New energy station high-altitude dangerous operation state monitoring method based on multi-source data

Abstract

The invention discloses a new energy station high-altitude dangerous operation state monitoring method based on multi-source data, and belongs to the field of new energy station high-altitude operation. The method comprises the following steps: acquiring worker information and wearing data of a safety device, and acquiring operation authority of a worker according to the worker information and the wearing data of the safety device; when the operation authority of the worker meets a set condition, collecting state data of the current worker and current operation environment data, and performing fusion processing to obtain data after fusion processing; performing prediction processing on the data after fusion processing by adopting a recurrent neural network fused with an attention mechanism to obtain an operation danger value of the worker; and judging whether the operation danger value of the worker meets a set condition or not, and if yes, giving an alarm. According to the method, the recurrent neural network state prediction model fused with the attention mechanism is adopted for data fusion processing, comprehensive evaluation of dangerous state monitoring is formed, and the safety of workers is guaranteed in real time.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of new energy power station high-altitude operation, and in particular to a new energy power station high-altitude dangerous operation state monitoring method, system, device, medium and program based on multi-source data. BACKGROUND

[0002] The equipment of a new energy power station is often distributed in a large area, and the terrain of these areas is complex and diverse, which may include mountains, plateaus, deserts, coasts and other landforms. The distribution of equipment is also relatively scattered, which brings great difficulties to daily maintenance and management. Especially for the monitoring of high-altitude operation state, it has become a "hard bone" in the safety management of new energy power stations.

[0003] The state monitoring of high-altitude workers mainly relies on unmanned aerial vehicles to provide high-altitude perspectives, sensors or cameras. Unmanned aerial vehicles can provide a macro perspective for monitoring personnel by flying flexibly and patrolling the work area from the air; sensors can be installed on the equipment of workers or in the work environment to collect relevant data in real time; cameras can continuously capture images of the work site. These technologies have reduced the workload of manual inspection to some extent and improved the efficiency of monitoring.

[0004] However, the current state of workers cannot be fully covered, and it is difficult to judge the state of workers in complex environments in real time, and their safety cannot be fully guaranteed. At the same time, the existing data analysis model often needs to be retrained when facing different equipment, environment and climate conditions, which is difficult to be universal, increasing the complexity and cost of maintenance. How to accurately and effectively monitor the state of high-altitude workers in new energy power stations and meet the monitoring needs in complex environments is a problem to be solved. SUMMARY

[0005] In view of the problem in the prior art that the state of high-altitude workers in new energy power stations cannot be fully covered and it is difficult to judge the state of workers in real time in complex environments, the safety of workers cannot be fully guaranteed, the present application provides a new energy power station high-altitude dangerous operation state monitoring method based on multi-source data, which adopts a recursive neural network state prediction model with a fusion attention mechanism to fuse the environment, real-time physiological indicators and action information, form a comprehensive evaluation of dangerous state monitoring, and ensure the safety of workers in real time.

[0006] In order to achieve the above-mentioned purpose, the present application provides the following technical solutions.

[0007] In the first aspect, the present application provides a new energy power station high-altitude dangerous operation state monitoring method based on multi-source data, comprising: Collecting worker information and wearing data of safety devices, obtaining work permission of the worker according to the worker information and the wearing data of the safety devices; When the work permission of the worker meets a set condition, collecting state data and current work environment data of the current worker and performing fusion processing to obtain fusion-processed data; Using a recurrent neural network with a fusion attention mechanism to perform prediction processing on the fusion-processed data to obtain a work danger value of the worker; Judging whether the work danger value of the worker meets a set condition, and issuing a warning if it does.

[0008] As a further improvement of the present application, the collecting worker information and wearing data of safety devices, obtaining work permission of the worker according to the worker information and the wearing data of the safety devices, comprises: Collecting worker information to obtain worker facial information F; Collecting wearing data of safety devices of the worker to obtain wearing conditions S of safety devices of the worker; When the worker facial information F is not recognized by the system or the wearing conditions S of safety devices of the worker do not meet the specifications, the worker cannot obtain high-altitude work permission; When the worker facial information F is recognized by the system and the wearing conditions S of safety devices of the worker meet the specifications, the worker obtains high-altitude work permission.

[0009] As a further improvement of the present application, the collecting worker information and wearing data of safety devices, obtaining work permission of the worker according to the worker information and the wearing data of the safety devices, comprises: When the worker facial information F is not recognized by the system or the wearing conditions S of safety devices of the worker do not meet the specifications, the worker cannot obtain high-altitude work permission; Performing missing value processing and outlier processing on the collected state data and current work environment data of the worker to obtain processed data; Performing data fusion processing on the processed data to obtain fusion-processed data.

[0010] As a further improvement of the present application, the using a recurrent neural network with a fusion attention mechanism to perform prediction processing on the fusion-processed data to obtain a work danger value of the worker, comprises: Organizing sequence data according to a time window and extracting data features of each time step from the fusion-processed data; Inputting the data features into the recurrent neural network with a fusion attention mechanism for training to predict a danger value of each time step and obtain a work danger value of the worker ; In the formula, is the risk value of the time step t, is the weight of the output layer; is the bias of the output layer.

[0011] As a further improvement of the present application, the judgment of whether the work risk value of the worker meets the set condition, if it meets, a warning is issued, including: Judging whether the work risk value of the worker exceeds the threshold value, if it exceeds the threshold value, an alarm is issued.

[0012] As a further improvement of the present application, the threshold value includes:

[0013] wherein: is the risk score; represents physiological data; represents environmental data; represents action or posture data; α, β, γ are weight coefficients, satisfying ; The attention-enhanced RNN model is used to realize the nonlinear mapping of the state risk.

[0014] In a second aspect, the present application provides a high-altitude dangerous work state monitoring system for new energy field stations based on multi-source data, comprising: An acquisition work permission module: used for collecting worker information and safety device wearing data, and acquiring work permission of the worker according to the worker information and the safety device wearing data; A fusion processing module: used for collecting state data and current work environment data of the current worker when the work permission of the worker meets the set condition, and performing fusion processing to obtain the fusion-processed data; An acquisition risk value module: used for predicting the fusion-processed data by using a recurrent neural network with a fusion attention mechanism to obtain the work risk value of the worker; A judgment risk module: used for judging whether the work risk value of the worker meets the set condition, and issuing a warning if it meets.

[0015] In a third aspect, the present application provides an electronic device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the steps of the new energy field station high-altitude dangerous work state monitoring method based on multi-source data when executing the computer program.

[0016] In a fourth aspect, the present application provides a computer readable storage medium storing a computer program, which, when executed by a processor, implements the steps of the method for monitoring the high-altitude dangerous operation state of a new energy power station based on multi-source data.

[0017] In a fifth aspect, the present application provides a computer program product comprising computer instructions, which, when executed by a processor, implement the steps of the method for monitoring the high-altitude dangerous operation state of a new energy power station based on multi-source data.

[0018] Compared with the prior art, the present application has the following beneficial effects: The present method effectively solves the problem that the monitoring cannot be comprehensive in the prior art by collecting multi-source data such as worker information, safety device wearing data, current worker state data, and current operation environment data, and performing fusion processing thereon. It not only focuses on the state of the worker itself, but also comprehensively considers the important factor of the operation environment, monitors the personnel and the environment as an organic whole, ensures the all-around nature of the monitoring, and can capture more potential dangerous factors, thereby providing a more solid data foundation for the safety protection of the worker. In real-time judgment and safety protection, a recurrent neural network with a fusion attention mechanism is used to perform prediction processing on the fusion-processed data to obtain the operation dangerous value of the worker, and whether to issue a warning is determined according to the set conditions. The present method can quickly and accurately process a large amount of complex multi-source data, and realize real-time and accurate judgment of the state of the worker in a complex environment. Compared with the traditional method, it can more timely discover potential dangers and gain valuable response time for the worker, thereby greatly improving the timeliness and effectiveness of safety protection. At the same time, the method forms a comprehensive evaluation of the dangerous state monitoring, comprehensively considers multiple factors such as the environment faced, real-time physiological indicators, and action information, avoids the one-sidedness that may be caused by single-factor judgment, makes the evaluation of the operation dangerous value of the worker more scientific, reasonable, and accurate, and further improves the reliability of safety protection. BRIEF DESCRIPTION OF DRAWINGS

[0019] The drawings described herein are for illustrative purposes only and are not intended to limit the scope of the present application in any way. In the drawings: Figure 1 FIG. 1 is a flowchart of the method for monitoring the high-altitude dangerous operation state of a new energy power station based on multi-source data according to the present application; Figure 2 FIG. 2 is a general technical flowchart of the method for monitoring the high-altitude dangerous operation state of a new energy power station based on multi-source data according to the present application; Figure 3 FIG. 3 is a training flowchart of the attention-enhanced recurrent neural network model of the method according to the embodiment of the present application; Figure 4 It is a structure schematic view of a new energy station high-altitude dangerous operation state monitoring system based on multi-source data according to the present application; Figure 5 It is an electronic device schematic view in the embodiment of the present application. DETAILED DESCRIPTION

[0020] In order for those skilled in the art to better understand the technical solutions in the present application, the technical solutions in the present application will be clearly and completely described below in combination with the drawings in the present application, and the described embodiments are only some of the embodiments of the present application, not all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative labor should belong to the scope of protection of the present application.

[0021] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which the present application belongs. The terms used in the specification of the present application are only for the purpose of describing the specific embodiments and are not intended to limit the present application. The term "and / or" used herein includes any and all combinations of one or more related listed items.

[0022] In view of the problems in the prior art that the monitoring of the state of the new energy station high-altitude worker cannot be fully covered, it is difficult to judge in real time in a complex environment, and the safety of the personnel is insufficient, the present application provides a new energy station high-altitude dangerous operation state monitoring method based on multi-source data, as shown in the figure, the method comprises the following steps: Figure 1 S100: collecting worker information and safety device wearing data, and obtaining the operation permission of the worker according to the worker information and the safety device wearing data; S200: when the operation permission of the worker meets the set condition, collecting the state data of the current worker and the current operation environment data, and performing fusion processing to obtain the fusion processed data; S300: using a recursive neural network with a fusion attention mechanism to perform prediction processing on the fusion processed data to obtain the operation danger value of the worker; S400: judging whether the operation danger value of the worker meets the set condition, and if so, issuing a warning.

[0023] The method uses a recursive neural network state prediction model with a fusion attention mechanism to fuse the environment, real-time physiological indicators and action information, form a comprehensive evaluation of the dangerous state monitoring, and real-time guarantee the safety of the worker.

[0024] The present application will be further explained and described below in combination with specific drawings. ​

[0025] As Figure 2 shown, the application is a new energy station high-altitude dangerous operation state monitoring method based on multi-source data, which specifically comprises: S1: Determine the high-altitude operation threshold to identify whether the working environment is accompanied by a large risk; Specifically, the high-altitude operation threshold is determined by occupational safety regulations and the relative height of the tower. According to the existing standard regulations in China, high-altitude operation is usually divided into four levels: Level 1 high-altitude operation: height exceeding 3 meters; Level 2 high-altitude operation: height exceeding 10 meters; Level 3 high-altitude operation: height exceeding 25 meters; Level 4 high-altitude operation: height exceeding 50 meters.

[0026] S2: Obtain the face information of the workers, the wearing situation of the safety devices, and determine the operation authority; Specifically, the operation authority refers to the collection of face information and identity information of all high-altitude workers in the new energy station. Before the high-altitude operation tower, the face information of the workers and the wearing situation of the safety devices are identified through image recognition. The wearing of safety devices includes: safety belt, safety helmet, smart bracelet. The safety belt and safety helmet are equipped with height sensors and posture sensors. The smart bracelet is equipped with heart rate and blood pressure monitoring functions.

[0027] The specific judgment rules are as follows: If , the personnel cannot enter the high-altitude operation area; If , the personnel enters the high-altitude operation area.

[0028] S3: Intelligent devices and Beidou satellites collect current worker state data and current operation environment data, and pre-process them for fusion; Specifically, the state data includes: personnel state information and working environment information collected by intelligent devices and Beidou satellites.

[0029] Intelligent devices include: smart bracelet, sensor, camera. The state data monitored by the intelligent device is composed of the heart rate and blood pressure of the worker, the body inclination angle of the worker, and the working time of the worker.

[0030] Current operation environment data includes: the height of the worker, the position of the worker based on Beidou satellite, wind speed, temperature, air pressure and weather conditions.

[0031] Data preprocessing includes missing value processing, outlier processing and data fusion. Specifically as follows: Mark the missing data in the collected data to determine the missing position; use K nearest neighbor (KNN) to fill in the missing values; After the missing value processing of the collected data, a new detection formula combining statistical indicators and machine learning prediction residual is proposed for abnormal value identification and processing. Let a data point The corresponding statistical standardized score is:

[0032] and are the mean and standard deviation of a data point , respectively.

[0033] At the same time, the predicted value is obtained by training the model, and the comprehensive abnormality index is defined as:

[0034] wherein is the weight coefficient. If exceeds the preset threshold, it is considered that the data point is abnormal.

[0035] The processed data from different data sources is integrated to form a complete and effective high-altitude worker state monitoring data set.

[0036] Let the data collected by each data source at time t be (i=1, 2, …, N), and the initial weight be The corresponding real-time error (or confidence index) is The improved fusion data is calculated as:

[0037] wherein λ is a positive adjustment parameter.

[0038] S4: Adopting a recurrent neural network with a fusion attention mechanism to predict the worker's work state risk value; Specifically, as shown in Figure 3 , the data formed after preprocessing is converted into a format available for the model, and is organized into sequence data according to the time window, and the data features of each time step are extracted. According to the feature input, the risk value of the work state is predicted, and the work data is input into the RNN model for training to predict the risk value of each time step.

[0039] Each time step has a hidden layer state, which is constantly updated with the input time step, and can gradually “remember” the previous state information. That is:

[0040] wherein is the input data for the current time step, is the hidden state of the previous time step, and is the weight matrix, is the bias term, is the activation function.

[0041] The dangerous value prediction formula is:

[0042] wherein, is the dangerous value at time step t, and is the weight and bias of the output layer.

[0043] To improve the accuracy of state prediction, we introduce an attention mechanism based on the traditional RNN to enhance the model's attention to key moments. Its update formula is:

[0044] wherein, is the activation function.

[0045] After introducing the attention weight , the weighted sum of the entire time series outputs the prediction value as follows:

[0046]

[0047] The obtained time series dangerous prediction value is weighted and fused with physiological data, environmental data and action data to construct a unified risk score .

[0048]

[0049] wherein: represents physiological data (such as heart rate, blood pressure), represents environmental data (such as temperature, wind speed, vehicle driving environment information), represents action or posture data, and α, β, γ are weight coefficients satisfying . The aforementioned attention-enhanced RNN model is used to realize the nonlinear mapping of state risk.

[0050] S5: Determine whether the state dangerous value at this moment and in the near future exceeds the dangerous threshold, and the corresponding reminder is issued by the danger empathy workwear.

[0051] Specifically, the dangerous empathy work clothes give direct feedback to the workers according to the high-altitude state danger value.

[0052] In the safe state, the work clothes remain in the normal state. When the dangerous state is about to exceed the threshold value, the clothes give a slight vibration or color change to remind. When the high-risk state occurs, the work clothes "lock" part of the action (such as limiting dangerous postures).

[0053] In summary, the present application defines the height threshold of high-altitude work, which helps enterprises or units to more accurately assess the risk of high-altitude work and optimize resource allocation, so as to formulate targeted prevention and control measures. Secondly, the data sources of the present method are extensive, and the reliability of each data source may fluctuate with environmental changes. An adaptive multi-source data weighted fusion formula is proposed, which uses an exponential decay function to weight each sensor data according to the real-time error, which can automatically reduce its influence when the data quality decreases, and improve the robustness of data fusion. Moreover, the present application proposes a recurrent neural network (RNN) state prediction model with attention mechanism. The attention mechanism enables the model to automatically identify the most important time steps in the predicted state, improving the interpretability and accuracy of the prediction. In the scenarios of vehicle operating state, driver behavior monitoring, etc., this model can more accurately capture long-term dependencies and sudden changes, achieving better risk warning. Finally, the present application fuses the environment faced by the high-altitude worker, the real-time physiological indicators and the action information to form a comprehensive evaluation of the dangerous state monitoring, and uses the dangerous empathy work clothes with visual perception as the warning medium, which improves the accuracy of risk assessment and reduces the accident risk of high-altitude workers during high-altitude work.

[0054] The present application will be further explained and described in conjunction with specific embodiments.

[0055] Embodiment As Figure 2 shown, the present application provides a new energy station high-altitude dangerous work state monitoring method based on multi-source data, which includes the following steps: S1: Determine the high-altitude work threshold to identify whether the working environment is accompanied by a larger risk.

[0056] High-altitude work threshold It is determined by occupational safety regulations and relative height of tower. According to the existing standards in China, high-altitude work is usually divided into four levels: First-level high-altitude work: height exceeding 3 meters; second-level high-altitude work: height exceeding 10 meters; third-level high-altitude work: height exceeding 25 meters; fourth-level high-altitude work: height exceeding 50 meters.

[0057] The present application sets the high-altitude work threshold to . If the height of the worker is If F = 1 and S = 1, it is determined that the work is high.

[0058] S2: Obtain the face information of the worker, the wearing situation of the safety device, and determine the work permission.

[0059] Collect the face information and identity information of all high-altitude workers in the new energy field station. Before high-altitude work, identify the face information F of the worker and the wearing situation S of the safety device through image recognition.

[0060] The safety device includes a safety belt, a safety helmet, and a smart bracelet. The safety belt and the safety helmet are equipped with height sensors and posture sensors, and the smart bracelet is equipped with heart rate and blood pressure monitoring functions.

[0061] If F = 0 and S = 0, the worker cannot enter the high-altitude work area. If F = 1 and S = 0, the worker enters the high-altitude work area.

[0062] S3: The intelligent device and the Beidou satellite collect the state data of the current worker and the current work environment data, and pre-process them for fusion.

[0063] The state data includes the state information of the worker and the work environment information collected by the intelligent device and the Beidou satellite. Specifically as follows: Worker state data: heart rate HR, blood pressure BP, body inclination angle , work time Twork.

[0064] Work environment data: the height of the worker , the position of the worker based on the Beidou satellite , wind speed , temperature , air pressure , weather condition R.

[0065] Data preprocessing includes missing value processing, outlier processing and data fusion: Missing value processing: mark the collected data for missing values, and use K-Nearest Neighbor (KNN) to fill in the missing values. In this embodiment, for a large proportion of missing values (95% missing values), a direct deletion method is adopted; for important variables, K-Nearest Neighbor (KNN) is used to fill in the missing values.

[0066] Outlier processing: in addition to using the traditional Z-Score method to detect outliers, a new detection formula combining statistical indicators and machine learning prediction residuals is used. Let a data point The corresponding statistical standardized score is:

[0067] ​​ and For each data point The mean and standard deviation of .

[0068] At the same time, the predicted value is obtained by training the model (For example, a regression model built using historical data), then define a comprehensive anomaly indicator for:

[0069] in, is the weight coefficient. If the value exceeds the preset threshold, the data point is considered abnormal.

[0070] Data fusion: Traditional weighted fusion methods generally use fixed weights, but in practical applications (such as high-altitude work or vehicle safety monitoring), the reliability of each data source may fluctuate with environmental changes. To this end, an adaptive weighted fusion formula is used, which dynamically adjusts the weights based on the real-time error of each sensor. Assume that at time t, the data collected by each data source is (i=1,2,…,N), whose initial weight is , the corresponding real-time error (or confidence index) is , then the improved fusion data The calculation formula is:

[0071] Here, λ is a positive tuning parameter.

[0072] S4: Use recurrent neural network (RNN) to predict the risk value of workers' working status.

[0073] like Figure 3 As shown in the figure, the preprocessed data is converted into a format that can be used by the model and organized into sequence data according to time windows. The data features of each time step are extracted. The hazard value of the job status is predicted based on the feature input. The job data is input into the RNN model for training to predict the hazard value of each time step.

[0074] Each time step has a hidden layer state, which is continuously updated with the input time step, and can gradually "remember" the previous state information. That is:

[0075] Where, is the input data of the current time step; is the hidden state of the previous time step; and is the weight matrix, is the bias term, is an activation function.

[0076] The dangerous value prediction formula is:

[0077] wherein, is the dangerous value at time step t, and are the weights and bias of the output layer.

[0078] To improve the accuracy of state prediction, an attention mechanism is introduced based on the traditional RNN to enhance the model's attention to key moments. The update formula is:

[0079] wherein, is an activation function.

[0080] The attention weight is introduced After that, the weighted sum of the entire time series outputs the prediction value As follows:

[0081]

[0082] The obtained time series dangerous prediction value is weighted and fused with physiological data, environmental data and action data to construct a unified risk score .

[0083]

[0084] wherein: represents physiological data (such as heart rate, blood pressure), represents environmental data (such as temperature, wind speed, vehicle driving environment information), represents action or posture data, and α, β, γ are weight coefficients satisfying . The aforementioned attention enhanced RNN model is used to realize the nonlinear mapping of state risk.

[0085] S5: Determine whether the dangerous value exceeds the threshold value, and the danger empathy work clothes issue a reminder.

[0086] In this embodiment, the dangerous degree threshold value is 0.6. If , the danger empathy work clothes issue a reminder: Safe state: , the work clothes remain normal state (green).

[0087] Approaching dangerous state: , the work clothes discolor (yellow) and vibrate slightly.

[0088] The above merely describes preferred embodiments of the present application, but the protection scope of the present application is not limited thereto, and any modification, equivalent replacement, improvement, etc. made within the technical scope disclosed in the present application should be included in the protection scope of the present application.

[0089] A second object of the present application is to provide a new energy station high-altitude dangerous operation state monitoring system based on multi-source data, as shown in Figure 4 The system comprises: The work permission acquisition module 100 is configured to collect worker information and safety device wearing data, and acquire work permission of the worker according to the worker information and the safety device wearing data. The fusion processing module 200 is configured to collect state data and current work environment data of the worker when the work permission of the worker meets a set condition, and perform fusion processing to obtain fusion-processed data. The dangerous value acquisition module 300 is configured to use a recurrent neural network with a fusion attention mechanism to perform prediction processing on the fusion-processed data to obtain a work dangerous value of the worker. The dangerous judgment module 400 is configured to judge whether the work dangerous value of the worker meets a set condition, and issue a warning if the set condition is met.

[0090] As shown in Figure 5 A third object of the present application is to provide an electronic device, which comprises a processor 501, a memory 502 and a display screen 503. The memory 502 and the display screen 503 are connected to the processor 501, for example, through a bus 504. Optionally, the electronic device can further comprise a transceiver 505. It should be noted that the transceiver 505 is not limited to one in actual application, and the structure of the electronic device does not constitute a limitation on the embodiments of the present application.

[0091] The processor 501 can be a CPU (Central Processing Unit), a general-purpose processor, a DSP (Digital Signal Processor), an ASIC (Application Specific Integrated Circuit), an FPGA (Field Programmable Gate Array) or other programmable logic device, transistor logic device, hardware component, or any combination thereof. It can implement or execute the various exemplary logical blocks, modules and circuits described in connection with the disclosure. The processor 501 can also be a combination of computing functions, such as a combination of one or more microprocessors, a combination of a DSP and a microprocessor, etc.

[0092] The bus 504 can include a path for transmitting information between the above-mentioned components. The bus 504 can be a PCI (Peripheral Component Interconnect) bus or an EISA (Extended Industry Standard Architecture) bus, etc. The bus 504 can be divided into an address bus, a data bus, a control bus, etc.

[0093] The memory 502 can be a ROM (Read Only Memory) or other type of static storage device that can store static information and instructions, a RAM (Random Access Memory) or other type of dynamic storage device that can store information and instructions, an EEPROM (Electrically Erasable Programmable Read Only Memory), a CD-ROM (Compact Disc Read Only Memory) or other optical disk storage, a magnetic disk storage medium or other magnetic storage device, or any other medium that can be used to carry or store desired program code in the form of instructions or data structures and that can be accessed by a computer, but is not limited thereto.

[0094] The memory 502 is used to store application program code for implementing the scheme of the present application, and is controlled by the processor 501 to execute. The processor 501 is used to execute the application program code stored in the memory 502 to realize the content shown in the foregoing method embodiments.

[0095] Figure 5 The electronic device shown is merely an example and should not bring any limitation to the function and use range of the embodiments of the present application.

[0096] A fourth object of the present application is to provide a computer readable storage medium storing a computer program, the computer readable storage medium storing a computer program, the computer program being executed by a processor to implement each process of the method embodiments shown in the above Figure 1 and Figure 2 The memory including instructions executable by the processor of the electronic device to perform the above method is for example.

[0097] The computer readable storage medium can be a tangible device that maintains and stores instructions for use by an instruction execution device. The computer readable storage medium can be, but is not limited to, an electronic storage device, a magnetic storage device, an optical storage device, an electromagnetic storage device, a semiconductor storage device, or any combination thereof. Specifically, the computer readable storage medium can be a portable computer diskette, a hard disk, a U disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), a static random access memory (SRAM), a portable compact disc read-only memory (CD-ROM), a digital versatile disk (DVD), a memory stick, a floppy disk, an optical disk, a magnetic disk, a mechanical encoding device, and any combination thereof.

[0098] A fifth object of the present application is to provide a computer program product comprising computer instructions, which, when executed by a processor, implement each process of the method embodiments shown in the above Figure 1 and Figure 2 and achieve the same technical effects. To avoid repetition, they will not be described here. The scope of the present teachings should not be determined with reference to the above description, but should be determined with reference to the above claims and the full scope of equivalents to which such claims are entitled. For the purpose of completeness, all articles and references, including patent applications and publications, are incorporated herein by reference. The omission of any aspect of the subject matter disclosed herein from the above claims does not constitute a disclaimer of such subject matter, nor should it be construed as a disclaimer or disavowal of such subject matter as part of the disclosed application.

[0099] Many embodiments and many applications other than those described herein will be apparent to those skilled in the art from consideration of the specification and practice of the teachings disclosed herein. Therefore, the scope of the present teachings should be determined by the broadest interpretation of the above claims and the full scope of equivalents to which such claims are entitled. For the purposes of the U.S.A., in the event of an infringement of the patent or the claims, the inventor may grant licenses under the patent or the claims and seek damages for any infringement of the patent or claims, including subsequent infringement. For the purposes of the U.S.A., the inventor may grant licenses under the patent or the claims and seek damages for any infringement of the patent or claims, including subsequent infringement.

[0100] The above is a further detailed description of the present application, which cannot be deemed to limit the specific embodiments of the present application to the above, and for those skilled in the art to which the present application belongs, without departing from the concept of the present application, a number of simple deductions or substitutions can be made, which should be deemed to belong to the present application determined by the submitted claims.

Claims

1. A method for monitoring the state of high-altitude dangerous operation of a new energy station based on multi-source data, characterized in that, The method comprises the following steps: Collecting worker information and wearing data of safety devices, obtaining work permission of the worker according to the worker information and the wearing data of the safety devices; When the work permission of the worker meets a set condition, collecting state data and current work environment data of the worker and performing fusion processing to obtain fusion-processed data; Using a recurrent neural network with a fusion attention mechanism to perform prediction processing on the fusion-processed data to obtain a work danger value of the worker; Determining whether the work danger value of the worker meets a set condition, and issuing a warning if the condition is met.

2. The method according to claim 1, wherein, The collecting worker information and wearing data of safety devices, obtaining work permission of the worker according to the worker information and the wearing data of the safety devices, comprises: Collecting worker information to obtain worker facial information F; Collecting wearing data of safety devices of the worker to obtain wearing conditions S of safety devices of the worker; When the worker facial information F is not recognized by the system or the wearing conditions S of safety devices of the worker do not meet the specifications, the worker cannot obtain high-altitude work permission; When the worker facial information F is recognized by the system and the wearing conditions S of safety devices of the worker meet the specifications, the worker obtains high-altitude work permission.

3. The method according to claim 1, wherein, The collecting state data and current work environment data of the worker and performing fusion processing to obtain fusion-processed data when the work permission of the worker meets a set condition, comprises: Collecting state data and current work environment data of the worker when the worker obtains high-altitude work permission; Performing missing value processing and abnormal value processing on the collected state data and current work environment data of the worker to obtain processed data; Performing data fusion processing on the processed data to obtain fusion-processed data.

4. The method according to claim 1, wherein, The using a recurrent neural network with a fusion attention mechanism to perform prediction processing on the fusion-processed data to obtain a work danger value of the worker, comprises: Organizing sequence data according to a time window on the fusion-processed data to extract data features of each time step; The data features are input into a recurrent neural network with fusion attention mechanism for training to predict the risk value at each time step, and the work risk value of the worker is obtained ; wherein is the risk value for time step t, is the weight of the output layer; is the bias of the output layer.

5. The method of claim 1, wherein the method is characterized by: The determining whether the work danger value of the worker meets a set condition, and issuing a warning if the condition is met, comprises: Determining whether the work danger value of the worker exceeds a threshold value, and issuing an alarm if the threshold value is exceeded.

6. The method according to claim 5, wherein, The threshold value comprises: wherein: is a risk score; denotes physiological data; denotes environmental data; denotes action or posture data; a, b, g are weight factors, satisfying ; The attention-enhanced RNN model is used to realize nonlinear mapping of state risk.

7. A new energy station high-altitude dangerous operation state monitoring system based on multi-source data, characterized in that, The method comprises the following steps: A work permission acquisition module is configured to collect worker information and wearing data of safety devices, and obtain work permission of the worker according to the worker information and the wearing data of the safety devices; A fusion processing module is configured to collect state data and current work environment data of the worker and perform fusion processing to obtain fusion-processed data when the work permission of the worker meets a set condition; A danger value acquisition module is configured to use a recurrent neural network with a fusion attention mechanism to perform prediction processing on the fusion-processed data to obtain a work danger value of the worker; A danger judgment module is configured to determine whether the work danger value of the worker meets a set condition, and issue a warning if the condition is met.

8. An electronic device, comprising: The computer readable storage medium stores a computer program, and the computer program is executed by the processor to realize the steps of the method for monitoring the high-altitude dangerous operation state of the new energy power station based on multi-source data according to any one of claims 1-6.

9. A computer-readable storage medium, characterized in that, The computer readable storage medium stores a computer program, and the computer program is executed by the processor to realize the steps of the method for monitoring the high-altitude dangerous operation state of the new energy power station based on multi-source data according to any one of claims 1-6.

10. A computer program product, characterised in that, The computer readable storage medium stores a computer program, and the computer program is executed by the processor to realize the steps of the method for monitoring the high-altitude dangerous operation state of the new energy power station based on multi-source data according to any one of claims 1-6.

Images