A dam fire-fighting early warning method and fire-fighting system based on big data

Abstract

The application provides a dam fire-fighting early warning method and a fire-fighting system based on big data. The method comprises the following steps: for each region of the dam, using an edge server to detect whether the parameters of each sensing data in the region are within the corresponding preset range; based on the sensing information and the local historical information corresponding to the warning region, obtaining the warning level and the first pre-diagnosis strategy information of the dam; obtaining the target pre-diagnosis strategy information according to the first pre-diagnosis strategy information; taking the sensing information as the first state information; generating the target pre-diagnosis strategy information based on the first pre-diagnosis strategy information and the second pre-diagnosis strategy information; and sending the target pre-diagnosis strategy information to the on-site user equipment. The sensing information in the fire-fighting system is used to monitor each region of the dam in real time, and the cloud server and the edge server are used to perform hierarchical processing and analysis on the sensor data, so that reliable and efficient early warning of the dam fire-fighting state is realized.

Description

Technical Field

[0001] This application relates to the technical fields of big data service technology and fault pre-diagnosis, and in particular to a dam fire early warning method and fire protection system based on big data. Background Technology

[0002] Dams are crucial water source regulation and flood control measures in water conservancy projects; therefore, their fire early warning systems are of paramount importance. However, existing technologies are primarily designed for relatively simple environments such as factories or industrial areas, employing a single-layer, flat information processing approach. There are discrepancies when applying these technologies to dam fire protection systems.

[0003] For example, taking a fire early warning system based on fire correlation disclosed in application number CN202110073419.X as an example, it includes an explicit fire information collection module, a implicit fire information collection module, and an identification and early warning module. It can execute fire early warnings when a fire is likely to occur, thus identifying specific fire hazard information even when there is only a fire risk, and then notifying maintenance personnel to eliminate the fire hazard, reducing the probability of a fire from its source and achieving precise fire prevention. However, it also uses a flat fault pre-diagnosis method, which limits its efficiency in complex environments.

[0004] For example, taking application number CN202310579735.3 as an example, a smart park fire early warning management system and method based on the Internet of Things is disclosed. This system predicts the current fire situation by collecting environmental data in real time and matching it with historical data, providing early warnings and fire alerts for firefighting work. Simultaneously, the system also helps firefighters determine whether the cause of a fire is related to electrical faults, thereby enabling early fire prevention and control, improving the accuracy of fire assessment, saving time, and reducing fire losses. However, it also employs a flat fault pre-diagnosis method. In the complex environment of a dam fire protection system, the prediction module needs to predict multiple dam areas across multiple dams, which limits its reliability and efficiency.

[0005] In other words, current technologies that rely solely on on-site servers or cloud servers for fire early warning systems cannot simultaneously guarantee the required reliability and efficiency when applied to dam fire early warning systems. Therefore, this application aims to provide a big data-based dam fire early warning method and fire protection system to address the aforementioned issues. Summary of the Invention

[0006] The purpose of this application is to provide a dam fire early warning method and fire protection system based on big data, so as to solve the problem of unsatisfactory safety and stability of dam fire early warning.

[0007] The objective of this application is achieved through the following technical solution:

[0008] This application provides a dam fire early warning method based on big data. The dam fire early warning method is used in a dam's fire protection system, which includes a cloud server, an edge server, an information acquisition module, a communication module, and on-site user equipment. The dam fire early warning method includes:

[0009] The information acquisition module is used to acquire sensor information in real time, including sensor data and area identifiers for different areas of the dam.

[0010] For each region, the edge server is used to detect whether the parameters of each sensor data in the region are within its corresponding preset range;

[0011] When at least N sensor data points within the area are detected to be outside their corresponding preset range, the area is designated as an early warning unit; N is a positive integer not less than 3.

[0012] One or more early warning units are used as early warning areas. Based on the sensor information and the local historical information corresponding to the early warning areas, the early warning level and the first pre-diagnosis strategy information of the dam are obtained.

[0013] When the warning level is lower than the preset warning level, the target pre-diagnosis strategy information is obtained according to the first pre-diagnosis strategy information;

[0014] When the warning level is not lower than the preset warning level, the sensing information is used as the first state information, and the first pre-diagnosis strategy information, the first state information and the semantic information corresponding to the warning area are sent to the cloud server through the communication module.

[0015] Based on the semantic information and the first state information, the second pre-diagnosis strategy information of the dam is obtained using the cloud server, and target pre-diagnosis strategy information is generated based on the first pre-diagnosis strategy information and the second pre-diagnosis strategy information.

[0016] The target pre-diagnosis strategy information is used to send to the on-site user equipment, enabling the user to perform fault pre-diagnosis of the dam's early warning area based on the target pre-diagnosis strategy information.

[0017] Further, the step of obtaining the second pre-diagnosis strategy information of the dam using the cloud server based on the semantic information and the first state information, and generating target pre-diagnosis strategy information based on the first and second pre-diagnosis strategy information, includes:

[0018] Based on the first pre-diagnosis strategy information and the historical pre-diagnosis strategy information set stored on the cloud server, the target pre-diagnosis strategy information is obtained;

[0019] Based on the semantic information and the first state information, the second pre-diagnosis strategy information of the dam is obtained;

[0020] Based on the second pre-diagnosis strategy information, obtain the credibility of the first pre-diagnosis strategy information;

[0021] When the confidence level is obtained within a preset time period and the confidence level is lower than the preset confidence level, the target pre-diagnosis strategy information is updated according to the second pre-diagnosis strategy information and sent to the on-site user equipment.

[0022] When the credibility is obtained within a preset time period and the credibility is not lower than the preset credibility, the target pre-diagnosis strategy information obtained according to the first pre-diagnosis strategy information and the historical pre-diagnosis strategy information set will be sent to the on-site user equipment.

[0023] If the confidence level is not obtained within the preset time period, the target pre-diagnosis strategy information obtained based on the first pre-diagnosis strategy information and the historical pre-diagnosis strategy information set will be sent to the on-site user equipment.

[0024] Furthermore, the method for obtaining the historical pre-diagnosis strategy information set includes:

[0025] According to preset rules, multiple historical data groups are acquired to form a historical pre-diagnosis strategy information set; each of the historical data groups includes warning level information, sensor data, warning area information, diagnosis strategy information, and processing feedback information.

[0026] The preset rules include at least one of the following:

[0027] Multiple historical data sets within a predetermined time range from the current time are used as a historical pre-diagnosis strategy information set;

[0028] Multiple historical data sets that are at the same level as the current early warning level of the dam are used as a historical pre-diagnosis strategy information set.

[0029] Further, based on the first pre-diagnosis strategy information and the historical pre-diagnosis strategy information set stored on the cloud server, target pre-diagnosis strategy information is obtained, including:

[0030] Obtain the diagnostic strategy information that has the highest similarity to the first pre-diagnosis strategy information from the historical pre-diagnosis strategy information set, and use it as the strategy information to be selected;

[0031] When the similarity between the strategy information to be selected and the first pre-diagnosis strategy information is greater than a preset similarity, the strategy information to be selected is taken as the target pre-diagnosis strategy information;

[0032] When the similarity between the selected strategy information and the first pre-diagnosis strategy information is not greater than a preset similarity, the first pre-diagnosis strategy information is used as the target pre-diagnosis strategy information.

[0033] Furthermore, according to preset rules, multiple historical data sets are acquired to form a historical pre-diagnosis strategy information set; each of the historical data sets includes warning level information, sensor data, warning area information, diagnosis strategy information, and processing feedback information.

[0034] The preset rules include at least one of the following:

[0035] Multiple historical data sets within a predetermined time range from the current time are used as a historical pre-diagnosis strategy information set;

[0036] Multiple historical data sets that are at the same level as the current early warning level of the dam are used as a historical pre-diagnosis strategy information set.

[0037] Further, obtaining the second pre-diagnosis strategy information for the dam based on the semantic information and the first state information includes:

[0038] Based on the semantic information, the danger index of the dam is obtained;

[0039] Based on the risk index and the first state information, the second pre-diagnosis strategy information is obtained.

[0040] Further, obtaining the second pre-diagnosis strategy information based on the risk index and the first state information includes:

[0041] The first state information is used as input, and the predicted policy information is obtained through the policy generation model.

[0042] The risk index is used as a correction factor to adjust the prediction strategy information, and the adjusted prediction strategy information is used as the second pre-diagnosis strategy information.

[0043] Secondly, this application also provides a fire protection system for a dam, which includes an information acquisition module, an edge server, a cloud server, and a communication module;

[0044] The information acquisition module is used to acquire sensor information in real time, including sensor data and area identifiers for different areas of the dam.

[0045] The edge server receives sensor information acquired by the information acquisition module and, for each region, detects whether the parameters of each sensor data within that region are within their corresponding preset range. When at least N sensor data within a region are detected to be outside their corresponding preset range, the region is designated as an early warning unit; N is a positive integer not less than 3. One or more acquired early warning units are designated as early warning regions. Based on the sensor information and the local historical information corresponding to the early warning region, the early warning level and first pre-diagnosis strategy information of the dam are obtained. When the early warning level is lower than a preset early warning level, target pre-diagnosis strategy information is obtained according to the first pre-diagnosis strategy information. When the early warning level is not lower than the preset early warning level, the sensor information is designated as first status information, and the first pre-diagnosis strategy information, the first status information, and the semantic information corresponding to the early warning region are sent to the cloud server via the communication module.

[0046] The cloud server is used to receive and obtain the second pre-diagnosis strategy information of the dam based on the semantic information and the first state information, and generate target pre-diagnosis strategy information based on the first pre-diagnosis strategy information and the second pre-diagnosis strategy information.

[0047] The target pre-diagnosis strategy information is used to send to the on-site user equipment, enabling the user to perform fault pre-diagnosis of the dam's early warning area based on the target pre-diagnosis strategy information.

[0048] Thirdly, this application also provides a computer-readable storage medium storing a computer program that, when executed by at least one processor, implements the steps of the method described in any of the first aspects.

[0049] Fourthly, this application also provides a computer program product comprising a computer program that, when executed by at least one processor, implements the steps of the control method described in any of the first aspects.

[0050] The beneficial effects of this application are: real-time monitoring of various areas of the dam through sensor information in the fire protection system, and hierarchical processing and analysis of sensor data through cloud servers and edge servers, thereby realizing early warning and fault pre-diagnosis of the dam status.

[0051] Edge servers can perform preliminary processing and analysis of sensor data. The processing results can serve as the first-level preprocessing results before being transmitted to cloud servers for further analysis. Cloud servers (i.e., remote servers) can perform more in-depth analysis and processing of data from local devices (edge ​​servers). On the one hand, by acquiring sensor information from various areas of the dam in real time, anomalies can be detected promptly. When sensor data in a certain area is abnormal, an early warning signal can be issued in a timely manner, helping to prevent accidents. On the other hand, it can automatically identify warning areas and generate pre-diagnosis strategies, reducing the burden of manual processing and improving efficiency. By combining the processing of edge servers and cloud servers, multi-level (hierarchical) pre-diagnosis can be achieved, improving the accuracy and reliability of diagnosis. Furthermore, sending target pre-diagnosis strategy information to on-site user equipment allows users to execute or judge the accuracy of fault pre-diagnosis, enhancing the interactivity and transparency of fault pre-diagnosis. When the warning level is lower than the preset warning level, it is not necessary to use cloud servers for fault pre-diagnosis; the problem can be solved solely through local judgment, making it more efficient. On the other hand, information is only sent to the cloud server when the warning level is not lower than the preset warning level. This reduces the information processing load on the cloud server, thereby improving the cloud server's processing speed for the received semantic information and the first state information. This is especially effective when the cloud server may be connected to edge servers of multiple dams. For cases where the warning level is not lower than the preset warning level, the pre-diagnosis results obtained through the dual-layer judgment of the edge server and the cloud server are more reliable. Attached Figure Description

[0052] The present application will be further described below with reference to the accompanying drawings and embodiments.

[0053] Figure 1 This is a flowchart illustrating a dam fire early warning method based on big data, provided in an embodiment of this application.

[0054] Figure 2 This is a flowchart illustrating a method for generating target pre-diagnosis strategy information provided in an embodiment of this application;

[0055] Figure 3 This is a flowchart illustrating a method for obtaining second pre-diagnosis strategy information provided in an embodiment of this application;

[0056] Figure 4 This is a structural block diagram of a fire protection system provided in an embodiment of this application;

[0057] Figure 5 This is a schematic diagram of the structure of a computer program product provided in an embodiment of this application. Detailed Implementation

[0058] The present application will now be further described in conjunction with the accompanying drawings and specific embodiments. It should be noted that, without conflict, the various embodiments or technical features described below can be arbitrarily combined to form new embodiments. The implementation process of the present application will be described below with reference to the accompanying drawings and preferred embodiments. Those skilled in the art can easily understand other advantages and effects of the present application from the content disclosed in this specification. The present application can also be implemented or applied through other different specific implementation procedures, and the details in this specification can also be modified or changed based on different viewpoints and applications without departing from the spirit of the present application. It should be understood that the preferred embodiments are only for illustrating the present application and not for limiting the scope of protection of the present application.

[0059] The following is a brief description of the technical field and related terms of the embodiments of this application, so as to facilitate understanding by those skilled in the art.

[0060] Fault prediction refers to the process of monitoring, analyzing, and evaluating equipment status and performance data to identify potential signs of equipment failure or abnormality in advance, and to diagnose and predict possible failures. Its aim is to prevent equipment failures or disasters from actually occurring by taking preventative maintenance measures, thereby reducing the likelihood of loss. Fault prediction can involve using sensor technology, data acquisition, data analysis, machine learning, and other technologies, combined with the experience of professionals, to monitor and analyze equipment status in real time to identify potential failure modes and provide targeted maintenance recommendations or measures.

[0061] Fire early warning, as an application of fault pre-diagnosis technology, refers to the use of data collected by sensors in fire protection systems to detect signs that may lead to fires or other safety risks in advance, and to diagnose and predict potential problems.

[0062] Compared to other fields, the safety and stability of dams are crucial to the safety of the surrounding environment and personnel. Therefore, dam fire protection systems need to be more reliable and efficient to cope with potential fire risks and emergencies. This application provides a dam fire early warning method and fire protection system based on big data. It addresses the problem that existing fault pre-diagnosis technologies, typically designed for relatively simple, single-layer, flat environments such as factories, employ single-layer information processing methods and cannot effectively utilize the advantages of big data processing. This application uses sensor information within the fire protection system to monitor various areas of the dam in real time and performs hierarchical processing and analysis of sensor data through cloud servers and edge servers, thereby achieving early warning of the dam's (fire) status.

[0063] Edge servers are computing devices located at the network edge or close to the data source, used for processing and storing data, as well as performing computing tasks. Compared to traditional cloud servers, edge servers are closer to where data is generated, thus providing faster response times and lower network latency. Cloud servers are one of the core infrastructures of the big data era. Based on cloud computing technology, they virtualize and abstract physical hardware resources such as computing, storage, and network resources, enabling large-scale computing and storage resources to be allocated and used on demand, providing reliable and efficient basic support for processing massive amounts of data.

[0064] Regarding the technical solution provided in this application, the edge server can perform preliminary processing and analysis on sensor data. The processing results can serve as the first-level preprocessing results, which are then transmitted to the cloud server for further analysis. The cloud server can perform more in-depth analysis and processing on data from the local end (edge ​​server), such as using machine learning algorithms for fault diagnosis and predictive analysis. The following section will first describe the dam fire early warning method based on big data, and then describe the system, etc.

[0065] Method implementation examples.

[0066] See Figure 1 , Figure 1 This is a flowchart illustrating a dam fire early warning method based on big data, provided in an embodiment of this application.

[0067] This embodiment provides a big data-based dam fire early warning method. The method is used in the dam's fire protection system, which includes a cloud server, an edge server, an information acquisition module, a communication module, and on-site user equipment. The cloud server communicates with multiple edge servers at the dam, and the edge servers are connected to the information acquisition module to obtain sensor information from each area of ​​the dam. The method includes steps S101 to S107.

[0068] Step S101: Use the information acquisition module to acquire sensor information in real time. The sensor information includes sensor data and area identifiers for different areas of the dam.

[0069] As an example, the dam is divided into a living area, an office area, a dam body area, and a power generation equipment area. The area labels are 1#, 2#, 3#, and 4#, respectively. Sensor information includes indicators such as [1#, smoke sensor smoke concentration is 30 ppm, temperature sensor temperature is 25°C], [2#, flame sensor did not detect a flame (parameter is 0), temperature sensor temperature is 25°C], etc.

[0070] Step S102: For each region, use an edge server to detect whether the parameters of each sensor data in the region are within its corresponding preset range.

[0071] It can be assumed that the preset ranges for the same type of sensor are different in different areas. For example, the preset range of the smoke sensor in the living area is [0ppmm, 50ppm], while the preset range of the smoke sensor in the office area is [0ppmm, 70ppm].

[0072] Step S103: When at least N sensor data points within the region are detected to be outside their corresponding preset ranges, the region is designated as an early warning unit; N is a positive integer not less than 3. As an example, the value of N could be 3, 4, 6, or 10.

[0073] Step S104: Using one or more acquired early warning units as early warning areas, based on the sensor information and the local historical information corresponding to the early warning areas, obtain the early warning level and first pre-diagnosis strategy information of the dam.

[0074] Warning levels are indicated by numbers such as A, B, Level 1, and Level 3. Different warning levels correspond to different policy information. The first type of pre-diagnostic policy information might include audible and visual alarms to alert people to fire prevention, reminders for personnel to check alarm sources in the corresponding area, or recommendations to continue monitoring.

[0075] Step S105: When the warning level is lower than the preset warning level, obtain the target pre-diagnosis strategy information according to the first pre-diagnosis strategy information.

[0076] As an example, when the warning level is selected from A, B, C, D, and E, A can indicate a high probability of a fire accident, while E indicates no fire accident will occur. The preset warning level is C. When the warning level is E, the target pre-diagnosis strategy information is obtained based on the first pre-diagnosis strategy information.

[0077] Step S106: When the warning level is not lower than the preset warning level, the sensing information is used as the first state information, and the first pre-diagnosis strategy information, the first state information and the semantic information corresponding to the warning area are sent to the cloud server through the communication module.

[0078] Step S107: Based on the semantic information and the first state information, the second pre-diagnosis strategy information of the dam is obtained using the cloud server, and target pre-diagnosis strategy information is generated based on the first pre-diagnosis strategy information and the second pre-diagnosis strategy information.

[0079] Semantic information can be obtained by processing the inspection information recorded by patrol personnel during their inspections of each area. Inspection information may include the operational status, maintenance details, repair history, fault repair records, and visual condition of fire-fighting equipment. After processing, this inspection information can be incorporated into the corpus as part of the semantic information, used to gain a deeper understanding of the dam's condition and potential problems.

[0080] The target pre-diagnosis strategy information includes the area identifier and recommended processing scheme corresponding to each early warning unit. The target pre-diagnosis strategy information is used to send to the on-site user equipment so that the user can perform fault pre-diagnosis of the early warning area of ​​the dam according to the target pre-diagnosis strategy information.

[0081] This embodiment uses sensor information from the fire protection system to monitor various areas of the dam in real time, and performs hierarchical processing and analysis of the sensor data through cloud servers and edge servers, thereby achieving early warning and fault prediction of the dam's condition. Compared to related technologies that only use a single-level identification and early warning module for all data processing and result prediction, this embodiment provides multi-level (hierarchical) pre-diagnosis (fire early warning).

[0082] Specifically, the information acquisition module uses real-time sensors in different areas of the dam to acquire data such as smoke and temperature. The edge server checks whether the parameters of the sensor data for each area are within a preset range. When at least a certain number (N) of the sensor data in a certain area are outside the preset range, that area is marked as a warning unit. The selection of warning units can be flexibly adjusted according to different values ​​of N to adapt to warning needs under different circumstances. The warning level of the dam is determined based on sensor information and local historical information, and first pre-diagnosis strategy information is generated. The generation of warning level and strategy information can be evaluated based on preset standards and historical data. When the warning level is lower than the warning level, target pre-diagnosis strategy information is generated based on the first pre-diagnosis strategy information. When the warning level is not lower than the warning level, sensor information, first pre-diagnosis strategy information, and semantic information are sent to the cloud server for further analysis and processing. The cloud server uses semantic information and first state information to acquire second pre-diagnosis strategy information for the dam and integrates it with the first pre-diagnosis strategy information to generate target pre-diagnosis strategy information. The acquisition and generation of second pre-diagnosis strategy information can utilize big data analytics technology, combining more data and information for deeper pre-diagnosis. The target pre-diagnosis strategy information will also be sent to the user equipment, and the user will use this information to perform pre-diagnosis of the dam's early warning area.

[0083] The advantages of this approach are as follows: Firstly, by acquiring sensor information from various areas of the dam in real time, anomalies can be detected promptly. When sensor data in a certain area is abnormal, an early warning signal can be issued in a timely manner, helping to prevent accidents. Secondly, it can automatically identify warning areas and generate pre-diagnosis strategies, reducing the burden of manual processing and improving efficiency. By combining the processing of edge servers and cloud servers, multi-level (hierarchical) pre-diagnosis can be achieved, improving the accuracy and reliability of diagnosis. Furthermore, sending the target pre-diagnosis strategy information to on-site user equipment allows users to execute or judge the accuracy of fault pre-diagnosis, enhancing the interactivity and transparency of fault pre-diagnosis. When the warning level is lower than the preset warning level, there is no need to use the cloud server for fault pre-diagnosis; the problem can be solved solely through local judgment, making it more efficient. Moreover, information is only sent to the cloud server when the warning level is not lower than the preset warning level, reducing the information processing load on the cloud server and thus improving the cloud server's processing speed for received semantic information and the first state information. This is especially effective when the cloud server may connect to edge servers of multiple dams. For situations where the warning level is not lower than the warning level, the pre-diagnosis results obtained through a two-layer judgment using edge servers and cloud servers are more reliable.

[0084] See Figure 2 , Figure 2 This is a flowchart illustrating a method for generating target pre-diagnosis strategy information provided in an embodiment of this application.

[0085] In some optional implementations, the step of obtaining the second pre-diagnosis strategy information of the dam using the cloud server based on the semantic information and the first state information, and generating target pre-diagnosis strategy information based on the first pre-diagnosis strategy information and the second pre-diagnosis strategy information (i.e., step S107) includes steps S201 to S206.

[0086] Step S201: Obtain target pre-diagnosis strategy information based on the first pre-diagnosis strategy information and the historical pre-diagnosis strategy information set stored on the cloud server.

[0087] The acquisition time for historical pre-diagnosis strategy information sets can be set to one day, one week, or one month, depending on the needs.

[0088] As an example, the historical pre-diagnostic strategy information set mainly refers to a collection of multiple sets of data obtained from information sent to the cloud server by multiple dams or the current dam itself over a period of time, based on warning levels not lower than the preset warning level, and the diagnostic strategies obtained through the cloud server. Each set of data includes, for example: warning level information, recording the warning level at each warning and the specific conditions or events that triggered the warning; sensor data, including real-time data from sensors that triggered the warning, such as temperature, smoke concentration, and flame detection sensors; warning area information, identifying the specific area or location involved in each warning; diagnostic strategy information, recording the diagnostic strategies generated by the cloud server based on the received sensor data and warning information, including predicted fault types, possible causes, and suggested handling methods; and processing feedback, recording user evaluation scores after each warning. The historical pre-diagnostic strategy information set does not contain all information about the dam, only recording information related to warning levels not lower than the preset warning level, avoiding the recording of a large amount of data unrelated to the warning level, reducing data redundancy and storage costs. Centralized recording of situations with higher warning levels and the corresponding diagnostic strategies from the cloud server helps to focus on analyzing data at critical moments and utilize big data technology to uncover patterns and characteristics in fault diagnosis.

[0089] Therefore, by collecting and analyzing historical pre-diagnosis strategy information sets, it is possible to optimize and improve the performance of the fault pre-diagnosis system, improve the accuracy and timeliness of early warnings, reduce false alarms and missed alarms, and thus improve the reliability and efficiency of the dam fire protection system.

[0090] Step S202: Based on the semantic information and the first state information, obtain the second pre-diagnosis strategy information of the dam.

[0091] Step S203: Obtain the credibility of the first pre-diagnosis strategy information based on the second pre-diagnosis strategy information.

[0092] The confidence level of the first pre-diagnosis strategy information can range from [0,1], representing the degree of trust in the first pre-diagnosis strategy information. For example, a confidence level of 0.8 indicates an 80% trust level in the first pre-diagnosis strategy information.

[0093] This embodiment does not limit the method of obtaining the credibility of the first pre-diagnosis strategy information based on the second pre-diagnosis strategy information. It can be a credibility score generated by weighted summation of the similarity of each group of similar data in the first and second pre-diagnosis strategy information.

[0094] As an example, the first and second pre-diagnostic strategy information respectively include: warning level, warning area, and diagnostic strategy information. The corresponding data in both sets of information are compared one by one, their similarity is calculated, and a weighted sum is performed based on the similarity to obtain a reliability score. The method for obtaining the similarity can be found in the section below on obtaining the similarity between the first pre-diagnostic strategy information and the diagnostic strategy information in the historical pre-diagnostic strategy information set; it will not be repeated here. The results obtained through the above method are as follows:

[0095] The similarity of warning levels is 0.8, with a weight of 0.4; the similarity of warning areas is 0.85, with a weight of 0.3; and the similarity of diagnostic strategy information is 0.75, with a weight of 0.3. Finally, through weighted summation, the credibility score is: Credibility = (0.8 × 0.4) + (0.85 × 0.3) + (0.75 × 0.3) = 0.8. It can be assumed that the sum of all weights is 1. Limiting the sum of weights to 1 ensures that the magnitudes of the weights remain consistent, making comparisons and interpretations between weights easier.

[0096] Step S204: When the confidence level is obtained within a preset time period and the confidence level is lower than the preset confidence level, the target pre-diagnosis strategy information is updated according to the second pre-diagnosis strategy information and sent to the on-site user equipment.

[0097] The preset confidence level is a pre-defined threshold used to determine whether the confidence level of the first pre-diagnosis strategy information meets the requirements for updating the target pre-diagnosis strategy information. For example, a preset confidence level of 0.7 means that the target pre-diagnosis strategy information will only be updated when the confidence level is greater than or equal to 0.7. The preset confidence level can range from (0,1).

[0098] The preset duration refers to the time frame within which the credibility of the first pre-diagnostic strategy information is tested. For example, it can be set to 10 minutes, meaning that the credibility is tested within 10 minutes to see if it reaches the preset value.

[0099] Step S205: When the confidence level is obtained within a preset time period and the confidence level is not lower than the preset confidence level, the target pre-diagnosis strategy information obtained based on the first pre-diagnosis strategy information and the historical pre-diagnosis strategy information set is sent to the on-site user equipment. In this case, the cloud server may send the target pre-diagnosis strategy information to the edge server, and the edge server may then send the target pre-diagnosis strategy information to the on-site user equipment.

[0100] Step S206: If the confidence level is not obtained within the preset time period, the target pre-diagnosis strategy information obtained based on the first pre-diagnosis strategy information and the historical pre-diagnosis strategy information set is sent to the on-site user equipment. In this case, the cloud server may send the target pre-diagnosis strategy information to the edge server, and the edge server may then send the target pre-diagnosis strategy information to the on-site user equipment. As an example, the target pre-diagnosis strategy information can be sent to the on-site user equipment through different methods such as pop-ups, emails, and app push notifications.

[0101] On-site user equipment can be various devices related to the fire protection system, such as fire prevention broadcasting equipment used to broadcast target pre-diagnosis strategy information via voice. Other examples include mobile terminal devices (such as mobile phones and tablets) with dedicated fire warning and fault pre-diagnosis applications installed to receive and view warning information and pre-diagnosis strategies corresponding to the target pre-diagnosis strategy information, and to provide corresponding operational guidance and feedback.

[0102] By acquiring historical pre-diagnosis strategy information sets for the dam, the response measures and effects taken in similar past situations, provided by big data, can offer a reference for processing current dam early warning information. Based on semantic information and first-state information, second-level pre-diagnosis strategy information for the dam is acquired to analyze and evaluate the current early warning situation from more angles and dimensions. Based on the second-level pre-diagnosis strategy information, the credibility of the first-level pre-diagnosis strategy information is assessed. This determines both the accuracy and reliability of the first-level pre-diagnosis strategy information and whether updating the target pre-diagnosis strategy information is necessary. Based on the credibility assessment results, a decision is made whether to update the target pre-diagnosis strategy information, and the updated information can be sent to on-site user equipment. This implementation method ensures that on-site users of the dam receive reliable early warning information, enabling them to take timely and appropriate countermeasures.

[0103] Therefore, by comprehensively considering historical, semantic, and real-time status information on the cloud server, the overall analysis capability of early warning information is improved, resulting in higher accuracy. Based on the information obtained from the second pre-diagnosis strategy, early warning strategies can be optimized in a timely manner, improving the accuracy and targeting of early warnings. By updating the target pre-diagnosis strategy information promptly, on-site users are ensured to receive the latest early warning information in a timely manner, effectively addressing potential risks and problems.

[0104] In some optional implementations, the step of obtaining the target pre-diagnosis strategy information based on the first pre-diagnosis strategy information and the historical pre-diagnosis strategy information set stored on the cloud server (i.e., step S201) includes:

[0105] Obtain the diagnostic strategy information that has the highest similarity to the first pre-diagnosis strategy information from the historical pre-diagnosis strategy information set, and use it as the strategy information to be selected;

[0106] When the similarity between the strategy information to be selected and the first pre-diagnosis strategy information is greater than a preset similarity, the strategy information to be selected is taken as the target pre-diagnosis strategy information;

[0107] When the similarity between the selected strategy information and the first pre-diagnosis strategy information is not greater than a preset similarity, the first pre-diagnosis strategy information is used as the target pre-diagnosis strategy information.

[0108] It can be considered that the diagnostic strategy information in the first pre-diagnostic strategy information (including the fault type predicted by the edge server, possible causes, and suggested handling methods) is compared with the similarity of each diagnostic strategy information in the historical pre-diagnostic strategy information set. The diagnostic strategy information with the highest similarity to the first pre-diagnostic strategy information is selected as the candidate strategy information. The similarity comparison can be achieved by converting both into feature vectors. Specifically, one-hot encoding can be used to map each feature to a vector, where the dimension of the vector equals the number of values ​​the feature can take. The vectors of each feature are then merged into a single feature vector. Alternatively, the vectors of each feature can be concatenated to form a longer vector. Standardizing the feature vectors ensures that the value ranges of each feature are the same, which helps improve the accuracy of the similarity calculation.

[0109] As an example, diagnostic strategy information A in the first pre-diagnostic strategy information includes "Fault type: circuit fault; Possible cause: short circuit; Suggested handling method: check if the circuit connection is loose," and diagnostic strategy information B in the historical pre-diagnostic strategy information set includes "Fault type: sensor fault; Possible cause: sensor damage; Suggested handling method: replace the faulty sensor." This information is converted into feature vectors, and then one-hot encoding is used to represent the fault type features. Since the fault type has two possible values ​​(circuit fault and sensor fault), the fault type feature vector for diagnostic strategy information A is [1,0], and the fault type feature vector for diagnostic strategy information B is [0,1]. Using a bag-of-words model, each word is mapped to a feature, and word frequencies are counted. The cause feature vector can be represented as [1,1,0,0,0,...], where the first element indicates that "short circuit" occurred once, the second element indicates that "sensor damage" did not occur, and so on. All feature vectors are merged into a single overall feature vector. Assuming the feature vector length is 7, including one-hot encoding of the fault type and a bag-of-words model representation of possible causes / suggested handling methods, the feature vector of diagnostic strategy information A can be represented as [1,0,1,0,0,1,0], and the feature vector of diagnostic strategy information B can be represented as [0,1,0,1,1,0,0]. Then, methods such as cosine similarity are used to confirm the similarity between the two, ultimately obtaining the diagnostic strategy information with the highest similarity to the first pre-diagnostic strategy information.

[0110] Therefore, by comparing the similarity with historical pre-diagnosis strategy information, a more comprehensive consideration of handling schemes in similar past situations can be made, improving the reliability of decision-making. Selecting historical pre-diagnosis strategy information with high similarity as the target pre-diagnosis strategy information is beneficial because historical information may have already been verified as effective, helping to reduce the risk of handling faults. The resulting target pre-diagnosis strategy information relies more on statistical analysis of historical data, extracting regularities and judgment criteria from historical events, which is conducive to the reliability of early warnings regarding dam fire conditions.

[0111] In some optional implementations, the historical pre-diagnosis strategy information set is obtained in the following ways:

[0112] According to preset rules, multiple historical data groups are acquired to form a historical pre-diagnosis strategy information set; each of the historical data groups includes warning level information, sensor data, warning area information, diagnosis strategy information, and processing feedback information.

[0113] The preset rules include at least one of the following:

[0114] Multiple historical data sets within a predetermined time range from the current time are used as a historical pre-diagnosis strategy information set;

[0115] Multiple historical data sets that are at the same level as the current early warning level of the dam are used as a historical pre-diagnosis strategy information set.

[0116] A historical pre-diagnosis strategy information set can be created by acquiring multiple historical data sets within a given time frame. Alternatively, multiple historical data sets with the same warning level can be directly acquired to form a historical pre-diagnosis strategy information set. Acquiring this historical pre-diagnosis strategy information set allows for a better understanding of past events and corresponding response strategies, providing a reference for current fault pre-diagnosis. By analyzing and summarizing measures taken in similar past situations using historical data, the accuracy and reliability of current pre-diagnosis strategies can be assessed through big data analysis.

[0117] See Figure 3 , Figure 3 This is a flowchart illustrating a method for obtaining second pre-diagnosis strategy information provided in an embodiment of this application.

[0118] In some optional implementations, obtaining the second pre-diagnosis strategy information of the dam based on the semantic information and the first state information (i.e., step S202) includes steps S301 and S302.

[0119] Step S301: Based on the semantic information, obtain the danger index of the dam.

[0120] A comprehensive assessment of the dam is conducted using semantic information to determine its current hazard index. Semantic information may include inspection records of various areas of the dam by patrol personnel, as well as other relevant information. Analysis of this information allows for the determination of the dam's current safety status, i.e., its hazard index.

[0121] Step S302: Based on the danger index and the first state information, obtain the second pre-diagnosis strategy information.

[0122] The purpose of using a hazard index in this embodiment is to transform complex semantic information into a quantifiable indicator, thereby facilitating further analysis and processing. The hazard index comprehensively considers factors such as inspection records, converting this information into a unified indicator that reflects the dam's current overall safety status and potential risk level from a perspective different from sensor data. The hazard index can be used in the model to more accurately assess the dam's safety condition.

[0123] Therefore, by comprehensively considering semantic information and first-state information, the safety status of the dam can be assessed more comprehensively, improving the accuracy and reliability of early warning and pre-diagnosis. Based on the comprehensive analysis of information from multiple sources, potential dam failures or anomalies can be more accurately identified, enabling timely early warnings and corresponding measures to be taken, thus improving the effectiveness of early warnings.

[0124] The risk index of the dam can be obtained based on the semantic information through a risk assessment matrix.

[0125] As an example, the warning zone for the dam is the area containing power generation equipment. The semantic information for this area corresponds to corpus content such as, "Fire-fighting equipment has undergone inspection and maintenance; smoke sensors are functioning well; a fire recently occurred but has been effectively controlled; temperature sensor #3 is malfunctioning; a fire drill was conducted within the past week," etc. The method for obtaining the dam's hazard index is as follows:

[0126] Fire frequency 4 5 20 Firefighting equipment status 5 2 10 Emergency evacuation situation 2 3 6

[0127] In this case, the danger index is the product of the probability and severity values. It can be assumed that by combining semantic information to obtain the probability values ​​of fire frequency, fire spread rate, fire equipment status, and emergency evacuation situations, and combining these with a preset severity value, the danger index is obtained.

[0128] In some optional implementations, step S302, obtaining the second pre-diagnosis strategy information based on the hazard index and the first state information, includes:

[0129] The first state information is used as input, and the predicted policy information is obtained through the policy generation model.

[0130] The risk index is used as a correction factor to adjust the prediction strategy information, and the adjusted prediction strategy information is used as the second pre-diagnosis strategy information.

[0131] The training process of the policy generation model includes:

[0132] Obtain a training set, which includes multiple training data, each of which includes sample state information and labeled data of policy information corresponding to the sample state information;

[0133] For each training data point in the training set, the following processing is performed:

[0134] The sample state information in the training data is input into a preset deep learning model to obtain the predicted data of the policy information corresponding to the sample state information.

[0135] Based on the predicted data and labeled data of the policy information corresponding to the sample state information, the model parameters of the deep learning model are updated;

[0136] The system checks whether the preset training termination condition is met; if so, the trained deep learning model is used as the policy information model; if not, the deep learning model is trained again using the next set of training data.

[0137] The training set for the strategy generation model is obtained from the cloud-based historical information of the dam stored on the cloud server. The cloud-based historical information includes past fault records, maintenance status, operation history (i.e., historical semantic information), historical state information, and historical pre-diagnosis strategy information of multiple dams.

[0138] In some optional implementations, the risk index is used as a correction factor to adjust the prediction strategy information using a correction formula. The correction formula may be:

[0139] When the risk index is higher than the preset index range, the recommended handling method in the prediction strategy information will be increased.

[0140] When the risk index is below the preset index range, the recommended handling method in the prediction strategy information is adjusted downwards.

[0141] When the risk index is within the preset index value range, the processing method recommended in the prediction strategy information will not be adjusted.

[0142] As an example, if a fire prediction model forecasts a potential fire type as electrical equipment failure, the recommended course of action is to inspect the electrical equipment and immediately disconnect the power. For instance, if the hazard index is lower than the preset hazard index, it indicates that the likelihood of an electrical equipment failure at the dam is low, and the course of action for this prediction can be adjusted, such as not immediately disconnecting the power but instead arranging for further verification.

[0143] On the other hand, if the danger index is not lower than the preset danger index, it means that the dam is in an extremely dangerous state. The recommendations for handling electrical equipment failures can be revised and strengthened, such as immediately cutting off the power supply and taking emergency measures.

[0144] In some optional implementations, the historical pre-diagnosis strategy information set is obtained in the following ways:

[0145] According to preset rules, multiple historical data groups are acquired to form a historical pre-diagnosis strategy information set; each of the historical data groups includes warning level information, sensor data, warning area information, diagnosis strategy information, and processing feedback information.

[0146] The preset rules include at least one of the following:

[0147] Multiple historical data sets within a predetermined time range from the current time are used as a historical pre-diagnosis strategy information set;

[0148] Multiple historical data sets that are at the same level as the current early warning level of the dam are used as a historical pre-diagnosis strategy information set.

[0149] A historical pre-diagnosis strategy information set can be created by acquiring multiple historical data sets within a given time frame. Alternatively, multiple historical data sets with the same warning level can be directly acquired to form a historical pre-diagnosis strategy information set. Acquiring this historical pre-diagnosis strategy information set allows for a better understanding of past events and corresponding response strategies, providing a reference for current fault pre-diagnosis. By analyzing and summarizing measures taken in similar past situations using historical data, the accuracy and reliability of current pre-diagnosis strategies can be assessed through big data analysis.

[0150] In a specific application, a dam fire early warning method based on big data and semantic processing is provided. This method is used in a dam's fire protection system, which includes a cloud server, an edge server, an information acquisition module, a communication module, and on-site user equipment. The dam fire early warning method includes:

[0151] The information acquisition module is used to acquire sensor information in real time, including sensor data and area identifiers for different areas of the dam.

[0152] Edge servers are used to detect sensor data. For each region, the edge server is used to detect whether the parameters of each sensor data point within that region are within their corresponding preset range.

[0153] When at least N sensor data points within the area are detected to be outside their corresponding preset range, the area is designated as an early warning unit; N is a positive integer not less than 3.

[0154] One or more early warning units are used as early warning areas. Based on the sensor information and the local historical information corresponding to the early warning areas, the early warning level and the first pre-diagnosis strategy information of the dam are obtained.

[0155] When the warning level is lower than the preset warning level, the target pre-diagnosis strategy information is obtained according to the first pre-diagnosis strategy information;

[0156] When the warning level is not lower than the preset warning level, the sensing information is used as the first state information, and the first pre-diagnosis strategy information, the first state information and the semantic information corresponding to the warning area are sent to the cloud server through the communication module.

[0157] Based on the first pre-diagnosis strategy information and the historical pre-diagnosis strategy information set stored on the cloud server, the target pre-diagnosis strategy information is obtained;

[0158] Obtain the diagnostic strategy information that has the highest similarity to the first pre-diagnosis strategy information from the historical pre-diagnosis strategy information set, and use it as the strategy information to be selected;

[0159] When the similarity between the strategy information to be selected and the first pre-diagnosis strategy information is greater than a preset similarity, the strategy information to be selected is taken as the target pre-diagnosis strategy information;

[0160] When the similarity between the selected strategy information and the first pre-diagnosis strategy information is not greater than a preset similarity, the first pre-diagnosis strategy information is used as the target pre-diagnosis strategy information.

[0161] Based on the semantic information, the danger index of the dam is obtained;

[0162] The first state information is used as input, and the predicted policy information is obtained through the policy generation model.

[0163] The risk index is used as a correction factor to adjust the prediction strategy information, and the adjusted prediction strategy information is used as the second pre-diagnosis strategy information.

[0164] Based on the second pre-diagnosis strategy information, obtain the credibility of the first pre-diagnosis strategy information;

[0165] When the confidence level is obtained within a preset time period and the confidence level is lower than the preset confidence level, the target pre-diagnosis strategy information is updated according to the second pre-diagnosis strategy information and sent to the on-site user equipment.

[0166] When the credibility is obtained within a preset time period and the credibility is not lower than the preset credibility, the target pre-diagnosis strategy information obtained according to the first pre-diagnosis strategy information and the historical pre-diagnosis strategy information set will be sent to the on-site user equipment.

[0167] If the confidence level is not obtained within the preset time period, the target pre-diagnosis strategy information obtained based on the first pre-diagnosis strategy information and the historical pre-diagnosis strategy information set will be sent to the on-site user equipment, so that the user can perform fault pre-diagnosis of the early warning area of ​​the dam based on the target pre-diagnosis strategy information.

[0168] The methods for obtaining the historical pre-diagnosis strategy information set include:

[0169] According to preset rules, multiple historical data groups are acquired to form a historical pre-diagnosis strategy information set; each of the historical data groups includes warning level information, sensor data, warning area information, diagnosis strategy information, and processing feedback information.

[0170] The preset rules include at least one of the following:

[0171] Multiple historical data sets within a predetermined time range from the current time are used as a historical pre-diagnosis strategy information set;

[0172] Multiple historical data sets that are at the same level as the current early warning level of the dam are used as a historical pre-diagnosis strategy information set.

[0173] Each of the historical data groups includes warning level information, sensor data, warning area information, diagnostic strategy information, and processing feedback information; the preset rules include at least one of the following:

[0174] Multiple historical data sets within a predetermined time range from the current time are used as a historical pre-diagnosis strategy information set;

[0175] Multiple historical data sets that are at the same level as the current early warning level of the dam are used as a historical pre-diagnosis strategy information set.

[0176] System Implementation Example.

[0177] See Figure 4 , Figure 4 This is a structural block diagram of a fire protection system provided in an embodiment of this application.

[0178] This embodiment provides a fire protection system for a dam, which includes an information acquisition module, an edge server, a cloud server, and a communication module.

[0179] The information acquisition module is used to acquire sensor information in real time, including sensor data and area identifiers for different areas of the dam.

[0180] The edge server receives sensor information acquired by the information acquisition module and, for each region, detects whether the parameters of each sensor data within that region are within their corresponding preset range. When at least N sensor data within a region are detected to be outside their corresponding preset range, the region is designated as an early warning unit. One or more acquired early warning units are designated as early warning regions. Based on the sensor information and the local historical information corresponding to the early warning regions, the early warning level and first pre-diagnosis strategy information of the dam are obtained. When the early warning level is lower than a preset early warning level, target pre-diagnosis strategy information is obtained according to the first pre-diagnosis strategy information. When the early warning level is not lower than the preset early warning level, the sensor information is designated as first status information, and the first pre-diagnosis strategy information, the first status information, and the semantic information corresponding to the early warning region are sent to the cloud server through the communication module.

[0181] The cloud server is used to obtain the second pre-diagnosis strategy information of the dam based on the semantic information and the first state information, and to generate target pre-diagnosis strategy information based on the first pre-diagnosis strategy information and the second pre-diagnosis strategy information.

[0182] The target pre-diagnosis strategy information is used to send to the on-site user equipment, enabling the user to perform fault pre-diagnosis of the dam's early warning area based on the target pre-diagnosis strategy information.

[0183] In some optional implementations, the cloud server is used to receive and obtain second pre-diagnosis strategy information for the dam based on the semantic information and the first state information, and to generate target pre-diagnosis strategy information based on the first and second pre-diagnosis strategy information, including:

[0184] The cloud server obtains the target pre-diagnosis strategy information based on the first pre-diagnosis strategy information and the historical pre-diagnosis strategy information set;

[0185] Based on the semantic information and the first state information, the second pre-diagnosis strategy information of the dam is obtained;

[0186] Based on the second pre-diagnosis strategy information, obtain the credibility of the first pre-diagnosis strategy information;

[0187] When the confidence level is obtained within a preset time period and the confidence level is lower than the preset confidence level, the target pre-diagnosis strategy information is updated according to the second pre-diagnosis strategy information and sent to the on-site user equipment.

[0188] When the credibility is obtained within a preset time period and the credibility is not lower than the preset credibility, the target pre-diagnosis strategy information obtained according to the first pre-diagnosis strategy information and the historical pre-diagnosis strategy information set will be sent to the on-site user equipment.

[0189] If the confidence level is not obtained within the preset time period, the target pre-diagnosis strategy information obtained based on the first pre-diagnosis strategy information and the historical pre-diagnosis strategy information set will be sent to the on-site user equipment.

[0190] In some optional implementations, obtaining the second pre-diagnosis strategy information for the dam based on the semantic information and the first state information includes:

[0191] Based on the semantic information, the danger index of the dam is obtained;

[0192] Based on the risk index and the first state information, the second pre-diagnosis strategy information is obtained.

[0193] In some optional implementations, the historical pre-diagnosis strategy information set is obtained in the following ways:

[0194] According to preset rules, multiple historical data groups are acquired to form a historical pre-diagnosis strategy information set; each of the historical data groups includes warning level information, sensor data, warning area information, diagnosis strategy information, and processing feedback information.

[0195] The preset rules include at least one of the following:

[0196] Multiple historical data sets within a predetermined time range from the current time are used as a historical pre-diagnosis strategy information set;

[0197] Multiple historical data sets that are at the same level as the current early warning level of the dam are used as a historical pre-diagnosis strategy information set.

[0198] In some optional implementations, obtaining the target pre-diagnosis strategy information based on the first pre-diagnosis strategy information and the historical pre-diagnosis strategy information set stored on the cloud server includes:

[0199] Obtain the diagnostic strategy information that has the highest similarity to the first pre-diagnosis strategy information from the historical pre-diagnosis strategy information set, and use it as the strategy information to be selected;

[0200] When the similarity between the strategy information to be selected and the first pre-diagnosis strategy information is greater than a preset similarity, the strategy information to be selected is taken as the target pre-diagnosis strategy information;

[0201] When the similarity between the selected strategy information and the first pre-diagnosis strategy information is not greater than a preset similarity, the first pre-diagnosis strategy information is used as the target pre-diagnosis strategy information.

[0202] In some optional implementations, each of the historical data groups includes warning level information, sensor data, warning area information, diagnostic strategy information, and processing feedback information; the preset rules include at least one of the following:

[0203] Multiple historical data sets within a predetermined time range from the current time are used as a historical pre-diagnosis strategy information set;

[0204] Multiple historical data sets that are at the same level as the current early warning level of the dam are used as a historical pre-diagnosis strategy information set.

[0205] In some optional implementations, obtaining the second pre-diagnosis strategy information for the dam based on the semantic information and the first state information includes:

[0206] Based on the semantic information, the danger index of the dam is obtained;

[0207] Based on the risk index and the first state information, the second pre-diagnosis strategy information is obtained.

[0208] In some optional implementations, obtaining the second pre-diagnosis strategy information based on the hazard index and the first state information includes:

[0209] The first state information is used as input, and the predicted policy information is obtained through the policy generation model.

[0210] The risk index is used as a correction factor to adjust the prediction strategy information, and the adjusted prediction strategy information is used as the second pre-diagnosis strategy information.

[0211] Storage medium example.

[0212] This embodiment provides a computer-readable storage medium having instructions stored thereon, which, when executed by a processor, cause the processor to implement any of the methods described in the method embodiment.

[0213] Example of a program product.

[0214] See Figure 5 , Figure 5 This is a schematic diagram of the structure of a computer program product provided in an embodiment of this application.

[0215] This embodiment provides a computer program product, which includes a computer program that, when executed by at least one processor, implements the steps of any of the methods described in the method embodiment.

[0216] The computer program product may be in the form of a portable compact disc read-only memory (CD-ROM) and include program code, and may run on a terminal device, such as a personal computer. However, the computer program product of this application is not limited thereto, and the computer program product may be in any combination of one or more computer-readable media.

[0217] In the various embodiments described in this application, the sequence number of each process does not imply the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of this application.

[0218] The terms “first,” “second,” “third,” etc. (if present) in the specification, claims, and accompanying drawings of this application are used to distinguish similar objects and are not necessarily used to describe a particular order or sequence. It should be understood that such data can be interchanged where appropriate so that embodiments of the application described herein can be implemented, for example, in orders other than those illustrated or described herein. Furthermore, the terms “comprising” and “corresponding to,” and any variations thereof, are intended to cover a non-exclusive inclusion; for example, a process, method, system, product, or apparatus that comprises a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or apparatus.

[0219] This application describes the invention from the perspectives of purpose, performance, progress, and novelty, and it meets the functional enhancement and use requirements emphasized by the Patent Law. The above description and drawings are merely preferred embodiments of this application and are not intended to limit this application. Therefore, all structures, devices, features, etc., that are similar to or identical to those of this application, i.e., all equivalent substitutions or modifications made in accordance with the scope of this patent application, shall fall within the scope of protection of this patent application.

Claims

1. A dam fire early warning method based on big data, characterized in that, The dam fire early warning method is used in the dam's fire protection system, which includes a cloud server, an edge server, an information acquisition module, a communication module, and on-site user equipment. The aforementioned dam fire early warning methods include: The information acquisition module is used to acquire sensor information in real time, including sensor data and area identifiers for different areas of the dam. For each region, the edge server is used to detect whether the parameters of each sensor data in the region are within its corresponding preset range; When at least N sensor data points within the area are detected to be outside their corresponding preset range, the area is designated as an early warning unit; N is a positive integer not less than 3. One or more early warning units are used as early warning areas. Based on the sensor information and the local historical information corresponding to the early warning areas, the early warning level and the first pre-diagnosis strategy information of the dam are obtained. When the warning level is lower than the preset warning level, the target pre-diagnosis strategy information is obtained according to the first pre-diagnosis strategy information; When the warning level is not lower than the preset warning level, the sensing information is used as the first state information, and the first pre-diagnosis strategy information, the first state information and the semantic information corresponding to the warning area are sent to the cloud server through the communication module. Based on the semantic information and the first state information, the second pre-diagnosis strategy information of the dam is obtained using the cloud server, and target pre-diagnosis strategy information is generated based on the first pre-diagnosis strategy information and the second pre-diagnosis strategy information. The target pre-diagnosis strategy information is used to send to the on-site user equipment, so that the user can perform fault pre-diagnosis of the early warning area of ​​the dam according to the target pre-diagnosis strategy information; the semantic information is obtained by processing the inspection information recorded by the patrol personnel during the inspection of the early warning area, and the inspection information includes the operating status, maintenance status, maintenance history, fault repair records and on-site visual status of fire-fighting equipment.

2. The dam fire early warning method according to claim 1, characterized in that, The step of obtaining second pre-diagnosis strategy information for the dam using the cloud server based on the semantic information and the first state information, and generating target pre-diagnosis strategy information based on the first and second pre-diagnosis strategy information, includes: Based on the first pre-diagnosis strategy information and the historical pre-diagnosis strategy information set stored on the cloud server, the target pre-diagnosis strategy information is obtained; Based on the semantic information and the first state information, the second pre-diagnosis strategy information of the dam is obtained; Based on the second pre-diagnosis strategy information, obtain the credibility of the first pre-diagnosis strategy information; When the confidence level is obtained within a preset time period and the confidence level is lower than the preset confidence level, the target pre-diagnosis strategy information is updated according to the second pre-diagnosis strategy information and sent to the on-site user equipment. When the credibility is obtained within a preset time period and the credibility is not lower than the preset credibility, the target pre-diagnosis strategy information obtained according to the first pre-diagnosis strategy information and the historical pre-diagnosis strategy information set will be sent to the on-site user equipment. If the confidence level is not obtained within the preset time period, the target pre-diagnosis strategy information obtained based on the first pre-diagnosis strategy information and the historical pre-diagnosis strategy information set will be sent to the on-site user equipment.

3. The dam fire early warning method according to claim 2, characterized in that, The methods for obtaining the historical pre-diagnosis strategy information set include: According to preset rules, multiple historical data groups are acquired to form a historical pre-diagnosis strategy information set; each of the historical data groups includes warning level information, sensor data, warning area information, diagnosis strategy information, and processing feedback information. The preset rules include at least one of the following: Multiple historical data sets within a predetermined time range from the current time are used as a historical pre-diagnosis strategy information set; Multiple historical data sets that are at the same level as the current early warning level of the dam are used as a historical pre-diagnosis strategy information set.

4. The dam fire early warning method according to claim 3, characterized in that, The step of obtaining target pre-diagnosis strategy information based on the first pre-diagnosis strategy information and the historical pre-diagnosis strategy information set stored on the cloud server includes: Obtain the diagnostic strategy information that has the highest similarity to the first pre-diagnosis strategy information from the historical pre-diagnosis strategy information set, and use it as the strategy information to be selected; When the similarity between the strategy information to be selected and the first pre-diagnosis strategy information is greater than a preset similarity, the strategy information to be selected is taken as the target pre-diagnosis strategy information; When the similarity between the selected strategy information and the first pre-diagnosis strategy information is not greater than a preset similarity, the first pre-diagnosis strategy information is used as the target pre-diagnosis strategy information.

5. The dam fire early warning method according to claim 2, characterized in that, The step of obtaining the second pre-diagnosis strategy information for the dam based on the semantic information and the first state information includes: Based on the semantic information, the danger index of the dam is obtained; Based on the risk index and the first state information, the second pre-diagnosis strategy information is obtained.

6. The dam fire early warning method according to claim 5, characterized in that, The step of obtaining the second pre-diagnosis strategy information based on the risk index and the first state information includes: The first state information is used as input, and the predicted policy information is obtained through the policy generation model. The risk index is used as a correction factor to adjust the prediction strategy information, and the adjusted prediction strategy information is used as the second pre-diagnosis strategy information.

7. A fire protection system for a dam, characterized in that, The fire protection system includes an information acquisition module, an edge server, a cloud server, and a communication module; The information acquisition module is used to acquire sensor information in real time, including sensor data and area identifiers for different areas of the dam. The edge server is used to receive the sensor information acquired by the information acquisition module, and for each area, detect whether the parameters of each sensor data in the area are within its corresponding preset range; when at least N sensor data in the area are detected to be outside their corresponding preset range, the area is used as an early warning unit. N is a positive integer not less than 3; One or more early warning units are used as early warning areas. Based on the sensor information and the local historical information corresponding to the early warning areas, the early warning level and the first pre-diagnosis strategy information of the dam are obtained. When the warning level is lower than the preset warning level, the target pre-diagnosis strategy information is obtained according to the first pre-diagnosis strategy information; When the warning level is not lower than the preset warning level, the sensor information is used as the first state information, and the first pre-diagnosis strategy information, the first state information and the semantic information corresponding to the warning area are sent to the cloud server through the communication module. The cloud server is used to receive and obtain the second pre-diagnosis strategy information of the dam based on the semantic information and the first state information, and generate target pre-diagnosis strategy information based on the first pre-diagnosis strategy information and the second pre-diagnosis strategy information. The target pre-diagnosis strategy information is used to send to on-site user equipment, enabling users to perform fault pre-diagnosis of the dam's early warning area based on the target pre-diagnosis strategy information.

8. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a computer program that, when executed by at least one processor, implements the steps of the method according to any one of claims 1-6.

9. A computer program product, characterized in that, The computer program product includes a computer program that, when executed by at least one processor, implements the steps of the method according to any one of claims 1-6.

Images