Electric power operation safety compliance management and control system integrated with multi-modal artificial intelligence technology

Abstract

The invention provides an electric power operation safety compliance management and control system fused with a multi-mode artificial intelligence technology, and the system comprises the steps: carrying out the multi-scale frequency domain decomposition of an operation field image, obtaining different frequency components, carrying out the quantum drying and noise reduction of the different frequency components, and obtaining a first quantum state component after drying and noise reduction; mapping the electric data to a quantum bit state, and performing phase adjustment to obtain a second quantum state component which is not interfered by noise; performing quantum entanglement fusion on the first quantum state component, the second quantum state component and the quantum code of the magnetic data according to a strong magnetic combination, a hidden combination and a tool combination, and inputting the fused feature vectors into a safety compliance evaluation model; and obtaining a strong electromagnetic field interference result, a hidden discharge detection result and an insulation tool state detection result. According to the embodiment of the invention, the quantum pulse coding is applied to the safety management and control of the power production operation site, so that the response speed and accuracy of the operation safety compliance management and control system are effectively improved.

Description

Technical Field

[0001] The present invention relates to the field of electric power technology, and in particular to an electric power operation safety compliance management and control system integrating multimodal artificial intelligence technology. Background Art

[0002] Safety management and control at power production operation sites is a crucial component of power operation safety. The unique characteristics of power production operation sites present significant challenges to traditional safety management methods. Improving safety and reducing the risk of operational errors are core requirements of power production operation safety management systems. Improving operational efficiency and reducing or even eliminating human error are also essential. Furthermore, real-time monitoring and data analysis capabilities can help identify and address issues promptly. The system uses monitoring modules, protection modules, and automation control modules to monitor the operating status of power systems (equipment) in real time, identifying equipment anomalies and taking appropriate measures.

[0003] Therefore, there is an urgent need for a power operation safety compliance management and control system. Summary of the Invention

[0004] The present invention provides a power operation safety compliance management and control system that integrates multimodal artificial intelligence technology. Its main purpose is to combine quantum pulse coding to effectively improve the response speed and accuracy of the power operation safety compliance management and control system.

[0005] An embodiment of the present invention provides a power operation safety and compliance management system that integrates multimodal artificial intelligence technology. The system includes optical sensors, electrical sensors, and magnetic sensors. A processor of the power operation safety and compliance management system receives work site images captured by the optical sensors, electrical data collected by the electrical sensors, and magnetic data received by the magnetic sensors, and processes the data through the following steps:

[0006] Performing multi-scale frequency domain decomposition on the work site image to obtain different frequency components, performing quantum dry denoising on the different frequency components to obtain dry denoised first quantum state components, converting the first quantum state components into a denoised work image, and performing a preliminary safety assessment of the workers based on the denoised work image;

[0007] Mapping the electrical data to a quantum bit state and performing phase adjustment to obtain a second quantum state component that is not interfered with by noise, and obtaining a preliminary operating state of the power device based on the second quantum state component;

[0008] The first quantum state component, the second quantum state component and the quantum encoding of the magnetic data are quantum entangled and fused according to the strong magnetic combination, the hidden discharge combination and the tool combination, and the fused feature vectors are respectively input into the safety compliance assessment model to obtain the strong electromagnetic field interference results, the hidden discharge detection results and the insulating tool status detection results.

[0009] Furthermore, the steps of performing multi-scale frequency domain decomposition on the work site image to obtain different frequency components, performing quantum drying denoising on the different frequency components to obtain first quantum state components of the drying denoise, and then converting the first quantum state components into the denoised work image include:

[0010] Converting the work site image from the spatial domain to the frequency domain, decomposing it to obtain different frequency components at multiple levels, and obtaining image data corresponding to the different frequency components;

[0011] Mapping image data corresponding to different frequency components to a first quantum bit state, and performing phase adjustment on the first quantum bit state to obtain a first quantum state component for drying and denoising;

[0012] The first quantum state component is measured to obtain the values of different frequency components after noise elimination, and the values of different frequency components after noise elimination are combined to obtain the operation image after noise elimination.

[0013] Furthermore, the steps of mapping the electrical data to a quantum bit state and performing phase adjustment to obtain a second quantum state component not interfered with by noise include:

[0014] Encoding the electrical data using three qubits and selecting the first three bits to obtain a second qubit state;

[0015] A noise model in the electrical data is established, and a phase adjustment strategy is designed based on the noise model. A quantum gate is applied based on the phase adjustment strategy to change the phase of the second quantum bit state through the quantum gate to obtain a second quantum state component that is not interfered with by noise.

[0016] Furthermore, the quantum encoding of the first quantum state component, the second quantum state component and the magnetic data is quantum entangled and fused according to a strong magnetic combination, a hidden release combination and an insulating combination, respectively, and the steps include:

[0017] According to the strong magnetic combination, the visible light code and the near infrared code in the first quantum state component are extracted, and quantum entangled and fused with the quantum code of the magnetic data to obtain a strong magnetic feature;

[0018] According to the concealment and release combination, the ultraviolet code and shortwave infrared code in the first quantum state component are extracted, the UHF code of the second quantum state component is extracted, and quantum entanglement fusion is performed to obtain the concealment and release characteristics;

[0019] According to the insulating combination, the short-wave infrared code and the visible light code in the first quantum state component are extracted, and quantum entanglement fusion is performed to obtain the insulating characteristics.

[0020] Furthermore, the fused feature vectors are input into the safety compliance assessment model respectively to obtain the strong electromagnetic field interference result, the hidden discharge detection result and the insulation tool status detection result, the steps include:

[0021] Inputting the strong magnetic characteristics into the safety compliance assessment model, and determining the strong electromagnetic field interference result through multi-spectral cross-validation;

[0022] Inputting the hidden discharge characteristics into the safety compliance assessment model, confirming the surface discharge of the power equipment through the ultraviolet coding, confirming the internal discharge of the power equipment through the UHF, and confirming the temperature rise of the power equipment through the short-wave infrared, and finally obtaining the hidden discharge detection result;

[0023] The insulation characteristics are input into the safety compliance assessment model, internal defects of the tool are confirmed by the short-wave infrared coding, and surface defects of the tool are confirmed by the visible light coding, and finally the insulation tool status detection result is obtained.

[0024] Furthermore, the steps of performing a preliminary safety assessment on the operator based on the noise-removed operation image include:

[0025] Performing target detection on the noise-removed operation image to define the position range of the operator and the position range of the power equipment, thereby calculating the actual distance between the operator and the power equipment, and determining whether the operator is safe based on the actual distance;

[0026] Performing posture estimation on the operator, extracting key action features of the operator from the posture estimation results, and comparing the key action features with pre-defined safety standard operating procedures to determine whether the operator's operation violates regulations;

[0027] Based on the noise-removed work image, the insulating suit, helmet, and insulating gloves are identified to determine whether the worker's protective equipment is complete;

[0028] If there is at least one violation among the operator, the operator's operation, or the operator's protective equipment, an early warning is issued.

[0029] Furthermore, the step of obtaining the preliminary operating state of the power equipment according to the second quantum state component includes:

[0030] extracting electrical characteristics corresponding to the operating state of the power device and dynamic characteristics of the power device from the second quantum state component;

[0031] Determine the indicator range of the power equipment under normal operating conditions according to the type, specifications and operating requirements of the power equipment;

[0032] The electrical characteristics and the dynamic characteristics are compared with the indicator range of normal operation. If at least one indicator is not within the indicator range, an early warning is issued.

[0033] Furthermore, the optical sensor includes a visible light camera, a near infrared camera, a short wave infrared detector and an ultraviolet sensor, and the electrical sensor includes a power frequency electric field sensor and a radio frequency sensor;

[0034] Among them, the visible light camera is installed on the top of the gantry, the operating platform of the insulated boom truck and the corner of the GIS room; the near-infrared camera is installed in the observation window of the circuit breaker arc extinguishing chamber; the short-wave infrared detector is installed at the handle of the insulating tool and the heat sink area of the transformer; and the ultraviolet sensor is installed at the flange of the GIS equipment and at both ends of the insulator string;

[0035] The power frequency electric field sensor is installed on the top of the operator's safety helmet, the radio frequency sensor is installed on the grounding down conductor of the transformer equipment, and the ultra-high frequency partial discharge sensor is installed on the flange of the GIS equipment housing.

[0036] This invention proposes a power operation safety and compliance management system that integrates multimodal artificial intelligence technology. By installing various types of optical and electrical sensors at power operation sites and processing the collected optical and electrical data, this embodiment of the invention offers the following advantages over traditional solutions:

[0037] (1) Perform multi-scale frequency domain decomposition on the work site image to obtain different frequency components, and then perform dry denoising on the different frequency components to obtain the first quantum state component. Multi-scale frequency domain decomposition combined with quantum dry denoising can better balance detail preservation and noise removal in the work site image. Through multi-scale frequency domain decomposition, the different frequency components of the work site image can be separated. High-frequency components usually contain detailed information of the image, such as minor damage to equipment, subtle changes in personnel movements, etc., while low-frequency components contain basic information such as the outline of the image. Quantum dry denoising processes different frequency components in the quantum state space, which can specifically remove the quantum state corresponding to the noise and reduce the damage to the detailed information.

[0038] (2) For electrical data, after mapping it to a quantum bit state, phase adjustment is performed to obtain a second quantum state component that is not interfered with by noise. The quantum denoising method of mapping electrical data to a quantum bit state and performing phase adjustment can achieve high-precision noise suppression at the quantum level. In the quantum state, characteristics such as quantum coherence can be used to more finely process noise, effectively improving the signal-to-noise ratio.

[0039] (3) The quantum encoding of the first quantum state component, the second quantum state component, and the magnetic data is subjected to feature extraction according to the strong magnetic combination, the hidden discharge combination, and the tool combination. Only the required features are extracted for quantum entanglement fusion. This not only reduces the amount of data processing and improves the response speed of the power operation safety compliance management system, but also explores the complex correlation between different modal data through quantum entanglement, thereby improving the accuracy of subsequent strong electromagnetic field interference, hidden discharge, and insulation tool status detection. Compared with the direct splicing of features or weighted fusion in traditional solutions, it can improve the accuracy of behavior recognition; at the same time, it also helps to further improve the real-time performance of the entire system.

[0040] (4) Based on the noise-removed operation image, preliminary safety inspections are conducted on the operator's location, operation procedures, and safety equipment. Based on the second quantum state component, the preliminary operating status of the power equipment is inspected. Preliminary inspections using single-mode sensor data can improve the system's response speed. For ultra-high frequency power production systems, quantum entanglement fusion can be performed by fusing data from different types of sensors to more accurately detect problems such as strong electromagnetic field interference, hidden discharges, and insulating tools, solving the difficulty of controlling ultra-high frequency power production systems. BRIEF DESCRIPTION OF THE DRAWINGS

[0041] Figure 1 A flowchart of the processing process of a power operation safety and compliance management system that integrates multimodal artificial intelligence technology provided by an embodiment of the present invention;

[0042] Figure 2 A block diagram of a data processing process provided by an embodiment of the present invention.

[0043] The purpose, features and advantages of the present invention will be further described with reference to the accompanying drawings and in conjunction with the embodiments. DETAILED DESCRIPTION

[0044] The embodiments of the present application are described in detail below. Examples of the embodiments are shown in the accompanying drawings, wherein the same or similar reference numerals throughout represent the same or similar elements or elements having the same or similar functions. The embodiments described below with reference to the accompanying drawings are exemplary and are only used to explain the present application and are not to be construed as limiting the present application.

[0045] In order to enable those skilled in the art to better understand the solutions of the present application, the technical solutions in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without making creative efforts are within the scope of protection of this application.

[0046] In the embodiments of the present application, at least one refers to one or more; a plurality refers to two or more. In the description of the present application, words such as "first", "second", and "third" are only used for the purpose of distinguishing descriptions, and cannot be understood as indicating or implying relative importance, nor can they be understood as indicating or implying order. In addition, the terms "first" and "second" are only used for descriptive purposes, and cannot be understood as indicating or implying relative importance or implicitly indicating the number of technical features indicated. Therefore, the features defined as "first" and "second" may explicitly or implicitly include one or more of the features. In the description of the present application, the meaning of "plurality" is two or more, unless otherwise clearly and specifically defined.

[0047] References to "one embodiment" or "some embodiments" in this specification mean that a particular feature, structure, or characteristic described in conjunction with that embodiment is included in one or more embodiments of the present application. Thus, the terms "including," "comprising," "having," and their variations in this specification all mean "including but not limited to," unless otherwise specifically stated.

[0048] It should be noted that in the embodiments of the present application, "connection" can be understood as electrical connection, and the connection between two electrical components can be a direct or indirect connection between the two electrical components. For example, the connection between A and B can be either a direct connection between A and B or an indirect connection between A and B through one or more other electrical components.

[0049] The present invention provides a power operation safety management and control system primarily for use at ultra-high frequency (UHF) power production sites. These complex, high-risk environments present significant challenges for traditional safety management methods. These sites can be subject to strong electromagnetic interference, which can dramatically reduce the signal-to-noise ratio of power equipment and increase false alarm rates. Furthermore, the dense reflections from metal equipment can mask the true operating status.

[0050] like Figure 1 and Figure 2As shown, the power operation safety compliance management and control system includes optical sensors, electrical sensors, and magnetic sensors. The processor of the power operation safety compliance management and control system receives the work site images captured by the optical sensors, the electrical data collected by the electrical sensors, and the magnetic data received by the magnetic sensors, and processes them through the following steps:

[0051] The optical sensors include visible light cameras, near infrared cameras, short wave infrared detectors and ultraviolet sensors, and the electrical sensors include power frequency electric field sensors and radio frequency sensors.

[0052] Among them, the visible light camera is installed on the top of the gantry, the operating platform of the insulated boom truck and the corner of the GIS room; the near-infrared camera is installed in the observation window of the circuit breaker arc extinguishing chamber; the short-wave infrared detector is installed at the handle of the insulating tool and the heat sink area of the transformer; and the ultraviolet sensor is installed at the flange of the GIS equipment and at both ends of the insulator string;

[0053] The power frequency electric field sensor is installed on the top of the operator's safety helmet, the radio frequency sensor is installed on the grounding down conductor of the transformer equipment, and the ultra-high frequency partial discharge sensor is installed on the flange of the GIS equipment housing.

[0054] Furthermore, the magnetic sensor is a magnetic field anisotropy probe installed at the boundary of the working area.

[0055] Among them, the industrial frequency electric field sensor installed on the top of the operator's safety helmet can monitor the industrial frequency electric field strength around the operator in real time. By monitoring the industrial frequency electric field strength, it can determine whether the operator is approaching a high electric field area, thereby avoiding harm to the operator due to excessively high electric fields, such as the risk of electric shock.

[0056] Magnetic field anisotropy probes installed at the perimeter of the operating area can detect changes in the magnetic field. During the operation of power equipment, abnormal changes in the magnetic field may indicate a malfunction or potential safety hazard. For example, an internal short circuit in a transformer or other equipment may cause an abnormal magnetic field.

[0057] The RF sensor located at the grounding down conductor of the transformer equipment can capture RF signals. Transformers and other power equipment will generate RF signals when problems such as insulation aging and partial discharge occur. The RF sensor can detect these potential faults in advance so that timely measures can be taken to prevent the fault from expanding.

[0058] Ultra High Frequency (UHF) partial discharge sensors, installed on the housing flange of GIS (Gas Insulated Switchgear) equipment, monitor partial discharge (PD) within the equipment. Insulation failures within GIS equipment often precede partial discharge (PD). UHF partial discharge sensors can sensitively capture these weak discharge signals and assess the insulation condition of the GIS equipment.

[0059] Visible light cameras installed on the gantry top, on the insulated boom truck operating platform, and in the corners of the GIS room capture visible light images of the work site. These images provide a visual display of personnel operations, equipment status, and surrounding environmental conditions. For example, they can monitor whether workers are following operating procedures and whether there is any visible damage or foreign matter attached to the equipment.

[0060] A near-infrared sensor installed in the circuit breaker's arc chamber observation window detects near-infrared information from the chamber. The near-infrared spectrum can reveal the temperature distribution of the equipment. During power equipment operation, abnormally high temperatures often occur at faulty locations, and near-infrared sensors can promptly detect these overheated areas.

[0061] Shortwave infrared sensors installed on the handles of insulating tools and transformer heat sinks monitor shortwave infrared information. For insulating tools, shortwave infrared can detect aging and moisture, as these factors alter the infrared characteristics of insulating tools. For transformer heat sinks, shortwave infrared can help monitor uniform temperature distribution and ensure proper heat dissipation.

[0062] UV sensors installed at the GIS equipment flange and at both ends of the insulator strings can detect UV information. Corona discharge in power equipment generates UV light, which the UV sensors can detect to determine if corona discharge is occurring. Corona discharge can degrade the insulation performance of the equipment, leading to failures.

[0063] The embodiment of the present invention installs various optical sensors, electrical sensors, and magnetic sensors, and processes the data collected by these sensors to perform compliance management and control on the safety of power operations.

[0064] S10, performing multi-scale frequency domain decomposition on the work site image to obtain different frequency components, performing quantum dry denoising on the different frequency components to obtain dry denoised first quantum state components, converting the first quantum state components into a denoised work image, and performing a preliminary safety assessment of the workers based on the denoised work image;

[0065] S10.1, converting the work site image from the spatial domain to the frequency domain, decomposing it to obtain different frequency components at multiple levels, and obtaining image data corresponding to the different frequency components;

[0066] The work site image is converted from the spatial domain to the frequency domain and its frequency components are analyzed at different scales. Common methods include wavelet transform. For an image of power equipment at a work site, wavelet transform can decompose it into low-frequency sub-bands (containing basic information such as the image outline) and high-frequency sub-bands (containing image details, edges, etc.). This allows the image's frequency components to be obtained at different scales, yielding image data corresponding to different frequency components.

[0067] S10.2, mapping image data corresponding to different frequency components to a first quantum bit state, and performing phase adjustment on the first quantum bit state to obtain a first quantum state component for drying and denoising;

[0068] Image data with different frequency components is mapped to the first qubit state. This can be achieved through quantum encoding, for example, by encoding information such as the grayscale value of an image pixel into the qubit state. Phase adjustments and other quantum gate operations are then performed on the first qubit state to obtain the first quantum state component for drying and denoising. These operations are based on pre-designed quantum algorithms designed to identify and suppress quantum states corresponding to noise. For example, quantum rotation gates are used to adjust the qubit phase, reducing the coherence of the noise state and, in turn, reducing the noise.

[0069] S10.3, measuring the first quantum state component to obtain the values of different frequency components after noise elimination, and combining the values of different frequency components after noise elimination to obtain a noise elimination operation image.

[0070] The state of the first qubit, after quantum drying and denoising, is measured to obtain the values of the different frequency components after denoising. These frequency components are then combined using methods such as the inverse discrete wavelet transform, following the previously described multi-scale frequency domain decomposition method. Using the wavelet transform as an example, the denoised low-frequency and high-frequency subbands are combined using the inverse decomposition process, ultimately yielding a denoised image of the worksite.

[0071] In embodiments of the present invention, multi-scale frequency domain decomposition combined with quantum drying denoising can achieve a better balance between detail preservation and noise removal. Multi-scale frequency domain decomposition can separate the different frequency components of an image. High-frequency components typically contain detailed image information, such as minor damage to equipment and subtle changes in human movement, while low-frequency components contain basic information, such as the image's outline. Quantum drying denoising processes different frequency components in quantum state space, specifically removing the quantum states corresponding to the noise and minimizing the damage to detailed information.

[0072] For example, when detecting tiny cracks in electrical equipment, traditional filtering methods may blur the edges of the cracks. This quantum denoising method can remove noise interference while retaining the detailed features of the cracks more clearly, thereby more accurately judging whether the equipment appearance is safe.

[0073] S10.4, performing target detection on the de-noised work image to define the position range of the worker and the position range of the power equipment, thereby calculating the actual distance between the worker and the power equipment, and determining whether the worker is safe based on the actual distance;

[0074] Computer vision techniques (such as object detection algorithms) are used to identify workers in de-noised work images. This can be achieved using trained deep learning models (such as Faster R-CNN and YOLO). These models can delineate the location of workers in the image and distinguish them from the background and other equipment.

[0075] Calculate the distance between workers and electrical equipment. This can be done by using pixel distances in the image, combined with known information about the actual dimensions of the equipment and scene (obtained through methods such as camera calibration), to estimate the true distance. For example, this can determine whether workers maintain a safe distance from high-voltage equipment to avoid the risk of electric shock.

[0076] Determine the operator's position relative to the equipment, such as whether they are working from the front, side, or back of the equipment. Certain equipment operations may only be permitted from specific locations to ensure safety. For example, if equipment has cooling vents or ventilation openings, operators should avoid operating near these openings, which could create sparks and prevent fires.

[0077] S10.5, performing posture estimation on the operator, extracting key motion features of the operator from the posture estimation results, comparing the key motion features with pre-defined safety standard operating procedures, and determining whether the operator's operation violates regulations;

[0078] After determining the operator's position, posture estimation is further performed to understand the operator's body movements. This can be achieved using deep learning-based posture estimation models, such as OpenPose. These models construct a human posture skeleton by detecting and identifying key points on the human body, such as joints. For example, by identifying the position and angle of key parts such as a person's arms and legs, it can be determined whether the person is operating equipment, walking, or in some other state.

[0079] Key features of the operator's actions are extracted from the posture estimation results. For example, when operating electrical equipment, the contact between the operator's hand and the equipment's operating components (such as switches and knobs) is observed, including the contact location and the direction of the hand movement (such as rotation or pressing). The coordination between the operator's posture and the operator's actions is also considered, for example, whether the body maintains balance and stability when working at height.

[0080] The extracted operational action features are compared with predefined safety standard operating procedures. These safety standard operating procedures can be action sequences and requirements established by power companies based on equipment operating manuals and safety regulations. For example, when closing a circuit breaker, the standard procedure might include a person standing in a specified position, grasping the closing handle with their right hand, and pushing it vertically upward. By comparing the actual actions with the standard procedures, any violations can be determined.

[0081] S10.6, based on the noise-removed work image, identify the insulating suit, safety helmet, and insulating gloves to determine whether the worker's protective equipment is complete;

[0082] Identify the safety protective equipment worn by workers, such as insulating clothing, helmets, and insulating gloves, in the denoised work images. This can also be done using an object detection algorithm, by training a specialized protective equipment detection model to identify features such as the type, color, and shape of the equipment.

[0083] Compare the identified protective equipment with safety requirements for electrical work. Check that the protective equipment is complete and correctly worn. For example, details such as whether the chin strap of the helmet is fastened tightly and whether the insulating gloves cover the wrists are crucial to the effectiveness of personnel protection.

[0084] S10.7: If there is at least one violation among the operator, the operator's operation, or the operator's protective equipment, an early warning will be issued.

[0085] If at least one of the following three conditions exists: the operator is unsafe, the operator's operation is in violation of regulations, or the operator's protective equipment is incomplete, an early warning can be issued; if the operator is safe, the operator's operation is not in violation of regulations, and the operator's protective equipment is complete, it means that there is no risk related to the operator.

[0086] An embodiment of the present invention provides a method for processing work site images using quantum coding. Through quantum drying denoising, image data generated by electromagnetic interference can be removed. In addition, a preliminary safety assessment is performed on the denoised work images. Here, a single-modal visual image is used to evaluate the position of the workers, their operations, and their equipment, thereby reducing the amount of data processing and simplifying the safety management process.

[0087] S20, mapping the electrical data to a quantum bit state and performing phase adjustment to obtain a second quantum state component that is not interfered with by noise, and obtaining a preliminary operating state of the power equipment based on the second quantum state component;

[0088] S20.1, encoding the electrical data using three qubits and selecting the first three bits to obtain a second qubit state;

[0089] Select an appropriate quantum encoding scheme based on the characteristics of the electrical data and the requirements of quantum computing. Common options include binary encoding and angle encoding. Taking binary encoding as an example, if you want to map a normalized electrical data value (say, 0.6) to the state of the second qubit, and use three qubits for encoding, you can convert 0.6 into binary (e.g., 0.10011...), select the first three digits (0.10), and then initialize the qubit to the corresponding state through a quantum gate operation. For example, the |010> state, where | and > represent quantum states, and 0 and 1 represent the two ground states of the qubit.

[0090] S20.2, establishing a noise model in the electrical data, and designing a phase adjustment strategy based on the noise model, and applying a quantum gate based on the phase adjustment strategy, thereby changing the phase of the second quantum bit state through the quantum gate to obtain a second quantum state component that is not interfered with by noise.

[0091] In quantum computing, quantum gates are the basic units for operating on qubits, similar to logic gates in classical computers. For phase adjustment, commonly used quantum gates include phase gates (such as P-gates). Phase gates can change the phase of a qubit without changing its probability amplitude.

[0092] To remove noise interference, it is necessary to establish a noise model for the electrical data. Noise in quantum states can manifest as random phase jitter or small changes in probability amplitude. Based on the noise model, a phase adjustment strategy can be designed. For example, if noise causes the phase of a qubit to shift with a certain probability, the phase can be corrected by periodically applying a phase gate. Assuming that noise causes the phase to shift by π / 4 with a certain probability, correction can be achieved by applying a reverse phase gate of -π / 4.

[0093] After the phase of the second quantum bit state is adjusted, a measurement is required to obtain the second quantum state component that is not disturbed by noise.

[0094] S20.3, extracting electrical characteristics corresponding to the operating state of the power device and dynamic characteristics of the power device from the second quantum state component;

[0095] Electrical characteristics closely related to the operating status of the power equipment are extracted from the second quantum state component. For example, for the quantum state component representing the current, information such as the current amplitude and phase can be extracted; for the quantum state component related to the electric field strength, parameters such as the magnitude and direction of the electric field strength can be obtained. These parameters can be extracted through quantum state measurement and inverse mapping, that is, converting the quantum state information back into a representation of electrical physical quantities.

[0096] The quantum noise reduction method of this invention, which maps electrical data to quantum bit states and performs phase adjustment, can achieve high-precision noise suppression at the quantum level. In the quantum state, properties such as quantum coherence can be exploited to more precisely process noise, effectively improving the signal-to-noise ratio.

[0097] Quantum state changes are extremely sensitive to tiny variations in physical quantities. During the operation of power equipment, some early signs of failure often manifest as small changes in electrical parameters. The second quantum state component obtained through quantum noise removal can better capture these subtle changes.

[0098] For example, during the initial aging of insulation materials within a device, electrical parameters such as the dielectric constant undergo subtle changes, which are reflected in the quantum state components. Traditional electrical data processing methods may overlook these subtle changes, resulting in a failure to detect potential equipment failures in a timely manner. Quantum noise reduction methods can help detect these issues in advance, providing more timely assessments of the device's safe operation.

[0099] S20.4. Determine the indicator range for the normal operation of the power equipment based on the type, specifications, and operating requirements of the power equipment;

[0100] Determine the index ranges for various electrical parameters and dynamic characteristics under normal operating conditions based on the type, specifications, and operating requirements of the power equipment. For example, for transformer equipment, parameters such as no-load current and load loss have certain standard ranges during normal operation. These ranges can be determined from the equipment manual, industry standards, or historical normal operating data.

[0101] In addition to static electrical parameters, it's also necessary to analyze the dynamic characteristics of power equipment operation. By observing how quantum state components change over time, dynamic characteristics such as frequency and rate of change can be extracted. For example, by analyzing the frequency variations of current quantum state components, it's possible to determine whether the equipment is operating at the normal operating frequency (50Hz or 60Hz) or whether there are abnormalities such as harmonics.

[0102] S20.5: Compare the electrical characteristics and the dynamic characteristics with the normal operating indicator range. If at least one indicator is outside the indicator range, issue an early warning.

[0103] The extracted electrical parameters and dynamic characteristics are compared with the normal operating range. If all parameters are within the normal range, the device can be preliminarily judged to be operating normally. For example, if parameters such as current amplitude, frequency, and device temperature all meet the normal operating standards for a transformer, the transformer can be considered to be operating normally.

[0104] The embodiment of the present invention also performs a simple analysis on the single-mode electrical data to quickly screen out abnormal indicators and issue an early warning.

[0105] S30, performing quantum entanglement fusion on the first quantum state component, the second quantum state component and the quantum encoding of the magnetic data according to the strong magnetic combination, the hidden discharge combination and the tool combination, respectively, and inputting the fused feature vectors into the safety compliance assessment model respectively to obtain the strong electromagnetic field interference result, the concealed discharge detection result and the insulating tool status detection result.

[0106] Ultra-high-frequency power production operations can be subject to extremely strong electromagnetic interference, which can dramatically reduce the signal-to-noise ratio of monitoring equipment and increase false alarm rates. Furthermore, dense reflections from metal equipment can mask the true operating status. High-frequency partial discharge (PD) and insulation defects can also interfere with detection by generating noise above 300 MHz. Single-modal data cannot directly identify phenomena such as EMI, metal equipment reflections, and high-frequency partial discharge. Therefore, multimodal data fusion is necessary to identify these phenomena.

[0107] S30.1, extracting the visible light code and near-infrared code from the first quantum state component according to the strong magnetic combination, and performing quantum entanglement fusion with the quantum code of the magnetic data to obtain a strong magnetic feature;

[0108] Among them, the visible light encoding is the encoding in the first quantum state component corresponding to the visible light data, the near-infrared encoding is the encoding in the first quantum state component corresponding to the near-infrared data, and the quantum encoding of the magnetic data is obtained. The visible light encoding, near-infrared encoding and the quantum encoding of the magnetic data are quantum entangled and fused to obtain strong magnetic characteristics.

[0109] S30.2, according to the concealment and release combination, extract the ultraviolet code and shortwave infrared code from the first quantum state component, extract the UHF code from the second quantum state component, and perform quantum entanglement fusion to obtain the concealment and release characteristics;

[0110] Among them, the ultraviolet code is the code in the first quantum state component corresponding to the ultraviolet data, the shortwave infrared code is the code in the first quantum state component corresponding to the shortwave infrared data, and the UHF code is the code in the second quantum state component corresponding to the ultrahigh frequency partial discharge data. The ultraviolet code, shortwave infrared code and UHF code are quantum entangled and fused to obtain the hidden release characteristics.

[0111] S30.3, according to the insulation combination, extract the short-wave infrared code and visible light code in the first quantum state component, and perform quantum entanglement fusion to obtain insulation characteristics.

[0112] The visible light code corresponds to the short-wave infrared code in the first quantum state component. The short-wave infrared code and the visible light code are quantum entangled to obtain the insulation characteristics.

[0113] Multimodal fusion methods based on quantum entanglement can fully utilize quantum effects to capture the complex correlations between modalities. Compared with traditional feature splicing or weighted fusion, they can improve the accuracy of behavior recognition. At the same time, the parallelism of quantum computing also helps to improve the real-time performance of the entire system.

[0114] S30.4, inputting the strong magnetic characteristics into the safety compliance assessment model, and determining the strong electromagnetic field interference result through multi-spectral cross-validation;

[0115] By performing multi-spectral cross-validation on strong magnetic features, it is possible to determine whether there is strong electromagnetic field interference. Through multimodal data fusion, the shortcomings of traditional filtering algorithms that cause signal delays and lose transient danger signals are solved.

[0116] S30.5, inputting the concealed discharge characteristics into the safety compliance assessment model, confirming surface discharge of the power equipment using the UV coding, confirming internal discharge of the power equipment using the UHF coding, and confirming temperature rise of the power equipment using the shortwave infrared coding, ultimately obtaining the concealed discharge detection result;

[0117] Ultraviolet coding is used to determine whether there is discharge on the surface of the power equipment, UHF data is used to determine whether there is discharge inside the power equipment, and short-wave infrared data is used to determine whether the internal temperature of the power equipment rises. If at least one of surface discharge, internal discharge, and internal temperature rise is confirmed, then it can be confirmed that the power equipment has hidden discharge. Otherwise, it can be confirmed that the power equipment does not have hidden discharge.

[0118] S30.6, inputting the insulation characteristics into the safety compliance assessment model, confirming the internal defects of the tool through the short-wave infrared coding, confirming the surface defects of the tool through the visible light coding, and finally obtaining the insulation tool status detection result.

[0119] Internal defects of the insulating tool are discovered through short-wave infrared, and surface defects of the insulating tool are checked through visible light data. The status inspection result of the insulating tool is determined based on these two results. If there are no internal defects or surface defects, the status inspection result of the insulating tool is normal. Otherwise, the insulating tool is faulty.

[0120] The quantum drying denoising process involves quantum state manipulation, which has the potential to enhance image information to a certain extent. In quantum state space, through properties such as quantum entanglement, hidden correlations between different frequency components of an image can be discovered. This correlation information may be difficult to detect in traditional image space, but it can be of great value in security assessments.

[0121] Initial detection using single-mode sensor data can improve the system's response speed. For ultra-high-frequency power production operations, quantum entanglement fusion, by fusing data from different sensor types, can more accurately detect problems such as strong electromagnetic interference, hidden discharges, and insulated tools, thus resolving the difficulty of controlling ultra-high-frequency power production operations.

[0122] Those skilled in the art will clearly understand that for the sake of convenience and brevity of description, only the division of the above-mentioned functional units and modules is used as an example. In actual applications, the above-mentioned functions can be distributed and completed by different functional units and modules as needed, that is, the internal structure of the device can be divided into different functional units or modules to complete all or part of the functions described above.

[0123] The embodiments described above are only used to illustrate the technical solutions of the present invention, rather than to limit the same. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. These modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the various embodiments of the present invention, and should all be included in the scope of protection of the present invention.

Claims

1. A power operation safety compliance management and control system integrating multimodal artificial intelligence technology, characterized by: The system includes installing optical sensors, electrical sensors, and magnetic sensors. The processor of the power operation safety compliance management system receives the work site images captured by the optical sensors, the electrical data collected by the electrical sensors, and the magnetic data received by the magnetic sensors, and processes the data through the following steps: Performing multi-scale frequency domain decomposition on the work site image to obtain different frequency components, performing quantum dry denoising on the different frequency components to obtain dry denoised first quantum state components, converting the first quantum state components into a denoised work image, and performing a preliminary safety assessment of the workers based on the denoised work image; Mapping the electrical data to a quantum bit state and performing phase adjustment to obtain a second quantum state component that is not interfered with by noise, and obtaining a preliminary operating state of the power device based on the second quantum state component; The first quantum state component, the second quantum state component and the quantum encoding of the magnetic data are quantum entangled and fused according to the strong magnetic combination, the hidden discharge combination and the tool combination, and the fused feature vectors are respectively input into the safety compliance assessment model to obtain the strong electromagnetic field interference results, the hidden discharge detection results and the insulating tool status detection results.

2. The electric power operation safety compliance management and control system integrating multimodal artificial intelligence technology according to claim 1 is characterized in that: The steps of performing multi-scale frequency domain decomposition on the work site image to obtain different frequency components, performing quantum drying denoising on the different frequency components to obtain first quantum state components subjected to drying denoising, and then converting the first quantum state components into a denoised work image include: Converting the work site image from the spatial domain to the frequency domain, decomposing it to obtain different frequency components at multiple levels, and obtaining image data corresponding to the different frequency components; Mapping image data corresponding to different frequency components to a first quantum bit state, and performing phase adjustment on the first quantum bit state to obtain a first quantum state component for drying and denoising; The first quantum state component is measured to obtain the values of different frequency components after noise elimination, and the values of different frequency components after noise elimination are combined to obtain the operation image after noise elimination.

3. The electric power operation safety compliance management and control system integrating multimodal artificial intelligence technology according to claim 1 is characterized in that: The steps of mapping the electrical data to a quantum bit state and performing phase adjustment to obtain a second quantum state component that is not interfered with by noise include: Encoding the electrical data using three qubits and selecting the first three bits to obtain a second qubit state; A noise model in the electrical data is established, and a phase adjustment strategy is designed based on the noise model. A quantum gate is applied based on the phase adjustment strategy to change the phase of the second quantum bit state through the quantum gate to obtain a second quantum state component that is not interfered with by noise.

4. The electric power operation safety compliance management and control system integrating multimodal artificial intelligence technology according to claim 1 is characterized in that: The quantum encoding of the first quantum state component, the second quantum state component and the magnetic data is quantum entangled and fused according to a strong magnetic combination, a hidden release combination and an insulating combination, respectively, comprising: According to the strong magnetic combination, the visible light code and the near infrared code in the first quantum state component are extracted, and quantum entangled and fused with the quantum code of the magnetic data to obtain a strong magnetic feature; According to the concealment and release combination, the ultraviolet code and shortwave infrared code in the first quantum state component are extracted, the UHF code of the second quantum state component is extracted, and quantum entanglement fusion is performed to obtain the concealment and release characteristics; According to the insulating combination, the short-wave infrared code and the visible light code in the first quantum state component are extracted, and quantum entanglement fusion is performed to obtain the insulating characteristics.

5. The electric power operation safety compliance management and control system integrating multimodal artificial intelligence technology according to claim 4 is characterized in that: The steps of inputting the fused feature vectors into the safety compliance assessment model respectively to obtain the strong electromagnetic field interference result, the hidden discharge detection result and the insulation tool status detection result include: Inputting the strong magnetic characteristics into the safety compliance assessment model, and determining the strong electromagnetic field interference result through multi-spectral cross-validation; Inputting the hidden discharge characteristics into the safety compliance assessment model, confirming the surface discharge of the power equipment through the ultraviolet coding, confirming the internal discharge of the power equipment through the UHF, and confirming the temperature rise of the power equipment through the short-wave infrared, and finally obtaining the hidden discharge detection result; The insulation characteristics are input into the safety compliance assessment model, internal defects of the tool are confirmed by the short-wave infrared coding, and surface defects of the tool are confirmed by the visible light coding, and finally the insulation tool status detection result is obtained.

6. The electric power operation safety compliance management and control system integrating multimodal artificial intelligence technology according to claim 1 is characterized in that: The steps of performing a preliminary safety assessment on the operator based on the noise-removed operation image include: Performing target detection on the noise-removed operation image to define the position range of the operator and the position range of the power equipment, thereby calculating the actual distance between the operator and the power equipment, and determining whether the operator is safe based on the actual distance; Performing posture estimation on the operator, extracting key action features of the operator from the posture estimation results, and comparing the key action features with pre-defined safety standard operating procedures to determine whether the operator's operation violates regulations; Based on the noise-removed work image, the insulating suit, helmet, and insulating gloves are identified to determine whether the worker's protective equipment is complete; If there is at least one violation among the operator, the operator's operation, or the operator's protective equipment, an early warning is issued.

7. The electric power operation safety compliance management and control system integrating multimodal artificial intelligence technology according to claim 1 is characterized in that: The step of obtaining a preliminary operating state of the power equipment according to the second quantum state component comprises: extracting electrical characteristics corresponding to the operating state of the power device and dynamic characteristics of the power device from the second quantum state component; Determine the indicator range of the power equipment under normal operating conditions according to the type, specifications and operating requirements of the power equipment; The electrical characteristics and the dynamic characteristics are compared with the indicator range of normal operation. If at least one indicator is not within the indicator range, an early warning is issued.

8. The electric power operation safety compliance management and control system integrating multimodal artificial intelligence technology according to claim 1 is characterized in that: The optical sensors include visible light cameras, near infrared cameras, short wave infrared detectors and ultraviolet sensors, and the electrical sensors include power frequency electric field sensors and radio frequency sensors; Among them, the visible light camera is installed on the top of the gantry, the operating platform of the insulated boom truck and the corner of the GIS room; the near-infrared camera is installed in the observation window of the circuit breaker arc extinguishing chamber; the short-wave infrared detector is installed at the handle of the insulating tool and the heat sink area of the transformer; and the ultraviolet sensor is installed at the flange of the GIS equipment and at both ends of the insulator string; The power frequency electric field sensor is installed on the top of the operator's safety helmet, the radio frequency sensor is installed on the grounding down conductor of the transformer equipment, and the ultra-high frequency partial discharge sensor is installed on the flange of the GIS equipment housing.

Images