Fire-fighting water storage facility anti-icing detection method and system

Abstract

The present application relates to the technical field of anti-icing, and particularly relates to a fire-fighting water storage facility anti-icing detection method and system. The method comprises the following steps: obtaining installation environment parameters and water storage basic information corresponding to a fire-fighting water storage facility, deploying an adapted temperature sensing component and completing signal calibration to obtain a calibrated temperature sensing data acquisition link; collecting water temperature data in the fire-fighting water storage facility in real time through the temperature sensing data acquisition link, synchronously collecting facility surrounding environment temperature data, and generating water temperature and environment temperature correlation data sets; setting multi-level anti-icing early warning thresholds based on the correlation data sets, and constructing a water temperature change trend prediction model; dynamically analyzing the real-time collected water temperature data by using the water temperature change trend prediction model, and outputting anti-icing detection results and corresponding intervention prompt information in combination with the multi-level anti-icing early warning thresholds. The present application can improve the anti-icing detection efficiency of the fire-fighting water storage facility.

Description

Technical Field

[0001] This invention relates to the field of anti-icing technology, and in particular to a method and system for detecting anti-icing in fire-fighting water storage facilities. Background Technology

[0002] Fire water storage facilities refer to fire pools, elevated fire water tanks, and elevated fire water hydrants. Elevated fire water hydrants, for example, are high-altitude structures that directly supply water to fire extinguishing systems for initial firefighting needs. In indoor fire hydrant systems, they are typically installed outdoors on rooftops, where freezing is possible in winter. Existing fire safety standards require daily temperature checks of fire water storage facilities in winter. If freezing occurs or the indoor temperature drops below 5°C, measures must be taken to prevent freezing and maintain a room temperature above 5°C. This method relies on the management level of the fire protection facility maintenance unit. Whether or not freezing is checked is also limited by the behavior of maintenance personnel. Furthermore, some elevated fire water hydrants are located in locations that hinder maintenance, requiring steel ladders for access, making daily checks ineffective in practice.

[0003] In addition, similar patents such as CN115416854A disclose an icing detection device and method based on temperature measurement. This method uses a temperature sensor array combined with appropriate theoretical calculations to replace conventional icing sensor arrays, significantly reducing the size of the icing detection device and lowering system costs. Furthermore, using the detection device and method of this application, while detecting ice thickness, it can also calculate cloud and fog parameters: water content (LWC) and median droplet diameter (MVD), solving the technical problem of difficult cloud and fog parameter detection. The detection device of this application can be made into a universal component, equivalent to a new type of icing sensor. This invention improves the comprehensiveness and economy of icing detection. However, this solution focuses on icing detection in cloud and fog environments and is not optimized for the application scenarios of fire-fighting water storage facilities. It cannot adjust the detection strategy according to the differentiated conditions such as facility type, water storage volume, and installation environment; it does not construct a correlation analysis model between water temperature and ambient temperature, making it difficult to capture the icing evolution law of fire-fighting water storage facilities, thus reducing the efficiency of anti-icing detection. Summary of the Invention

[0004] To address the aforementioned technical problems in existing anti-icing detection processes for water storage facilities, this invention provides a method for detecting anti-icing in fire-fighting water storage facilities. This method involves accurately collecting information on the installation environment and water storage of the facility, deploying a calibrated temperature sensing link to ensure reliable temperature data acquisition, simultaneously collecting water temperature and ambient temperature data to generate a correlated dataset, setting multi-level early warning thresholds for different scenarios, constructing a trend prediction model to dynamically adapt detection strategies, dynamically analyzing real-time data using the model, and outputting detection results and intervention prompts. This replaces daily manual inspections, avoids the risks of high-altitude maintenance, improves detection efficiency, and simultaneously captures the icing evolution pattern, providing early warnings of icing hazards, ensuring the normal operation of fire-fighting water storage facilities in winter, and thus strengthening fire safety capabilities. The method includes the following steps:

[0005] Step S1: Obtain the installation environment parameters and basic water storage information corresponding to the fire water storage facility, deploy the appropriate temperature sensing components and complete the signal calibration to obtain the calibrated temperature sensing data acquisition link;

[0006] Step S2: Collect water temperature data in the fire-fighting water storage facility in real time through the temperature sensing data acquisition link, and simultaneously collect ambient temperature data around the facility to generate a correlation dataset of water temperature and ambient temperature.

[0007] Step S3: Based on the associated dataset, set multi-level anti-icing early warning thresholds and construct a water temperature change trend prediction model;

[0008] Step S4: Utilize the water temperature change trend prediction model to dynamically analyze the real-time collected water temperature data, and combine the multi-level anti-icing early warning threshold to output the anti-icing detection results and corresponding intervention prompts.

[0009] This invention acquires installation environment parameters and basic water storage information, deploys suitable temperature sensing components, and completes signal calibration, ensuring the stability and reliability of the temperature data acquisition link. This lays the foundation for subsequent accurate detection and avoids data deviations caused by poor sensor component compatibility. Simultaneously collecting water temperature and surrounding environmental temperature data to generate a correlated dataset overcomes the limitation of similar patents that lack a correlation analysis model, enabling comprehensive capture of key environmental factors affecting the freezing of water storage facilities. Furthermore, based on the correlated dataset, multi-level anti-icing warning thresholds are set, and a water temperature change trend prediction model is constructed. This achieves dynamic adaptation of the detection strategy according to differentiated conditions such as facility type and water storage volume, capturing the freezing evolution pattern and solving the problems of inflexible traditional fixed threshold warnings and the inability of similar patents to adapt to fire-fighting water storage scenarios. By using a model to dynamically analyze real-time water temperature data and combining multi-level thresholds to output detection results and intervention prompts, the traditional manual daily inspection mode has been replaced. This completely avoids the problems of difficult maintenance of high-level facilities and inconvenient personnel access, significantly improving inspection efficiency. At the same time, by providing early warning of icing hazards, it provides maintenance personnel with sufficient intervention time, ensuring the normal operation of fire-fighting water storage facilities in winter, effectively protecting fire safety, reducing operation and maintenance management costs, and meeting the strict requirements of fire protection technical standards for winter anti-icing inspection.

[0010] Preferably, step S1 includes the following steps:

[0011] Collect the volume parameters, installation height parameters, open exposure area parameters, and winter climate characteristics parameters of the area where the fire-fighting water storage facility is located. At the same time, record the material properties and insulation structure information of the fire-fighting water storage facility to form a basic information set of the water storage facility.

[0012] The number and location of temperature sensing components are determined based on the basic information set of the water storage facility. Sensing units adapted to low-temperature environments are selected and fixedly installed at different water depths and key parts of the wall inside the water storage facility.

[0013] Zero-point calibration and range calibration are performed on each sensing unit, the calibration deviation value is recorded, a signal transmission link between the sensing unit and the data receiver is constructed, signal strength detection and anti-interference test are performed on the signal transmission link, and a link test report is generated.

[0014] Based on the calibration deviation value and the link detection report, the signal transmission link is optimized and adjusted to obtain the calibrated temperature sensing data acquisition link.

[0015] This invention first systematically collects basic parameters such as the volume and installation height of the water storage facility, as well as winter climate characteristics, to clarify the facility's materials and insulation structure, forming a complete set of basic information. This provides a scientific basis for subsequent sensor deployment and avoids blind installation. Based on this basic information, the number and location of sensor units are precisely determined. Low-temperature adaptable sensor units are selected and installed at different water depths and key parts of the wall surface, ensuring temperature collection covers the entire facility area. This overcomes the limitations of similar patented universal designs that cannot adapt to the diverse scenarios of fire-fighting water storage. Through zero-point and range calibration, signal link detection, and anti-interference testing, a link detection report is generated. The link is optimized based on calibration deviations to ensure the stability and reliability of the data acquisition link, avoiding the impact of data deviations or link failures on detection results. This step establishes a standardized data acquisition foundation adapted to fire-fighting water storage scenarios, providing high-quality data support for subsequent accurate monitoring and early warning, while reducing the blindness of sensor deployment and lowering operation and maintenance costs.

[0016] Preferably, the following steps are included before optimizing the signal transmission link:

[0017] The signal transmission link is monitored in real time, and the signal transmission strength, data transmission delay and power supply status data of the sensing unit corresponding to the signal transmission link are collected to generate a link status monitoring dataset.

[0018] The link status monitoring dataset is analyzed to determine whether there is an anomaly in the signal transmission link. If an anomaly is found, the anomaly type and location are determined.

[0019] For different types and locations of anomalies, corresponding link repair solutions are generated, and backup sensing units and backup transmission links are activated to ensure uninterrupted water temperature data acquisition.

[0020] The link anomaly information and link repair plan are synchronized to the display terminal in the fire control room, a link anomaly early warning prompt is generated, and maintenance personnel are reminded to handle the problem in a timely manner based on the link anomaly early warning prompt.

[0021] This invention monitors link transmission strength, latency, and sensor power supply status in real time, generating a monitoring dataset to enable early detection of link anomalies and avoid monitoring blind spots caused by data acquisition interruptions. Data dataset analysis accurately determines the type and location of anomalies, generating targeted repair plans. Simultaneously, backup sensor units and transmission links are activated to ensure uninterrupted water temperature acquisition, overcoming the limitations of traditional manual inspections, such as delayed fault detection and inability to promptly retest. Anomaly information and repair plans are synchronized to the fire control room terminal, generating early warning prompts to remind maintenance personnel to handle issues promptly, ensuring rapid resolution of link faults and improving operational response efficiency. This process constructs a closed-loop link protection system of "monitoring-early warning-emergency-repair," ensuring the continuous and stable operation of the temperature acquisition link, providing reliable protection for continuous monitoring against icing in winter, avoiding missed detection of icing hazards due to link failures, and strengthening the continuity and stability of fire safety management.

[0022] Preferably, step S2 includes the following steps:

[0023] Through the calibrated temperature sensing data acquisition link, the water temperature data corresponding to each sensing unit is collected in real time according to the preset acquisition cycle, and the acquisition time and sensing location information corresponding to each water temperature data are marked to form the original water temperature dataset.

[0024] Install ambient temperature sensors at a predetermined distance around the fire-fighting water storage facility to collect ambient temperature data synchronously and mark the collection time of the ambient temperature data to ensure that it is synchronized with the collection time of the water temperature data.

[0025] The original water temperature dataset is subjected to outlier removal processing to remove invalid data caused by sensor interference and retain valid water temperature data.

[0026] The synchronized and matched ambient temperature data and effective water temperature data are linked and integrated according to the collection time. Key parameters in the basic information set of water storage facilities are added as association tags to generate a correlation dataset of water temperature and ambient temperature.

[0027] This invention collects water temperature data at preset intervals and marks the time and location to form a raw dataset, achieving comprehensive and time-series recording of water temperature data, replacing manual random detection, and ensuring the comprehensiveness and continuity of the data. Ambient temperature is simultaneously collected around the facility and the collection time is precisely matched to ensure the spatiotemporal consistency of the temperature sensing data, laying the foundation for subsequent correlation analysis. Outlier removal removes invalid data, improving the quality of water temperature data and preventing interference from affecting the analysis results. Ambient temperature data and valid water temperature data are correlated and integrated by time, supplementing basic information as labels to generate a correlated dataset, overcoming the limitation of similar patents that do not construct a temperature sensing correlation model, and comprehensively capturing the correlation patterns between water temperature and ambient temperature. This step achieves the standardization and structured integration of fire-fighting water storage temperature data, providing complete data support for subsequent threshold setting and trend prediction, while improving data usability, avoiding the problem of scattered data being unable to support in-depth analysis, and strengthening the scientific nature of anti-icing detection.

[0028] Preferably, step S3 includes the following steps:

[0029] Extract the valid water temperature data and corresponding ambient temperature data from the associated dataset, statistically analyze the water temperature change data under the same historical climatic conditions, and construct a historical water temperature change database.

[0030] Trend analysis was performed on the data in the historical water temperature change database to determine the correlation coefficient between the rate of water temperature decrease and the ambient temperature, and the critical water temperature change values ​​corresponding to different ambient temperature ranges were divided.

[0031] Based on the correlation coefficient and critical water temperature change value, and combined with the material insulation performance parameters of the fire-fighting water storage facility, a first-level warning benchmark value, a second-level warning benchmark value and an emergency warning benchmark value are set to form a multi-level anti-icing warning threshold.

[0032] A portion of the data in the associated dataset is selected as training samples. With ambient temperature, initial water temperature, and water temperature drop rate as input features, and whether it is close to freezing as the output label, a water temperature change trend prediction model is trained.

[0033] The water temperature change trend prediction model is validated using the remaining data in the associated dataset, and the model parameters are adjusted until the model prediction accuracy meets the preset requirements.

[0034] This invention extracts temperature sensing data from a correlated dataset and constructs a database by combining it with historical water temperature change data corresponding to the same climate period. Trend analysis clarifies the correlation coefficient between the rate of water temperature decrease and ambient temperature, and delineates critical change values ​​for different temperature ranges, providing a scientific basis for threshold setting. Multi-level early warning benchmark values ​​are set based on parameters such as facility materials and insulation performance, achieving precise adaptation of early warning thresholds to the characteristics of the facilities themselves and environmental conditions, overcoming the limitations of inflexible traditional single-threshold early warning systems. A water temperature change trend prediction model is trained using correlated data, with multi-dimensional feature inputs ensuring the model's ability to capture the evolution of icing patterns. After validation and optimization with remaining data, the model's predictive reliability is improved. This step constructs a scenario-based and differentiated early warning and prediction system, addressing the pain point of similar patents not being optimized for fire-fighting water storage scenarios. It makes anti-icing detection more aligned with practical application needs, providing precise support for subsequent dynamic early warning and enhancing the scientific rigor and foresight of anti-icing detection.

[0035] Preferably, step S4 includes the following steps:

[0036] The calibrated temperature sensor data acquisition link continuously acquires real-time water temperature data and real-time ambient temperature data, and inputs them into the water temperature change trend prediction model to obtain the water temperature prediction curve for a future preset time period.

[0037] Extract key parameters from the water temperature prediction curve, including the predicted minimum water temperature, the rate of water temperature decrease, and the predicted time to reach the warning benchmark value;

[0038] The predicted minimum water temperature is compared with the multi-level anti-icing warning threshold, and the degree of water temperature deviation from the warning benchmark value is calculated by combining the difference between real-time water temperature data and the warning benchmark value.

[0039] The level of the anti-icing detection result is determined based on the degree of water temperature deviation from the warning benchmark, the rate of water temperature drop, and the predicted time to reach the warning benchmark.

[0040] Based on the level of the anti-icing detection results, corresponding intervention prompts are generated, including the type of intervention measure, implementation priority, and operation procedure guidance.

[0041] This invention inputs real-time temperature sensing data into a prediction model to generate a water temperature prediction curve, extracting key parameters such as the predicted minimum value and the rate of temperature drop, enabling early prediction of icing risks and overcoming the passive approach of traditional "post-event detection." By comparing the predicted parameters with multi-level warning thresholds and combining the difference between the real-time water temperature and the warning baseline value, the detection result level is accurately determined, ensuring the comprehensiveness of the risk assessment. Based on the level, intervention prompts including the type of measure, priority, and operation guidelines are generated, allowing maintenance personnel to quickly understand the direction of action and avoid the expansion of icing hazards due to improper or untimely intervention. This step overcomes the limitations of similar patents, such as low detection efficiency and lack of intervention guidance, achieving intelligent management of the entire process from risk prediction and level determination to intervention guidance, significantly improving the efficiency of anti-icing response. It is particularly suitable for scenarios where the maintenance of high-level water storage facilities is difficult, reducing the frequency of personnel working at heights, providing maintenance personnel with sufficient time for handling, and ensuring the safe operation of fire-fighting water storage facilities in winter.

[0042] Preferably, after the step of combining the multi-level anti-icing early warning threshold to output the anti-icing detection result and the corresponding intervention prompt information, the method further includes the following steps:

[0043] Obtain the anti-icing detection results and corresponding intervention prompts, and construct an anti-icing detection log by combining the real-time water temperature data of the fire water storage facility and the ambient temperature data.

[0044] The anti-icing detection logs are categorized and organized, and archived according to detection time, detection result level, and implementation of intervention measures to form a historical detection database;

[0045] The data in the historical detection database are periodically statistically analyzed to calculate the triggering frequency of different warning levels, the effectiveness of intervention measures, and the water temperature change pattern, so as to generate statistical analysis results.

[0046] Based on the statistical analysis results, the parameters of the multi-level anti-icing early warning threshold and water temperature change trend prediction model are optimized to improve the accuracy and timeliness of anti-icing detection.

[0047] The statistical analysis results and parameter optimization schemes are generated into a report and synchronized to the fire protection facility maintenance management system to provide data support for the formulation of subsequent maintenance plans.

[0048] This invention integrates detection results, intervention information, and real-time temperature sensing data to construct a detection log, which is then categorized and archived according to standards to form a historical database. This enables traceability and queryability of data throughout the entire anti-icing process, meeting the data standardization requirements of fire management. Historical data is periodically analyzed to calculate early warning trigger frequency, intervention effects, and water temperature change patterns, providing data support for system optimization. Based on the analysis results, multi-level early warning thresholds and model parameters are optimized to achieve dynamic iteration of the detection system, continuously improving detection accuracy and timeliness. Analysis results and optimization plans are generated into a report and synchronized to the maintenance management system, providing a scientific basis for maintenance planning and promoting a shift in anti-icing management from "passive response" to "proactive prediction." This step constructs a closed-loop management system of "detection-archiving-analysis-optimization," overcoming the limitations of traditional management lacking a continuous optimization mechanism and solving the problem of similar patents' inability to iteratively improve detection effects. This allows the anti-icing detection system to continuously adapt to environmental and facility changes, ensuring the long-term safe operation of fire-fighting water storage facilities during winter.

[0049] Preferably, the process of classifying and organizing the anti-icing detection logs includes the following steps:

[0050] The archiving fields of the anti-icing detection log are determined, including the collection time, water temperature data of each sensor unit, ambient temperature data, detection result level, intervention prompt information, intervention measure implementation time, implementation personnel, and implementation effect feedback;

[0051] The anti-icing detection logs are classified and archived according to a preset time period, and historical detection data sub-databases are established in units of months, quarters, and years. Each sub-database is labeled with the climate season and temperature change characteristics to which it belongs.

[0052] Extract key data from each historical detection data sub-database, including the number of times different warning levels are triggered, the range of ambient temperature at the time of triggering, the distribution of water temperature drop rate, and the success rate of intervention measures, to form a basic dataset for statistical analysis;

[0053] The statistical analysis dataset is processed using data statistical analysis methods to calculate the triggering frequency, average triggering duration, and effectiveness of intervention measures for different warning levels, and to summarize the water temperature change patterns under different climatic conditions, thereby forming a historical detection database.

[0054] This invention clearly defines the core archiving fields of the detection logs, covering key information such as collection time, temperature sensing data, warning level, and intervention implementation status, ensuring the integrity and standardization of the archived data and avoiding incomplete analysis due to missing core data. The data is archived by monthly, quarterly, and annual cycles, and labeled with climatic and seasonal characteristics, establishing a hierarchical historical data sub-database to achieve time-series and scenario-based data management, facilitating the accurate extraction of water temperature change patterns under different climatic conditions. Key data is extracted to form a basic dataset for statistical analysis. By quantitatively calculating indicators such as warning trigger frequency and intervention effectiveness, water temperature change patterns are summarized, and a historical detection database for the system is constructed. This step overcomes the limitations of traditional manual data recording, which results in chaotic data and an inability to systematically analyze it. It achieves standardized archiving and in-depth mining of data throughout the entire anti-icing process, providing high-quality data support for subsequent threshold optimization and model iteration, while also meeting the data traceability requirements of fire management and improving the precision of anti-icing management.

[0055] Preferably, after processing the basic dataset using statistical analysis methods to calculate the triggering frequency, average triggering duration, and effectiveness of intervention measures for different warning levels, and summarizing the water temperature change patterns under different climatic conditions, the method further includes the following steps:

[0056] Based on the aforementioned statistical analysis dataset, it is determined whether the current multi-level anti-icing warning threshold is suitable for the water temperature change pattern under different climatic conditions. If there is a situation where the warning threshold is too high, resulting in missed reports, or too low, resulting in false reports, then the direction of threshold adjustment is determined.

[0057] Extract water temperature change data after different warning levels are triggered from historical detection data, and combine the implementation effect of intervention measures to determine the threshold adjustment range, and correct the benchmark values ​​for the first-level warning, the second-level warning, and the emergency warning.

[0058] For the water temperature change trend prediction model, new water temperature change data obtained from statistical analysis are selected as supplementary training samples, the input feature weights of the model are adjusted, and the model is retrained.

[0059] The revised multi-level anti-icing early warning threshold and the optimized water temperature change trend prediction model were verified. The latest anti-icing detection data were used for testing to determine whether the early warning accuracy and prediction accuracy met the preset standards, and a parameter optimization file was generated.

[0060] This invention assesses the suitability of existing warning thresholds based on historical statistical data, accurately identifying issues such as missed detections due to excessively high thresholds or false alarms due to excessively low thresholds. It clarifies adjustment directions, ensuring that thresholds accurately match water temperature change patterns under different climatic conditions. By combining water temperature change data after warning triggering with intervention effects, the threshold adjustment range is scientifically determined, and multi-level warning benchmark values ​​are corrected, improving the targeting and flexibility of warnings. New water temperature change data is added to train the model, and input feature weights are adjusted to make the model more closely reflect actual icing evolution patterns, overcoming the limitations of similar patented models that are fixed and unable to adapt to changing scenarios. The optimization effect is verified using the latest detection data, and a parameter optimization archive is created to ensure the reliability of the optimized thresholds and model. This step constructs a dynamic optimization mechanism of "analysis-adjustment-optimization-verification," allowing the anti-icing detection system to continuously adapt to environmental and facility changes, improving detection accuracy, avoiding missed detections of icing hazards due to outdated thresholds or models, and strengthening the timeliness and reliability of fire safety assurance.

[0061] Preferably, the verification of the modified multi-level anti-icing early warning threshold and the optimized water temperature change trend prediction model, using the latest anti-icing detection data, includes the following steps:

[0062] Collect validation data of the revised multi-level anti-icing warning threshold and the optimized water temperature change trend prediction model, including warning accuracy, prediction accuracy, false alarm rate and false alarm rate, to form a parameter optimization validation dataset.

[0063] The parameter optimization validation dataset is analyzed and compared with the performance metrics before optimization to evaluate the effect of parameter optimization.

[0064] If the parameter optimization effect meets the preset requirements, the corrected multi-level anti-icing early warning threshold and the optimized water temperature change trend prediction model will be determined as the formal operating parameters and updated to the anti-icing detection system.

[0065] If the parameter optimization effect does not meet the preset requirements, return to re-analyze the statistical data, adjust the threshold correction range and model training parameters until the optimization effect meets the requirements;

[0066] The entire process of parameter optimization and the verification results are recorded in the historical detection database to form a parameter optimization archive.

[0067] This invention collects core verification data such as early warning accuracy and false negative rate to form a parameter optimization verification dataset. By comparing the dataset with the pre-optimization indicators, the actual effect of threshold correction and model optimization is objectively evaluated, avoiding blind optimization that could lead to a decline in detection performance. Clear criteria for judging the optimization effect are established: if the criteria are met, the system operating parameters are updated; if not, the process is returned for readjustment, ensuring the reliability and applicability of the final operating parameters and forming a closed-loop management system of "optimization-verification-re-optimization". The entire optimization process and verification results are recorded and archived to form a complete parameter optimization archive, enabling traceability and review of the optimization process and providing a reference for subsequent continuous optimization. This step overcomes the limitations of traditional optimization methods that lack verification and cannot quantify effects, ensuring the scientific validity and effectiveness of threshold and model optimization. It allows the anti-icing detection system to continuously adapt to dynamic factors such as environmental changes and facility aging, maintaining a high level of detection performance over the long term. This solves the pain point of similar patents' inability to iteratively improve detection effects, further strengthening the safety guarantee for the winter operation of fire-fighting water storage facilities.

[0068] Preferably, the analysis of the parameter optimization verification dataset and the comparison with the performance metrics before optimization includes the following steps:

[0069] An evaluation index system for the effect of parameter optimization is set, including early warning accuracy, prediction accuracy, false alarm rate, false alarm rate and data processing delay, and a corresponding qualified threshold is set for each evaluation index.

[0070] Extract the values ​​of each evaluation indicator in the parameter optimization verification dataset and compare them one by one with the set qualified threshold.

[0071] If the values ​​of all evaluation indicators reach or exceed the corresponding qualified thresholds, the parameter optimization effect is deemed to meet the preset requirements.

[0072] If at least one evaluation indicator fails to meet the acceptable threshold, the parameter optimization effect is deemed not to meet the preset requirements, and the evaluation indicator that fails to meet the standard and the corresponding numerical difference are recorded.

[0073] This invention establishes an evaluation index system covering core dimensions such as early warning accuracy and prediction accuracy, and sets clear qualification thresholds for each index, constructing an objective and quantifiable benchmark for performance evaluation. This avoids the subjective bias of judging optimization effects solely based on experience. The values ​​of each index in the verification data are extracted and compared one by one with the qualification thresholds. The "full compliance equals qualification" rule clarifies whether the optimization effect meets the preset requirements. Simultaneously, the indicators that fail to meet the standards and their numerical differences are recorded, providing a clear direction for subsequent targeted adjustments. This step overcomes the limitations of traditional optimization's lack of standardized performance evaluation methods, achieving standardized and quantitative evaluation of parameter optimization effects. This ensures that the optimized thresholds and models truly improve anti-icing detection performance. Both early warning accuracy and data processing efficiency can be verified through unified standards, avoiding performance shortcomings after optimization that could lead to missed or misjudged icing hazards. This provides crucial assurance for the reliable operation of winter anti-icing detection in fire-fighting water storage facilities.

[0074] Preferably, if the parameter optimization effect does not meet the preset requirements, the process of returning to re-analyze the statistical data, adjusting the threshold correction magnitude and model training parameters until the optimization effect meets the target includes the following steps:

[0075] When the parameter optimization effect fails to meet the preset requirements, extract the relevant data corresponding to the evaluation indicators that fail to meet the standards, analyze the reasons for the failure of the indicators, and determine whether the threshold correction range is unreasonable or the model training parameters are set improperly.

[0076] If the threshold correction range is unreasonable, then combine the corresponding water temperature change cases in the historical detection data to recalculate a reasonable threshold correction range and correct the multi-level anti-icing warning threshold again.

[0077] If the model training parameters are set incorrectly, adjust the model's learning rate, number of iterations, and regularization parameters, and retrain the model using high-quality data from the supplementary training samples.

[0078] The revised warning threshold and the retrained model were re-validated, new validation data were collected, and the evaluation process was repeated until all evaluation indicators met the standards.

[0079] In this invention, when the optimization effect fails to meet the target, the relevant data of the non-compliant indicators are first extracted, and the reasons are accurately analyzed to determine whether it is due to an unreasonable threshold correction range or improper model training parameter settings, thus avoiding blind adjustments that lead to ineffective optimization. Differentiated rectification measures are taken for different root causes of problems. If the threshold is unreasonable, the correction range is recalculated based on historical cases; if the model parameters are inappropriate, key parameters such as the learning rate and number of iterations are adjusted and the system is retrained to ensure that the rectification measures directly address the core issues. By repeating the verification and evaluation process until all indicators meet the target, a closed-loop iterative mechanism of "analysis-rectification-verification" is formed, breaking through the limitation of traditional optimization where it is difficult to effectively correct problems after they occur. This step ensures the final effect of parameter optimization, ensuring that the optimized threshold and model can adapt to the differentiated scenarios of fire-fighting water storage facilities, improving the reliability and stability of anti-icing detection, avoiding performance defects in the detection system due to incomplete optimization, and further strengthening the effectiveness of fire safety protection.

[0080] Preferably, if the threshold correction range is unreasonable, then by combining the corresponding water temperature change cases in historical detection data, a reasonable threshold correction range is recalculated, and the multi-level anti-icing warning threshold is corrected again, including the following steps:

[0081] The relevant data corresponding to the evaluation indicators that did not meet the standards were screened to remove invalid data and abnormal interference data, and to retain representative and valid data samples.

[0082] The valid data samples were classified according to climate conditions, water storage facility type, and water temperature change stage, and the characteristics of the data samples under different classifications were analyzed.

[0083] For valid data samples of different categories, calculate the corresponding critical water temperature value, rate of change threshold, and early warning response time requirement, and determine the reasonable range of threshold correction based on these data characteristics;

[0084] Based on the reasonable range of threshold correction, and combined with the numerical differences of the evaluation indicators that did not meet the standards, the magnitude of the correction is accurately calculated to ensure that the corrected threshold can adapt to the water temperature change requirements under different classification scenarios.

[0085] This invention first screens data related to non-compliant indicators, removing invalid and interfering data, and retaining representative and valid samples to ensure that subsequent analysis is based on high-quality data and avoids deviations in correction direction caused by interfering data. Valid samples are categorized according to dimensions such as climate conditions and facility type, and the characteristics of different categories are analyzed to accurately grasp the water temperature change patterns and early warning needs under different scenarios, providing a scientific basis for differentiated threshold correction. The reasonable range for threshold correction is determined by combining the water temperature critical values ​​and change rates of the categorized samples, and then the correction magnitude is calculated based on the numerical differences of the non-compliant indicators, ensuring that the corrected threshold can adapt to the needs of various scenarios. This step breaks through the limitations of the traditional "one-size-fits-all" threshold correction, achieving refined threshold adjustment based on scenario differences. It avoids the problem of delayed early warnings or false alarms in some scenarios due to a single correction magnitude failing to adapt to the water temperature change patterns of different climates and different types of facilities. Scenario-based correction makes the threshold more aligned with actual application needs, further improving the targeting and effectiveness of anti-icing detection and ensuring the safe operation of various fire-fighting water storage facilities in winter.

[0086] Preferably, after accurately calculating the magnitude of the second correction based on the reasonable range of the threshold correction and the numerical differences of the evaluation indicators that did not meet the standards, ensuring that the corrected threshold can adapt to the water temperature change requirements under different classification scenarios, the following steps are also included:

[0087] Based on the determined threshold, the benchmark values ​​for Level 1, Level 2, and Emergency Warnings are adjusted again, and the threshold values ​​before and after the adjustment and the basis for the adjustment are recorded.

[0088] Valid data samples from different classification scenarios were selected as test data. The adjusted warning threshold was applied to the detection and analysis of the test data to obtain the warning accuracy, false alarm rate and false alarm rate data after the test.

[0089] The evaluation index data after the test is compared with the set pass threshold to determine whether the standard is met;

[0090] If the target is met, the threshold is adjusted again; if the target is still not met, the characteristics of the valid data samples are re-analyzed, the threshold adjustment range is adjusted, and the testing process is repeated until the evaluation indicators corresponding to the test data meet the target.

[0091] The process of revising the threshold, the basis for the adjustment, and the test results are recorded in detail in the parameter optimization file to improve the traceability of the optimization process.

[0092] This invention adjusts multi-level early warning benchmark values ​​according to a determined correction range, meticulously recording the values ​​before and after adjustment and the basis for each adjustment. This ensures the standardization and traceability of threshold adjustments, avoiding threshold confusion caused by adjustments without a basis. Valid samples from different classification scenarios are selected as test data. The adjusted thresholds are applied to detection analysis to obtain core indicators such as early warning accuracy and false alarm rate. The correction effect is verified by comparing with qualified thresholds, ensuring that the thresholds can adapt to water temperature changes in various scenarios. If the standard is not met, the sample characteristics are re-analyzed, the correction range is adjusted, and the test is repeated, forming a closed-loop verification mechanism of "adjustment-testing-re-adjustment." This overcomes the limitation of insufficient verification after traditional threshold correction, ensuring the reliability and applicability of the final threshold. The correction process, basis, and test results are recorded in the parameter optimization file, further improving the traceability of the optimization process and providing a reference for subsequent threshold iteration optimization. This step ensures the accuracy and comprehensiveness of threshold correction, avoiding missed detections or false alarms of icing hazards due to insufficient adaptation to some scenarios, improving the overall stability of the anti-icing detection system, and providing a solid guarantee for the safe operation of fire-fighting water storage facilities in winter.

[0093] Preferably, the present invention also provides a fire-fighting water storage facility anti-icing detection system, the system being implemented based on the fire-fighting water storage facility anti-icing detection method described above, the fire-fighting water storage facility anti-icing detection system comprising:

[0094] The data acquisition and calibration module is used to acquire the installation environment parameters and basic information of the fire water storage facility, deploy the appropriate temperature sensing components and complete the signal calibration to obtain the calibrated temperature sensing data acquisition link.

[0095] The associated data generation module is used to collect water temperature data in the fire-fighting water storage facility in real time through the temperature sensing data acquisition link, and simultaneously collect ambient temperature data around the facility to generate an associated dataset of water temperature and ambient temperature.

[0096] The early warning model construction module is used to set multi-level anti-icing early warning thresholds based on the associated dataset and construct a water temperature change trend prediction model.

[0097] The dynamic detection and early warning module is used to dynamically analyze the real-time collected water temperature data using the water temperature change trend prediction model, and output the anti-icing detection results and corresponding intervention prompts in combination with the multi-level anti-icing early warning threshold.

[0098] In this invention, the data acquisition and calibration module accurately acquires facility information, deploys compatible sensing components, and completes calibration to build a stable and reliable data acquisition link, providing high-quality data support for subsequent detection and overcoming the limitations of traditional manual data collection, which is characterized by non-standardized and inefficient data. The associated data generation module simultaneously collects water temperature and ambient temperature data and integrates them to generate an associated dataset, compensating for the lack of a temperature-sensing association model in similar patents and providing a complete data foundation for capturing the evolution of icing. The early warning model construction module sets multi-level early warning thresholds and builds a prediction model based on the associated dataset, enabling scenario-based adaptation of detection strategies and solving the problem of not being able to adjust strategies according to the differentiated conditions of fire-fighting water storage. The dynamic detection and early warning module uses the model to dynamically analyze real-time data and output detection results and intervention prompts, replacing daily manual inspections, avoiding the difficulties of high-level facility maintenance, and significantly improving detection efficiency and response timeliness. Each module performs its own function while working collaboratively to build a fully intelligent anti-icing detection system, achieving automated management from data acquisition and analysis to early warning intervention, fully adapting to the application scenarios of fire-fighting water storage facilities, and strengthening fire safety assurance capabilities.

[0099] The present invention has the following specific beneficial effects:

[0100] (1) By comprehensively acquiring the installation environment parameters and basic information of the fire-fighting water storage facility, including key data such as volume, installation height, material and insulation structure, a scientific basis is provided for the deployment of sensing components, avoiding data collection failure caused by blind installation. Based on this information, temperature sensing units suitable for low-temperature environments are selected and precisely deployed in key parts of the facility at different water depths and on the walls, ensuring that temperature collection covers the entire facility area, overcoming the limitations of similar patented universal designs that cannot adapt to the differentiated scenarios of fire-fighting water storage. Through zero-point and range calibration, signal link detection and anti-interference testing, a stable and reliable temperature sensing data acquisition link is formed after optimization and adjustment, ensuring the continuity and accuracy of subsequent data acquisition and avoiding the impact of data deviation on the detection effect. This step constructs a standardized data acquisition foundation adapted to the fire-fighting water storage scenario, providing high-quality data support for the whole process of anti-icing detection, while reducing the blindness of sensor deployment, reducing operation and maintenance costs, and solving the pain point of traditional manual detection lacking system data support.

[0101] (2) Based on the calibrated sensor data acquisition link, water temperature data of each sensor unit is collected in real time according to a preset cycle, and the collection time and location information are marked to realize the full-domain and time-series recording of water temperature data, replacing manual random detection and ensuring the comprehensiveness and continuity of the data. Ambient temperature data is collected synchronously around the facility and the collection time is accurately matched to ensure the spatiotemporal consistency of temperature sensing data and lay the foundation for subsequent correlation analysis. Invalid data generated by sensor interference is removed by outlier removal processing to improve the quality of water temperature data and avoid interference data from affecting the analysis results. The synchronously matched ambient temperature data and valid water temperature data are integrated by time correlation, and the basic facility information is supplemented as a correlation label to generate a correlation dataset. This breaks through the limitation of similar patents that have not built a temperature sensing correlation model, and comprehensively captures the intrinsic correlation between key environmental factors affecting the freezing of water storage facilities and water temperature changes, providing complete data support for subsequent threshold setting and trend prediction.

[0102] (3) By extracting effective temperature sensing data based on the associated dataset and combining it with historical water temperature change data under the same climatic conditions, the correlation coefficient between the rate of water temperature drop and the ambient temperature is analyzed, and the critical change values ​​of water temperature corresponding to different ambient temperature ranges are divided. Combined with the differentiated parameters such as the material insulation performance of fire-fighting water storage facilities, first-level, second-level and emergency warning benchmark values ​​are set to form a multi-level anti-icing warning threshold, so as to achieve accurate adaptation of the warning threshold with facility characteristics and environmental conditions, and break through the limitation of the inflexibility of traditional single-threshold warning. Select some data in the associated dataset as training samples, and use ambient temperature, initial water temperature value and water temperature drop rate as input features to train and build a water temperature change trend prediction model. After verification and optimization with the remaining data, the model's ability to capture the icing evolution law is improved. This step constructs a scenario-based and differentiated warning and prediction system, solves the pain point of similar patents not being optimized for fire-fighting water storage scenarios, makes anti-icing detection more in line with actual application needs, provides scientific support for subsequent dynamic warning, and improves the foresight and effectiveness of anti-icing detection.

[0103] (4) Input the real-time collected water temperature data and ambient temperature data into the water temperature change trend prediction model, dynamically analyze and obtain the water temperature prediction curve for the future preset time period, extract key parameters such as the predicted minimum water temperature value and the rate of water temperature drop, realize the early prediction of icing risk, and get rid of the passive situation of traditional "post-event detection". Compare the prediction parameters with the multi-level anti-icing warning threshold, and combine the difference between the real-time water temperature data and the warning benchmark value to accurately determine the anti-icing detection result level, and ensure the comprehensiveness of risk assessment. Generate corresponding intervention prompt information based on the detection result level, clarify the type of intervention measures, implementation priority and operation steps, so that operation and maintenance personnel can quickly clarify the direction of disposal and avoid the expansion of icing hazards due to improper or untimely intervention. This step breaks through the limitations of traditional manual detection relying on human behavior and the difficulty of high-level facility maintenance, replaces the daily manual inspection mode, greatly improves detection efficiency, and at the same time provides operation and maintenance personnel with sufficient intervention time, especially suitable for facilities that are inconvenient to reach, such as high-level fire water tanks, effectively ensuring the normal operation of fire water storage facilities in winter and strengthening fire safety protection capabilities. Attached Figure Description

[0104] Other features, objects, and advantages of the invention will become more apparent from the following detailed description of non-limiting embodiments with reference to the accompanying drawings:

[0105] Figure 1 This is a flowchart illustrating the steps of the fire-fighting water storage facility anti-icing detection method of the present invention;

[0106] Figure 2 This is a structural schematic diagram of the fire-fighting water storage facility of the present invention. Detailed Implementation

[0107] The present invention will be further described below with reference to the accompanying drawings and embodiments, but this should not be construed as limiting the present invention.

[0108] To achieve the above objectives, please refer to Figure 1 Embodiment 1 of the present invention provides a method for detecting anti-icing of fire-fighting water storage facilities, comprising the following steps:

[0109] Step S1: Obtain the installation environment parameters and basic water storage information corresponding to the fire water storage facility, deploy the appropriate temperature sensing components and complete the signal calibration to obtain the calibrated temperature sensing data acquisition link;

[0110] In this embodiment of the invention, environmental parameter acquisition tools are used to obtain information about fire-fighting water storage facilities (such as...). Figure 2The installation environment parameters (as shown) were used to determine the basic information of the water storage facility, including installation height, exposed area, and winter climate characteristics of the region. This information was obtained through on-site surveys and literature review, covering facility volume, material properties, insulation structure type, and insulation layer thickness. These two types of information were integrated to form a complete basic information set. Based on the volume size, height distribution, and exposed area in the basic information set, the number and specific locations of temperature sensing components were determined, and sensing units with low-temperature resistance, water vapor erosion resistance, and water flow impact resistance were selected. A bolt-fixed installation method combined with sealant was used to fix the sensing units at different water depths (surface, middle, and bottom layers) and at key areas on the wall prone to icing within the water storage facility, ensuring full contact between the sensing units and the water and secure installation. Zero-point and range calibrations were performed on each sensing unit using a standard temperature source, and the calibration deviation values ​​were recorded. Shielded cables were used to construct the signal transmission link between the sensing units and the data receiver. The cables were laid along the facility wall and waterproofed. Signal strength was tested using a signal detector, and anti-interference tests were conducted simulating the on-site electromagnetic environment, generating a link test report. By combining the calibration deviation value and the link test report, the installation angle of the sensing unit was adjusted, the cable laying path was optimized, the cable section with severe signal attenuation was replaced, and signal amplifiers were added in the weak signal area. The signal transmission link was optimized and adjusted to obtain the calibrated temperature sensing data acquisition link, ensuring stable data transmission and accurate acquisition.

[0111] Step S2: Collect water temperature data in the fire-fighting water storage facility in real time through the temperature sensing data acquisition link, and simultaneously collect ambient temperature data around the facility to generate a correlation dataset of water temperature and ambient temperature.

[0112] In this embodiment of the invention, a calibrated temperature sensor data acquisition link is activated, and shielded cables within the link collect water temperature data from each sensor unit in real time according to a preset acquisition cycle. A time stamping tool is used to label the acquisition time and corresponding sensor location information for each set of water temperature data, ensuring data traceability. An ambient temperature sensor, equipped with a windproof, waterproof, and low-temperature resistant structure, is fixedly installed at a preset distance around the fire-fighting water storage facility. This sensor simultaneously collects ambient temperature data from the surrounding area, and a time synchronization module ensures that the acquisition time of the ambient temperature data perfectly matches the acquisition time of the water temperature data. An outlier removal algorithm is used to process the collected raw water temperature data, identifying and removing invalid data caused by sensor interference or water flow impact, retaining valid water temperature data that conforms to normal water temperature change patterns. A data association tool integrates the synchronized and matched ambient temperature data with the valid water temperature data according to the acquisition time dimension, forming a temperature data combination with a consistent time series. Key parameters such as material and insulation structure from the basic water storage information are added as association tags to generate a complete water temperature and ambient temperature association dataset, providing data support for subsequent early warning threshold setting and model construction.

[0113] Step S3: Based on the associated dataset, set multi-level anti-icing early warning thresholds and construct a water temperature change trend prediction model;

[0114] In this embodiment of the invention, effective water temperature data and corresponding ambient temperature data are extracted from a correlated dataset of water temperature and ambient temperature using a data extraction tool. Historical water temperature change data under similar climatic conditions in the same region are retrieved using a meteorological data integration tool. These two types of data are then categorized and integrated according to climatic characteristics and time dimension to construct a historical water temperature change database. Trend analysis tools are used to perform trend fitting analysis on the data in the historical water temperature change database, identifying the pattern of water temperature decrease rate changing with ambient temperature and determining the correlation coefficient between the two. Simultaneously, different intervals are divided according to ambient temperature, and a data clustering method is used to determine the critical water temperature change value corresponding to each interval. Based on the obtained correlation coefficient and critical water temperature change value, combined with the thermal conductivity of the fire-fighting water storage facility's material and the insulation performance parameters of the insulation structure, a threshold setting tool is used to set multi-level anti-icing warning thresholds, including a first-level warning benchmark value, a second-level warning benchmark value, and an emergency warning benchmark value, clarifying the triggering conditions for different warning levels. 70% of the valid data in the associated dataset was selected as training samples. Ambient temperature, initial water temperature, and rate of water temperature decrease were used as input features, and the approximation of freezing was used as the output label. A water temperature change trend prediction model was constructed using a model training tool. The remaining 30% of the valid data in the associated dataset was selected as validation samples and input into the model for prediction testing. The prediction results were compared with the actual water temperature changes. The model's internal weights and iteration count were adjusted using a parameter tuning tool until the model's prediction accuracy reached the preset requirements, thus completing the construction and optimization of the water temperature change trend prediction model.

[0115] Step S4: Utilize the water temperature change trend prediction model to dynamically analyze the real-time collected water temperature data, and combine the multi-level anti-icing early warning threshold to output the anti-icing detection results and corresponding intervention prompts.

[0116] In this embodiment of the invention, real-time water temperature data within the fire-fighting water storage facility and real-time ambient temperature data are continuously acquired via a calibrated temperature sensor data acquisition link. These two types of real-time data are simultaneously input into an optimized water temperature change trend prediction model. Based on the input data, the model generates a water temperature prediction curve for a preset future time period through internal algorithm calculations, visually presenting the water temperature change trend over time. A curve analysis tool is used to extract key parameters from the water temperature prediction curve, including the predicted minimum water temperature, the rate of water temperature decrease, and the predicted time to reach each warning benchmark value. A data comparison tool is used to compare the predicted minimum water temperature with the set multi-level anti-icing warning thresholds one by one, while simultaneously calculating the difference between the real-time water temperature data and each warning benchmark value. The degree of water temperature deviation from the warning benchmark value is quantitatively analyzed through the difference. Combining the degree of water temperature deviation, the rate of water temperature decrease, and the predicted time to reach the warning benchmark value, a risk level determination tool is used to determine the level of the anti-icing detection result, classifying it into four levels: normal, level one warning, level two warning, and emergency warning. Based on different detection result levels, corresponding intervention prompts are generated through the prompt information generation tool: Level 1 warning corresponds to insulation and reinforcement prompts, Level 2 warning corresponds to heating preparation prompts, and emergency warning corresponds to immediate activation of the heating device. All prompts include the type of intervention measure, implementation priority, and detailed operation steps. The detection results and intervention prompts are synchronized to the fire control room display terminal through the data transmission link for on-duty personnel to view and perform the corresponding intervention operations.

[0117] Furthermore, step S1 includes the following steps:

[0118] Collect the volume parameters, installation height parameters, open exposure area parameters, and winter climate characteristics parameters of the area where the fire-fighting water storage facility is located. At the same time, record the material properties and insulation structure information of the fire-fighting water storage facility to form a basic information set of the water storage facility.

[0119] The number and location of temperature sensing components are determined based on the basic information set of the water storage facility. Sensing units adapted to low-temperature environments are selected and fixedly installed at different water depths and key parts of the wall inside the water storage facility.

[0120] Zero-point calibration and range calibration are performed on each sensing unit, the calibration deviation value is recorded, a signal transmission link between the sensing unit and the data receiver is constructed, signal strength detection and anti-interference test are performed on the signal transmission link, and a link test report is generated.

[0121] Based on the calibration deviation value and the link detection report, the signal transmission link is optimized and adjusted to obtain the calibrated temperature sensing data acquisition link.

[0122] In this embodiment of the invention, the volume parameters, installation height parameters, and exposed area parameters of the fire-fighting water storage facility are collected using measuring tools. Winter climate characteristic parameters of the region are retrieved from meteorological data. Simultaneously, the material properties and insulation structure information of the water storage facility are recorded through visual inspection and document review. All parameters are integrated to form a basic information set for the water storage facility. Based on the volume, height, and exposed area parameters in the basic information set, the number and location of temperature sensing components are determined. Sensing units suitable for low-temperature environments are selected, possessing low-temperature resistance and water vapor erosion resistance, enabling stable operation in low-temperature environments. Sensing units are fixedly installed inside the water storage facility at different water depths and in key areas of the wall (areas prone to freezing) using bolt fixing and sealant sealing, ensuring full contact between the sensing units and the water body and secure installation to prevent displacement due to water flow impact. Zero-point calibration and range calibration are performed on each sensing unit. The output signal of the sensing unit is compared using a standard temperature source, and the calibration deviation value is recorded. A signal transmission link is established by connecting the sensing unit and the data receiver with cables. Shielded cables are used to reduce external interference. Signal strength and anti-interference tests are performed on the signal transmission link. The signal strength of the link is collected using a signal detector, and anti-interference tests are conducted to simulate the on-site electromagnetic environment, generating a link test report that includes signal strength and interference tolerance. Based on the calibration deviation value and the link test report, the installation angle of the sensing unit and the cable laying path are adjusted, severely attenuated cable sections are replaced, and signal amplifiers are added to enhance the signal in weak areas, resulting in a calibrated temperature sensing data acquisition link that ensures stable link transmission and accurate data.

[0123] Furthermore, before optimizing and adjusting the signal transmission link, the following steps are also included:

[0124] The signal transmission link is monitored in real time, and the signal transmission strength, data transmission delay and power supply status data of the sensing unit corresponding to the signal transmission link are collected to generate a link status monitoring dataset.

[0125] The link status monitoring dataset is analyzed to determine whether there is an anomaly in the signal transmission link. If an anomaly is found, the anomaly type and location are determined.

[0126] For different types and locations of anomalies, corresponding link repair solutions are generated, and backup sensing units and backup transmission links are activated to ensure uninterrupted water temperature data acquisition.

[0127] The link anomaly information and link repair plan are synchronized to the display terminal in the fire control room, a link anomaly early warning prompt is generated, and maintenance personnel are reminded to handle the problem in a timely manner based on the link anomaly early warning prompt.

[0128] In this embodiment of the invention, before optimizing and adjusting the signal transmission link, a link status monitoring device is activated to collect real-time data on signal transmission strength, data transmission delay, and power supply status of the sensing units. The collected data is then organized according to a time series to generate a link status monitoring dataset. This dataset is analyzed using data analysis tools, comparing it with preset signal strength thresholds, transmission delay thresholds, and power supply voltage thresholds to determine if there are any anomalies in the signal transmission link, such as signal attenuation, excessive transmission delay, or unstable power supply. If anomalies are found, the anomaly type and specific location are determined through segmented detection. Corresponding link repair solutions are generated for different anomaly types and locations: for signal attenuation anomalies, the shielded cable is replaced or a signal amplifier is added; for excessive transmission delay, the data transmission path is optimized; for unstable power supply, the power supply line connection is checked and faulty power supply components are replaced. Simultaneously, a backup sensing unit and backup transmission link are activated. The signal from the backup sensing unit is connected to the data receiver, and data transmission is switched to the backup transmission link to ensure uninterrupted water temperature data acquisition. The data transmission module synchronizes link anomaly information and link repair plans to the display terminal in the fire control room. The display terminal pops up an anomaly prompt window and issues an audible and visual warning, generating a link anomaly warning prompt to remind maintenance personnel to handle the faulty link in a timely manner.

[0129] Furthermore, step S2 includes the following steps:

[0130] Through the calibrated temperature sensing data acquisition link, the water temperature data corresponding to each sensing unit is collected in real time according to the preset acquisition cycle, and the acquisition time and sensing location information corresponding to each water temperature data are marked to form the original water temperature dataset.

[0131] Install ambient temperature sensors at a predetermined distance around the fire-fighting water storage facility to collect ambient temperature data synchronously and mark the collection time of the ambient temperature data to ensure that it is synchronized with the collection time of the water temperature data.

[0132] The original water temperature dataset is subjected to outlier removal processing to remove invalid data caused by sensor interference and retain valid water temperature data.

[0133] The synchronized and matched ambient temperature data and effective water temperature data are linked and integrated according to the collection time. Key parameters in the basic information set of water storage facilities are added as association tags to generate a correlation dataset of water temperature and ambient temperature.

[0134] In this embodiment of the invention, water temperature data corresponding to each sensing unit is collected in real time according to a preset collection cycle via a calibrated temperature sensing data acquisition link. A data tagging module marks the acquisition time and sensing location information corresponding to each water temperature data point, and all collected data are integrated to form a raw water temperature dataset. An ambient temperature sensing device is fixedly installed at a preset distance around the fire-fighting water storage facility using a bracket. This device has a windproof and waterproof structure and can accurately collect the surrounding ambient temperature. The ambient temperature data is collected synchronously, and a time synchronization module marks the acquisition time of the ambient temperature data to ensure synchronization with the water temperature data acquisition time. An outlier removal algorithm is used to process the raw water temperature dataset, identifying and removing invalid data caused by sensor interference or water flow impact, retaining valid water temperature data that conforms to normal water temperature change patterns. By using data association tools, the synchronized and matched ambient temperature data and effective water temperature data are associated and integrated according to the collection time to form a temperature data combination with consistent time dimension. Key parameters such as material properties, insulation structure, and regional climate characteristics in the basic information set of water storage facilities are added as association tags to generate a data set of association between water temperature and ambient temperature. This data set provides complete data support for subsequent icing risk assessment and can be directly transmitted to the display terminal in the fire control room for on-duty personnel to view.

[0135] Furthermore, step S3 includes the following steps:

[0136] Extract the valid water temperature data and corresponding ambient temperature data from the associated dataset, statistically analyze the water temperature change data under the same historical climatic conditions, and construct a historical water temperature change database.

[0137] Trend analysis was performed on the data in the historical water temperature change database to determine the correlation coefficient between the rate of water temperature decrease and the ambient temperature, and the critical water temperature change values ​​corresponding to different ambient temperature ranges were divided.

[0138] Based on the correlation coefficient and critical water temperature change value, and combined with the material insulation performance parameters of the fire-fighting water storage facility, a first-level warning benchmark value, a second-level warning benchmark value and an emergency warning benchmark value are set to form a multi-level anti-icing warning threshold.

[0139] A portion of the data in the associated dataset is selected as training samples. With ambient temperature, initial water temperature, and water temperature drop rate as input features, and whether it is close to freezing as the output label, a water temperature change trend prediction model is trained.

[0140] The water temperature change trend prediction model is validated using the remaining data in the associated dataset, and the model parameters are adjusted until the model prediction accuracy meets the preset requirements.

[0141] In this embodiment of the invention, valid water temperature data and corresponding ambient temperature data are extracted from the associated dataset of water temperature and ambient temperature using a data filtering tool. After removing invalid and interfering data, historical water temperature change data under the same climatic conditions in the same region are retrieved using a meteorological data docking tool. The two types of data are classified and integrated according to the time dimension and the climatic characteristic dimension to construct a historical water temperature change database. The database completely preserves records of water temperature fluctuations under different climatic conditions. Trend analysis tools are used to perform linear and nonlinear trend fitting on the data in the historical water temperature change database to analyze the law of water temperature decrease rate changing with ambient temperature, determine the correlation coefficient between water temperature decrease rate and ambient temperature, and divide different ambient temperature intervals according to the degree of coldness of the climate. The critical water temperature change value corresponding to each interval is determined by data clustering method, clarifying the water temperature change range that needs to be guarded against under different ambient temperatures. Based on the obtained correlation coefficients and critical water temperature change values, combined with the material properties and insulation performance parameters such as insulation layer thickness of the fire-fighting water storage facilities, a multi-level anti-icing warning threshold was formed by setting primary, secondary, and emergency warning benchmark values ​​using a threshold setting tool. The emergency warning benchmark value is linked to the 4℃ alarm threshold in the briefing content to ensure accurate triggering of warnings when approaching the risk of icing. 70% of the effective data in the correlated dataset was selected as training samples. Ambient temperature, initial water temperature, and water temperature drop rate were used as input features, and whether or not icing was approaching was used as the output label. A water temperature change trend prediction model was constructed using a model training tool. The model learns the correspondence between input features and output labels to capture the water temperature change pattern. The remaining 30% of the data in the correlated dataset was used to validate the water temperature change trend prediction model. The model's prediction results were compared with actual water temperature changes. The model's internal weights and iteration count were adjusted using a parameter adjustment tool until the model's prediction accuracy met the preset requirements, ensuring the reliability of the model's predictions.

[0142] Furthermore, step S4 includes the following steps:

[0143] The calibrated temperature sensor data acquisition link continuously acquires real-time water temperature data and real-time ambient temperature data, and inputs them into the water temperature change trend prediction model to obtain the water temperature prediction curve for a future preset time period.

[0144] Extract key parameters from the water temperature prediction curve, including the predicted minimum water temperature, the rate of water temperature decrease, and the predicted time to reach the warning benchmark value;

[0145] The predicted minimum water temperature is compared with the multi-level anti-icing warning threshold, and the degree of water temperature deviation from the warning benchmark value is calculated by combining the difference between real-time water temperature data and the warning benchmark value.

[0146] The level of the anti-icing detection result is determined based on the degree of water temperature deviation from the warning benchmark, the rate of water temperature drop, and the predicted time to reach the warning benchmark.

[0147] Based on the level of the anti-icing detection results, corresponding intervention prompts are generated, including the type of intervention measure, implementation priority, and operation procedure guidance.

[0148] In this embodiment of the invention, real-time water temperature data from each sensing unit inside the water storage facility and real-time ambient temperature data from the surrounding environmental temperature sensing devices are continuously acquired via a calibrated temperature sensing data acquisition link and signal transmission cable. These two types of real-time data are synchronously input into a trained water temperature change trend prediction model through a data transmission interface. Based on the input real-time data, the model generates a water temperature prediction curve for a preset future time period through internal algorithms. The curve visually presents the water temperature change trend over time. A curve analysis tool extracts key parameters from the water temperature prediction curve, including the predicted minimum water temperature, the rate of water temperature decrease, and the predicted time to reach each warning benchmark value, ensuring no core risk parameters are omitted. A data comparison tool compares the predicted minimum water temperature with multi-level anti-icing warning thresholds one by one, and simultaneously calculates the difference between the real-time water temperature data and each warning benchmark value. The degree of water temperature deviation from the warning benchmark value is quantitatively analyzed through the difference. Combining the degree of water temperature deviation, the rate of water temperature decrease, and the predicted time to reach the warning benchmark value, a risk level determination tool determines the level of the anti-icing detection result, which is divided into normal, level one warning, level two warning, and emergency warning. Based on different anti-icing detection result levels, corresponding intervention prompts are generated through the prompt information generation tool. The prompts for emergency warnings clearly include specific measures such as temperature increase intervention and insulation reinforcement, with the highest implementation priority marked. At the same time, detailed operation steps are listed, such as the specific procedures for turning on the heat tracing device and the key points for adding insulation layers, to ensure that the personnel on duty in the fire control room can quickly handle the situation according to the instructions, which is in line with the requirement of "taking anti-icing measures in a timely manner" in the briefing content.

[0149] Furthermore, after the step of combining the multi-level anti-icing early warning threshold to output the anti-icing detection result and the corresponding intervention prompt information, the following steps are also included:

[0150] Obtain the anti-icing detection results and corresponding intervention prompts, and construct an anti-icing detection log by combining the real-time water temperature data of the fire water storage facility and the ambient temperature data.

[0151] The anti-icing detection logs are categorized and organized, and archived according to detection time, detection result level, and implementation of intervention measures to form a historical detection database;

[0152] The data in the historical detection database are periodically statistically analyzed to calculate the triggering frequency of different warning levels, the effectiveness of intervention measures, and the water temperature change pattern, so as to generate statistical analysis results.

[0153] Based on the statistical analysis results, the parameters of the multi-level anti-icing early warning threshold and water temperature change trend prediction model are optimized to improve the accuracy and timeliness of anti-icing detection.

[0154] The statistical analysis results and parameter optimization schemes are generated into a report and synchronized to the fire protection facility maintenance management system to provide data support for the formulation of subsequent maintenance plans.

[0155] In this embodiment of the invention, anti-icing detection results and corresponding intervention prompts are obtained through data integration tools. Combined with real-time water temperature data, real-time ambient temperature data, and detection time information from fire-fighting water storage facilities, an anti-icing detection log is constructed in a fixed format, completely recording all information from each detection. A classification and archiving tool is used to organize the anti-icing detection logs, classifying and archiving them according to the order of detection time, the level of detection results, whether intervention measures were implemented, and their effects. The archived data is imported into a historical detection database, which adopts a hierarchical storage structure to ensure convenient data retrieval. A periodic scheduling tool is used to set a monthly cycle for statistical analysis of the data in the historical detection database. Statistical analysis tools are used to calculate the trigger frequency of different warning levels, clarifying high-incidence warning periods and climatic conditions; assessing the water temperature recovery effect after implementing different intervention measures, summarizing effective intervention methods; and analyzing water temperature change patterns under different seasons and climatic conditions to generate comprehensive statistical analysis results. Based on the statistical analysis results, parameter optimization tools are used to adjust the specific values ​​of multi-level anti-icing warning thresholds, correcting the internal parameters of the water temperature change trend prediction model, improving the model's adaptability to local climate and facility characteristics, and enhancing the accuracy and timeliness of anti-icing detection. The statistical analysis results and parameter optimization schemes are compiled into standardized reports by the report generation tool and synchronized to the fire protection facility maintenance management system through the data transmission link. The reports clearly identify the key periods, critical facilities and optimization suggestions for subsequent maintenance, providing accurate data support for the formulation of maintenance plans, further reducing the manual inspection burden of maintenance personnel in winter and improving the level of fire water supply safety.

[0156] Furthermore, the classification and organization of the anti-icing detection logs includes the following steps:

[0157] The archiving fields of the anti-icing detection log are determined, including the collection time, water temperature data of each sensor unit, ambient temperature data, detection result level, intervention prompt information, intervention measure implementation time, implementation personnel, and implementation effect feedback;

[0158] The anti-icing detection logs are classified and archived according to a preset time period, and historical detection data sub-databases are established in units of months, quarters, and years. Each sub-database is labeled with the climate season and temperature change characteristics to which it belongs.

[0159] Extract key data from each historical detection data sub-database, including the number of times different warning levels are triggered, the range of ambient temperature at the time of triggering, the distribution of water temperature drop rate, and the success rate of intervention measures, to form a basic dataset for statistical analysis;

[0160] The statistical analysis dataset is processed using data statistical analysis methods to calculate the triggering frequency, average triggering duration, and effectiveness of intervention measures for different warning levels, and to summarize the water temperature change patterns under different climatic conditions, thereby forming a historical detection database.

[0161] In this embodiment of the invention, the archiving fields of the anti-icing detection log are first clearly defined. These fields cover the collection time, water temperature data of each sensor unit, ambient temperature data, detection result level, intervention prompts, intervention implementation time, implementing personnel, and feedback on implementation effects, ensuring that the archived information completely covers the entire detection and response process. A time-segmentation tool is used to classify and archive the anti-icing detection log according to a preset cycle, dividing the archiving cycle into natural months, natural quarters, and natural years. Monthly, quarterly, and annual historical detection data sub-databases are established for each sub-database. Each sub-database is labeled with its corresponding climate season (winter, transition season, etc.) and temperature change characteristics (severe cold, mild, etc.), achieving accurate data classification by time and climate dimensions. Data extraction tools are used to extract key data from each historical detection data sub-database, specifically including the number of triggers for different warning levels, the ambient temperature range at the time of triggering, the distribution of water temperature decline rates, and the success rate of intervention measures. The extracted key data is integrated in a unified format to form a basic dataset for statistical analysis. Data statistical analysis tools were used to process the basic dataset for statistical analysis. The triggering frequency of different warning levels was calculated using frequency statistics, the average triggering duration of each level was calculated using duration statistics, the effectiveness of intervention measures was calculated using effect accounting, and the water temperature change patterns under different climatic conditions were summarized using trend fitting. The processed statistical results were integrated with the original key data to form a well-structured and complete historical detection database, providing data support for subsequent parameter optimization.

[0162] Furthermore, after processing the basic dataset using statistical analysis methods to calculate the triggering frequency, average triggering duration, and effectiveness of intervention measures for different warning levels, and summarizing the water temperature change patterns under different climatic conditions, the process also includes the following steps:

[0163] Based on the aforementioned statistical analysis dataset, it is determined whether the current multi-level anti-icing warning threshold is suitable for the water temperature change pattern under different climatic conditions. If there is a situation where the warning threshold is too high, resulting in missed reports, or too low, resulting in false reports, then the direction of threshold adjustment is determined.

[0164] Extract water temperature change data after different warning levels are triggered from historical detection data, and combine the implementation effect of intervention measures to determine the threshold adjustment range, and correct the benchmark values ​​for the first-level warning, the second-level warning, and the emergency warning.

[0165] For the water temperature change trend prediction model, new water temperature change data obtained from statistical analysis are selected as supplementary training samples, the input feature weights of the model are adjusted, and the model is retrained.

[0166] The revised multi-level anti-icing early warning threshold and the optimized water temperature change trend prediction model were verified. The latest anti-icing detection data were used for testing to determine whether the early warning accuracy and prediction accuracy met the preset standards, and a parameter optimization file was generated.

[0167] In this embodiment of the invention, based on a statistical analysis dataset, a threshold adaptability analysis tool is used to compare the matching degree between the current multi-level anti-icing warning thresholds and the water temperature change patterns under different climatic conditions. The focus is on verifying whether there are situations where warning thresholds are too high, leading to missed icing risks, or too low, leading to frequent false alarms. If such problems exist, the direction of threshold adjustment is determined (lowering if too high, raising if too low). Water temperature change data after different warning levels are triggered is extracted from historical detection data using data filtering tools. Combined with information such as the water temperature recovery effect after intervention measures are implemented and the timeliness of risk response, a threshold correction calculation method is used to determine the threshold adjustment range. Precise corrections are made to the first-level warning baseline value, the second-level warning baseline value, and the emergency warning baseline value respectively, ensuring that the corrected thresholds accurately match the actual water temperature change risk. For the water temperature change trend prediction model, new water temperature change data obtained from statistical analysis (such as water temperature fluctuation data under newly added climatic conditions and temperature range change data not yet covered) were selected as supplementary training samples. The weight allocation of the model input features (ambient temperature, initial water temperature, and rate of water temperature decrease) was adjusted using a feature weight adjustment tool to strengthen the influence of water temperature change characteristics under key climatic conditions. The water temperature change trend prediction model was then retrained using a model training tool. The revised multi-level anti-icing warning thresholds and the optimized water temperature change trend prediction model were validated using the latest collected anti-icing detection data. The accuracy of the threshold warnings was assessed using a warning accuracy calculation tool, and the model's prediction performance was evaluated using a prediction deviation analysis tool. It was determined whether the warning accuracy and prediction accuracy met the preset standards. The details of the threshold correction, the model optimization process, and the validation results were compiled and archived to form a parameter optimization archive.

[0168] Furthermore, the verification of the revised multi-level anti-icing early warning threshold and the optimized water temperature change trend prediction model, using the latest anti-icing detection data for testing, includes the following steps:

[0169] Collect validation data of the revised multi-level anti-icing warning threshold and the optimized water temperature change trend prediction model, including warning accuracy, prediction accuracy, false alarm rate and false alarm rate, to form a parameter optimization validation dataset.

[0170] The parameter optimization validation dataset is analyzed and compared with the performance metrics before optimization to evaluate the effect of parameter optimization.

[0171] If the parameter optimization effect meets the preset requirements, the corrected multi-level anti-icing early warning threshold and the optimized water temperature change trend prediction model will be determined as the formal operating parameters and updated to the anti-icing detection system.

[0172] If the parameter optimization effect does not meet the preset requirements, return to re-analyze the statistical data, adjust the threshold correction range and model training parameters until the optimization effect meets the requirements;

[0173] The entire process of parameter optimization and the verification results are recorded in the historical detection database to form a parameter optimization archive.

[0174] In this embodiment of the invention, verification data of the corrected multi-level anti-icing early warning threshold and the optimized water temperature change trend prediction model are collected using data collection tools. Core verification indicators include early warning accuracy, prediction accuracy, false negative rate, and false positive rate. These indicator data are categorized and organized according to verification time and test scenario to form a parameter optimization verification dataset. Data comparison and analysis tools are used to process the parameter optimization verification dataset, comparing the optimized early warning accuracy, prediction accuracy, false negative rate, and false positive rate with their corresponding performance indicators before optimization. Difference analysis is used to evaluate the parameter optimization effect and clarify the improvement of each indicator after optimization. If the comparison results show that the parameter optimization effect meets the preset requirements (early warning accuracy and prediction accuracy meet the standards, false negative rate and false positive rate are lower than the limit standards), then the corrected multi-level anti-icing early warning threshold and the optimized water temperature change trend prediction model are determined as the official operating parameters using parameter update tools and synchronously updated to the core parameter module of the anti-icing detection system to ensure that the system conducts detection work according to the optimized parameters. If the parameter optimization effect does not meet the preset requirements, the system returns to the historical detection database to reanalyze the statistical data. The threshold adjustment range is corrected using the threshold adjustment range calibration tool, and parameters such as the number of iterations and initial feature weights are adjusted using the model parameter adjustment tool. The threshold correction and model training are then repeated until the optimization effect meets the requirements. The entire parameter optimization process (including data source, adjustment basis, correction details, and training steps) and verification results (comparison data of various indicators, and compliance status) are completely recorded in the historical detection database using the archiving tool, forming a traceable and searchable parameter optimization archive.

[0175] Furthermore, the analysis of the parameter optimization verification dataset and the comparison with the performance metrics before optimization include the following steps:

[0176] An evaluation index system for the effect of parameter optimization is set, including early warning accuracy, prediction accuracy, false alarm rate, false alarm rate and data processing delay, and a corresponding qualified threshold is set for each evaluation index.

[0177] Extract the values ​​of each evaluation indicator in the parameter optimization verification dataset and compare them one by one with the set qualified threshold.

[0178] If the values ​​of all evaluation indicators reach or exceed the corresponding qualified thresholds, the parameter optimization effect is deemed to meet the preset requirements.

[0179] If at least one evaluation indicator fails to meet the acceptable threshold, the parameter optimization effect is deemed not to meet the preset requirements, and the evaluation indicator that fails to meet the standard and the corresponding numerical difference are recorded.

[0180] In this embodiment of the invention, an evaluation index system for parameter optimization effects is constructed, with the core evaluation indicators being early warning accuracy, prediction accuracy, false negative rate, false positive rate, and data processing delay. Each evaluation indicator has a corresponding qualified threshold set based on the actual needs of anti-icing detection in fire-fighting water storage facilities, ensuring accurate matching between the indicators and actual application scenarios. Data extraction tools are used to extract specific values ​​corresponding to each evaluation indicator from the parameter optimization verification dataset. During the extraction process, data integrity and consistency are checked using data verification tools, and invalid data caused by transmission interference is removed to ensure the reliability of the extracted values. Data comparison tools are used to compare each extracted indicator value with the set qualified thresholds one by one. The comparison process is carried out according to indicator type: first, the early warning accuracy and prediction accuracy are compared; then, the false negative rate and false positive rate are compared; and finally, the data processing delay is compared. If all evaluation indicators reach or exceed the corresponding qualified threshold, it indicates that the optimized parameters can accurately adapt to the anti-icing detection requirements, and the parameter optimization effect meets the preset requirements. If at least one evaluation indicator fails to reach the qualified threshold, it indicates that the optimized parameters have an adaptation defect, and the parameter optimization effect does not meet the preset requirements. At the same time, the name of the evaluation indicator that failed to meet the standard and the specific difference between the indicator value and the qualified threshold are recorded in detail through the difference recording tool, providing a clear direction for subsequent problem investigation.

[0181] Furthermore, if the parameter optimization effect does not meet the preset requirements, the process of re-analyzing statistical data, adjusting the threshold correction magnitude and model training parameters until the optimization effect meets the standards includes the following steps:

[0182] When the parameter optimization effect fails to meet the preset requirements, extract the relevant data corresponding to the evaluation indicators that fail to meet the standards, analyze the reasons for the failure of the indicators, and determine whether the threshold correction range is unreasonable or the model training parameters are set improperly.

[0183] If the threshold correction range is unreasonable, then combine the corresponding water temperature change cases in the historical detection data to recalculate a reasonable threshold correction range and correct the multi-level anti-icing warning threshold again.

[0184] If the model training parameters are set incorrectly, adjust the model's learning rate, number of iterations, and regularization parameters, and retrain the model using high-quality data from the supplementary training samples.

[0185] The revised warning threshold and the retrained model were re-validated, new validation data were collected, and the evaluation process was repeated until all evaluation indicators met the standards.

[0186] In this embodiment of the invention, when the optimization effect of the judgment parameters fails to meet the preset requirements, a data filtering tool is used to extract relevant data corresponding to the substandard evaluation indicators, including the detection time, ambient temperature, water temperature change data, early warning trigger records, and model prediction results corresponding to the indicator. A causal analysis tool is used to perform in-depth analysis of the extracted data to identify the core reasons for the indicators failing to meet the standards. The analysis focuses on whether the early warning deviation is caused by an unreasonable multi-level anti-icing early warning threshold correction range, or by improper training parameter settings in the water temperature change trend prediction model. If the analysis results show that the threshold correction range is unreasonable (e.g., too high a threshold leads to missed reports, too low a threshold leads to false reports), then the corresponding water temperature change cases in the historical detection database are retrieved. Case analysis tools are used to analyze the threshold adaptation patterns under different climatic conditions and different water temperature change rates, and a reasonable threshold correction range is recalculated. Based on this range, the first-level early warning benchmark value, the second-level early warning benchmark value, and the emergency early warning benchmark value are further corrected. If the analysis results indicate that the model training parameters are improperly set (e.g., an excessively high learning rate leading to overfitting, or insufficient iterations resulting in low prediction accuracy), then the learning rate, iteration count, and regularization parameters of the model should be adjusted using parameter adjustment tools. Simultaneously, high-quality data with even distribution and strong feature representativeness should be selected from the supplementary training samples using data filtering tools, and the model should be retrained using model training tools. After completing the correction and retraining, the latest anti-icing detection data should be collected as new validation data. The previous evaluation process should be repeated, comparing each evaluation indicator with the acceptable threshold one by one until all evaluation indicators reach the acceptable threshold, ensuring that the optimized parameters meet the actual detection requirements.

[0187] Furthermore, if the threshold correction range is unreasonable, then by combining the corresponding water temperature change cases in historical detection data, a reasonable threshold correction range is recalculated, and the multi-level anti-icing warning threshold is corrected again, including the following steps:

[0188] The relevant data corresponding to the evaluation indicators that did not meet the standards were screened to remove invalid data and abnormal interference data, and to retain representative and valid data samples.

[0189] The valid data samples were classified according to climate conditions, water storage facility type, and water temperature change stage, and the characteristics of the data samples under different classifications were analyzed.

[0190] For valid data samples of different categories, calculate the corresponding critical water temperature value, rate of change threshold, and early warning response time requirement, and determine the reasonable range of threshold correction based on these data characteristics;

[0191] Based on the reasonable range of threshold correction, and combined with the numerical differences of the evaluation indicators that did not meet the standards, the magnitude of the correction is accurately calculated to ensure that the corrected threshold can adapt to the water temperature change requirements under different classification scenarios.

[0192] In this embodiment of the invention, data cleaning tools are used to filter relevant data corresponding to evaluation indicators that fail to meet the standards. An outlier detection algorithm identifies and removes invalid and abnormal interference data caused by sensor malfunctions and signal interference, retaining valid data samples that truly reflect the water temperature change patterns and early warning response. Data classification tools are used to classify the valid data samples in multiple dimensions: by climate conditions (e.g., severe cold, cold, temperate); by water storage facility type (e.g., elevated fire water tanks, fire water pools); and by water temperature change stage (e.g., rapid decline, slow decline, stable fluctuation). This classification clearly presents the differences in data characteristics under different scenarios. For each type of valid data sample, feature extraction tools are used to calculate the corresponding water temperature critical value (the temperature threshold close to freezing), the rate of change threshold (the rate of water temperature decline that triggers an early warning), and the early warning response time requirement (the longest allowable time from detection to early warning). Combining these data characteristics, a threshold range calculation tool is used to determine a reasonable range for threshold correction, ensuring that the correction range covers all reasonable requirements for that type of scenario. Based on the determined reasonable range of threshold correction, and combined with the numerical differences of the evaluation indicators that did not meet the standards (such as the extent of the false alarm rate exceeding the standard, the extent of the false alarm rate exceeding the standard), the threshold correction calculation tool is used to accurately calculate the magnitude of the correction. The setting of the correction magnitude is based on the core principle of being able to make up for the numerical differences and adapt to the data characteristics, so as to ensure that the corrected threshold can not only meet the anti-icing detection requirements in this type of scenario, but also be compatible with the adaptation requirements of other classification scenarios, thus achieving accurate adaptation in different classification scenarios.

[0193] Furthermore, after accurately calculating the magnitude of the second correction based on the reasonable range of the threshold correction and the numerical differences of the evaluation indicators that did not meet the standards, ensuring that the corrected threshold can adapt to the water temperature change requirements under different classification scenarios, the following steps are also included:

[0194] Based on the determined threshold, the benchmark values ​​for Level 1, Level 2, and Emergency Warnings are adjusted again, and the threshold values ​​before and after the adjustment and the basis for the adjustment are recorded.

[0195] Valid data samples from different classification scenarios were selected as test data. The adjusted warning threshold was applied to the detection and analysis of the test data to obtain the warning accuracy, false alarm rate and false alarm rate data after the test.

[0196] The evaluation index data after the test is compared with the set pass threshold to determine whether the standard is met;

[0197] If the target is met, the threshold is adjusted again; if the target is still not met, the characteristics of the valid data samples are re-analyzed, the threshold adjustment range is adjusted, and the testing process is repeated until the evaluation indicators corresponding to the test data meet the target.

[0198] The process of revising the threshold, the basis for the adjustment, and the test results are recorded in detail in the parameter optimization file to improve the traceability of the optimization process.

[0199] In this embodiment of the invention, the threshold adjustment range is further adjusted based on the determined thresholds. A threshold adjustment tool precisely adjusts the primary warning benchmark, secondary warning benchmark, and emergency warning benchmark values ​​respectively. During the adjustment process, a data recording tool records the threshold values ​​before and after adjustment in real time, while detailing the adjustment basis, including the names of the evaluation indicators that did not meet the standards, the numerical differences, the characteristics of the corresponding valid data samples, and the calculation logic of the correction range, ensuring the traceability of the adjustment process. A data sampling tool extracts representative data as test data from valid data samples of different classification scenarios (high-level fire water tanks in frigid climates, fire water pools in cold climates, open-air water storage facilities in temperate climates, etc.). The test data needs to cover different water temperature change rates and different ambient temperature ranges in each scenario. The adjusted warning thresholds are imported into a detection and analysis tool, and the test data are input into the tool to conduct simulated detection and analysis. The tool makes a warning judgment based on the water temperature changes in the test data according to the adjusted thresholds, and outputs the warning accuracy rate, false alarm rate, and false alarm rate data after the test. A data comparison tool is used to compare the data of each evaluation indicator after the test with the preset qualified thresholds one by one to verify whether each indicator meets the qualified standard. If all indicators meet the standards, the threshold correction is confirmed to be compliant, and the threshold correction is completed. If any indicators still fail to meet the standards, the characteristics of the valid data samples are re-analyzed, with a focus on checking for any uncovered water temperature change scenarios or feature extraction biases. The correction magnitude is readjusted using the threshold correction magnitude calibration tool, and the above test process is repeated until all evaluation indicators corresponding to the test data meet the standards. The entire process of threshold correction (including the initial adjustment magnitude, the adjusted threshold, the source of test data, and the test process), the basis for adjustment, and the test results (the compliance status of each indicator and the improvement measures when they fail) are recorded in detail in the parameter optimization file using the file improvement tool, supplementing the complete optimization trajectory and further improving the traceability of the optimization process.

[0200] Furthermore, Embodiment 2 of the present invention also provides a fire-fighting water storage facility anti-icing detection system. The system is implemented based on the fire-fighting water storage facility anti-icing detection method described above. The fire-fighting water storage facility anti-icing detection system includes:

[0201] The data acquisition and calibration module is used to acquire the installation environment parameters and basic information of the fire water storage facility, deploy the appropriate temperature sensing components and complete the signal calibration to obtain the calibrated temperature sensing data acquisition link.

[0202] The associated data generation module is used to collect water temperature data in the fire-fighting water storage facility in real time through the temperature sensing data acquisition link, and simultaneously collect ambient temperature data around the facility to generate an associated dataset of water temperature and ambient temperature.

[0203] The early warning model construction module is used to set multi-level anti-icing early warning thresholds based on the associated dataset and construct a water temperature change trend prediction model.

[0204] The dynamic detection and early warning module is used to dynamically analyze the real-time collected water temperature data using the water temperature change trend prediction model, and output the anti-icing detection results and corresponding intervention prompts in combination with the multi-level anti-icing early warning threshold.

[0205] Therefore, the embodiments should be considered as exemplary and non-limiting in all respects, and the scope of the invention is defined by the appended claims rather than the foregoing description. Thus, all variations falling within the meaning and scope of the equivalents of the application are intended to be included within the invention.

[0206] The above description is merely a specific embodiment of the present invention, enabling those skilled in the art to understand or implement the invention. 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 the invention. Therefore, the present invention 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 of the invention herein.

Claims

1. A method for detecting ice formation in fire-fighting water storage facilities, characterized in that, Includes the following steps: Step S1: Obtain the installation environment parameters and basic water storage information corresponding to the fire water storage facility, deploy the appropriate temperature sensing components and complete the signal calibration to obtain the calibrated temperature sensing data acquisition link; Step S2: Collect water temperature data in the fire-fighting water storage facility in real time through the temperature sensing data acquisition link, and simultaneously collect ambient temperature data around the facility to generate a correlation dataset of water temperature and ambient temperature. Step S3: Based on the associated dataset, set multi-level anti-icing early warning thresholds and construct a water temperature change trend prediction model; Step S4: Utilize the water temperature change trend prediction model to dynamically analyze the real-time collected water temperature data, and combine the multi-level anti-icing early warning threshold to output the anti-icing detection results and corresponding intervention prompts.

2. The method for detecting anti-icing of fire-fighting water storage facilities according to claim 1, characterized in that, Step S1 includes the following steps: Collect the volume parameters, installation height parameters, open exposure area parameters, and winter climate characteristics parameters of the area where the fire-fighting water storage facility is located. At the same time, record the material properties and insulation structure information of the fire-fighting water storage facility to form a basic information set of the water storage facility. The number and location of temperature sensing components are determined based on the basic information set of the water storage facility. Sensing units adapted to low-temperature environments are selected and fixedly installed at different water depths and key parts of the wall inside the water storage facility. Zero-point calibration and range calibration are performed on each sensing unit, the calibration deviation value is recorded, a signal transmission link between the sensing unit and the data receiver is constructed, signal strength detection and anti-interference test are performed on the signal transmission link, and a link test report is generated. Based on the calibration deviation value and the link detection report, the signal transmission link is optimized and adjusted to obtain the calibrated temperature sensing data acquisition link.

3. The method for detecting anti-icing of fire-fighting water storage facilities according to claim 1, characterized in that, Step S2 includes the following steps: Through the calibrated temperature sensing data acquisition link, the water temperature data corresponding to each sensing unit is collected in real time according to the preset acquisition cycle, and the acquisition time and sensing location information corresponding to each water temperature data are marked to form the original water temperature dataset. Install ambient temperature sensors at a predetermined distance around the fire-fighting water storage facility to collect ambient temperature data synchronously and mark the collection time of the ambient temperature data to ensure that it is synchronized with the collection time of the water temperature data. The original water temperature dataset is subjected to outlier removal processing to remove invalid data caused by sensor interference and retain valid water temperature data. The synchronized and matched ambient temperature data and effective water temperature data are linked and integrated according to the collection time. Key parameters in the basic information set of water storage facilities are added as association tags to generate a correlation dataset of water temperature and ambient temperature.

4. The method for detecting anti-icing of fire-fighting water storage facilities according to claim 1, characterized in that, After the step of combining the multi-level anti-icing early warning threshold to output the anti-icing detection result and the corresponding intervention prompt information, the following steps are also included: Obtain the anti-icing detection results and corresponding intervention prompts, and construct an anti-icing detection log by combining the real-time water temperature data of the fire water storage facility and the ambient temperature data. The anti-icing detection logs are categorized and organized, and archived according to detection time, detection result level, and implementation of intervention measures to form a historical detection database; The data in the historical detection database are periodically statistically analyzed to calculate the triggering frequency of different warning levels, the effectiveness of intervention measures, and the water temperature change pattern, so as to generate statistical analysis results. Based on the statistical analysis results, the parameters of the multi-level anti-icing early warning threshold and water temperature change trend prediction model are optimized to improve the accuracy and timeliness of anti-icing detection. The statistical analysis results and parameter optimization schemes are generated into a report and synchronized to the fire protection facility maintenance management system to provide data support for the formulation of subsequent maintenance plans.

5. The method for detecting anti-icing of fire-fighting water storage facilities according to claim 4, characterized in that, The process of classifying and organizing the anti-icing detection logs includes the following steps: The archiving fields of the anti-icing detection log are determined, including the collection time, water temperature data of each sensor unit, ambient temperature data, detection result level, intervention prompt information, intervention measure implementation time, implementation personnel, and implementation effect feedback; The anti-icing detection logs are classified and archived according to a preset time period, and historical detection data sub-databases are established in units of months, quarters, and years. Each sub-database is labeled with the climate season and temperature change characteristics to which it belongs. Extract key data from each historical detection data sub-database, including the number of times different warning levels are triggered, the range of ambient temperature at the time of triggering, the distribution of water temperature drop rate, and the success rate of intervention measures, to form a basic dataset for statistical analysis; The statistical analysis dataset is processed using data statistical analysis methods to calculate the triggering frequency, average triggering duration, and effectiveness of intervention measures for different warning levels, and to summarize the water temperature change patterns under different climatic conditions, thereby forming a historical detection database.

6. The method for detecting anti-icing of fire-fighting water storage facilities according to claim 5, characterized in that, The process of processing the basic statistical analysis dataset using data statistical analysis methods to calculate the triggering frequency, average triggering duration, and effectiveness of intervention measures for different warning levels, and summarizing the water temperature change patterns under different climatic conditions, also includes the following steps: Based on the aforementioned statistical analysis dataset, it is determined whether the current multi-level anti-icing warning threshold is suitable for the water temperature change pattern under different climatic conditions. If there is a situation where the warning threshold is too high, resulting in missed reports, or too low, resulting in false reports, then the direction of threshold adjustment is determined. Extract water temperature change data after different warning levels are triggered from historical detection data, and combine the implementation effect of intervention measures to determine the threshold adjustment range, and correct the benchmark values ​​for the first-level warning, the second-level warning, and the emergency warning. For the water temperature change trend prediction model, new water temperature change data obtained from statistical analysis are selected as supplementary training samples, the input feature weights of the model are adjusted, and the model is retrained. The revised multi-level anti-icing early warning threshold and the optimized water temperature change trend prediction model were verified. The latest anti-icing detection data were used for testing to determine whether the early warning accuracy and prediction accuracy met the preset standards, and a parameter optimization file was generated.

7. The method for detecting anti-icing of fire-fighting water storage facilities according to claim 6, characterized in that, The verification of the revised multi-level anti-icing early warning threshold and the optimized water temperature change trend prediction model, using the latest anti-icing detection data, includes the following steps: Collect validation data of the revised multi-level anti-icing warning threshold and the optimized water temperature change trend prediction model, including warning accuracy, prediction accuracy, false alarm rate and false alarm rate, to form a parameter optimization validation dataset. The parameter optimization validation dataset is analyzed and compared with the performance metrics before optimization to evaluate the effect of parameter optimization. If the parameter optimization effect meets the preset requirements, the corrected multi-level anti-icing early warning threshold and the optimized water temperature change trend prediction model will be determined as the formal operating parameters and updated to the anti-icing detection system. If the parameter optimization effect does not meet the preset requirements, return to re-analyze the statistical data, adjust the threshold correction range and model training parameters until the optimization effect meets the requirements; The entire process of parameter optimization and the verification results are recorded in the historical detection database to form a parameter optimization archive.

8. The method for detecting anti-icing of fire-fighting water storage facilities according to claim 7, characterized in that, The analysis of the parameter optimization verification dataset and the comparison with the performance metrics before optimization include the following steps: An evaluation index system for the effect of parameter optimization is set, including early warning accuracy, prediction accuracy, false alarm rate, false alarm rate and data processing delay, and a corresponding qualified threshold is set for each evaluation index. Extract the values ​​of each evaluation indicator in the parameter optimization verification dataset and compare them one by one with the set qualified threshold. If the values ​​of all evaluation indicators reach or exceed the corresponding qualified thresholds, the parameter optimization effect is deemed to meet the preset requirements. If at least one evaluation indicator fails to meet the acceptable threshold, the parameter optimization effect is deemed not to meet the preset requirements, and the evaluation indicator that fails to meet the standard and the corresponding numerical difference are recorded.

9. The method for detecting anti-icing of fire-fighting water storage facilities according to claim 8, characterized in that, If the parameter optimization effect does not meet the preset requirements, the process involves re-analyzing the statistical data, adjusting the threshold correction magnitude and model training parameters until the optimization effect meets the target, including the following steps: When the parameter optimization effect fails to meet the preset requirements, extract the relevant data corresponding to the evaluation indicators that fail to meet the standards, analyze the reasons for the failure of the indicators, and determine whether the threshold correction range is unreasonable or the model training parameters are set improperly. If the threshold correction range is unreasonable, then combine the corresponding water temperature change cases in the historical detection data to recalculate a reasonable threshold correction range and correct the multi-level anti-icing warning threshold again. If the model training parameters are set incorrectly, adjust the model's learning rate, number of iterations, and regularization parameters, and retrain the model using high-quality data from the supplementary training samples. The revised warning threshold and the retrained model were re-validated, new validation data were collected, and the evaluation process was repeated until all evaluation indicators met the standards.

10. A fire-fighting water storage facility anti-icing detection system, characterized in that, The system is implemented based on the anti-icing detection method for fire-fighting water storage facilities as described in any one of claims 1-9 above, and the anti-icing detection system for fire-fighting water storage facilities includes: The data acquisition and calibration module is used to acquire the installation environment parameters and basic information of the fire water storage facility, deploy the appropriate temperature sensing components and complete the signal calibration to obtain the calibrated temperature sensing data acquisition link. The associated data generation module is used to collect water temperature data in the fire-fighting water storage facility in real time through the temperature sensing data acquisition link, and simultaneously collect ambient temperature data around the facility to generate an associated dataset of water temperature and ambient temperature. The early warning model construction module is used to set multi-level anti-icing early warning thresholds based on the associated dataset and construct a water temperature change trend prediction model. The dynamic detection and early warning module is used to dynamically analyze the real-time collected water temperature data using the water temperature change trend prediction model, and output the anti-icing detection results and corresponding intervention prompts in combination with the multi-level anti-icing early warning threshold.

Images