Safety early warning analysis method and early warning system for power transmission line overwater operation platform
By establishing atypical datasets and a bidirectional long short-term memory network model, the shortcomings of long-term operational status monitoring of waterborne operation platforms were addressed, enabling the prediction and early warning of safety status and improving construction safety.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-11-27
- Publication Date
- 2026-03-27
AI Technical Summary
Existing technologies for monitoring the safety status of waterborne operation platforms mainly rely on real-time data, lacking effective management and analysis of long-term operational stability, and failing to effectively extract non-data-related operational description text, resulting in a lack of effective support for changes in the long-term operational status of the platform.
An atypical dataset of safety status for power transmission line waterborne operation platforms was established. Through zero-value filling and sliding processing, a correlation data feature matrix was constructed. A bidirectional long short-term memory network model was used for safety status prediction and analysis. Safety early warning was then provided in conjunction with real-time data from the data acquisition terminal.
It enables effective monitoring and early warning of the long-term operational status of the water-based work platform, improves the predictability and stability of safety management, and enhances construction safety.
Smart Images
Figure CN121747285A_ABST
Abstract
Description
Technical Field
[0001] This invention belongs to the technical field of power transmission line construction operation management system, and particularly relates to a safety early warning analysis method and early warning system for power transmission line water operation platform. Background Technology
[0002] To avoid the high costs and environmental burdens associated with traditional methods such as dike construction, earth filling, and cofferdam building, overhead power transmission line projects in river networks, lakes, tidal flats, and economically important water areas can utilize standardized floating platform assemblies for material transport, foundation construction, and tower resistance monitoring. However, the dynamic changes in buoyancy, water flow, and waves, as well as dynamic loads from drilling, transportation, and hoisting, will exacerbate platform instability, affecting operational safety and quality. Therefore, monitoring the safety status of the floating platform is crucial.
[0003] Safety monitoring of floating platforms mainly relies on the analysis and processing of data from various sensors. Targeted monitoring is conducted based on preset anomaly factors. However, in the actual operation and use of floating platforms, many safety factors or signs are often mixed in with the specific construction process, including a large amount of non-data-based operation descriptions and engineering records. Further analysis of these non-data-based contents can further uncover the intrinsic factors and changing patterns of the safety status of floating platforms, providing a reference for the safety management of floating platforms. However, current detection data is mainly used for short-term, temporary monitoring of real-time operation status and does not provide effective support for the long-term stable operation of the platform. There is a lack of effective management and analysis of the changing status of the platform during long-term operation. Summary of the Invention
[0004] The purpose of this invention is to provide an early warning analysis method and early warning system that can further enrich the safety status monitoring and management measures for power transmission line waterway operation platforms, based on actual needs.
[0005] To achieve the above objectives, the present invention adopts the following technical solution.
[0006] A safety early warning method for power transmission line waterworks platforms includes:
[0007] Establish an atypical dataset of safety status for power transmission line waterborne operation platforms, and determine the analysis items and safety status types related to the safety status of power transmission line waterborne operation platforms;
[0008] The atypical data processing of the safety status of the power transmission line water operation platform includes zero-value filling and sliding processing to obtain the processed atypical data.
[0009] Based on the analysis items and safety status types related to the safety status of power transmission line waterborne operation platforms, the safety status level is determined. Through data decomposition, different related data items in the processed atypical data under different installation status levels are obtained. A data feature matrix of related data and a predictive analysis model of the safety status of power transmission line waterborne operation platforms based on bidirectional long short-term memory networks are established.
[0010] The processed atypical dataset is used to allocate training and test sets for training until the model error meets the requirements. The safety status data sequence features within a period of time before the current time point are used as input to train the model to obtain the safety status prediction results of the power transmission line waterway operation platform.
[0011] Based on the predicted security status level, a corresponding security warning is issued, and corresponding security measures are taken according to the preset platform security handling rules.
[0012] Further improvements or specific implementations of the aforementioned safety early warning method for power transmission line waterborne operation platforms include establishing an atypical dataset of safety status for power transmission line waterborne operation platforms, including:
[0013] Based on the safety concerns of power transmission line waterworks platforms, identify the relevant analysis items for the safety status of power transmission line waterworks platforms; based on the safety matters involved in the relevant analysis items, determine the type of safety status that needs to be analyzed for each analysis item;
[0014] The aforementioned safety status types are evaluated based on the safety status evaluation index of the power transmission line waterway operation platform. The safety status level is divided based on the evaluation results and the actual safety status of different analysis items. The safety status level is used to characterize the urgency of the safety warning.
[0015] By setting up a data acquisition terminal on the power transmission line waterway operation platform, atypical data corresponding to the state type of each analysis item under different safety status levels are collected to establish an atypical dataset of the installation status of the power transmission line waterway operation platform. The atypical dataset is established separately according to the state type of each analysis item.
[0016] Further improvements or specific implementations of the aforementioned safety early warning method for power transmission line waterborne operation platforms, including the processing of atypical data on the safety status of power transmission line waterborne operation platforms, specifically include:
[0017] Based on the data type corresponding to the atypical dataset of safety status of power transmission line waterway operation platform, daily datasets of historical safety status of power transmission line waterway operation platform are extracted from the original data of safety project operations of power transmission line waterway operation platform, and zero-value filling and sliding processing are performed.
[0018] Zero-value filling refers to filling the daily dataset of historical safety status data for power transmission line waterway operation platforms with 0 data for the normal status, i.e., the daily dataset that does not contain any platform safety status level.
[0019] Sliding processing refers to using a fixed-length time window to slide from the beginning to the end of each atypical dataset of the installation status of power transmission line waterworks platforms. The time window stores the latest T historical data of the safety status of power transmission line waterworks platforms. During the data update process, the window queue is continuously updated with the latest data, while discarding the data at the first position of the time window. The data sequence of the safety status of power transmission line waterworks platforms within the time window is used as the basic data of the safety status of power transmission line waterworks platforms.
[0020] Further improvements or specific implementations of the aforementioned safety early warning method for power transmission line waterworks platforms involve establishing a data feature matrix of associated data, implemented in the following manner:
[0021] For the atypical dataset of the installation status of power transmission line waterborne operation platforms, data decomposition is used to obtain different key data items in the atypical data, as well as all related data items containing the key data items. A related dataset is then constructed by sorting the related data items according to their analytical weights and the number of key data items in the related data from least to most. ,in Indicates the first The first atypical chain dataset Atypical related data; for data containing the same number of related data, sort them according to the sum of the analysis weights of all key data items in the related data;
[0022] Calculate related data correlation Its meaning is related data. The previous atypical association dataset Frequency of occurrence , Refers to related data In the previous associated dataset The number of times it appears in Refers to associated datasets Total number of related data items;
[0023] Calculate related data support level Its meaning is related data. In terms of support Related data Support of all key data items ratio ,in , Refers to the data that constitutes the association The Support for key data items Refers to the data that constitutes the association The total number of key data items;
[0024] Calculate related data urgency Its meaning refers to related data. Security status levels of atypical datasets in the installation status of the operating platform. The number of times it appears The total number of times this security status level occurs ratio ;
[0025] Establish related data based on the aforementioned indicators. Data feature matrix for ;
[0026] Further improvements or specific implementations of the aforementioned safety early warning method for power transmission line waterborne operation platforms, including the establishment of a safety status prediction and analysis model for power transmission line waterborne operation platforms based on a bidirectional long short-term memory network, refer to: establishing a prediction and analysis model composed of a forward LSTM network module, a backward LSTM network module, and an omnidirectional connection module. The omnidirectional connection module includes three fully connected layers. Activation function, and adopt The function discards some nodes from the first two levels;
[0027] The forward LSTM module acquires the current safety status data of the power transmission line waterway operation platform in chronological order for forward analysis, while the backward LSTM network module performs reverse analysis from the current safety status of the power transmission line waterway operation platform.
[0028] The forward LSTM network module takes the feature matrix of the security state data sequence at the current sampling point or the feature matrix of the security state data sequence before the current sampling point as its initial input, while the backward LSTM network module takes the feature matrix of the security state data sequence after the current sampling point as its initial input and outputs the number of times each security state level occurs in future periods. The network optimizer performs optimization training; the input to the activation function is... The sample output predicted value is Sample output error ;in For the first The weights of each sample, No. The tensor of the output is calculated for each sample. This is the actual value.
[0029] In a further improvement or specific implementation of the aforementioned safety early warning method for power transmission line waterborne operation platforms, the analysis items are formulated based on the structure or function of the power transmission line waterborne operation platform. Each analysis item operates as an independent entity and does not directly affect the safety status of the others. The status type refers to the status type that directly affects the safe operation of each analysis item, and refers to the cause that leads to safety hazards in each analysis item.
[0030] This application also provides a safety early warning system for power transmission line waterworks platforms, including:
[0031] A module is established to create an atypical dataset of safety status for power transmission line waterborne operation platforms, determine the analysis items and safety status types related to the safety status of power transmission line waterborne operation platforms, and process the atypical data of safety status for power transmission line waterborne operation platforms, including zero-value filling and sliding processing, to obtain the processed atypical data.
[0032] The determination module is used to determine the safety status level based on the analysis items and safety status types related to the safety status of the power transmission line waterway operation platform. Through data decomposition, different related data items in the processed atypical data under different installation status levels are obtained. A data feature matrix of the related data and a safety status prediction and analysis model of the power transmission line waterway operation platform based on a bidirectional long short-term memory network are established. The processed atypical dataset is used to allocate training and test sets for training until the model error meets the requirements. The safety status data sequence features within a period of time before the current time point are used as the input to train the model to obtain the safety status prediction results of the power transmission line waterway operation platform.
[0033] The execution module is used to issue corresponding security warnings based on the security status level corresponding to the prediction results, and to carry out corresponding security actions according to the preset platform security handling rules.
[0034] In a further improved or preferred embodiment of the aforementioned safety early warning system for power transmission line waterborne operation platforms, the establishment module is specifically used to determine the analysis items related to the safety status of the power transmission line waterborne operation platform based on the safety status concern targets; determine the safety status type that needs to be analyzed for each analysis item based on the safety matters involved in the relevant analysis items; evaluate the aforementioned safety status types according to the safety status evaluation indicators of the power transmission line waterborne operation platform, and classify the safety status level based on the evaluation results and the actual safety status of different analysis items; the safety status level is used to characterize the urgency of the safety early warning; and collect atypical data corresponding to the status type of each analysis item under different safety status levels by setting up a data acquisition terminal on the power transmission line waterborne operation platform, and establish an atypical dataset of the installation status of the power transmission line waterborne operation platform, wherein the atypical dataset is established separately according to the status type of each analysis item.
[0035] In a further improved or preferred implementation of the aforementioned safety early warning system for power transmission line waterway operation platforms, the establishment module is specifically used to extract daily datasets of historical safety status data for power transmission line waterway operation platforms from the original data of safety project operations, based on the data type corresponding to the atypical dataset of safety status of power transmission line waterway operation platforms, and to perform zero-value filling and sliding processing. Zero-value filling means that for the daily datasets of historical safety status data of power transmission line waterway operation platforms, the normal state, i.e., the daily datasets that do not contain any platform safety status level, are filled with 0 data. Sliding processing means that for each atypical dataset of installation status of power transmission line waterway operation platforms, a fixed-length time window is used to slide from the beginning to the end of the sequence. The time window stores the latest T historical data of the historical safety status of power transmission line waterway operation platforms. During the data update process, the window queue is continuously updated with the latest data, while discarding the data at the first position of the time window. The sequence of safety status data of power transmission line waterway operation platforms within the time window is used as the basic data of safety status of power transmission line waterway operation platforms.
[0036] In a further improvement or preferred embodiment of the aforementioned safety early warning system for power transmission line waterborne operation platforms, the establishment module establishes a data feature matrix of associated data, based on the following method:
[0037] For the atypical dataset of the installation status of power transmission line waterborne operation platforms, data decomposition is used to obtain different key data items in the atypical data, as well as all related data items containing the key data items. A related dataset is then constructed by sorting the related data items according to their analytical weights and the number of key data items in the related data from least to most. ,in Indicates the first The first atypical chain dataset Atypical related data; for data containing the same number of related data, sort them according to the sum of the analysis weights of all key data items in the related data;
[0038] Calculate related data correlation Its meaning is related data. The previous atypical association dataset Frequency of occurrence , Refers to related data In the previous associated dataset The number of times it appears in Refers to associated datasets Total number of related data items;
[0039] Calculate related data support level Its meaning is related data. In terms of support Related data Support of all key data items ratio ,in , Refers to the data that constitutes the association The Support for key data items Refers to the data that constitutes the association The total number of key data items;
[0040] Calculate related data urgency Its meaning refers to related data. Security status levels of atypical datasets in the installation status of the operating platform. The number of times it appears The total number of times this security status level occurs ratio ;
[0041] Establish related data based on the aforementioned indicators. Data feature matrix for . Attached Figure Description
[0042] Figure 1 This is a schematic diagram of a power transmission line water-based operation platform;
[0043] Figure 2 This is a schematic diagram of the data acquisition system for the safety early warning system of the power transmission line water operation platform. Detailed Implementation
[0044] The present invention will be described in detail below with reference to specific embodiments.
[0045] The main objective of this invention is to provide a method and system for improving the safety status monitoring and early warning of power transmission line construction platforms on water. This method is used to extract characteristic data representing changes in safety status from non-data files such as process documents and management documents related to safety status during the operation of the water platform. Based on traditional monitoring data, this method further improves the dimensions of power transmission line safety status monitoring and early warning, uncovers hidden status elements, and provides a guarantee for further improving the safety of power transmission line construction on water.
[0046] This application discloses a safety early warning method for a power transmission line waterborne operation platform, which mainly includes the following steps:
[0047] Step S1: Establish an atypical dataset of safety status for power transmission line waterborne operation platforms; specifically including:
[0048] A1. Based on the safety status concerns of power transmission line waterworks platforms, determine the analysis items related to the safety status of power transmission line waterworks platforms; based on the safety matters involved in different analysis items, determine the safety status items that need to be analyzed for each analysis item;
[0049] like Figure 1 As shown, the analysis items of the power transmission line waterborne operation platform are formulated according to the structure or function of the power transmission line waterborne operation platform. Each analysis item is an independent entity during operation and does not directly affect the safety status of each other. In this application, the analysis items of the power transmission line waterborne operation platform mainly include: the main platform of the power transmission line waterborne operation, the platform connecting section, the mobile transfer platform, and the platform operation equipment.
[0050] Among them, the state type refers to the state type that directly affects the safe operation of each analysis item. It usually refers to the cause that leads to the safety hazards of each analysis item. In this application, it mainly includes the breakage of the connecting pin between the pile foundation and the platform, the overall tilting of several pile foundations, the instability of the hydraulic system, and the large-scale shaking under the coupling effect of wind, waves and current.
[0051] A2. Determine the safety status level corresponding to different analysis items based on the safety status evaluation index of the power transmission line waterway operation platform; the safety status level is used to characterize the urgency of safety warnings, and is classified according to the actual safety status of different analysis items.
[0052] The atypical data targeted in this application does not have typical source atypical data that can be clearly compared and analyzed directly. It is usually atypical data on safety status that is related to safety status but does not have clear indicators or parameters to describe and define it. This includes, but is not limited to, comprehensive descriptive texts of various types of atypical data. Since comprehensive descriptive texts are usually only used to describe a certain stage or to comprehensively describe the details of safety status, in actual analysis, it is usually necessary to combine the descriptions and contents of multiple atypical data for comprehensive judgment. Each atypical data can usually only represent the trend of the safety status corresponding to the atypical data, but cannot be directly determined. In order to facilitate the determination of the safety status of atypical data, combined with actual operational needs, a fuzzy non-indicator safety status level is also adopted. The safety status levels in this application include: general emergency, relatively emergency, emergency, and extremely emergency.
[0053] A3. By setting up a data acquisition terminal on the power transmission line waterway operation platform, collect the detection atypical data corresponding to the state type of each analysis item, and establish an atypical dataset of the safety status of the power transmission line waterway operation platform. The atypical dataset is established according to the state type of each analysis item, and each atypical dataset is randomly allocated to the training set and the test set according to a preset ratio.
[0054] To achieve automated analysis of safety early warning for power transmission line waterborne operation platforms based on atypical data from PC devices, this application adopts a machine learning approach to learn and process the features of atypical data within atypical data. To this end, it is necessary to combine the aforementioned foundation and train and maintain the machine learning algorithm using atypical data from the operation of power transmission line waterborne operation platforms.
[0055] Currently, the construction and management of power transmission line waterway operation platforms have entered a stable mode. The patterns of platform safety status and the characteristic manifestations of their occurrence are similar. In particular, during the processing of atypical data, the characteristic change trend analysis of platform safety status data based on historical cycles can effectively determine the safety status of power transmission line waterway operation platforms. These atypical data characteristics that match historical data can be simulated by establishing a long short-term memory network, thereby establishing a predictive analysis model for the safety status of power transmission line waterway operation platforms.
[0056] When performing the above processing, it may be found that not all safety status events of the operating platform will occur within the current historical data period, or the frequency of safety status events of some operating platforms may be low. This results in a large amount of scattered data in the original data, which cannot be effectively utilized and leads to a decrease in the computational efficiency of the predictive analysis model. Therefore, necessary data processing is required.
[0057] Specifically, step S2 involves processing atypical data on the safety status of the power transmission line's waterborne operation platform.
[0058] Zero-value padding: For the daily dataset of historical safety status data of power transmission line water operation platforms, the daily dataset of normal status, that is, the daily dataset that does not contain any platform safety status level, is filled with 0 data.
[0059] The sliding processing method is used for each atypical dataset of the safety status of power transmission line waterworks platforms. A fixed-length time window slides from the beginning to the end of the sequence. The time window stores the latest T historical data of the safety status of power transmission line waterworks platforms. During the data update process, the window queue is continuously updated with the latest data, while the data at the first position of the time window is discarded. The data sequence of the safety status of power transmission line waterworks platforms within the time window is used as the basic data of the safety status of power transmission line waterworks platforms.
[0060] Step S3: Construct a feature matrix based on atypical data
[0061] For the atypical dataset of safety status of power transmission line waterborne operation platforms, data decomposition is used to obtain different key data items in the atypical data, as well as all related data items containing the key data items. The related dataset is then constructed by sorting the related data items according to their analytical weights and the number of key data items in the related data from least to most. ,in Indicates the first The first atypical chain dataset Atypical related data; for data containing the same number of related data, the data is sorted according to the sum of the analytical weights of all key data items in the related data.
[0062] Calculate related data correlation Its meaning is related data. The previous atypical association dataset Frequency of occurrence , Refers to related data In the previous associated dataset The number of times it appears in Refers to associated datasets Total number of related data items;
[0063] Calculate related data support level Its meaning is related data. In terms of support Related data Support of all key data items ratio ,in , Refers to the data that constitutes the association The Support for key data items Refers to the data that constitutes the association The total number of key data items;
[0064] Calculate related data urgency Its meaning refers to related data. Each security status level in the atypical dataset of work platform security status The number of times it appears The total number of times this security status level occurs ratio ;
[0065] Establish related data based on the aforementioned indicators. Data feature matrix for ;
[0066] A safety status prediction and analysis model for power transmission line waterborne operation platforms based on bidirectional long short-term memory networks is established, such as... Figure 2 As shown, the model consists of a forward LSTM network module, a backward LSTM network module, and an omnidirectional connection module. The omnidirectional connection module includes three fully connected layers and employs... Activation function, and adopt The function discards some nodes from the first two levels;
[0067] The forward LSTM module acquires the current safety status data of the power transmission line's waterborne operation platform in chronological order for forward analysis, while the backward LSTM network module performs reverse analysis starting from the current safety status of the power transmission line's waterborne operation platform. The forward LSTM network module uses the feature matrix of the safety status data sequence at the current sampling point or the safety status data sequence feature matrix before the current sampling point as its initial input, while the backward LSTM network module uses the feature matrix of the safety status data sequence feature matrix after the current sampling point as its initial input. For the raw safety status data sequence feature matrix, it is first normalized, and the output uses the frequency of occurrence of each safety status level within future periods. The network optimizer performs optimization training; the input to the activation function is... The sample output predicted value is Sample output error ;in For the first The weights of each sample, No. The tensor of the output is calculated for each sample. The actual value;
[0068] The safety status prediction and analysis model of the power transmission line waterway operation platform based on the bidirectional long short-term memory network was trained using the training and test sets allocated from the atypical dataset of the safety status of the power transmission line waterway operation platform until the model error met the requirements.
[0069] Using the characteristics of the safety status data sequence over a period of time prior to the current point in time as input, the frequency of occurrence of each safety status level in future periods is predicted to obtain the safety status prediction results of the power transmission line waterway operation platform.
[0070] To meet the above requirements, this application also provides a safety early warning system for power transmission line water operation platforms, including a hardware layer consisting of a cloud server, a central control terminal, and a data acquisition terminal;
[0071] This application also provides a safety early warning system for power transmission line waterworks platforms, including:
[0072] A module is established to create an atypical dataset of safety status for power transmission line waterborne operation platforms, determine the analysis items and safety status types related to the safety status of power transmission line waterborne operation platforms, and process the atypical data of safety status for power transmission line waterborne operation platforms, including zero-value filling and sliding processing, to obtain the processed atypical data.
[0073] The module is specifically used to determine the relevant analysis items for the safety status of the power transmission line waterborne operation platform based on the safety concerns of the platform; determine the safety status type that needs to be analyzed for each analysis item based on the safety matters involved in the relevant analysis items; evaluate the aforementioned safety status types according to the safety status evaluation indicators of the power transmission line waterborne operation platform; and classify the safety status level based on the evaluation results and the actual safety status of different analysis items. The safety status level is used to characterize the urgency of safety warnings. By setting up a data acquisition terminal on the power transmission line waterborne operation platform, atypical data corresponding to the status type of each analysis item under different safety status levels are collected to establish an atypical dataset of the installation status of the power transmission line waterborne operation platform. The atypical dataset is established separately according to the status type of each analysis item.
[0074] like Figure 2 As shown, the data acquisition terminal includes:
[0075] UWB personnel positioning base stations, including those deployed on water work platforms and at the corners of platform passageways;
[0076] RTK positioning mobile stations, including those deployed on water-based work platforms and transfer platforms;
[0077] Accelerometer acquisition components include accelerometers deployed on the platform underframe or platform pontoon that are connected to the pile foundation of the floating platform;
[0078] Inclination acquisition components include inclination sensors deployed on the platform underframe or platform pontoons in the working area and edge area of the water-based work platform;
[0079] Temperature, humidity and wind direction and speed acquisition components, including temperature, humidity and wind direction and speed sensors deployed on the edge or frame of the water work platform;
[0080] Image acquisition components, including those installed in the working area of the waterborne operation platform;
[0081] The strain acquisition component includes a strain data acquisition device installed in the connection area of the floating work platform;
[0082] The determination module is used to determine the safety status level based on the analysis items and safety status types related to the safety status of power transmission line waterway operation platforms. Through data decomposition, it obtains different related data items in the processed atypical data under different installation status levels, establishes a data feature matrix of the related data, and a safety status prediction and analysis model for power transmission line waterway operation platforms based on a bidirectional long short-term memory network. The processed atypical dataset is used to allocate training and testing sets for training until the model error meets the requirements. The safety status data sequence characteristics within a period prior to the current time point are used as input to the training model to obtain the safety status prediction results for power transmission line waterway operation platforms. Specifically, it is used to extract data from the original data of power transmission line waterway operation platform safety projects based on the data type corresponding to the atypical dataset of power transmission line waterway operation platform safety status. The daily datasets of historical safety status data for power transmission line waterway operation platforms are collected and subjected to zero-value filling and sliding processing. Zero-value filling means that for the daily datasets of historical safety status data for power transmission line waterway operation platforms, the normal state (i.e., the daily datasets that do not contain any platform safety status level) are filled with 0 data. Sliding processing means that for each atypical dataset of installation status of power transmission line waterway operation platforms, a fixed-length time window is used to slide from the beginning to the end of the sequence. The time window stores the latest T historical data of the safety status of power transmission line waterway operation platforms. During the data update process, the window queue is continuously updated with the latest data, while the data at the first position of the time window is discarded. The data sequence of safety status of power transmission line waterway operation platforms within the time window is used as the basic data for the safety status of power transmission line waterway operation platforms.
[0083] A data feature matrix of the related data is established based on the following method:
[0084] For the atypical dataset of the installation status of power transmission line waterborne operation platforms, data decomposition is used to obtain different key data items in the atypical data, as well as all related data items containing the key data items. A related dataset is then constructed by sorting the related data items according to their analytical weights and the number of key data items in the related data from least to most. ,in Indicates the first The first atypical chain dataset Atypical related data; for data containing the same number of related data, sort them according to the sum of the analysis weights of all key data items in the related data;
[0085] Calculate related data correlation Its meaning is related data. The previous atypical association dataset Frequency of occurrence , Refers to related data In the previous associated dataset The number of times it appears in Refers to associated datasets Total number of related data items;
[0086] Calculate related data support level Its meaning is related data. In terms of support Related data Support of all key data items ratio ,in , Refers to the data that constitutes the association The Support for key data items Refers to the data that constitutes the association The total number of key data items;
[0087] Calculate related data urgency Its meaning refers to related data. Security status levels of atypical datasets in the installation status of the operating platform. The number of times it appears The total number of times this security status level occurs ratio ;
[0088] Establish related data based on the aforementioned indicators. Data feature matrix for .
[0089] It also includes an execution module, which issues corresponding security warnings based on the security status level corresponding to the prediction results, and performs corresponding security actions according to the preset platform security handling rules.
[0090] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and are not intended to limit the scope of protection of the present invention. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the essence and scope of the technical solutions of the present invention.
Claims
1. A safety early warning method for a power transmission line waterborne operation platform, characterized in that, include: Establish an atypical dataset of safety status for power transmission line waterborne operation platforms, and determine the analysis items and safety status types related to the safety status of power transmission line waterborne operation platforms; The atypical data processing of the safety status of the power transmission line water operation platform includes zero-value filling and sliding processing to obtain the processed atypical data. Based on the analysis items and safety status types related to the safety status of power transmission line waterborne operation platforms, the safety status level is determined. Through data decomposition, different related data items in the processed atypical data under different installation status levels are obtained. A data feature matrix of related data and a predictive analysis model of the safety status of power transmission line waterborne operation platforms based on bidirectional long short-term memory networks are established. The processed atypical dataset is used to allocate training and test sets for training until the model error meets the requirements. The safety status data sequence features within a period of time before the current time point are used as input to train the model to obtain the safety status prediction results of the power transmission line waterway operation platform. Based on the predicted security status level, a corresponding security warning is issued, and corresponding security measures are taken according to the preset platform security handling rules.
2. The safety early warning method for power transmission line waterborne operation platforms according to claim 1, characterized in that, An atypical dataset of safety status for power transmission line waterborne operation platforms was established, including: Based on the safety concerns of power transmission line waterworks platforms, identify the relevant analysis items for the safety status of power transmission line waterworks platforms; based on the safety matters involved in the relevant analysis items, determine the type of safety status that needs to be analyzed for each analysis item; The aforementioned safety status types are evaluated based on the safety status evaluation index of the power transmission line waterway operation platform. The safety status level is divided based on the evaluation results and the actual safety status of different analysis items. The safety status level is used to characterize the urgency of the safety warning. By setting up a data acquisition terminal on the power transmission line waterway operation platform, atypical data corresponding to the state type of each analysis item under different safety status levels are collected to establish an atypical dataset of the installation status of the power transmission line waterway operation platform. The atypical dataset is established separately according to the state type of each analysis item.
3. The safety early warning method for power transmission line waterborne operation platforms according to claim 2, characterized in that, Atypical data processing for the safety status of power transmission line waterborne operation platforms specifically includes: Based on the data type corresponding to the atypical dataset of safety status of power transmission line waterway operation platform, daily datasets of historical safety status of power transmission line waterway operation platform are extracted from the original data of safety project operations of power transmission line waterway operation platform, and zero-value filling and sliding processing are performed. Zero-value filling refers to filling the daily dataset of historical safety status data for power transmission line waterway operation platforms with 0 data for the normal status, i.e., the daily dataset that does not contain any platform safety status level. Sliding processing refers to using a fixed-length time window to slide from the beginning to the end of each atypical dataset of the installation status of power transmission line waterworks platforms. The time window stores the latest T historical data of the safety status of power transmission line waterworks platforms. During the data update process, the window queue is continuously updated with the latest data, while discarding the data at the first position of the time window. The data sequence of the safety status of power transmission line waterworks platforms within the time window is used as the basic data of the safety status of power transmission line waterworks platforms.
4. The safety early warning method for power transmission line waterborne operation platforms according to claim 2, characterized in that, A data feature matrix of the related data is established based on the following method: For the atypical dataset of the installation status of power transmission line waterborne operation platforms, data decomposition is used to obtain different key data items in the atypical data, as well as all related data items containing the key data items. A related dataset is then constructed by sorting the related data items according to their analytical weights and the number of key data items in the related data from least to most. ,in Indicates the first The first atypical chain dataset Atypical related data; for data containing the same number of related data, sort them according to the sum of the analysis weights of all key data items in the related data; Calculate related data correlation Its meaning is related data. The previous atypical association dataset Frequency of occurrence , Refers to related data In the previous associated dataset The number of times it appears in Refers to associated datasets Total number of related data items; Calculate related data support level Its meaning is related data. In terms of support Related data Support of all key data items ratio ,in , Refers to the data that constitutes the association The Support for key data items Refers to the data that constitutes the association The total number of key data items; Calculate related data urgency Its meaning refers to related data. Security status levels of atypical datasets in the installation status of the operating platform. The number of times it appears The total number of times this security status level occurs ratio ; Establish related data based on the aforementioned indicators. Data feature matrix for .
5. The safety early warning method for power transmission line waterborne operation platforms according to claim 2, characterized in that, Establishing a safety status prediction and analysis model for power transmission line waterborne operation platforms based on bidirectional long short-term memory networks refers to: establishing a prediction and analysis model composed of a forward LSTM network module, a backward LSTM network module, and an omnidirectional connection module. The omnidirectional connection module includes three fully connected layers and employs... Activation function, and adopt The function discards some nodes from the first two levels; The forward LSTM module acquires the current safety status data of the power transmission line waterway operation platform in chronological order for forward analysis, while the backward LSTM network module performs reverse analysis from the current safety status of the power transmission line waterway operation platform. The forward LSTM network module takes the feature matrix of the security state data sequence at the current sampling point or the feature matrix of the security state data sequence before the current sampling point as its initial input, while the backward LSTM network module takes the feature matrix of the security state data sequence after the current sampling point as its initial input and outputs the number of times each security state level occurs in future periods. The network optimizer performs optimization training; the input to the activation function is... ; Sample output predicted value Sample output error ;in For the first The weights of each sample, No. The tensor of the output is calculated for each sample. This is the actual value.
6. The safety early warning method for power transmission line waterborne operation platforms according to claim 1, characterized in that, The analysis items are formulated based on the structure or function of the power transmission line waterway operation platform. Each analysis item operates as an independent entity and does not directly affect the safety status of each other. The status type refers to the status type that directly affects the safe operation of each analysis item, and refers to the cause that leads to safety hazards in each analysis item.
7. A safety early warning system for a power transmission line waterborne operation platform, characterized in that, include: A module is established to create an atypical dataset of the safety status of power transmission line waterborne operation platforms, and to determine the analysis items and safety status types related to the safety status of power transmission line waterborne operation platforms. The atypical data processing of the safety status of the power transmission line water operation platform includes zero-value filling and sliding processing to obtain the processed atypical data. The determination module is used to determine the safety status level based on the analysis items and safety status types related to the safety status of the power transmission line waterway operation platform. Through data decomposition, different related data items in the processed atypical data under different installation status levels are obtained. A data feature matrix of the related data and a safety status prediction and analysis model of the power transmission line waterway operation platform based on a bidirectional long short-term memory network are established. The processed atypical dataset is used to allocate training and test sets for training until the model error meets the requirements. The safety status data sequence features within a period of time before the current time point are used as the input to train the model to obtain the safety status prediction results of the power transmission line waterway operation platform. The execution module is used to issue corresponding security warnings based on the security status level corresponding to the prediction results, and to carry out corresponding security actions according to the preset platform security handling rules.
8. The safety early warning system for power transmission line waterborne operation platforms according to claim 7, characterized in that, The establishment module is specifically used to determine the analysis items related to the safety status of the power transmission line waterway operation platform based on the safety status concern targets; determine the safety status type that needs to be analyzed for each analysis item based on the safety matters involved in the relevant analysis items; evaluate the aforementioned safety status types according to the safety status evaluation index of the power transmission line waterway operation platform; and classify the safety status level based on the evaluation results and the actual safety status of different analysis items. The security status level is used to characterize the urgency of the security warning; By setting up a data acquisition terminal on the power transmission line waterway operation platform, atypical data corresponding to the state type of each analysis item under different safety status levels are collected to establish an atypical dataset of the installation status of the power transmission line waterway operation platform. The atypical dataset is established separately according to the state type of each analysis item.
9. The safety early warning system for power transmission line waterborne operation platforms according to claim 8, characterized in that, The aforementioned module is specifically used to extract daily datasets of historical safety status data for power transmission line waterway operation platforms from the original data of safety project operations, based on the data type corresponding to the atypical dataset of safety status of power transmission line waterway operation platforms, and to perform zero-value filling and sliding processing. Zero-value filling means that for the daily datasets of historical safety status data of power transmission line waterway operation platforms, the normal state, i.e., the daily datasets that do not contain any platform safety status level, are filled with 0 data. Sliding processing means that for each atypical dataset of installation status of power transmission line waterway operation platforms, a fixed-length time window is used to slide from the beginning to the end of the sequence. The time window stores the latest T historical data of the historical safety status of power transmission line waterway operation platforms. During the data update process, the window queue is continuously updated with the latest data, while discarding the data at the first position of the time window. The safety status data sequence of the power transmission line waterway operation platform within the time window is used as the basic data for the safety status of the power transmission line waterway operation platform.
10. The safety early warning system for power transmission line waterborne operation platforms according to claim 9, characterized in that, The establishment module establishes a data feature matrix of the associated data, based on the following method: For the atypical dataset of the installation status of power transmission line waterborne operation platforms, data decomposition is used to obtain different key data items in the atypical data, as well as all related data items containing the key data items. A related dataset is then constructed by sorting the related data items according to their analytical weights and the number of key data items in the related data from least to most. ,in Indicates the first The first atypical chain dataset Atypical related data; for data containing the same number of related data, sort them according to the sum of the analysis weights of all key data items in the related data; Calculate related data correlation Its meaning is related data. The previous atypical association dataset Frequency of occurrence , Refers to related data In the previous associated dataset The number of times it appears in Refers to associated datasets Total number of related data items; Calculate related data support level Its meaning is related data. In terms of support Related data Support of all key data items ratio ,in , Refers to the data that constitutes the association The Support for key data items Refers to the data that constitutes the association The total number of key data items; Calculate related data urgency Its meaning refers to related data. Security status levels of atypical datasets in the installation status of the operating platform. The number of times it appears The total number of times this security status level occurs ratio ; Establish related data based on the aforementioned indicators. Data feature matrix for .