A power plant cable trench intelligent robot inspection and adjustment method, system, device and medium

Abstract

The application discloses a power plant cable trench intelligent robot inspection and adjustment method, system, equipment and medium, and belongs to the technical field of robot control inspection systems, and comprises the following steps: collecting environment data, performing data fusion to output fused environment data vectors, and constructing an inspection control system; based on input data, using vibration and microphone array sound wave analysis arranged by the inspection network control system, performing environment initial simulation, generating a three-dimensional map with an abnormal point marker for path planning, the inspection network control system performing torque control of an execution mechanism, the inspection network control system issuing torque commands of the execution mechanism, performing inspection end trajectory planning and hybrid control, the inspection network control system issuing execution movement, and establishing multi-inspection control system communication for inspection cooperation. The application generates a stable and safe inspection sequence, realizes flexible and stable movement in a narrow space, improves the adaptability and energy efficiency of complex terrains, and realizes abnormal cooperative response and dynamic task allocation.

Description

Technical Field

[0001] This invention relates to the field of robot control and inspection system technology, specifically to a method, system, equipment, and medium for intelligent robot inspection and adjustment of power plant cable trenches. Background Technology

[0002] As a key channel for cable laying, the internal environment of power plant cable trenches is complex and variable. They often have high temperature, high humidity, dense obstacles, and abnormal phenomena caused by cable aging, such as local overheating, insulation damage, or mechanical loosening.

[0003] Currently, traditional inspection methods present safety risks associated with manual inspections. For example, cable trenches are often in narrow, hot, and humid environments, and may contain toxic gases (such as SF6 and CO) and pose a risk of electric shock. Manual inspections can easily lead to poisoning, suffocation, or accidents. Manual inspections rely on experience and are prone to missing hidden defects (such as partial discharge and insulation degradation). Furthermore, the complex structure of the trenches (such as corners and multi-layered cable installations) makes comprehensive coverage difficult. Traditional wheeled or track-based robots cannot adapt to the complex terrain of cable trenches (such as stairs, gravel, and water accumulation), and their stability is insufficient in environments with high electromagnetic interference and high temperature and humidity. Manual recording is prone to errors, lacks real-time data analysis and early warning, and has a fault response time that can be as long as several hours, potentially leading to power outages or fires. Existing robots have low levels of intelligence, poor positioning accuracy, and unstable motion control, making it difficult to achieve autonomous inspections and collaborative work.

[0004] Furthermore, relying solely on temperature or humidity data cannot fully reflect the overall condition of cable trenches. If vibration and acoustic spectrum parameters are not integrated for analysis, it is difficult to accurately identify hidden anomalies such as cable loosening or voids, which can easily lead to missed or false alarms. Existing SLAM technology often does not integrate anomaly point markers in its 3D maps, and path planning often adopts a single-layer strategy, which cannot achieve bidirectional collaboration between local and global aspects, resulting in the robot being unable to dynamically avoid high-risk areas or unknown obstacles.

[0005] To address these shortcomings, this invention significantly improves the comprehensiveness, adaptability, and efficiency of inspection by using multi-sensor data fusion to output environmental vectors, a three-dimensional map of anomaly markers based on sound waves and vibrations, bidirectional path planning for both local and global environments, joint torque control constrained by minimum turning radius, hybrid control of foot trajectory, and intelligent collaboration among multiple robots. Summary of the Invention

[0006] In view of the above-mentioned problems, the present invention provides a method, system, equipment and medium for intelligent robot inspection and adjustment of power plant cable trenches.

[0007] Therefore, this invention aims to achieve autonomous, efficient, and safe inspection of power plant cable trenches through an intelligent legged robot. By integrating multi-sensor fusion, real-time navigation control, and intelligent decision-making, it solves the pain points of manual inspection and traditional robot inspection, and improves inspection efficiency and reliability.

[0008] To solve the above-mentioned technical problems, the present invention provides the following technical solution: a method for intelligent robot inspection and adjustment of power plant cable trenches, comprising,

[0009] Sensors are deployed to collect environmental data, and the data is fused to output a fused environmental data vector, including temperature, humidity, obstacle location, and acoustic spectrum parameters, to construct an inspection control system. Based on the input data, acoustic wave analysis is performed using vibration and microphone arrays deployed by the inspection network control system to conduct initial environmental simulation, construct a 3D map, and mark anomalies. Based on the 3D map with anomaly markers, local and global bidirectional path planning is performed to determine the inspection path sequence. Based on the inspection path, the inspection network control system uses the minimum turning radius as a constraint to perform torque control on the actuators. The inspection network control system issues torque commands to the actuators, including the target torque and angle of each actuator. Based on the actuator torque commands and the 3D map, the end-point trajectory planning and hybrid control are performed. The inspection network control system issues the executed motion, establishes communication between multiple inspection control systems for inspection collaboration, and feeds back the executed motion to the communication network.

[0010] As a preferred embodiment of the intelligent robot inspection and adjustment method for power plant cable trenches described in this invention, the environmental data collection includes deploying temperature sensors, humidity sensors, ultrasonic sensors, thermal imaging sensors, vibration sensors, and microphone arrays in the power plant cable trenches to construct an inspection and control system.

[0011] The raw data collected is preprocessed, including filtering and denoising. The preprocessed data stream is received, and a weighted average algorithm is used for fusion. The sensor data weights are automatically adjusted.

[0012] The fused data is formatted into an environmental data vector, including time series and spatial coordinates, and transmitted to the inspection and control system via API for initial environmental simulation. All motion systems are connected to the inspection and control system for unified management.

[0013] As a preferred embodiment of the intelligent robot inspection and adjustment method for power plant cable trenches described in this invention, the initial environmental simulation includes preprocessing the acoustic signal and vibration data V in the fused environmental data vector, including bandpass filtering and noise reduction, outputting cleaned data, and using a fast Fourier transform to convert the acoustic signal into a spectrum. Extract frequency domain features;

[0014] Key characteristic parameters, including peak frequency, low-frequency energy ratio, and high-frequency attenuation coefficient, are extracted from vibration data from the amplitude spectrum.

[0015] Based on the preprocessed spectrum and vibration data, an initial environmental simulation is performed to calculate the spectrum variation. Using historical data under normal conditions as the baseline spectrum, the difference between the current spectrum and the baseline spectrum is used. This indicates the difference in the current spectrum and calculates the anomaly probability. ;

[0016] If P ≥ 0.5, it is marked as an anomaly and further classified according to predefined rules. > If the type is uniformly defined as void; if > The type is uniformly defined as loose, among which, The threshold for spectral variation. The vibration amplitude threshold. The average amplitude of the vibration;

[0017] All abnormal locations are marked and compiled into an anomaly list, including anomaly type, location, and anomaly probability. A 3D map is then constructed. The independent inspection and control system uses positioning and mapping technology, integrating LiDAR and visual sensor data, to generate a 3D map of the cable trench within the inspection network control system.

[0018] Anomalies are marked with different colors according to their type: red indicates cavities, and yellow indicates looseness. A 3D map is output to the inspection and control system's visualization interface and terminal.

[0019] As a preferred embodiment of the intelligent robot inspection and adjustment method for power plant cable trenches described in this invention, the method of determining the inspection path sequence includes, before generating a three-dimensional map with color markings, deploying edge computing sensors at the motion system and fixed nodes of the trench, and, based on the three-dimensional map, performing a two-layer inspection path planning layout for the inspection network control system, including a local planning layer and a global planning layer.

[0020] The global planning layer calculates the optimal path and the path cost function. The A* algorithm uses As a heuristic function, it outputs a global path sequence. , These are path point coordinates, based on global path cost and abnormal probability path risk avoidance rules;

[0021] Once the global path sequence is planned, the distance to obstacles from the edge computing sensors is transmitted in real time. Calculate the risk change and risk gradient of the local risk gradient assessment path points. Calculated as risk value Partial derivative with respect to position x The final optimal inspection path sequence is calculated based on the local and global planning layers, taking into account the stability of the global path and the impact of local risk gradients.

[0022]

[0023] in: For the final path sequence, To adjust the step size coefficient, This is the attenuation coefficient.

[0024] As a preferred embodiment of the intelligent robot inspection and adjustment method for power plant cable trenches described in this invention, the torque control of the actuator includes: after path planning is completed, the inspection network control system issues a start command, and the inspection control system receives the command and controls the motion system to perform inspection according to the planned path, corresponding to the inspection control system using the minimum turning radius. To constrain the system, adjust the torque of the actuator.

[0025] An interpolation algorithm is used to generate a continuous desired body trajectory of the motion system from the path sequence. The output desired trajectory parameters include the desired body position. Expected body speed Expected body orientation Expected body angular velocity and expected turning radius ;

[0026] The calculation steps for calculating the robot's current body acceleration based on the desired trajectory parameters involve using a PID controller to calculate the desired acceleration.

[0027] The target angle of the actuator is calculated using an inverse kinematics model.

[0028] Based on the current status of the enforcement agency, including from the perspective of the enforcement agency. and actuator angular velocity The torque of the actuator is calculated by mapping the body acceleration to the torque of the actuator based on the principle of inverse dynamics.

[0029] The calculation results of the torque of each actuator They are combined into an actuator torque command sequence H.

[0030] As a preferred embodiment of the intelligent robot inspection and adjustment method for power plant cable trenches described in this invention, the inspection end trajectory planning and hybrid control includes, based on the planned path and combined with the markings on the three-dimensional map, the inspection network control system uses a permanent magnet synchronous motor and vector control technology for trajectory planning and global hybrid control.

[0031] The corresponding separate inspection control system is planned at a high level, and spline interpolation is used to generate inspection trajectories. :

[0032]

[0033] in, It is a time parameter, and spline represents a cubic spline interpolation function;

[0034] Based on the inspection trajectory and the torque command sequence of the actuator, low-level planning is performed to calculate the real-time force application point and torque output, and to control the height and depth of the actuator end effector. ,in, It is a vertical force. It's an altitude error. It is the stiffness coefficient. It is angular velocity;

[0035] The adjusted end position and torque are obtained and then vector controlled by the permanent magnet synchronous motor as follows: ,in, It is the q-axis current. It is the torque constant;

[0036] Upon receiving a current command, the end motor is driven to move.

[0037] As a preferred embodiment of the intelligent robot inspection and adjustment method for power plant cable trenches described in this invention, the inspection collaboration includes: establishing an overall communication network in the inspection control system of the power plant cable trench; all inspection control systems are self-organized and connected through a wireless Mesh network protocol; each inspection control system acts as an individual network node, receives responses from other nodes, periodically sends online signals during inspection to detect the liveness of network nodes, adjusts the routing table to cope with node movement or failure, and obtains network connection status and node list.

[0038] When the inspection and control system identifies an anomaly through the initial environment simulation, it generates a digital signal packet. The digital signal packet encapsulates the anomaly type, anomaly location, timestamp, the transmitting inspection and control system ID, and the current anomaly probability.

[0039] The digital signal packet transmission is based on the UDP retransmission mechanism, prioritizing the transmission of digital signal packets with high anomaly probability. Other inspection and control systems receive and analyze the digital signal packets, extract the source location and transmitting inspection and control system ID of the digital signal packet, and calculate the received signal strength indication.

[0040] Using a signal attenuation model, where the RSSI value is inversely proportional to the distance, and recording the reception timestamp, the decision module calculates the signal gradient. ;

[0041] if If the gradient change exceeds the preset threshold (set to 2dBm / m) and the task priority of the received inspection control system is lower than the abnormal response priority, then the system will autonomously decide to move in the reverse gradient direction.

[0042] When two or more motion systems decide to go to the same anomaly source, a graph search algorithm is used to replan the path sequence, aiming to minimize the total response time. Simultaneously, the distance between each motion system and the anomaly source is calculated, and priorities are assigned.

[0043] The system that is closest to the other system and has no overlapping path with other motion systems is given priority. The inspection and control system then controls the remaining motion systems to continue inspection according to the original path or adjust the path to avoid path conflicts. The updated path is broadcast to all motion systems through the communication network. After receiving the update, the system confirms synchronization and checks the synchronization status in real time. If synchronization fails, a retransmission mechanism is triggered.

[0044] This invention provides an intelligent robot inspection and adjustment system for power plant cable trenches.

[0045] To solve the above-mentioned technical problems, the present invention provides the following technical solution: an intelligent robot inspection and adjustment system for power plant cable trenches, comprising:

[0046] The data acquisition module deploys sensors to collect environmental data, performs data fusion, and outputs a fused environmental data vector, including temperature, humidity, obstacle location, and sound wave spectrum parameters, to build an inspection and control system.

[0047] The simulation module, based on the input data, uses the vibration and microphone array acoustic wave analysis deployed by the inspection network control system to perform initial environmental simulation, construct a three-dimensional map, and mark anomalies.

[0048] The inspection planning module performs bidirectional path planning, both local and global, based on a 3D map marked with anomaly points, to determine the inspection path sequence.

[0049] The control module, based on the inspection path inspection network control system, uses the minimum turning radius as a constraint to control the torque of the actuators. The inspection network control system issues torque commands to the actuators, including the target torque and angle of each actuator.

[0050] The hybrid control module performs end-of-line trajectory planning and hybrid control based on the actuator torque command and 3D map. The inspection network control system issues the execution motion, establishes communication between multiple inspection control systems for inspection collaboration, and feeds back the executed motion to the communication network.

[0051] The present invention provides a computer device, including a memory and a processor. The memory stores a computer program, and the processor executes the computer program to implement the steps of the intelligent robot inspection and adjustment method for power plant cable trenches.

[0052] The present invention provides a computer-readable storage medium having a computer program stored thereon, wherein the computer program, when executed by a processor, implements the steps of the intelligent robot inspection and adjustment method for power plant cable trenches.

[0053] The beneficial effects of this invention are as follows: The data fusion and automatic weight adjustment of this invention improve the accuracy and robustness of environmental data, reduce noise interference and false alarm risks, thereby enhancing the reliability of inspection decisions. Furthermore, the weighted average algorithm can dynamically adapt to environmental changes, ensuring data real-time performance and avoiding the problem of traditional single-sensor systems easily failing in complex environments.

[0054] Multimodal data analysis and anomaly probability calculation improve the accuracy and efficiency of anomaly detection, reducing missed or false detections. Color marking allows operators to intuitively identify anomaly types and locations, facilitating rapid response and maintenance. Meanwhile, SLAM technology, combined with acoustic and vibration data, enables reliable map building in low-light or complex terrain conditions, enhancing the robot's adaptability to harsh environments.

[0055] Bidirectional path planning achieves a balance between global efficiency and local safety, while risk gradient calculation enables the robot to respond to environmental changes in real time, improving the flexibility and robustness of the path. The introduction of mathematical formulas ensures the stability and predictability of path planning, reducing the probability of the robot getting stuck or out of control in narrow channels, thereby improving the overall inspection efficiency.

[0056] The robot can perform complex actions with a minimum turning radius, reducing energy consumption and mechanical wear. The PID controller and inverse dynamics model ensure smooth and accurate motion, avoiding the jitter or instability problems that occur when traditional control methods make sharp turns, thereby extending the robot's service life and improving its mobility in confined spaces.

[0057] Combining trajectory planning with torque output improves the robot's motion stability and adaptability on complex terrains, such as uneven surfaces in cable trenches; vector control optimizes motor efficiency, reduces energy loss, and extends battery life, while spline interpolation ensures smooth gait and reduces foot impact, thereby reducing damage to the robot and the environment.

[0058] The communication network enables intelligent collaboration and load balancing among multiple robots; signal gradient analysis and path replanning ensure that high-priority anomalies are responded to quickly, reducing the overall inspection response time; the graph search algorithm avoids task overlap and path conflicts between robots, improving system efficiency and reliability, thereby achieving full-coverage inspection in large power plant trenches. Attached Figure Description

[0059] To more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0060] Figure 1 This is a flowchart illustrating the overall process of an intelligent robot inspection and adjustment method for power plant cable trenches, as provided in one embodiment of the present invention. Detailed Implementation

[0061] To make this invention more easily understood, the specific embodiments of the invention will be described in detail below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of this invention, not all of them. Based on the embodiments of this invention, all other embodiments obtained by those skilled in the art without creative effort should fall within the protection scope of this invention.

[0062] Example 1, referring to Figure 1 This is one embodiment of the present invention, which provides a method for intelligent robot inspection and adjustment of power plant cable trenches, including:

[0063] Sensors are deployed to collect environmental data, and the data is fused to output a fused environmental data vector, including temperature, humidity, obstacle location, and sound wave spectrum parameters, to build an inspection and control system.

[0064] Based on the input data, the vibration and microphone array acoustic wave analysis deployed by the inspection network control system are used to perform initial environmental simulation, construct a three-dimensional map, and mark the abnormal points.

[0065] Based on the 3D map with anomaly markers, perform local and global bidirectional path planning to determine the inspection path sequence;

[0066] Based on the inspection path, the inspection network control system uses the minimum turning radius as a constraint to control the torque of the actuator. The inspection network control system issues joint torque commands including the target torque and angle of each joint.

[0067] Based on the joint torque command and the 3D map, the end-of-line trajectory planning and hybrid control of the inspection are carried out. The inspection network control system issues the execution motion, establishes communication between multiple inspection control systems for inspection collaboration, and feeds back the executed motion to the communication network.

[0068] The intelligent robot inspection and adjustment method for power plant cable trenches of this invention constructs a high-precision environmental data vector through multi-sensor data fusion and weighted processing, and realizes intelligent anomaly identification and 3D map visualization marking by utilizing acoustic spectrum and vibration analysis; based on a bidirectional path planning algorithm, it introduces local risk gradient dynamic obstacle avoidance while ensuring the global optimal path, generating a stable and safe inspection sequence; through network communication and signal gradient autonomous decision-making mechanism, it realizes anomaly collaborative response and dynamic task allocation, forming a closed-loop intelligent inspection system from perception, decision-making to execution, which significantly improves the accuracy, efficiency and safety of power plant cable trench inspection, while reducing manual intervention and operation and maintenance costs.

[0069] Example 2, an embodiment of the present invention, provides a method for intelligent robot inspection and adjustment of power plant cable trenches based on the previous embodiment, including:

[0070] S100: Deploy sensors to collect environmental data, perform data fusion, and output the fused environmental data vector, including temperature, humidity, obstacle location, and sound wave spectrum parameters, to build an inspection and control system.

[0071] Temperature sensors, humidity sensors, ultrasonic sensors, thermal imaging sensors, vibration sensors, and microphone arrays are deployed in the power plant's cable trenches to construct an inspection and control system. All robots are treated as separate motion systems and connected to their respective inspection and control systems. Multiple inspection and control systems are then combined into an inspection network control system for unified inspection and management.

[0072] The sensors are connected in a mesh topology and transmit data to the central processing unit via a wireless protocol. The raw data is preprocessed, including filtering and noise reduction.

[0073] The preprocessed data stream is received and fused using a weighted average algorithm.

[0074] The sensor data weights are automatically adjusted. Thermal imaging data is prioritized in high-temperature areas to detect overheated cables, while ultrasonic sensor weights are increased in narrow bends to improve obstacle avoidance accuracy.

[0075] The fused data is formatted into an environmental data vector, including time series and spatial coordinates, and transmitted to the inspection and control system via API for initial environmental simulation.

[0076] S200: Based on the input data, the vibration and microphone array deployed by the inspection network control system performs sound wave analysis, performs initial environmental simulation, constructs a three-dimensional map, and marks abnormal points.

[0077] The acoustic signal and vibration data V in the fused environmental data vector are preprocessed, including bandpass filtering and denoising, to remove environmental interference and output clean data. A fast Fourier transform is then used to convert the acoustic signal into a spectrum. Extract frequency domain features;

[0078] Vibration data is mainly used to verify the reliability of acoustic data. Key characteristic parameters, including peak frequency, low-frequency energy ratio and high-frequency attenuation coefficient, are extracted from the amplitude spectrum.

[0079] Wherein, the peak frequency is the frequency corresponding to the maximum value in the amplitude spectrum, the low-frequency energy ratio is the ratio of low-frequency energy to total energy, and the high-frequency attenuation coefficient... The degree of energy attenuation in the high-frequency band:

[0080]

[0081] in, The high-frequency threshold is set to 5kHz. It is the reference high-frequency energy. f is the frequency corresponding to the maximum value; the peak frequency, low-frequency energy ratio, and high-frequency attenuation coefficient are integrated and fused to calculate the final vibration amplitude.

[0082] Based on the preprocessed spectrum and vibration data, the inspection network control system performs an initial environmental simulation and calculates the spectrum variation. Using historical data under normal conditions as the baseline spectrum, the difference between the current spectrum and the baseline spectrum is used. This indicates the difference in the current spectrum and calculates the anomaly probability:

[0083]

[0084] in, This represents the probability of an anomaly. The average amplitude of the vibration. and Weighting coefficients are assigned to the spectrum and vibration data by training a machine learning model based on historical anomaly data.

[0085] If P ≥ 0.5, it is marked as an anomaly and further classified according to predefined rules. > The type is uniformly defined as void;

[0086] if > The type is uniformly defined as loose. In order to prevent omission of abnormal conditions, local sparseness is uniformly classified as loose in this invention.

[0087] in, The threshold value for spectral variation is set to 30Hz. The vibration amplitude threshold is set to 0.1 m / s².

[0088] In cable trench environments, the frequency spectrum variation threshold is relatively small under normal conditions, with ΔF typically distributed in the range of 10–20 Hz. Based on the typical acoustic frequency range of 1–10 kHz, the presence of voids can cause significant frequency spectrum changes due to acoustic reflection. During normal inspections, the vibration amplitude threshold is usually low, for example, below 0.05 m / s², due to the flat ground and stable cables. When loosened, mechanical foot contact can trigger stronger vibrations, rising to 0.1 m / s² or higher; therefore, it is set to 0.1.

[0089] All abnormal locations are marked and compiled into an anomaly list, including anomaly type, location, and anomaly probability. A 3D map is then constructed. The independent inspection and control system uses positioning and mapping technology, integrating LiDAR and visual sensor data, to generate a 3D map of the cable trench within the inspection network control system.

[0090] Anomalies are marked with different colors according to their type: red indicates cavities, and yellow indicates looseness. A 3D map is output to the inspection and control system's visualization interface and terminal.

[0091] S300. Based on the 3D map with anomaly markers, perform local and global bidirectional path planning to determine the inspection path sequence;

[0092] Before generating a color-coded 3D map, edge computing sensors are deployed at fixed nodes in the motion system and channels. Based on the 3D map, a two-layer inspection path planning layout for the inspection network control system is carried out, including a local planning layer and a global planning layer.

[0093] The global planning layer calculates the optimal path, and the path cost function is:

[0094]

[0095] in, It is the global path cost. It is the length of the j-th path segment. This is a risk amplification factor. Let be the probability of an anomaly near the i-th path segment. The number of path segments. The A* algorithm avoids linear weighting, which increases the cost of high-risk points and allows the path to avoid abnormal regions. As a heuristic function, it outputs a global path sequence. , These are the coordinates of the path points;

[0096] Based on global path cost and anomaly probability path risk avoidance rules:

[0097] The primary rule prioritizes the path with the lowest anomaly probability, not exceeding 0.3. When the anomaly probabilities are equal, the path with the lowest probability is selected. The lowest-risk path segment;

[0098] The secondary rule is that if all currently planned optimal paths have a global anomaly probability greater than 0.3 and not exceeding 0.8, then the anomaly probability should be absolutely avoided. For high-risk anomaly regions with an anomaly probability >0.8, select the path with the lowest weighted risk of the anomaly probability of all anomaly points;

[0099] The third-level rule is that if all currently planned optimal paths have a global anomaly probability greater than 0.8, the motion system sends an alarm to the corresponding inspection control system and sends its current position to other motion systems through the communication network. It then obtains all paths with a global anomaly probability greater than 0.8, forms an overall range, sends it to the inspection network control system, stops inspection, and waits for further instructions.

[0100] Once the global path sequence is planned, the distance to obstacles from the edge computing sensors is transmitted in real time. Calculate the risk change and risk gradient of the local risk gradient assessment path points. Calculated as risk value Partial derivative with respect to position x:

[0101]

[0102] Among them, risk value It is derived from the fusion of the anomaly probability P and the obstacle distance. , For obstacle weights, The distance scale factor is initially set to 1m;

[0103] The final optimal inspection path sequence is calculated based on the local and global planning layers, taking into account the stability of the global path and the impact of local risk gradients:

[0104]

[0105] in: For the final path sequence, To adjust the step size coefficient, The attenuation coefficient controls the impact of global cost on the adjustment amount: since high-risk paths have been avoided through global planning, the adjustment amount decreases when global cost is high and increases when global cost is low to address local risks.

[0106] S400, the inspection network control system based on the inspection path uses the minimum turning radius as a constraint to control the torque of the actuators. The torque command issued by the inspection network control system to the actuators includes the target torque and angle of each actuator.

[0107] Once path planning is complete, the inspection network control system issues a start command. Upon receiving the command, the inspection control system controls the motion system to perform inspections according to the planned path, corresponding to a minimum turning radius. To constrain the system, adjust the torque of the actuator.

[0108] An interpolation algorithm is used to generate a continuous desired body trajectory of the motion system from the path sequence. The output desired trajectory parameters include the desired body position. Expected body speed Expected body orientation Expected body angular velocity and expected turning radius Among them, the expected turning radius Calculated from path curvature: =1 / κ, where κ is the path curvature;

[0109] The calculation steps for calculating the body acceleration of the current motion system based on the desired trajectory parameters include calculating the desired acceleration using a PID controller.

[0110] The target angle of the actuator is calculated using an inverse kinematics model. ;

[0111] Based on the current status of the enforcement agency, including from the perspective of the enforcement agency. and actuator angular velocity The torque of the actuator is calculated by mapping body acceleration to the torque of the actuator based on the principle of inverse dynamics.

[0112]

[0113] in, This is the torque vector of the actuators, representing the output torque of each actuator. M is the transpose of the body Jacobian matrix, which maps the actuator spatial velocity to the body spatial velocity; M is the body mass matrix, representing the body inertia. For Coriolis and centrifugal force terms, G is the gravity term;

[0114] The calculation results of the torque of each actuator They are combined into an actuator torque command sequence H.

[0115] S500 performs end-of-line trajectory planning and hybrid control based on the actuator torque command and 3D map. The inspection network control system issues the execution motion, establishes communication between multiple inspection control systems for inspection collaboration, and feeds back the executed motion to the communication network.

[0116] Based on the planned path and combined with the markings on the 3D map, the inspection network control system uses permanent magnet synchronous motors and vector control technology for trajectory planning and global hybrid control.

[0117] The corresponding separate inspection control system is planned at a high level, and spline interpolation is used to generate inspection trajectories. :

[0118]

[0119] in, It is a time parameter, and spline represents a cubic spline interpolation function;

[0120] Based on the inspection trajectory and the torque command sequence of the actuator, the underlying planning is carried out to calculate the real-time force point and torque output, and to control the height and depth of the execution end;

[0121] Because power plant cable trenches can be interfered with by magnetic signals or other factors, when no signal is detected or the data is unusable, high-bandwidth communication and data acquisition are automatically suspended, and only basic motion control is maintained. Edge computing sensors are used as beacons, and the independent inspection and control system obtains absolute position information by reading edge computing sensors. The original path is planned according to distance order based on the absolute position information.

[0122] Furthermore, an overall communication network is established in the inspection and control system of the power plant cable trench. All inspection and control systems are self-organized and connected through the wireless Mesh network protocol. Each inspection and control system acts as an individual network node, receives responses from other nodes, and periodically sends online signals during inspection to detect the liveness of network nodes and adjusts the routing table to cope with node movement or failure. It also obtains the network connection status and node list, where the node list contains the ID and location of all online inspection and control systems.

[0123] When the inspection and control system identifies an anomaly through the initial environment simulation, it generates a digital signal packet. The digital signal packet encapsulates the anomaly type, anomaly location, timestamp, the transmitting inspection and control system ID, and the current anomaly probability.

[0124] Digital signal packet transmission is based on a UDP retransmission mechanism, prioritizing the transmission of digital signal packets with high anomaly probability. Other inspection and control systems receive and analyze these packets, extracting the source location and transmitting inspection and control system ID, and calculating the received signal strength indication.

[0125]

[0126] in, Where is the transmit power, and s is the path loss exponent (dimensionless). For distance, It is a random variable;

[0127] Using a signal attenuation model, where the RSSI value is inversely proportional to the distance, and recording the reception timestamp, the decision module calculates the signal gradient, i.e., the change in the RSSI rate of change relative to the distance.

[0128]

[0129] in, It is the change in RSSI at two consecutive time points, and Δd is the change in distance caused by receiver movement;

[0130] if If the gradient change exceeds a preset threshold (set to 2 dBm / m), and the task priority of the received inspection control system is lower than the anomaly response priority, then the system autonomously decides to move in the opposite gradient direction. If multiple motion systems decide to move towards the same anomaly source, a new path sequence is generated using a graph search algorithm, aiming to minimize the total response time. Simultaneously, the distance between each motion system and the anomaly source is calculated, and priorities are assigned.

[0131] The system that is closest and has no path overlap with other motion systems is prioritized. The inspection and control system then controls the remaining motion systems to continue inspection according to the original path or adjust the path to avoid path conflicts.

[0132] The updated path is broadcast to all moving systems via the communication network. Upon receiving the update, synchronization is confirmed, and the synchronization status is checked in real time. If synchronization fails, a retransmission mechanism is triggered.

[0133] Example 3 is an embodiment of the present invention, which provides an intelligent robot inspection and adjustment system for power plant cable trenches, comprising:

[0134] The data acquisition module deploys sensors to collect environmental data, performs data fusion, and outputs a fused environmental data vector, including temperature, humidity, obstacle location, and sound wave spectrum parameters, to build an inspection and control system.

[0135] The simulation module, based on the input data, uses the vibration and microphone array acoustic wave analysis deployed by the inspection network control system to perform initial environmental simulation, construct a three-dimensional map, and mark anomalies.

[0136] The inspection planning module performs bidirectional path planning, both local and global, based on a 3D map marked with anomaly points, to determine the inspection path sequence.

[0137] The control module, based on the inspection path inspection network control system, uses the minimum turning radius as a constraint to control the torque of the actuators. The inspection network control system issues torque commands to the actuators, including the target torque and angle of each actuator.

[0138] The hybrid control module performs end-of-line trajectory planning and hybrid control based on the actuator torque command and 3D map. The inspection network control system issues the execution motion, establishes communication between multiple inspection control systems for inspection collaboration, and feeds back the executed motion to the communication network.

[0139] This embodiment also provides an electronic device applicable to a method for intelligent robot inspection and adjustment of power plant cable trenches, comprising: a memory and a processor; the memory is used to store computer-executable instructions, and the processor is used to execute the computer-executable instructions to realize the method for intelligent robot inspection and adjustment of power plant cable trenches as proposed in the above embodiment.

[0140] This embodiment also provides a storage medium storing a computer program, which, when executed by a processor, implements a method for intelligent robot inspection and adjustment of power plant cable trenches as proposed in the above embodiments.

[0141] The storage medium proposed in this embodiment belongs to the same inventive concept as the method for intelligent robot inspection and adjustment of power plant cable trenches proposed in the above embodiments. Technical details not described in detail in this embodiment can be found in the above embodiments, and this embodiment has the same beneficial effects as the above embodiments.

[0142] Based on the above description of the implementation methods, those skilled in the art can clearly understand that the present invention can be implemented using software and necessary general-purpose hardware, and of course, it can also be implemented using hardware, but in many cases the former is a better implementation method. Based on this understanding, the technical solution of the present invention, or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product can be stored in a computer-readable storage medium, such as a computer floppy disk, read-only memory (ROM), random access memory (RAM), flash memory, hard disk, or optical disk, etc., including several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute the methods of the various embodiments of the present invention.

[0143] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and not to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention, and all such modifications or substitutions should be covered within the scope of the claims of the present invention.

Claims

1. A method for intelligent robot inspection and adjustment of power plant cable trenches, characterized in that: include, Sensors are deployed to collect environmental data, and the data is fused to output a fused environmental data vector, including temperature, humidity, obstacle location, and sound wave spectrum parameters, to build an inspection and control system. The collected environmental data includes deploying temperature sensors, humidity sensors, ultrasonic sensors, thermal imaging sensors, vibration sensors, and microphone arrays in the power plant cable trenches to build an inspection and control system; The raw data collected is preprocessed, including filtering and denoising. The preprocessed data stream is received, and a weighted average algorithm is used for fusion. The sensor data weights are automatically adjusted. The fused data is formatted into an environmental data vector, including time series and spatial coordinates, and transmitted to the inspection and control system via API for initial environmental simulation. All motion systems are connected to the inspection and control system for unified management. Based on the input data, the vibration sensors and microphone arrays deployed by the inspection network control system are used to perform sound wave analysis and initial environmental simulation, and a three-dimensional map is constructed and anomalies are marked. The initial environmental simulation includes preprocessing the acoustic signal and vibration data V in the fused environmental data vector, including bandpass filtering and noise reduction, outputting cleaned data, and using Fast Fourier Transform to convert the acoustic signal into a spectrum. Extract frequency domain features; Key characteristic parameters, including peak frequency, low-frequency energy ratio, and high-frequency attenuation coefficient, are extracted from vibration data from the amplitude spectrum. Based on the preprocessed spectrum and vibration data, an initial environmental simulation is performed to calculate the spectrum variation. Using historical data under normal conditions as the baseline spectrum, the difference between the current spectrum and the baseline spectrum is used. This indicates the difference in the current spectrum and calculates the anomaly probability. ; If P ≥ 0.5, it is marked as an anomaly and further classified according to predefined rules. > If the type is uniformly defined as void; if > The type is uniformly defined as loose, among which, The threshold for spectral variation. The vibration amplitude threshold. The average amplitude of the vibration; All abnormal locations are marked and compiled into an anomaly list, including anomaly type, location, and anomaly probability. A 3D map is then constructed. The separate inspection and control system uses positioning and mapping technology, integrating LiDAR and visual sensor data, to generate a 3D map of the cable trench within the inspection network control system. Anomalies are marked with different colors according to their type: red indicates voids, yellow indicates looseness, and a 3D map is output to the inspection and control system's visualization interface and terminal. Based on the 3D map with anomaly markers, local and global bidirectional path planning is performed to determine the inspection path sequence. The determination of the inspection path sequence includes deploying edge computing sensors at fixed nodes of the motion system and the channel before generating the 3D map with color markers, and performing a two-layer inspection path planning layout for the inspection network control system based on the 3D map, including a local planning layer and a global planning layer. The global planning layer calculates the optimal path and the path cost function. The A* algorithm uses As a heuristic function, it outputs a global path sequence. , These are path point coordinates, based on global path cost and abnormal probability path risk avoidance rules; Once the global path sequence is planned, the distance to obstacles from the edge computing sensors is transmitted in real time. Calculate the risk change and risk gradient of the local risk gradient assessment path points. Calculated as risk value Partial derivative with respect to position x The final optimal inspection path sequence is calculated based on the local and global planning layers, taking into account the stability of the global path and the impact of local risk gradients. ， in: This represents the final path sequence, where γ is the adjustment step size coefficient. The attenuation coefficient; The inspection network control system based on the inspection path uses the minimum turning radius as a constraint to control the torque of the actuators. The torque command issued by the inspection network control system to the actuators includes the target torque and angle of each actuator. Based on the torque command of the actuator and the 3D map, the end-of-line trajectory planning and hybrid control of the inspection are carried out. The inspection network control system issues the execution motion, establishes communication between multiple inspection control systems for inspection collaboration, and feeds back the executed motion to the communication network.

2. The intelligent robot inspection and adjustment method for power plant cable trenches as described in claim 1, characterized in that: The torque control of the actuator includes, after path planning is completed, the inspection network control system issues a start command, and the inspection control system receives the command and controls the motion system to perform inspection according to the planned path. The corresponding inspection control system uses the minimum turning radius. To constrain the system, adjust the torque of the actuator. An interpolation algorithm is used to generate a continuous desired body trajectory of the motion system from the path sequence. The output desired trajectory parameters include the desired body position. Expected body speed Expecting the body to face Expected body angular velocity and expected turning radius ; The calculation steps for calculating the robot's current body acceleration based on the desired trajectory parameters involve using a PID controller to calculate the desired acceleration. The target angle of the actuator is calculated using an inverse kinematics model. Based on the current status of the enforcement agency, including from the perspective of the enforcement agency. and actuator angular velocity The torque of the actuator is calculated by mapping the body acceleration to the torque of the actuator based on the principle of inverse dynamics. The calculation results of the torque of each actuator They are combined into an actuator torque command sequence H.

3. The intelligent robot inspection and adjustment method for power plant cable trenches as described in claim 2, characterized in that: The inspection end trajectory planning and hybrid control includes, based on the planned path and combined with the markings on the 3D map, the inspection network control system uses permanent magnet synchronous motors and vector control technology for trajectory planning and global hybrid control; The corresponding separate inspection control system is planned at a high level, and spline interpolation is used to generate inspection trajectories. : ， in, It is a time parameter, and spline represents a cubic spline interpolation function; Based on the inspection trajectory and the torque command sequence of the actuator, low-level planning is performed to calculate the real-time force application point and torque output, and to control the height and depth of the actuator end effector. ,in, It is a vertical force. It's an altitude error. It is the stiffness coefficient. It is angular velocity; The adjusted end position and torque are obtained and then vector controlled by the permanent magnet synchronous motor as follows: ,in, It is the q-axis current. It is the torque constant; Upon receiving a current command, the end motor is driven to move.

4. The intelligent robot inspection and adjustment method for power plant cable trenches as described in claim 3, characterized in that: The inspection collaboration includes establishing an overall communication network in the inspection control system of the power plant cable trench. All inspection control systems are self-organized and connected through the wireless Mesh network protocol. Each inspection control system acts as an individual network node, receives responses from other nodes, periodically sends online signals during inspection to detect the liveness of network nodes, adjusts the routing table to cope with node movement or failure, and obtains network connection status and node list. When the inspection and control system identifies an anomaly through the initial environment simulation, it generates a digital signal packet. The digital signal packet encapsulates the anomaly type, anomaly location, timestamp, the transmitting inspection and control system ID, and the current anomaly probability. The digital signal packet transmission is based on the UDP retransmission mechanism, prioritizing the transmission of digital signal packets with high anomaly probability. Other inspection and control systems receive and analyze the digital signal packets, extract the source location and transmitting inspection and control system ID of the digital signal packet, and calculate the received signal strength indication. Using a signal attenuation model, where the RSSI value is inversely proportional to the distance, and recording the reception timestamp, the decision module calculates the signal gradient. ; if If the gradient change exceeds the preset threshold (set to 2dBm / m) and the task priority of the received inspection control system is lower than the abnormal response priority, then the system will autonomously decide to move in the reverse gradient direction. When two or more motion systems decide to go to the same anomaly source, a graph search algorithm is used to replan the path sequence, aiming to minimize the total response time. Simultaneously, the distance between each motion system and the anomaly source is calculated, and priorities are assigned. The system that is closest to the other system and has no overlapping path with other motion systems is given priority. The inspection and control system then controls the remaining motion systems to continue inspection according to the original path or adjust the path to avoid path conflicts. The updated path is broadcast to all motion systems through the communication network. After receiving the update, the system confirms synchronization and checks the synchronization status in real time. If synchronization fails, a retransmission mechanism is triggered.

5. A smart robot inspection and adjustment system for power plant cable trenches, employing the smart robot inspection and adjustment method for power plant cable trenches as described in any one of claims 1 to 4, characterized in that, include: The data acquisition module deploys sensors to collect environmental data, performs data fusion, and outputs a fused environmental data vector, including temperature, humidity, obstacle location, and sound wave spectrum parameters, to build an inspection and control system. The simulation module, based on the input data, uses the vibration and microphone array acoustic wave analysis deployed by the inspection network control system to perform initial environmental simulation, construct a three-dimensional map, and mark anomalies. The inspection planning module performs bidirectional path planning, both local and global, based on a 3D map marked with anomaly points, to determine the inspection path sequence. The control module, based on the inspection path inspection network control system, uses the minimum turning radius as a constraint to control the torque of the actuators. The inspection network control system issues torque commands to the actuators, including the target torque and angle of each actuator. The hybrid control module performs end-of-line trajectory planning and hybrid control based on the actuator torque command and 3D map. The inspection network control system issues the execution motion, establishes communication between multiple inspection control systems for inspection collaboration, and feeds back the executed motion to the communication network.

6. A computer device comprising a memory and a processor, wherein the memory stores a computer program, characterized in that, When the processor executes the computer program, it implements the steps of the intelligent robot inspection and adjustment method for power plant cable trenches as described in any one of claims 1 to 4.

7. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, it implements the steps of the intelligent robot inspection and adjustment method for power plant cable trenches as described in any one of claims 1 to 4.

Images