

[0200] The normal spraying interval determination unit is used to calculate the average spraying volume and the mode of spraying volume based on the spraying volume of each calling nozzle, and to determine the normal spraying interval based on the average spraying volume and the mode of spraying volume.

[0201] The first maintenance early warning generation unit is used to record the calling nozzles whose spray volume does not belong to the normal spraying range as abnormal calling nozzles and generate maintenance early warnings for abnormal calling nozzles.

[0202] The abnormal water consumption value calculation unit is used to calculate the abnormal water consumption value based on the water consumption of the pipeline and the spraying volume of each sprinkler head.

[0203] The second maintenance early warning generation unit is used to generate a pipeline maintenance early warning when the abnormal water consumption value exceeds the preset consumption ratio.

[0204] In one embodiment, the fire alarm generation module 205 includes:

[0205] The fire parameter determination unit is used to determine the fire intensity change value using various fire images, and to determine the feasible value of fire handling based on device status information;

[0206] The fire warning generation unit is used to determine the fire control status based on the fire change value and the fire handling feasibility value, and generate a fire warning based on the fire control status.

[0207] In one embodiment, the fire parameter determination unit includes:

[0208] The remaining combustion range determination subunit is used to map the fire image at each fire time point to the fire space model to obtain the remaining combustion range at each fire time point;

[0209] The fire intensity change value calculation subunit is used to calculate the fire intensity change value based on each remaining burning range.

[0210] In one embodiment, the fire parameter determination unit includes:

[0211] The fire resource remaining value extraction subunit is used to extract the fire resource remaining value at each fire time point from the device status information;

[0212] The fire resource consumption ratio calculation subunit is used to calculate the fire range reduction ratio based on each remaining fire range, and to calculate the fire resource consumption ratio based on each remaining fire resource value.

[0213] The fire-fighting treatment feasibility value calculation subunit is used to calculate the fire-fighting treatment feasibility value based on the reduction ratio of the combustion range and the ratio of fire-fighting resource consumption.

[0214] In one embodiment, this application also provides a storage medium storing computer-readable instructions that, when executed by one or more processors, cause the one or more processors to perform the steps of the IoT-based remote fire early warning method for substation transformers as described in any of the above embodiments.

[0215] In one embodiment, this application also provides an IoT-based remote fire warning system for substation transformers. The IoT-based remote fire warning system for substation transformers stores computer-readable instructions. When the computer-readable instructions are executed by one or more processors, the one or more processors perform the steps of the IoT-based remote fire warning method for substation transformers as described in any of the above embodiments.

[0216] Indicatively, such as Figure 4 As shown, Figure 4 This is a schematic diagram of the internal structure of an IoT-based remote fire early warning system for substation transformers, provided as an embodiment of this application. This IoT-based remote fire early warning system 300 for substation transformers can be provided as a server. (Refer to...) Figure 4 The IoT-based remote fire warning system 300 for substation transformers includes a processing component 302, which further includes one or more processors, and memory resources represented by a memory 301 for storing instructions, such as applications, that can be executed by the processing component 302. The applications stored in the memory 301 may include one or more modules, each corresponding to a set of instructions. Furthermore, the processing component 302 is configured to execute instructions to perform the IoT-based remote fire warning method for substation transformers described in any of the above embodiments.

[0217] The IoT-based remote fire alarm system 300 for substation transformers may further include a power supply component 303 configured to perform power management for the IoT-based remote fire alarm system 300, a wired or wireless network interface 304 configured to connect the IoT-based remote fire alarm system 300 to a network, and an input / output (I / O) interface 305. The IoT-based remote fire alarm system 300 can operate on an operating system stored in memory 301, such as Windows Server™, Mac OS X™, Unix™, Linux™, Free BSD™, or similar.

[0218] Those skilled in the art will understand that Figure 4 The structure shown is merely a block diagram of a portion of the structure related to the present application and does not constitute a limitation on the IoT-based remote fire early warning system for substation transformers applied thereto. A specific IoT-based remote fire early warning system for substation transformers may include more or fewer components than shown in the figure, or combine certain components, or have different component arrangements.

[0219] Finally, it should be noted that in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Unless otherwise specified, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element. In this document, "a," "an," "the," "the," and "its" may also include plural forms unless the context clearly indicates otherwise. "Multiple" refers to at least two, such as 2, 3, 5, or 8, etc. "And / or" includes any and all combinations of the related listed items.

[0220] The various embodiments in this specification are described in a progressive manner. Each embodiment focuses on the differences from other embodiments. The various embodiments can be combined as needed, and the same or similar parts can be referred to each other.

[0221] The above description of the disclosed embodiments enables those skilled in the art to make or use this application. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of this application. Therefore, this application is not to be limited to the embodiments shown herein, but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A remote fire early warning method for substation transformers based on the Internet of Things, characterized in that, The method includes: Based on the spatial location of fire-fighting equipment and transformers in the substation, a three-dimensional fire-fighting space model is constructed. Fire risk monitoring is performed on the transformer, and when the transformer catches fire, the fire burning point corresponding to each fire time point and the direction of fire spread are determined based on the fire images of the transformer at multiple fire time points. Based on each of the fire ignition points and the direction of fire spread, the fire-fighting strategy is optimized in conjunction with the fire-fighting space model. The fire-fighting strategy is used for fire-fighting treatment. Using the device status information of the fire-fighting equipment in fire-fighting operations, a maintenance early warning for the fire-fighting equipment is generated; Record fire images at multiple fire time points during firefighting operations, and generate fire warnings based on each fire image and the device status information.

2. The method for remote fire early warning of substation transformers based on the Internet of Things according to claim 1, characterized in that, The steps for fire risk monitoring of the transformer include: Obtain the normal concentration and real-time concentration of each gas type within the operating range of the transformer, as well as the normal temperature and real-time temperature of the transformer. Calculate the gas concentration deviation based on the normal gas concentration and the real-time gas concentration of each gas type, and calculate the temperature deviation based on the normal temperature and the real-time temperature. Based on the gas concentration deviation and the temperature deviation, the transformer anomaly value is calculated; The fire risk of the transformer is predicted by using the transformer anomaly value and a preset transformer anomaly value threshold.

3. The method for remote fire early warning of substation transformers based on the Internet of Things according to claim 1, characterized in that, The step of determining the fire burning point corresponding to each fire time point and predicting the direction of fire spread based on fire images of multiple fire time points of the transformer includes: Obtain a normal operating image of the transformer; The fire image at each fire time point is compared with the normal working image to extract the fire burning point corresponding to each fire time point; The fire combustion point corresponding to each fire time point is mapped to the fire-fighting space model to determine the direction of fire spread.

4. The method for remote fire early warning of substation transformers based on the Internet of Things according to claim 3, characterized in that, The step of mapping the fire combustion point corresponding to each fire time point to the fire-fighting space model and determining the direction of fire spread includes: Map the fire burning point corresponding to each fire time point to the fire-fighting space model to obtain the burning point coordinates corresponding to each fire time point; Based on the coordinates of the combustion point corresponding to each fire time point, calculate the combustion midpoint corresponding to each fire time point, and calculate the change vector of the combustion midpoint corresponding to adjacent fire time points; The average value of the change vectors of each combustion midpoint is taken to obtain the fire spread vector, and the fire spread vector is normalized to obtain the fire spread direction.

5. The method for remote fire early warning of substation transformers based on the Internet of Things according to claim 1, characterized in that, The fire-fighting device includes the activation of sprinklers. The step of optimizing the fire-fighting strategy based on each fire ignition point and the direction of fire spread, combined with the fire-fighting space model, includes: The combustion zone is determined based on the coordinates of each fire point in the fire-fighting space model. Based on the device coordinates of the called sprinkler head in the fire-fighting space model and the coordinates of the combustion points adjacent to the called sprinkler head, the spray vector is calculated, and the spray angle is calculated using the spray vector. Calculate the cross product of the fire spread direction and the spray vector to obtain the atomization vector, and then perform calculations on the spray vector and the atomization vector to obtain the atomization angle; The fire-fighting strategy for the combustion zone is optimized by using the spray angle and the atomization angle.

6. The method for remote fire early warning of substation transformers based on the Internet of Things according to claim 1, characterized in that, The fire-fighting device includes a sprinkler head and a pipeline, and the device status information includes the spray volume of the sprinkler head and the water consumption of the pipeline. The step of generating a maintenance early warning for the fire-fighting equipment by utilizing the equipment status information of the fire-fighting equipment during fire-fighting operations includes: Based on the spray volume of each of the called nozzles, calculate the average spray volume and the mode of spray volume, and determine the normal spraying range based on the average spray volume and the mode of spray volume. Sprinklers whose spray volume does not fall within the normal spray range are recorded as abnormal sprinkler heads, and maintenance warnings for abnormal sprinkler heads are generated. Calculate the abnormal water consumption value based on the water consumption of the pipeline and the spray volume of each of the calling nozzles; When the abnormal water consumption value exceeds the preset consumption ratio, a maintenance warning for the pipeline is generated.

7. The method for remote fire early warning of substation transformers based on the Internet of Things according to claim 1, characterized in that, The step of generating a fire warning based on each of the fire images and the device status information includes: The fire intensity change value is determined using each of the aforementioned fire images, and the feasible value of fire handling is determined based on the device status information; Based on the fire intensity change value and the fire-fighting feasibility value, the fire intensity control status is determined, and the fire warning is generated using the fire intensity control status.

8. The method for remote fire early warning of substation transformers based on the Internet of Things according to claim 7, characterized in that, The step of determining the fire intensity change value using each of the fire images includes: The fire image at each fire time point is mapped to the fire space model to obtain the remaining burning range at each fire time point; Calculate the fire intensity change value based on the remaining combustion range of each item.

9. The method for remote fire early warning of substation transformers based on the Internet of Things according to claim 8, characterized in that, The step of determining the feasible value for fire handling based on the device status information includes: Extract the remaining value of fire-fighting resources for each fire-fighting time point from the device status information; Based on each remaining combustion range, calculate the combustion range reduction ratio, and based on each remaining fire-fighting resource value, calculate the fire-fighting resource consumption ratio. The feasibility value of the fire treatment is calculated based on the reduction ratio of the combustion range and the consumption ratio of fire-fighting resources.

10. A remote fire alarm system for substation transformers based on the Internet of Things, characterized in that, include: One or more processors, and memory; The memory stores computer-readable instructions, which, when executed by the one or more processors, perform the steps of the IoT-based remote fire early warning method for substation transformers as described in any one of claims 1 to 9.

Images