Electrical operator behavior analysis and safety risk prevention and control system

Abstract

The invention relates to the technical field of electrical safety, in particular to an electrical operator behavior analysis and safety risk prevention and control system which comprises a behavior data acquisition module, a data preprocessing module, a behavior analysis module, a historical risk database, a safety risk assessment module and a real-time early warning module. Wherein the behavior data acquisition module is used for acquiring behavior data and environment data of operating personnel; the behavior analysis module is used for constructing a behavior recognition model; the security risk assessment module is used for performing multi-dimensional risk assessment; the real-time early warning module is used for sending out early warning signals in various forms; according to the invention, through combination of accurate behavior feature identification, multi-dimensional risk assessment and a dynamic early warning technology, the problems of insufficient real-time monitoring of operation behaviors, incomplete risk assessment and untimely early warning response in the prior art are solved, and the safety and risk prevention and control capability in the electrical operation process are improved.

Description

Technical Field

[0001] The present invention relates to the field of electrical safety technology, and in particular to a behavior analysis and safety risk prevention and control system for electrical operators. Background Art

[0002] With the continuous development of the intelligence and complexity of power systems, the working environment and task requirements of electrical operations are becoming increasingly stringent. Electrical operators need to face high voltage, complex equipment and uncertain environmental conditions in daily operations, which significantly increases the operational risks. At the same time, the operators' irregular operations, the influence of environmental factors and potential equipment failures may cause serious safety accidents. Although some existing monitoring systems can monitor the working environment and equipment status to a certain extent, their real-time monitoring and risk identification capabilities for operators' behaviors are limited, especially the lack of fusion analysis of multi-dimensional data in complex environments. In addition, existing systems generally lack intelligent risk assessment and dynamic early warning functions, making it difficult to perceive and prevent safety hazards in advance.

[0003] In response to the above problems, the existing technology has obvious deficiencies in the following aspects: First, it is impossible to accurately analyze and judge the behavior of operators in real time, especially in the monitoring of complex actions or irregular behaviors, and lacks the recognition of behavioral spatial relationships and detailed characteristics; second, it fails to effectively combine real-time environmental data, historical risk event data and operational behavior data, making it difficult to form a comprehensive risk assessment system; third, there is a lack of a multi-level dynamic early warning mechanism, and corresponding intervention measures cannot be taken in a timely manner according to different risk levels. Summary of the invention

[0004] Based on the above objectives, the present invention provides a system for analyzing electrical worker behavior and preventing and controlling safety risks.

[0005] A behavior analysis and safety risk prevention and control system for electrical workers, including a behavior data acquisition module, a data preprocessing module, a behavior analysis module, a historical risk database, a safety risk assessment module and a real-time warning module; wherein:

[0006] Behavior data collection module: used to deploy sensors and monitoring equipment at the electrical operation site to collect the behavior data and environmental data of operators in real time. The behavior data includes movement trajectory and clothing status information, and the environmental data includes temperature, humidity and electromagnetic field strength;

[0007] Data preprocessing module: used to preprocess the data collected by the behavior data collection module, including noise removal, outlier removal and standardization operations, to generate standardized data streams;

[0008] Behavior analysis module: Based on the standardized data stream output by the data preprocessing module, a behavior recognition model is constructed using the capsule network algorithm to analyze whether the operator's action behavior meets the established operating standards;

[0009] Historical risk database module: used to store risk events in historical operations and provide reference data for the safety risk assessment module;

[0010] Safety risk assessment module: Based on the behavior analysis results output by the behavior analysis module, the real-time collected environmental data, and the risk events from the historical risk database module, it conducts multi-dimensional risk assessment and calculates the risk level of the current operation;

[0011] Real-time warning module: Based on the risk level calculated by the security risk assessment module, it is used to issue various forms of warning signals when the risk level exceeds the set threshold, including voice prompts, visual warnings, and sending alarm information to the management platform.

[0012] Optionally, the behavior data collection module includes a motion trajectory collection unit, a clothing status collection unit and an environment data collection unit; wherein:

[0013] Motion trajectory acquisition unit: used to collect the motion trajectory data of the operator in real time through the ultra-wideband positioning sensor and inertial measurement subunit installed at the work site. The ultra-wideband positioning sensor is used to locate the spatial position of the operator, and the inertial measurement subunit is used to collect the acceleration and angular velocity information of the operator's limbs, thereby generating a complete motion trajectory;

[0014] Dressing status acquisition unit: used to monitor the dressing status of workers in real time through a high-resolution camera and a wearable RFID recognition device. The high-resolution camera is used to collect images of workers wearing the equipment, and the RFID recognition device is used to identify the electronic tags of safety equipment.

[0015] Environmental data acquisition unit: used to monitor environmental data in real time through temperature and humidity sensors and electromagnetic interference detectors deployed at the work site. The temperature and humidity sensors are used to collect temperature and humidity information at the work site; the electromagnetic interference detector is used to detect the electromagnetic field strength at the work site in real time.

[0016] Optionally, the data preprocessing module includes a noise removal unit, an outlier removal unit and a standardization operation unit; wherein:

[0017] Noise removal unit: used to filter the noise of the raw data collected by the behavior data collection module, specifically using a combination of high-pass filter and Kalman filter, where the high-pass filter is used to remove low-frequency noise, and the Kalman filter is used to dynamically estimate the noise part of the signal to ensure that the data is consistent in time and space dimensions;

[0018] Outlier elimination unit: used to identify and eliminate outliers in the collected data. It judges each data point based on the Z-Score algorithm and IQR method. If a data point deviates from the normal data distribution, it is judged as an outlier and eliminated.

[0019] Standardization operation unit: used to standardize the data after noise removal and outlier elimination. Specifically, the Z-Score standardization method is used to standardize each data dimension, converting the data into a standard normal distribution with a mean of 0 and a variance of 1 to ensure the scale consistency between different data dimensions.

[0020] Optionally, the behavior analysis module includes a model building unit and a behavior determination unit; wherein:

[0021] Model building unit: used to build a behavior recognition model according to the capsule network algorithm, the behavior recognition model includes an input layer, a capsule layer and an output layer, wherein the input layer is used to receive a standardized data stream; the capsule layer is used to extract local features and overall features of the operator's actions; and the extracted features are transferred to the output layer through a dynamic routing mechanism to form a feature vector of the action behavior;

[0022] Behavior judgment unit: used to compare the feature vector generated by the capsule network model with the preset operation standard, and judge whether the operator's action behavior meets the established standard based on the similarity between the feature vector and the standard; the judgment result includes meeting the standard or not meeting the standard.

[0023] Optionally, the model building unit includes:

[0024] Input data encoding: The standardized data stream output by the data preprocessing module is input into the input layer of the capsule network, and the initial features are extracted through convolution operation to obtain the feature matrix;

[0025] Capsule layer construction: Input the feature matrix into the capsule layer to construct a feature set containing multiple capsules; each capsule represents a specified action or clothing feature in the form of a vector; the calculation formula for capsule output is: U ij =W ij ·V i , where U ij is the prediction vector from the i-th low-level capsule to the j-th high-level capsule; W ijis the weight matrix from the i-th low-level capsule to the j-th high-level capsule; V i is the output vector of the i-th low-level capsule;

[0026] Dynamic routing mechanism: A dynamic routing mechanism is used between capsule layers to determine the connection weight C between low-level capsules and high-level capsules. ij The dynamic routing mechanism achieves the optimized transmission of feature information by iteratively updating the connection probability, and the formula is: Among them, C ij is the connection weight from the i-th low-level capsule to the j-th high-level capsule; B ij is the routing weight in the current iteration; m is the index variable of the total number of low-level capsules;

[0027] High-level capsule output calculation: By accumulating the output vectors of low-level capsules and combining the routing weights, the output of high-level capsules is calculated. The formula is: V j =squash(∑ i C ij U ij ), where V j is the output vector of the jth advanced capsule; squash is a nonlinear compression function used to limit the modulus length of the vector to the range [0, 1]; C ij is the connection weight from the i-th low-level capsule to the j-th high-level capsule; U ij is the prediction vector from the i-th low-level capsule to the j-th high-level capsule;

[0028] Behavior feature output: output vector V of high-level capsule j As the final behavior feature vector, it is passed to the behavior judgment unit to determine whether the operator's action behavior meets the established operation standards.

[0029] Optionally, the behavior determination unit specifically includes:

[0030] Extract feature vectors and operate standard vectors: Get the behavior feature vector V output by the behavior recognition model j , and load the standard vector S of the corresponding operation from the preset operation standard database;

[0031] Calculate similarity: Use the cosine similarity algorithm to calculate the behavior feature vector V j The similarity between and the standard vector S is: Among them, V j S is the dot product of two vectors; ||V j || and ||S|| are V j and the modulus length of S;

[0032] Determine whether it meets the standard: The calculated similarity Sim(Vj , S) is compared with a preset similarity threshold T: If Sim(V j , S) ≥ T, it is determined that the operator's action behavior conforms to the established standard; if Sim(V j , S) < T, it is determined that the operator's action behavior does not conform to the established standard.

[0033] Optionally, the historical risk database module includes a risk event storage unit, a data association unit, and a classification storage unit; where:

[0034] Risk event storage unit: Used to record and store risk event information that occurred during historical operations, including the time, location, operator identification, operation steps, behavior characteristics, environmental conditions, and risk level of the event;

[0035] Data association unit: Used to establish associations between risk events, behavior characteristics, and environmental conditions, store them in the chronological order of risk event occurrence, and mark the corresponding operation steps and operator identification;

[0036] Classification storage unit: Used to classify and store risk events according to the risk level, group them into low risk, medium risk, and high risk; establish an index during the storage process to support quick retrieval by time, location, or operator identification.

[0037] Optionally, the safety risk assessment module includes a behavior risk assessment unit, an environmental risk assessment unit, a historical risk association unit, a comprehensive risk assessment unit, and a risk level determination unit; where:

[0038] Behavior risk assessment unit: Used to receive the behavior analysis results output by the behavior analysis module, extract the deviation degree between the operator's action characteristics and the operation standard, and quantify it as the behavior risk value R b ;

[0039] Environmental risk assessment unit: Used to analyze the real-time collected environmental data, calculate the risk value of each environmental parameter respectively, and obtain the environmental risk value R through the weighted average method e ;

[0040] Historical risk association unit: Used to extract risk events similar to the current operation from the historical risk database module, and calculate the historical risk value R according to the frequency and risk level of the similar risk events h ;

[0041] Comprehensive risk assessment unit: Used to comprehensively consider the behavior risk value, environmental risk value, and historical risk value, calculate the comprehensive risk value according to the set weights, and the comprehensive risk value calculation formula is: R c = w a ·R b +we ·R e +w h ·R h , where R c is the comprehensive risk value, w a ,w e ,w h are the weight coefficients of behavioral risk, environmental risk and historical risk respectively;

[0042] Risk level determination unit: used to determine the risk level according to the comprehensive risk value R c Determine the risk level of the current operation and classify the risk level into low risk, medium risk and high risk.

[0043] Optionally, the risk level determination unit includes:

[0044] Risk level calculation: receive the comprehensive risk value R calculated by the comprehensive risk assessment unit c , and compare it with the preset risk level threshold; the specific judgment formula is:

[0045]

[0046] Where L is the risk level, 1 represents low risk, 2 represents medium risk, and 3 represents high risk; T1 is the first threshold, representing the dividing line between low risk and medium risk; T2 is the second threshold, representing the dividing line between medium risk and high risk; R c is the comprehensive risk value;

[0047] Risk level output: Based on the judgment results, the risk level L is output to the real-time warning module to trigger the corresponding warning signal.

[0048] Optionally, the real-time warning module includes a warning signal generating unit, a multi-form warning output unit and a warning recording unit; wherein:

[0049] Warning signal generation unit: used to generate warning signals of different intensities according to the risk level L; specifically, when the risk level is medium risk, a medium-level warning signal is generated, including a visual warning signal and a voice prompt signal; when the risk level is high risk, a high-level warning signal is generated, including a visual alarm, a high-decibel voice warning, and a signal to notify the management platform in real time;

[0050] Multi-form warning output unit: used to send out the warning signal generated by the warning signal generation unit through the on-site warning device, remote notification device and wearable device reminder; the on-site warning device is used to control the indicator light and speaker at the work site to send corresponding visual and voice alarms; the remote notification device is used to send warning information to the management platform through the wireless network, and the information content includes the risk level, the specific description of the current operation and the real-time environmental parameters; the wearable device reminder is used to send vibration or voice warnings to the wearable device worn by the operator;

[0051] Early warning recording unit: used to record the information of each early warning, including the warning trigger time, risk level, warning signal type and issuance form; the recorded data is stored in the historical risk database module.

[0052] Beneficial effects of the present invention:

[0053] The present invention, by introducing a capsule network algorithm, can achieve accurate recognition of the behavioral characteristics of operators, and can comprehensively extract the spatial relationship and detailed characteristics of actions, thereby effectively judging whether the behavior of operators meets the established standards. Combined with the cosine similarity algorithm, it is used for accurate matching of behavioral feature vectors with operating standards, thereby achieving quantification of behavioral deviations. In addition, through risk assessment technology based on multidimensional data, it integrates behavioral characteristics, real-time environmental data, and historical risk event information, and can dynamically assess the comprehensive risk value of the current operation, providing data support for real-time response.

[0054] The present invention, through risk level classification and multi-form early warning mechanism, combines the comparison of risk value with preset threshold value, and timely generates alarm signals for different levels, including voice prompts, visual warnings, remote notifications and other forms. This technical solution can realize closed-loop management from risk identification to risk response, and greatly improve the safety assurance capability during electrical operations. BRIEF DESCRIPTION OF THE DRAWINGS

[0055] In order to more clearly illustrate the technical solutions in the present invention or the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings in the following description are only for the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0056] Figure 1 A schematic diagram of a behavior analysis and security risk prevention and control system according to an embodiment of the present invention;

[0057] Figure 2 Schematic diagram of a behavior analysis module according to an embodiment of the present invention. DETAILED DESCRIPTION

[0058] The present invention is described in detail below in conjunction with the accompanying drawings and specific embodiments. At the same time, it is explained here that in order to make the embodiments more detailed, the following embodiments are the best and preferred embodiments, and those skilled in the art may also adopt other alternatives to implement some known technologies; and the accompanying drawings are only for more specific description of the embodiments, and are not intended to specifically limit the present invention.

[0059] It should be noted that the references to "one embodiment", "an embodiment", "an exemplary embodiment", "some embodiments" and the like in the specification indicate that the embodiments described may include specific features, structures or characteristics, but not every embodiment may include the specific features, structures or characteristics. In addition, when a specific feature, structure or characteristic is described in conjunction with an embodiment, it should be within the knowledge of a person skilled in the art to implement such feature, structure or characteristic in conjunction with other embodiments (whether or not explicitly described).

[0060] In general, a term can be understood, at least in part, from its use in context. For example, depending, at least in part, on the context, the term "one or more" as used herein can be used to describe any feature, structure, or characteristic in the singular sense, or can be used to describe a combination of features, structures, or characteristics in the plural sense. Additionally, the term "based on" can be understood as not necessarily intended to convey an exclusive set of factors, but can instead, depending, at least in part, on the context, allow for the presence of other factors that are not necessarily explicitly described.

[0061] like Figure 1-Figure 2 As shown, a behavior analysis and safety risk prevention and control system for electrical workers includes a behavior data acquisition module, a data preprocessing module, a behavior analysis module, a historical risk database, a safety risk assessment module and a real-time warning module; wherein:

[0062] Behavior data collection module: used to deploy sensors and monitoring equipment at the electrical operation site to collect the behavior data and environmental data of operators in real time. The behavior data includes movement trajectory and clothing status information, and the environmental data includes temperature, humidity and electromagnetic field strength.

[0063] Data preprocessing module: used to preprocess the data collected by the behavior data collection module, including noise removal, outlier removal and standardization operations, to generate standardized data streams;

[0064] Behavior analysis module: Based on the standardized data stream output by the data preprocessing module, the capsule network algorithm is used to build a behavior recognition model to analyze whether the operator's action behavior meets the established operation standards;

[0065] Historical risk database module: used to store risk events in historical operations and provide reference data for the safety risk assessment module;

[0066] Safety risk assessment module: Based on the behavior analysis results output by the behavior analysis module, the real-time collected environmental data, and the risk events from the historical risk database module, it conducts multi-dimensional risk assessment and calculates the risk level of the current operation;

[0067] Real-time warning module: Based on the risk level calculated by the security risk assessment module, it is used to issue various forms of warning signals when the risk level exceeds the set threshold, including voice prompts, visual warnings, and sending alarm information to the management platform.

[0068] The behavior data collection module includes a motion trajectory collection unit, a clothing status collection unit, and an environmental data collection unit; wherein:

[0069] Motion trajectory acquisition unit: used to collect the motion trajectory data of the operator in real time through the ultra-wideband (UWB) positioning sensor and inertial measurement subunit (IMU) installed at the work site. The ultra-wideband positioning sensor is used to locate the spatial position of the operator, and the inertial measurement subunit is used to collect the acceleration and angular velocity information of the operator's limbs, thereby generating a complete motion trajectory;

[0070] Dressing status acquisition unit: used to monitor the dressing status of workers in real time through high-resolution cameras and wearable RFID identification devices. High-resolution cameras are used to collect images of workers wearing clothes and determine whether workers are wearing safety equipment that meets the standards through image recognition technology. RFID identification devices are used to identify the electronic tags of safety equipment to ensure the correctness of the types and quantities of safety equipment.

[0071] Environmental data acquisition unit: used to monitor environmental data in real time through temperature and humidity sensors and electromagnetic interference detectors deployed at the work site. The temperature and humidity sensors are used to collect temperature and humidity information at the work site; the electromagnetic interference detector is used to detect the electromagnetic field strength at the work site in real time; the above units are connected to the data preprocessing module, and output the motion trajectory, clothing status and environmental data respectively, to ensure the real-time and accuracy of the collected data, and provide a reliable input basis for subsequent data processing and analysis.

[0072] The data preprocessing module includes a noise removal unit, an outlier removal unit, and a standardization operation unit; wherein:

[0073] Noise removal unit: used to filter the noise of the raw data collected by the behavior data collection module, specifically using a combination of high-pass filter and Kalman filter, where the high-pass filter is used to remove low-frequency noise, and the Kalman filter is used to dynamically estimate the noise part of the signal to ensure that the data is consistent in time and space dimensions;

[0074] Outlier elimination unit: used to identify and eliminate outliers in the collected data. Each data point is judged based on the Z-Score algorithm and IQR (interquartile range) method. If a data point deviates from the normal data distribution, it is determined to be an outlier and eliminated to ensure the normality of the input data set and avoid affecting the subsequent analysis model.

[0075] Standardization operation unit: used to standardize the data after noise removal and outlier elimination. Specifically, the Z-Score standardization method is used to standardize each data dimension, and the data is converted into a standard normal distribution with a mean of 0 and a variance of 1, to ensure the scale consistency between different data dimensions, so that the data is suitable for subsequent deep learning algorithm analysis; the above-mentioned data preprocessing module effectively cleans and transforms the original data collected by the behavior data acquisition module, and generates a standardized data stream as the input data of the behavior analysis module to ensure the data quality and model training effect in the subsequent analysis process.

[0076] The behavior analysis module includes a model building unit and a behavior determination unit; wherein:

[0077] Model building unit: used to build a behavior recognition model according to the capsule network algorithm. The behavior recognition model includes an input layer, a capsule layer and an output layer. The input layer is used to receive standardized data streams; the capsule layer is used to extract local and overall features of the operator's actions; and the extracted features are transferred to the output layer through a dynamic routing mechanism to form a feature vector of the action behavior;

[0078] Behavior judgment unit: used to compare the feature vector generated by the capsule network model with the preset operating standard, and judge whether the operator's action behavior meets the established standard based on the similarity between the feature vector and the standard; the judgment result includes compliance with the standard or non-compliance with the standard; the above-mentioned behavior analysis module constructs a behavior recognition model through the capsule network algorithm, extracts the operator's action behavior characteristics, and judges whether it meets the operating standard, providing support for subsequent safety risk assessment.

[0079] The model building units include:

[0080] Input data encoding: The standardized data stream output by the data preprocessing module is input into the input layer of the capsule network, and the initial features are extracted through convolution operation to obtain the feature matrix F. The formula of convolution operation is: F = Conv(X, W) + b, where F is the initial feature matrix, X is the standardized data stream, W is the convolution kernel weight matrix, b is the bias term, and Conv is the convolution operation. This step is used to convert the standardized data into a high-dimensional feature representation to provide input for subsequent capsule processing.

[0081] Capsule layer construction: Input the feature matrix F into the capsule layer to construct a feature set containing multiple capsules; each capsule represents a specified action or clothing feature in the form of a vector; the calculation formula for capsule output is: U ij =W ij ·V i , where U ij is the prediction vector from the i-th low-level capsule to the k-th high-level capsule; W ij is the weight matrix from the i-th low-level capsule to the j-th high-level capsule; V i is the output vector of the i-th low-level capsule;

[0082] Dynamic routing mechanism: A dynamic routing mechanism is used between capsule layers to determine the connection weight C between low-level capsules and high-level capsules. ij ; The dynamic routing mechanism achieves the optimized transmission of feature information by iteratively updating the connection probability. The formula is: Among them, C ij is the connection weight from the i-th low-level capsule to the j-th high-level capsule; B ij is the routing weight in the current iteration; m is the index variable of the total number of low-level capsules;

[0083] High-level capsule output calculation: By accumulating the output vectors of low-level capsules and combining the routing weights, the output of high-level capsules is calculated. The formula is: V j =squash(∑ i C ij U ij ), where V j is the output vector of the jth advanced capsule; squash is a nonlinear compression function used to limit the modulus length of the vector to the range [0, 1]; C ij is the connection weight from the i-th low-level capsule to the j-th high-level capsule; U ij is the prediction vector from the i-th low-level capsule to the j-th high-level capsule;

[0084] Behavior feature output: output vector V of high-level capsule j As the final behavior feature vector, it is passed to the behavior judgment unit to determine whether the operator's action behavior meets the established operating standards; through the above steps, the model construction unit completes the construction of the behavior recognition model based on the capsule network, from the extraction of standardized data streams to the generation of the final behavior feature vector, realizing comprehensive feature modeling of actions and clothing status.

[0085] The behavior determination unit specifically includes:

[0086] Extract feature vectors and operate standard vectors: Get the behavior feature vector V output by the behavior recognition model j, and load the standard vector S corresponding to the operation from a pre-set operation standard database; both the behavior feature vector and the standard vector are represented as vectors of the same dimension, which are used for the comparison calculation between features;

[0087] Calculate the similarity: Use the cosine similarity algorithm to calculate the similarity between the behavior feature vector V j and the standard vector S. The formula is: where, V j ·S is the dot product of the two vectors; ||V j || and ||S|| are the norms of V j and S respectively, and the calculation formulas are:

[0088] is the value of the i-th dimension in the behavior feature vector V j s i is the value of the i-th dimension in the standard vector S, and n is the dimension of the vector;

[0089] Determine whether it meets the standard: Compare the calculated similarity Sim(V j , S) with the preset similarity threshold T: If Sim(V j , S) ≥ T, it is determined that the action behavior of the operator meets the established standard; if Sim(V j , S) < T, it is determined that the action behavior of the operator does not meet the established standard; where, T is a threshold set according to the operation complexity and actual requirements, and is pre-stored in the operation standard database.

[0090] The historical risk database module includes a risk event storage unit, a data association unit, and a classification storage unit; where:

[0091] Risk event storage unit: Used to record and store the risk event information that occurred during the historical operation process, including the time of the event occurrence, location, operator identification, operation steps, behavior features, environmental conditions, and risk level; where, the time represents the occurrence moment of the risk event; the location represents the geographical location or equipment location of the operation site; the operator identification is used to uniquely identify the involved personnel; the operation steps record the specific work links involved in the risk event; the behavior features are output by the behavior analysis module and are used to describe the action behavior features of the operator; the environmental conditions include the temperature, humidity, and electromagnetic interference intensity of the operation site; the risk level is calculated and generated by the safety risk assessment module and is used to identify the severity of the event;

[0092] Data association unit: Used to establish the association between risk events and behavior features, environmental conditions, store them in the order of the time of risk event occurrence, and mark the corresponding operation steps and operator identification to support subsequent risk assessment and trend analysis;

[0093] Classification storage unit: used to classify and store risk events according to risk levels, and group them into low risk, medium risk and high risk for easy query and management; indexes are also established during the storage process to support quick retrieval by time, location or operator identification; the above-mentioned historical risk database fully records risk event information through the risk event storage unit and the data association unit, and classifies and manages the data in combination with the classification storage unit, providing an accurate reference basis for the safety risk assessment module.

[0094] The safety risk assessment module includes a behavior risk assessment unit, an environmental risk assessment unit, a historical risk association unit, a comprehensive risk assessment unit, and a risk level determination unit; among which:

[0095] Behavior risk assessment unit: used to receive the behavior analysis results output by the behavior analysis module, extract the deviation between the operator's action characteristics and the operation standards, and quantify them into a behavior risk value R b ; The behavior risk value is obtained through the behavior deviation calculation formula: R b =1-Sim(V j , S), where the behavioral deviation is a supplement to the similarity of the operating standard. The larger the value, the higher the risk. The behavioral risk value R b Used to evaluate the degree of deviation between operator behavior and standard requirements;

[0096] Environmental risk assessment unit: used to analyze the environmental data collected in real time, calculate the risk value of each environmental parameter, and obtain the environmental risk value R through the weighted average method. e ; The environmental risk value calculation formula is: R e =w1·T r +w2·H r +w3·E r , where T r ,H r ,E r are the risk quantification values ​​of temperature, humidity, and electromagnetic interference intensity, respectively; w1, w2, and w3 are the corresponding weight coefficients, which are preset based on actual experience or statistical data;

[0097] Historical risk association unit: used to extract risk events similar to the current operation from the historical risk database module, and calculate the historical risk value R according to the frequency and risk level of similar risk events h ; The historical risk value calculation formula is: Among them, R i is the risk level of each similar historical risk event, and n1 is the number of similar risk events extracted;

[0098] Comprehensive risk assessment unit: used to integrate behavioral risk value, environmental risk value and historical risk value, and calculate the comprehensive risk value according to the set weights. The calculation formula for the comprehensive risk value is: R c =w a ·R b +w e ·R e +w h ·R h , where R c is the comprehensive risk value, w a ,w e ,w h They are the weight coefficients of behavioral risk, environmental risk and historical risk, which are preset according to actual operational needs;

[0099] Risk level determination unit: used to determine the risk level according to the comprehensive risk value R c Determine the risk level of the current operation and divide the risk level into low risk, medium risk and high risk; the above-mentioned safety risk assessment module realizes the quantitative assessment of multi-dimensional risks and the accurate calculation of the current operation risk level through the behavioral risk assessment unit, environmental risk assessment unit and historical risk association unit, combined with the comprehensive risk assessment unit and the risk level determination unit, to provide decision-making basis for the real-time warning module.

[0100] The risk level determination unit includes:

[0101] Risk level calculation: receive the comprehensive risk value R calculated by the comprehensive risk assessment unit c , and compare it with the preset risk level threshold; the specific judgment formula is:

[0102]

[0103] Where L is the risk level, 1 represents low risk, 2 represents medium risk, and 3 represents high risk; T1 is the first threshold, representing the dividing line between low risk and medium risk; T2 is the second threshold, representing the dividing line between medium risk and high risk; R c is the comprehensive risk value;

[0104] Risk level output: Based on the judgment result, the risk level L is output to the real-time warning module to trigger the corresponding warning signal; the above-mentioned risk level judgment unit accurately divides the risk level of the current operation by comparing the comprehensive risk value with the preset threshold, ensuring that the system triggers corresponding safety measures according to the actual risk situation.

[0105] The real-time warning module includes a warning signal generation unit, a multi-form warning output unit and a warning recording unit; wherein:

[0106] Warning signal generation unit: used to generate warning signals of different intensities according to the risk level L; specifically, when the risk level is medium risk, a medium-level warning signal is generated, including a visual warning signal (such as a yellow indicator light) and a voice prompt signal; when the risk level is high risk, a high-level warning signal is generated, including a visual alarm (such as a red flashing light), a high-decibel voice warning, and a signal to notify the management platform in real time;

[0107] Multi-form warning output unit: used to send out the warning signal generated by the warning signal generation unit through the on-site warning device, remote notification device and wearable device reminder; the on-site warning device is used to control the indicator light and speaker at the work site to issue corresponding visual and voice alarms; the remote notification device is used to send warning information to the management platform through the wireless network, and the information content includes the risk level, the specific description of the current operation and the real-time environmental parameters; the wearable device reminder is used to send vibration or voice warnings to the wearable device (such as a smart watch or helmet) worn by the operator to remind the operator to take necessary measures;

[0108] Warning recording unit: used to record the information of each warning, including warning trigger time, risk level, warning signal type and issuance form; the recorded data is stored in the historical risk database module for subsequent analysis and review; the above-mentioned real-time warning module realizes a multi-level and multi-form warning signal issuance mechanism through the warning signal generation unit, multi-form warning output unit and warning recording unit, ensuring that the safety risks at the work site can be responded to and handled in a timely and effective manner.

[0109] The present invention covers any substitution, modification, equivalent method and scheme made on the essence and scope of the present invention. In order to make the public have a thorough understanding of the present invention, specific details are described in detail in the following preferred embodiments of the present invention, but those skilled in the art can fully understand the present invention without the description of these details. In addition, in order to avoid unnecessary confusion about the essence of the present invention, well-known methods, processes, procedures, components and circuits are not described in detail.

[0110] The above is only a preferred embodiment of the present invention. It should be pointed out that for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the principle of the present invention. These improvements and modifications should also be regarded as the scope of protection of the present invention.

Claims

1. A system for analyzing electrical worker behavior and preventing and controlling safety risks, characterized in that: It includes behavior data collection module, data preprocessing module, behavior analysis module, historical risk database, security risk assessment module and real-time warning module; among which: Behavior data collection module: used to deploy sensors and monitoring equipment at the electrical operation site to collect the behavior data and environmental data of operators in real time. The behavior data includes movement trajectory and clothing status information, and the environmental data includes temperature, humidity and electromagnetic field strength; Data preprocessing module: used to preprocess the data collected by the behavior data collection module, including noise removal, outlier removal and standardization operations, to generate standardized data streams; Behavior analysis module: Based on the standardized data stream output by the data preprocessing module, a behavior recognition model is constructed using the capsule network algorithm to analyze whether the operator's action behavior meets the established operating standards; Historical risk database module: used to store risk events in historical operations and provide reference data for the safety risk assessment module; Safety risk assessment module: Based on the behavior analysis results output by the behavior analysis module, the real-time collected environmental data, and the risk events from the historical risk database module, it conducts multi-dimensional risk assessment and calculates the risk level of the current operation; Real-time warning module: Based on the risk level calculated by the security risk assessment module, it is used to issue various forms of warning signals when the risk level exceeds the set threshold, including voice prompts, visual warnings, and sending alarm information to the management platform.

2. The electrical worker behavior analysis and safety risk prevention and control system according to claim 1, characterized in that: The behavior data collection module includes a motion trajectory collection unit, a clothing status collection unit and an environmental data collection unit; wherein: Motion trajectory acquisition unit: used to collect the motion trajectory data of the operator in real time through the ultra-wideband positioning sensor and inertial measurement subunit installed at the work site. The ultra-wideband positioning sensor is used to locate the spatial position of the operator, and the inertial measurement subunit is used to collect the acceleration and angular velocity information of the operator's limbs, thereby generating a complete motion trajectory; Dressing status acquisition unit: used to monitor the dressing status of workers in real time through a high-resolution camera and a wearable RFID recognition device. The high-resolution camera is used to collect images of workers wearing the equipment, and the RFID recognition device is used to identify the electronic tags of safety equipment. Environmental data acquisition unit: used to monitor environmental data in real time through temperature and humidity sensors and electromagnetic interference detectors deployed at the work site. The temperature and humidity sensors are used to collect temperature and humidity information at the work site; the electromagnetic interference detector is used to detect the electromagnetic field strength at the work site in real time.

3. The electrical worker behavior analysis and safety risk prevention and control system according to claim 1, characterized in that: The data preprocessing module includes a noise removal unit, an outlier removal unit and a standardization operation unit; wherein: Noise removal unit: used to filter the noise of the raw data collected by the behavior data collection module, specifically using a combination of high-pass filter and Kalman filter, where the high-pass filter is used to remove low-frequency noise, and the Kalman filter is used to dynamically estimate the noise part of the signal to ensure that the data is consistent in time and space dimensions; Outlier elimination unit: used to identify and eliminate outliers in the collected data. It judges each data point based on the Z-Score algorithm and IQR method. If a data point deviates from the normal data distribution, it is judged as an outlier and eliminated. Standardization operation unit: used to standardize the data after noise removal and outlier elimination. Specifically, the Z-Score standardization method is used to standardize each data dimension, converting the data into a standard normal distribution with a mean of 0 and a variance of 1 to ensure the scale consistency between different data dimensions.

4. The electrical worker behavior analysis and safety risk prevention and control system according to claim 1, characterized in that: The behavior analysis module includes a model building unit and a behavior determination unit; wherein: Model building unit: used to build a behavior recognition model according to the capsule network algorithm, the behavior recognition model includes an input layer, a capsule layer and an output layer, wherein the input layer is used to receive a standardized data stream; the capsule layer is used to extract local features and overall features of the operator's actions; and the extracted features are transferred to the output layer through a dynamic routing mechanism to form a feature vector of the action behavior; Behavior judgment unit: used to compare the feature vector generated by the capsule network model with the preset operation standard, and judge whether the operator's action behavior meets the established standard based on the similarity between the feature vector and the standard; the judgment result includes meeting the standard or not meeting the standard.

5. The electrical worker behavior analysis and safety risk prevention and control system according to claim 4, characterized in that: The model building unit comprises: Input data encoding: The standardized data stream output by the data preprocessing module is input into the input layer of the capsule network, and the initial features are extracted through convolution operation to obtain the feature matrix; Capsule layer construction: Input the feature matrix into the capsule layer to construct a feature set containing multiple capsules; each capsule represents a specified action or clothing feature in the form of a vector; the calculation formula for capsule output is: U ij =W ij ·V i , where U ij is the prediction vector from the i-th low-level capsule to the j-th high-level capsule; W ij is the weight matrix from the i-th low-level capsule to the j-th high-level capsule; V i is the output vector of the i-th low-level capsule; Dynamic routing mechanism: A dynamic routing mechanism is used between capsule layers to determine the connection weight C between low-level capsules and high-level capsules. ij The dynamic routing mechanism achieves the optimized transmission of feature information by iteratively updating the connection probability, and the formula is: Among them, C ij is the connection weight from the i-th low-level capsule to the j-th high-level capsule; B ij is the routing weight in the current iteration; m is the index variable of the total number of low-level capsules; High-level capsule output calculation: By accumulating the output vectors of low-level capsules and combining the routing weights, the output of high-level capsules is calculated. The formula is: V j =squash(∑ i C ij U ij ), where V j is the output vector of the jth advanced capsule; squash is a nonlinear compression function used to limit the modulus length of the vector to the range [0, 1]; C ij is the connection weight from the i-th low-level capsule to the j-th high-level capsule; U ij is the prediction vector from the i-th low-level capsule to the j-th high-level capsule; Behavior feature output: output vector V of high-level capsule j As the final behavior feature vector, it is passed to the behavior judgment unit to determine whether the operator's action behavior meets the established operation standards.

6. The electrical worker behavior analysis and safety risk prevention and control system according to claim 5, characterized in that: The behavior determination unit specifically includes: Extract feature vectors and operate standard vectors: Get the behavior feature vector V output by the behavior recognition model j , and load the standard vector S of the corresponding operation from the preset operation standard database; Calculate similarity: Use the cosine similarity algorithm to calculate the behavior feature vector V j The similarity between and the standard vector S is: Among them, V j S is the dot product of two vectors; ∥V j ∥ and ∥S∥ are V j and the modulus length of S; Determine whether it meets the standard: Compare the calculated similarity Sim(V j , S) with the preset similarity threshold T: If Sim(V j , S) ≥ T, it is determined that the action behavior of the operator meets the established standard; if Sim(V j , S) < T, it is determined that the action behavior of the operator does not meet the established standard.

7. The electrical worker behavior analysis and safety risk prevention and control system according to claim 1, characterized in that: The historical risk database module includes a risk event storage unit, a data association unit and a classification storage unit; wherein: Risk event storage unit: used to record and store risk event information that occurred during historical operations, including the time, location, operator identification, operation steps, behavioral characteristics, environmental conditions, and risk level of the event; Data association unit: used to establish the association between risk events and behavioral characteristics and environmental conditions, store them in the chronological order of risk events, and mark the corresponding operation steps and operator identification; Classification storage unit: used to classify and store risk events according to risk levels, and group them into low risk, medium risk and high risk. Indexes are also created during the storage process to support quick retrieval by time, location or operator identification.

8. The electrical worker behavior analysis and safety risk prevention and control system according to claim 1, characterized in that: The safety risk assessment module includes a behavior risk assessment unit, an environmental risk assessment unit, a historical risk association unit, a comprehensive risk assessment unit and a risk level determination unit; wherein: Behavior risk assessment unit: used to receive the behavior analysis results output by the behavior analysis module, extract the deviation between the operator's action characteristics and the operation standards, and quantify them into a behavior risk value R b ; Environmental risk assessment unit: used to analyze the environmental data collected in real time, calculate the risk value of each environmental parameter, and obtain the environmental risk value R through the weighted average method. e ; Historical risk association unit: used to extract risk events similar to the current operation from the historical risk database module, and calculate the historical risk value R according to the frequency and risk level of similar risk events h ; Comprehensive risk assessment unit: used to integrate behavioral risk value, environmental risk value and historical risk value, and calculate the comprehensive risk value according to the set weights. The calculation formula for the comprehensive risk value is: R c =w a ·R b +w e ·R e +w h ·R h , where R c is the comprehensive risk value, w a ,w e ,w h are the weight coefficients of behavioral risk, environmental risk and historical risk respectively; Risk level determination unit: used to determine the risk level according to the comprehensive risk value R c Determine the risk level of the current operation and classify the risk level into low risk, medium risk and high risk.

9. The electrical worker behavior analysis and safety risk prevention and control system according to claim 8, characterized in that: The risk level determination unit comprises: Risk level calculation: receive the comprehensive risk value R calculated by the comprehensive risk assessment unit c , and compare it with the preset risk level threshold; the specific judgment formula is: Where L is the risk level, 1 represents low risk, 2 represents medium risk, and 3 represents high risk; T1 is the first threshold, representing the dividing line between low risk and medium risk; T2 is the second threshold, representing the dividing line between medium risk and high risk; R c is the comprehensive risk value; Risk level output: Based on the judgment results, the risk level L is output to the real-time warning module to trigger the corresponding warning signal.

10. The electrical worker behavior analysis and safety risk prevention and control system according to claim 9, characterized in that: The real-time warning module includes a warning signal generating unit, a multi-form warning output unit and a warning recording unit; wherein: Warning signal generation unit: used to generate warning signals of different intensities according to the risk level L; specifically, when the risk level is medium risk, a medium-level warning signal is generated, including a visual warning signal and a voice prompt signal; when the risk level is high risk, a high-level warning signal is generated, including a visual alarm, a high-decibel voice warning, and a signal to notify the management platform in real time; Multi-form warning output unit: used to send out the warning signal generated by the warning signal generation unit through the on-site warning device, remote notification device and wearable device reminder; the on-site warning device is used to control the indicator light and speaker at the work site to send corresponding visual and voice alarms; the remote notification device is used to send warning information to the management platform through the wireless network, and the information content includes the risk level, the specific description of the current operation and the real-time environmental parameters; the wearable device reminder is used to send vibration or voice warnings to the wearable device worn by the operator; Early warning recording unit: used to record the information of each early warning, including the warning trigger time, risk level, warning signal type and issuance form; the recorded data is stored in the historical risk database module.

Images