Inspection robot intelligent inspection control method and system based on edge computing

Abstract

The application provides an edge computing-based intelligent inspection control method and system for an inspection robot, wherein the method comprises: acquiring multi-dimensional parameters of the inspection robot during flight; inputting the multi-dimensional parameters into a pre-trained heat conduction model to output temperature evolution data of a circuit of the inspection robot, boundary parameters of the heat conduction model being obtained through multi-round iterative adjustment by a dynamic optimization algorithm; constructing a virtual model corresponding to the circuit of the inspection robot in a digital twin framework, and dynamically calibrating the temperature evolution data based on the virtual model; and generating a state monitoring atlas of the circuit of the inspection robot based on the calibrated temperature evolution data and the virtual model. The application improves the real-time performance and accuracy of thermal state monitoring of the circuit of the inspection robot.

Description

Technical Field

[0001] This application relates to the field of circuit status monitoring technology for inspection robots, and in particular to an intelligent inspection control method and system for inspection robots based on edge computing. Background Technology

[0002] As inspection robots undertake more long-duration flight missions in complex environments, their circuit systems face risks such as abnormal temperatures and localized overheating. A method is needed to monitor the thermal state of the circuits in real time and predict potential faults to ensure flight safety and equipment reliability.

[0003] Current solutions employ a heat conduction model based on fixed parameters for circuit temperature prediction. This involves collecting actual data from an onboard temperature sensor and statically comparing it with the model's output. An alarm is triggered when the temperature exceeds a preset threshold. This solution relies on preset heat conduction parameters and combines sensor data to achieve basic temperature anomaly detection.

[0004] The fixed parameters of the heat conduction model in this scheme make it difficult to adapt to dynamic changes in the flight environment, resulting in insufficient temperature prediction accuracy. At the same time, the static comparison method cannot make full use of the real-time nature of sensor data, resulting in a lag in anomaly detection and affecting the accuracy and timeliness of monitoring. Summary of the Invention

[0005] This application provides an intelligent inspection control method and system for inspection robots based on edge computing, which solves the problems of low real-time performance and poor accuracy in the thermal status monitoring of inspection robot circuits in the prior art.

[0006] In a first aspect, this application provides an intelligent inspection control method for inspection robots based on edge computing, comprising: Acquire multi-dimensional parameters of the inspection robot during its flight; The multidimensional parameters are input into a pre-trained heat conduction model, and the temperature evolution data of the inspection robot circuit is output. The boundary parameters of the heat conduction model are obtained by multiple rounds of iterative adjustment through a dynamic optimization algorithm. A virtual model corresponding to the circuitry of the inspection robot is constructed within a digital twin framework, and the temperature evolution data is dynamically calibrated based on the virtual model. Based on the calibrated temperature evolution data and the virtual model, a state monitoring map of the inspection robot circuit is generated.

[0007] Optionally, generating a state monitoring map of the inspection robot circuit based on the calibrated temperature evolution data and the virtual model includes: The calibrated temperature evolution data is mapped to the three-dimensional spatial topology mesh of the virtual model according to spatial coordinates; Extract the temperature change trajectory data corresponding to the monitoring period from each grid cell in the three-dimensional spatial topology grid; Based on the temperature change trajectory data, calculate the instantaneous slope of each grid cell; The grid cells with instantaneous slopes exceeding a set threshold are combined into a grid cell set; Visually label the set of grid cells in the three-dimensional spatial topological grid of the virtual model to form a status monitoring map.

[0008] Optionally, visually marking the set of grid cells in the three-dimensional spatial topological grid of the virtual model to form a state monitoring map includes: Calculate the temperature change of each grid cell in the grid cell set during the monitoring period; Based on the temperature change, and in conjunction with preset risk mapping rules, a risk level code is generated; In the three-dimensional spatial topological grid, the spatial coordinates of each grid cell are bound to the corresponding risk level code to generate data tags; A condition monitoring map is generated based on the data labels of all grid cells.

[0009] Optionally, generating a state monitoring map based on the data tags of all grid cells includes: Convert the spatial coordinates in the data markers of each grid cell into planar coordinates in a preset projected coordinate system; Based on the risk level code in the data markers, select symbols of the corresponding type from a preset visualization symbol library; In the projected coordinate system, the display position of each type of symbol is determined according to the planar coordinates, and the display color of each symbol is adjusted according to the risk level code; Each symbol is integrated into the circuit diagram template of the inspection robot according to its corresponding display position, display color and type to form a status monitoring diagram.

[0010] Optionally, the step of inputting the multidimensional parameters into a pre-trained heat conduction model and outputting temperature evolution data of the inspection robot circuit includes: Obtain a boundary parameter sequence, wherein each boundary parameter in the boundary parameter sequence is generated during the parameter adjustment process by a dynamic optimization algorithm; The multidimensional parameters are combined with the latest parameter value in the boundary parameter sequence to form the model input vector; The heat conduction model is used to calculate based on the model input vector to generate a set of predicted temperature values ​​for each location point of the inspection robot circuit within a future set time period. All the predicted temperature values ​​are organized in chronological order to form temperature evolution data.

[0011] Optionally, the step of calculating based on the model input vector using the heat conduction model to generate a set of predicted temperature values ​​for each location point of the inspection robot circuit within a future set time period includes: Extract airflow velocity parameters and light intensity parameters from the model input vector; Configure the input conditions for the heat conduction model based on the airflow velocity parameters and the light intensity parameters; In the heat conduction model, the geometric structure of the inspection robot circuit is spatially discretized to obtain multiple discrete nodes; The transient heat conduction equations of each discrete node under the input conditions are solved by numerical methods to obtain the temperature prediction values ​​of each discrete node at each preset time point within a future set time period. The temperature prediction values ​​of each discrete node at the same preset time point are organized into a temperature prediction value set according to their spatial location.

[0012] Optionally, the step of dynamically calibrating the temperature evolution data based on the virtual model includes: Obtain the actual measurement value collected by the temperature sensor bound to the virtual model at the target timestamp; Extract the calculated temperature value of the corresponding spatial location at the target timestamp from the temperature evolution data stream; Calculate the numerical deviation between the actual measured value and the calculated temperature value; When the numerical deviation exceeds the set tolerance, the temperature evolution data is dynamically calibrated to obtain calibrated temperature evolution data.

[0013] Secondly, this application provides an intelligent inspection control system for inspection robots based on edge computing, comprising: The acquisition module is used to acquire multi-dimensional parameters of the inspection robot during flight. The input module is used to input the multidimensional parameters into the pre-trained heat conduction model and output the temperature evolution data of the inspection robot circuit. The boundary parameters of the heat conduction model are obtained by multiple rounds of iterative adjustment through a dynamic optimization algorithm. A construction module is used to build a virtual model corresponding to the circuitry of the inspection robot within a digital twin framework, and to dynamically calibrate the temperature evolution data based on the virtual model; The generation module is used to generate a state monitoring map of the inspection robot circuit based on the calibrated temperature evolution data and the virtual model.

[0014] Thirdly, this application provides a computing device, including a processor and a memory, wherein the memory stores a computer program, and the processor is configured to run the computer program to execute the intelligent inspection control method for an inspection robot based on edge computing as described in any of the first aspects.

[0015] Fourthly, this application provides a computer storage medium storing computer program instructions thereon, wherein the computer program instructions, when executed by a processor, implement the intelligent inspection control method for an inspection robot based on edge computing as described in any one of the first aspects.

[0016] This application provides an intelligent inspection control method for an inspection robot based on edge computing. The method includes: acquiring multi-dimensional parameters of the inspection robot during flight; inputting the multi-dimensional parameters into a pre-trained heat conduction model and outputting temperature evolution data of the inspection robot circuit, wherein the boundary parameters of the heat conduction model are obtained through multiple rounds of iterative adjustment using a dynamic optimization algorithm; constructing a virtual model corresponding to the inspection robot circuit in a digital twin framework; dynamically calibrating the temperature evolution data based on the virtual model; and generating a state monitoring map of the inspection robot circuit based on the calibrated temperature evolution data and the virtual model.

[0017] The technical solution provided in this application has the following beneficial effects: This application collects flight environment data in real time, providing multi-dimensional input for circuit thermal state analysis. It improves temperature prediction accuracy through dynamically optimized boundary parameters, adapting to different flight conditions. A digital mapping of the circuit's physical entity is established, providing a foundation for virtual-real data interaction. Prediction biases are corrected using sensor data from the virtual model, enhancing data reliability. The location and risk level of circuit thermal anomalies are intuitively presented, supporting rapid decision-making.

[0018] Furthermore, this application also maps the calibrated temperature data onto a virtual model grid, extracts the temperature change trajectory of each grid cell and calculates the instantaneous slope, filters the set of cells with slopes exceeding the threshold for visualization marking, and forms a state monitoring map that integrates spatial location and risk trend.

[0019] Furthermore, by accurately capturing areas of abnormal temperature changes through slope analysis and intuitively locating risk locations by combining 3D spatial markings, the timeliness and accuracy of anomaly identification are improved, providing visual decision support for operation and maintenance.

[0020] These or other aspects of this application will become more apparent in the following description of the embodiments. Attached Figure Description

[0021] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0022] Figure 1 A flowchart illustrating an intelligent inspection control method for an inspection robot based on edge computing, provided as an embodiment of this application; Figure 2 A second flowchart of an intelligent inspection control method for an inspection robot based on edge computing, provided in an embodiment of this application; Figure 3 A third flowchart of an intelligent inspection control method for an inspection robot based on edge computing, provided in an embodiment of this application; Figure 4 The fourth flowchart of an intelligent inspection control method for an inspection robot based on edge computing, provided in an embodiment of this application; Figure 5 The fifth flowchart of an intelligent inspection control method for an inspection robot based on edge computing, provided in an embodiment of this application; Figure 6 The sixth flowchart of an intelligent inspection control method for an inspection robot based on edge computing, provided in an embodiment of this application; Figure 7 The seventh flowchart of an intelligent inspection control method for an inspection robot based on edge computing, provided in an embodiment of this application; Figure 8 A schematic diagram of the structure of an intelligent inspection control system for an inspection robot based on edge computing, provided in an embodiment of this application; Figure 9 This is a schematic diagram of the structure of a computing device provided in an embodiment of this application. Detailed Implementation

[0023] To enable those skilled in the art to better understand the present application, the technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the accompanying drawings.

[0024] In some of the processes described in the specification, claims, and accompanying drawings of this application, multiple operations appearing in a specific order are included. However, it should be clearly understood that these operations may not be executed in the order they appear herein, or may be executed in parallel. The operation numbers, such as 101, 102, etc., are merely used to distinguish different operations and do not themselves represent any execution order. Furthermore, these processes may include more or fewer operations, and these operations may be executed sequentially or in parallel. It should be noted that the descriptions such as "first," "second," etc., in this document are used to distinguish different messages, devices, modules, etc., and do not represent a chronological order, nor do they limit "first" and "second" to different types.

[0025] Existing circuit temperature monitoring solutions for inspection robots employ fixed-parameter heat conduction models, whose boundary conditions cannot be dynamically adjusted according to the flight environment. This leads to systematic deviations between predicted temperature values ​​and actual operating conditions. Furthermore, static calibration mechanisms rely solely on threshold comparisons, lacking spatial correlation analysis with circuit physical characteristics. This results in delayed anomaly detection and coarse localization, failing to meet the demands for accurate monitoring in highly dynamic flight scenarios. This limitation stems from the fragmented handling of environmental parameters, model calculations, and state assessments, restricting the real-time performance and reliability of the monitoring system.

[0026] To address the aforementioned shortcomings, this application proposes an intelligent inspection control method for inspection robots based on edge computing. This method reconstructs the temperature monitoring technology path by integrating multi-dimensional environmental parameters, dynamically optimizing the heat conduction model, and performing closed-loop calibration using virtual and real data. Specifically, a dynamic optimization algorithm is used to adjust the model's boundary parameters in real time, enabling temperature prediction to adapt to changes in the flight environment. Based on the spatial topological relationship of a digital twin virtual model, sensor data and predicted values ​​are dynamically calibrated for spatiotemporal alignment. Finally, slope analysis and 3D labeling are used to generate a monitoring map that integrates risk trends and positional accuracy. This method overcomes the rigidity of fixed-parameter models, establishing a closed-loop system for collaborative optimization of the environment, model, and virtual entity. This eliminates the response delay of traditional static calibration and achieves early and accurate localization of abnormal areas, providing a highly reliable solution for circuit status monitoring under complex operating conditions.

[0027] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0028] Figure 1 A flowchart illustrating an intelligent inspection control method for an inspection robot based on edge computing, as provided in this application embodiment, is shown below. Figure 1 As shown, the method includes: Step S101: Obtain multi-dimensional parameters of the inspection robot during flight.

[0029] In step S101, the multidimensional parameters refer to various environmental physical quantities collected by the inspection robot during flight, including altitude (air pressure data), airflow speed (anemometer data), and light intensity (photosensitive sensor data), which are used to characterize the external thermal environment in which the circuit is located.

[0030] In this embodiment, flight environment data is collected in real time through an airborne sensor network, wherein an altitude sensor acquires altitude data, an airflow sensor measures oncoming wind speed, and a photosensitive element records solar radiation intensity; the three types of data are packaged into a structured parameter set at fixed time intervals and transmitted to a data processing unit for timestamp alignment and format standardization to form a continuous multidimensional parameter sequence.

[0031] For example, when an inspection robot is performing an inspection task at location A, its onboard sensors collect environmental data every 5 seconds: altitude 1500 meters (converted from air pressure), airflow speed 8 meters / second (measured by an ultrasonic anemometer), and light intensity 1200 watts / square meter (output from a silicon photovoltaic cell). The data processing unit combines these three data points into a parameter vector of [1500, 8, 1200], adds a timestamp, and stores it in a cache queue.

[0032] Step S102: Input the multidimensional parameters into the pre-trained heat conduction model and output the temperature evolution data of the inspection robot circuit. The boundary parameters of the heat conduction model are obtained by multiple rounds of iterative adjustment through a dynamic optimization algorithm.

[0033] In step S102, the heat conduction model represents a numerical calculation model based on Fourier's law. Its inputs include environmental parameters and dynamic boundary parameters, and its output is a set of predicted values ​​of temperature changes at various locations in the circuit over time. Temperature evolution data refers to a data set generated by the heat conduction model, reflecting the temperature change trend of each location point in the inspection robot circuit over a set future time period. It includes the predicted temperature values ​​of each discrete grid node at different time points. Each data item consists of three elements: a location identifier, a time identifier, and a temperature value, used to characterize the spatiotemporal distribution characteristics of the circuit's thermal state under dynamic flight conditions. Specifically, it is represented by a sequence of predicted temperature values ​​organized in chronological order, such as the continuous data of node A's temperature value T1 at time t1 and temperature value T2 at time t2, which together constitute the time-varying evolution process of the circuit's temperature field. Boundary parameters represent the model's heat exchange condition parameters (convective heat transfer coefficient, radiation coefficient), which are adjusted through a dynamic optimization algorithm to match actual operating conditions.

[0034] In this embodiment, the multidimensional parameters and the latest boundary parameters generated by the dynamic optimization algorithm are combined into an input vector, which is then input into the heat conduction model for transient solution: First, the circuit geometry is spatially discretized to generate a computational grid, with each grid node associated with material thermal property parameters; then, the discretized heat conduction equation is solved, where airflow velocity is converted into convective heat transfer coefficient to participate in boundary condition calculation, and light intensity is converted into radiative heat flux density; finally, the temperature prediction sequence of all grid nodes in the future time period is output.

[0035] For example, in the input vector [1500,8,1200,25,0.85], 25 represents the current optimal convective heat transfer coefficient (watts / square meter·Kelvin), and 0.85 represents the radiation coefficient. The model discretizes the circuit board into 36 nodes and solves the equations using the finite difference method. Where ρ is the material density, c_p is the specific heat capacity, and k is the thermal conductivity. The temperature values ​​of each node within 30 seconds are calculated; for example, node 1 is 56.3℃ at 5 seconds.

[0036] Step S103: Construct a virtual model corresponding to the inspection robot circuit in the digital twin framework, and dynamically calibrate the temperature evolution data based on the virtual model; In step S103, the virtual model represents a three-dimensional mesh model of the circuit within the digital twin framework, with nodes bound to the actual circuit temperature sensor locations for virtual-real data fusion. Dynamic calibration refers to using measured sensor data from the virtual model to correct deviations in the model's predicted values.

[0037] In this embodiment, a three-dimensional mesh virtual model is constructed based on the circuit design drawings, and the installation coordinates of the actual sensors are mapped to the corresponding mesh nodes; real-time sensor data is read from the virtual model and compared with the predicted values ​​of temperature evolution data at the same time and location to calculate the numerical deviation; subsequent predicted values ​​are corrected through a weighted compensation algorithm, wherein the weights are dynamically adjusted according to historical deviation statistics.

[0038] For example, the virtual model binds node G-19 in the main control chip area to the actual PT100 sensor. When the sensor measures 62.3℃ at 85 seconds, while the model predicts 58.7℃, the calculated deviation is 3.6℃. The compensation formula T_calibration = T_original + 0.85 × ΔT (0.85 being the adaptive weight) is used to uniformly correct subsequent predicted values; for instance, the predicted value of 59.2℃ at 86 seconds is calibrated to 62.3℃.

[0039] Step S104: Based on the calibrated temperature evolution data and the virtual model, generate a state monitoring map of the inspection robot circuit.

[0040] In step S104, the status monitoring map is a two-dimensional visualization chart that marks the location and risk level of abnormal areas by using symbol colors and shapes.

[0041] In this embodiment, the calibrated temperature data is mapped to a virtual model mesh, and the slope of temperature change for each unit is calculated as k = ΔT / Δt. Units with k exceeding the threshold of 0.5℃ / second are selected, and risk levels are divided according to the absolute value of the temperature difference: ΔT ≤ 5℃ is Level 1 (circle symbol), 5℃ < ΔT ≤ 10℃ is Level 2 (square), and ΔT > 10℃ is Level 3 (triangle). Symbols are rendered on the two-dimensional projection base map and legends are added.

[0042] For example, node G-19 has a temperature difference ΔT = 26.2℃ between 250 and 255 seconds, with a slope of 0.76℃ / second, and is marked as a red triangle. The final map contains 7 tertiary risk symbols, 4 secondary symbols, and 17 primary symbols.

[0043] This method dynamically drives the optimization of the heat conduction model through environmental parameters, and combines it with closed-loop calibration using digital twin virtual and real data to achieve high-precision prediction and spatial positioning of circuit temperature anomalies in inspection robots. Compared with traditional static models, it improves the real-time performance and reliability of thermal state monitoring in complex environments, and provides visualized decision support for preventing circuit faults.

[0044] To address the challenge of accurately locating abnormal areas in the thermal status monitoring of inspection robots' circuits, some embodiments refer to... Figure 2 Step S104: Based on the calibrated temperature evolution data and the virtual model, the condition monitoring map of the inspection robot circuit is generated, including: Step S201: Map the calibrated temperature evolution data to the three-dimensional spatial topology mesh of the virtual model according to spatial coordinates.

[0045] In step S201, spatial coordinate mapping refers to establishing a one-to-one correspondence between the physical location of temperature data and the grid nodes of the virtual model. Each coordinate point contains three-dimensional location information and a temperature value sequence. The three-dimensional spatial topology grid of the virtual model originates from a virtual model corresponding to the inspection robot circuit constructed in the digital twin framework. This grid is generated in the digital twin framework through spatial discretization based on the physical structure geometry and component layout coordinates of the circuit board. The grid cell size and connection relationship strictly correspond to the physical structure and component position relationships of the actual circuit.

[0046] In this embodiment of the application, the calibrated temperature data is matched to the corresponding grid cell according to the location by using the node coordinate index table pre-stored in the virtual model. For example, the temperature data of the main control chip area is mapped to the G-12 to G-15 cells of grid layer 1.

[0047] Step S202: Extract the temperature change trajectory data corresponding to the monitoring period from each grid cell in the three-dimensional spatial topology grid.

[0048] In step S202, the temperature change trajectory data refers to the set of temperature values ​​of a single grid cell arranged in chronological order during a complete monitoring period, reflecting the thermal state evolution process at that location.

[0049] In this embodiment of the application, a time-stamp-aligned temperature sequence is extracted from each mapped grid cell, such as the temperature values ​​of grid G-12 at 5-second intervals within 0-300 seconds [56.2℃, 57.1℃, ..., 82.4℃].

[0050] Step S203: Calculate the instantaneous slope of each grid cell based on the temperature change trajectory data.

[0051] In step S203, the instantaneous slope is the rate of temperature change obtained by differential calculation, which characterizes the temperature rise and fall trend per unit time.

[0052] In the embodiments of this application, the slope value at adjacent time points is calculated using the central difference method. For example, the slope k of grid G-12 during the period of 255-260 seconds is (82.4-78.6) / 5 = 0.76℃ / second.

[0053] Step S204: Combine the grid cells whose instantaneous slope exceeds the set threshold into a grid cell set.

[0054] In step S204, the threshold is set as a critical rate of change based on the thermal resistance characteristics of the circuit material, which is used to screen for abnormal temperature rise areas.

[0055] In this embodiment of the application, when the slope exceeds the threshold of 0.5℃ / second, the corresponding grid cell is added to the anomaly set, such as G-12, G-19 and other 7 cells being marked.

[0056] Step S205: Visually mark the set of grid cells in the three-dimensional spatial topology grid of the virtual model to form a status monitoring map.

[0057] In step S205, visual marking is a technique for highlighting abnormal cells in a three-dimensional mesh using symbolic methods.

[0058] In this embodiment of the application, all abnormal units are marked with red triangles in the virtual model interface, and a visual map containing location coordinates and risk level is generated on the two-dimensional projection base map.

[0059] Here is a specific example: When an inspection robot performs an inspection task at location A, based on calibrated temperature evolution data, it first positions the calibrated temperature values ​​[82.4℃, 83.7℃, 85.2℃, 86.8℃, 88.5℃, 90.3℃] of node G-19 at 5-second intervals during the 250-300 second period to the layer 1 region of the 3D spatial topological grid at x=35 mm, y=22 mm, z=0 mm, according to the preset coordinate mapping relationship in the virtual model. After extracting the temperature change trajectory data from this grid cell, the instantaneous slope k=(T_{t+1}-T_{t-1}) / 2Δt is calculated using the central difference method, where T_{t+1} is the temperature value at time t+1, T_{t-1} is the temperature value at time t-1, and Δt is the 5-second time interval. The slope at 265 seconds is calculated to be k=(85.2-83.7) / 10=0. At 15℃ / second, k=(88.5-85.2) / 10=0.33℃ / second at 275 seconds, and k=(90.3-86.8) / 10=0.35℃ / second at 285 seconds; when the instantaneous slope k=(90.3-88.5) / 5=0.36℃ / second at 290 seconds exceeds the set threshold of 0.3℃ / second, the grid cell is added to the anomaly set; finally, the cell is marked as an orange square symbol in the virtual model, and its color depth is determined to be a level 2 risk based on the absolute value of the temperature difference ΔT=90.3-82.4=7.9℃, which is in the range of 5-10℃. Together with the marked red triangle level 3 risk cell G-19, it constitutes a complete monitoring map. When the map is displayed through two-dimensional projection, it contains one newly added orange square symbol and seven existing red triangle symbols to help operators identify thermal anomaly areas of different levels.

[0060] In this embodiment, the precise location and risk classification of the circuit thermal anomaly area are achieved through the collaborative processing of temperature data spatial mapping, trajectory slope analysis and abnormal unit visualization marking, providing an intuitive and reliable decision-making basis for the circuit maintenance of the inspection robot.

[0061] To address the issue of visualizing the risk level of thermal anomalies in the circuitry of inspection robots, some embodiments refer to... Figure 3 Step S205: Visually marking the set of grid cells in the three-dimensional spatial topological grid of the virtual model to form a state monitoring map includes: Step S301: Calculate the temperature change of each grid cell in the grid cell set during the monitoring period.

[0062] In step S301, the temperature change refers to the absolute difference between the final temperature value and the initial temperature value of the grid cell during the monitoring period. It is a single numerical indicator used to quantify the overall magnitude of temperature change. On the other hand, the "temperature change trajectory data" refers to the sequence of temperature values ​​recorded by the grid cell in chronological order throughout the entire monitoring period, reflecting the continuous process characteristics of temperature change over time. The difference between the two is that the former is the final-initial state calculation result of the latter, while the latter is a complete dataset that includes intermediate state changes.

[0063] In this embodiment of the application, the temperature values ​​at the beginning and end of the time point are extracted from the temperature change trajectory data of each grid cell and the difference is calculated. For example, if the temperature of a certain cell is 50°C at 0 seconds and 78°C at 300 seconds, the temperature change is 28°C.

[0064] Step S302: Based on the temperature change, and in conjunction with preset risk mapping rules, generate a risk level code.

[0065] In step S302, the preset risk mapping rule is a grading standard based on the thermal properties of the circuit materials, which converts the temperature change into discrete risk level codes. The risk level code is a numerical identifier based on the magnitude of the temperature change in the grid cells, used to characterize the degree of risk of circuit thermal anomalies. Level 1 codes represent the low-risk state with the smallest temperature change, Level 2 codes represent the medium-risk state, and Level 3 codes represent the high-risk state with the largest temperature change. The code value is positively correlated with the severity of the risk, providing a grading basis for subsequent visualization and labeling.

[0066] In this embodiment of the application, a three-level classification rule is adopted: when the temperature change does not exceed 10℃, it is a first-level code; 10-20℃ is a second-level code; and more than 20℃ is a third-level code. The coding results are used to control the style of subsequent visual tags.

[0067] Step S303: In the three-dimensional spatial topology grid, bind the spatial coordinates of each grid cell with the corresponding risk level code to generate a data tag.

[0068] In step S303, the data tag is a structured data unit containing spatial location information and risk level, used to associate physical location with risk level.

[0069] In this embodiment, the three-dimensional coordinates of the grid cell are combined with the risk level code into a quadruple data structure of {coordinate x, coordinate y, coordinate z, risk level}. For example, when the coordinates of a cell are (35, 22, 0) and the risk level is level 2, the label {35, 22, 0, 2} is generated.

[0070] Step S304: Generate a condition monitoring map based on the data tags of all grid cells.

[0071] In this embodiment, the symbol position is located on the two-dimensional projection plane based on the coordinate information in all data markers. The symbol shape and color are matched according to the coding level. The first level is displayed as a green circle, the second level as a yellow square, and the third level as a red triangle. Finally, a complete map including illustrations is synthesized.

[0072] Here is a specific example: When an inspection robot performs a monitoring task in area B, based on eight identified abnormal grid cells, including node G-19, it first calculates the temperature change of each cell within a complete 300-second monitoring period. Node G-19 has an initial temperature of 82.4℃ and a final temperature of 90.3℃, with a change ΔT = 90.3 - 82.4 = 7.9℃. Node G-23 has an initial temperature of 78.6℃ and a final temperature of 92.5℃, with a change ΔT = 92.5 - 78.6 = 13.9℃. According to a preset risk mapping rule, when the change is 5℃ < ΔT ≤ 1... A secondary code is generated at 0℃, and a tertiary code is generated when 10℃ < ΔT ≤ 20℃. Therefore, G-19 is marked as code 2, and G-23 is marked as code 3. These codes are bound to the unit spatial coordinates to generate data markers. For example, G-19 is marked as x=35 mm, y=22 mm, z=0 mm, code 2, and G-23 is marked as x=40 mm, y=18 mm, z=0 mm, code 3. When finally generating the status monitoring map, the code 2 unit is rendered as a yellow square symbol and the code 3 unit is rendered as a red triangle symbol in the virtual model.

[0073] In this embodiment of the application, through progressive processing of temperature change calculation, risk level mapping, spatial coordinate binding and symbolic presentation, the visualization and hierarchical positioning of the thermal anomaly risk of the inspection robot circuit is realized, enabling maintenance personnel to quickly identify key risk areas and take targeted measures.

[0074] Reference Figure 4 To address the visualization issue of the thermal status monitoring results of the inspection robot's circuits, in some embodiments, step S304: generating a status monitoring map based on the data tags of all grid cells includes: Step S401: Convert the spatial coordinates in the data markers of each grid cell into planar coordinates in a preset projected coordinate system.

[0075] In step S401, spatial coordinate transformation refers to the mathematical transformation process of mapping three-dimensional mesh coordinates to a two-dimensional display plane, the purpose of which is to maintain the spatial relationship of the circuit structure without distortion.

[0076] In this embodiment, an orthogonal projection algorithm is used to convert the three-dimensional coordinates x, y, z of the mesh cells into two-dimensional planar coordinates u, v, where the z-axis depth information is converted into the display priority of the u and v coordinates through a scaling factor.

[0077] Step S402: Select a symbol of the corresponding type from a preset visualization symbol library according to the risk level code in the data marker.

[0078] In step S402, the visual symbol library is a predefined collection of graphic elements, containing standard symbol styles corresponding to different risk levels.

[0079] In this embodiment of the application, the symbol library corresponds to a solid circle for level one risk, a solid square for level two risk, and a solid triangle for level three risk. Each symbol has a basic size and adjustable parameters.

[0080] Step S403: In the projected coordinate system, determine the display position of each type of symbol according to the planar coordinates, and adjust the display color of each symbol according to the risk level code.

[0081] In step S403, determining the display position refers to the process of accurately locating the center point of the symbol on the display canvas based on the converted planar coordinates.

[0082] In this embodiment of the application, the pixel position corresponding to each data marker is found through a coordinate mapping table, and the symbol size is adjusted according to the risk level. The higher the risk level, the larger the symbol display size.

[0083] Step S404: Integrate each symbol into the circuit diagram template of the inspection robot according to its corresponding display position, display color and type to form a status monitoring diagram.

[0084] In step S404, the map template is a blank background map with a circuit outline and a coordinate reference system, used to carry visualization symbols.

[0085] In this embodiment, each symbol is rendered onto the template according to its calculated display position and color, and a legend bar containing risk level descriptions is added to ultimately generate an interactive status monitoring map.

[0086] Here is a specific example: When an inspection robot performs a high-temperature environment flight test in area C, based on the generated data tags of 8 abnormal units, it first converts the three-dimensional coordinates into two-dimensional planar coordinates using the orthogonal projection formulas u=2.5x+50, v=2.5y+30, where x and y are the original coordinate values ​​in millimeters, 2.5 is the scaling factor, and 50 and 30 are the offsets. For example, the coordinates of unit G-19 (35,22,0) are converted to 137.5,85 pixels, and the coordinates of unit G-23 (40,18,0) are converted to 150,75 pixels. Then, according to the risk level code in the data tags, a symbol is selected from a preset symbol library... The symbols are designated as follows: Code 2 uses a yellow square with a side length of 6 pixels, and Code 3 uses a red triangle with a bottom side length of 8 pixels. On the generated 800×600 pixel circuit template, the center point of the symbol is located according to the converted coordinates. The coordinates of the upper left corner of the G-19 symbol are 134.5, 82, and the coordinates of the vertex of the G-23 symbol are 150, 71. At the same time, the color saturation of the symbols is adjusted according to the risk level. The red saturation of the third-level coded symbols is increased by 40% compared with the yellow of the second-level coded symbols. Finally, all symbols are rendered on the template, and a legend is added to the lower right corner: "Square: Medium risk (5-10℃) Triangle: High risk (10-20℃)" to form a complete status monitoring map. When this map is displayed through the ground station monitoring system, the operator can clearly see the distribution of the two yellow squares and six red triangles on the corresponding positions of the circuit board. Clicking on any symbol can view the detailed temperature change curve of that unit. Note: The projection formula parameter 2.5 is determined based on the ratio of the display pixel density to the actual circuit size. The offsets of 50 and 30 are used for centering the display. The color saturation adjustment adopts the HSV color space, where the saturation of the third-level risk symbol is set to 0.9 and the second-level to 0.5.

[0087] In this embodiment, a combination of technologies—maintaining spatial accuracy through coordinate transformation, highlighting risk differences through symbolic representation, and ensuring professionalism through templated synthesis—achieves a visual representation of the circuit thermal status monitoring results of the inspection robot, enabling maintenance personnel to intuitively and quickly grasp the distribution of circuit thermal anomalies.

[0088] To improve the accuracy of circuit temperature prediction for inspection robots, refer to Figure 5 In some embodiments, step S102: inputting the multidimensional parameters into a pre-trained heat conduction model and outputting temperature evolution data of the inspection robot circuit includes: Step S501: Obtain the boundary parameter sequence, wherein each boundary parameter in the boundary parameter sequence is generated during the parameter adjustment process by a dynamic optimization algorithm.

[0089] In step S501, the boundary parameter sequence refers to the set of historical values ​​of convective heat transfer coefficient and radiation coefficient generated by the dynamic optimization algorithm during the iterative optimization process, reflecting the adjustment trajectory of model parameters as the environment changes.

[0090] In this embodiment of the application, by recording the effective results of each parameter optimization in real time, a parameter sequence sorted by time is formed to ensure that the latest optimized parameters can be obtained for each calculation.

[0091] Step S502: Combine the multidimensional parameters with the latest parameter value in the boundary parameter sequence to form a model input vector.

[0092] In step S502, "latest" in the latest parameter value combination refers to the boundary parameter values ​​output by the algorithm in the last iteration within the current parameter adjustment cycle, relative to the iterative process of the dynamic optimization algorithm. These parameter values ​​are the set of valid calculation results that are closest to the current time and meet the preset convergence conditions during the continuous optimization process. The model input vector is structured input data formed by integrating environmental monitoring parameters with optimized boundary parameters.

[0093] In this embodiment, multi-dimensional parameters such as airflow velocity and light intensity are combined with the convective heat transfer coefficient and radiation coefficient at the end of the boundary parameter sequence in a fixed format to form a complete model input data structure.

[0094] Step S503: Calculate the temperature prediction values ​​of each location point of the inspection robot circuit within a future set time period using the heat conduction model based on the model input vector.

[0095] In step S503, the set of temperature prediction values ​​is the result of the heat conduction model calculating the temperature changes of the discretized circuit nodes over future time periods.

[0096] In this embodiment, based on the parameters in the input vector, the heat conduction equation is solved on the spatially discretized circuit model, and the predicted temperature value of each node at different time points is output.

[0097] Step S504: Organize all the temperature prediction values ​​in chronological order to form temperature evolution data.

[0098] In this embodiment of the application, the predicted node temperature values ​​at each time step are arranged in chronological order to form a complete temperature change dataset containing location identifiers and time identifiers.

[0099] Here is a specific example: When an inspection robot performs a high-temperature inspection task in area E, it first obtains the latest boundary parameter sequence generated by the dynamic optimization algorithm. The latest set of parameters is a convective heat transfer coefficient of 26 W / m² Kelvin and an emissivity of 0.86. These two parameters are the optimization results obtained by minimizing the temperature prediction error of the last 10 times, specifically calculated as argmin∑(T_predicted - T_measured)². Real-time environmental parameters are collected: altitude 1600 meters measured by a barometric pressure sensor, airflow velocity 7.5 m / s obtained by an airborne anemometer, and light intensity 1150 W / m² collected by a photoelectric sensor. These three parameters, combined with the optimized boundary parameters, form the model input vector [1600, 7.5, 1150, 26, 0.86]. When calculating using the heat conduction model, the circuit board is discretized into 40 nodes, based on the heat conduction equation... The solution is performed, where ρ = 1800 kg / m³ is the circuit board density, c_p = 1100 joules / kg Kelvin is the specific heat capacity, and k = 3.2 watts / m Kelvin is the thermal conductivity. The temperature values ​​of each node in the next 30 seconds are calculated. For example, the temperature of node 12 is 58.7℃ at 10 seconds and 61.2℃ at 20 seconds. Finally, the predicted temperature values ​​of all nodes are sorted in chronological order to form a temperature evolution dataset containing 1200 data items. Each data item contains three fields: node number, time point, and temperature value, providing basic data for subsequent virtual model calibration.

[0100] In this embodiment, by co-processing dynamically optimized boundary parameters and environmental parameters, the heat conduction model can adapt to changes in the flight environment, and the output temperature evolution data is more in line with the actual working conditions, providing a reliable data foundation for subsequent condition monitoring.

[0101] To improve the accuracy of circuit temperature prediction for inspection robots, refer to Figure 6 In some embodiments, step S503: the calculation based on the model input vector using the heat conduction model to generate a set of predicted temperature values ​​for each location point of the inspection robot circuit within a future set time period includes: Step S601: Extract airflow velocity parameters and light intensity parameters from the model input vector.

[0102] In step S601, the airflow velocity parameter and the light intensity parameter refer to the key thermal influence factors extracted from the multidimensional environmental parameters, which respectively characterize the air convection heat transfer intensity and the solar radiation heat load.

[0103] In this embodiment of the application, airflow velocity and light intensity data are separated from the input vector. These two parameters directly determine the heat exchange boundary conditions of the circuit surface.

[0104] Step S602: Configure the input conditions of the heat conduction model according to the airflow velocity parameters and the light intensity parameters.

[0105] In step S602, input condition configuration refers to the process of converting the original environmental parameters into physical boundary conditions that the heat conduction model can handle.

[0106] In this embodiment, the airflow velocity is converted into the convective heat transfer coefficient for boundary condition calculation, and the light intensity is converted into the radiative heat flux density as the surface heat source term.

[0107] Step S603: In the heat conduction model, the geometric structure of the inspection robot circuit is spatially discretized to obtain multiple discrete nodes.

[0108] In step S603, the geometric structure of the inspection robot circuit is derived from the circuit physical structure parameters built into the pre-trained heat conduction model. These parameters have been determined during the model training phase through circuit design drawings and physical measurement data, including fixed geometric features such as the circuit board dimensions, component layout, and material layering structure. Spatial discretization refers to the process of dividing a continuous circuit structure into a finite number of computational units.

[0109] In this embodiment of the application, the circuit is divided into several grid nodes according to the actual size of the circuit board and the component layout, and each node represents the temperature characteristics of a specific area.

[0110] Step S604: Solve the transient heat conduction equation of each discrete node under the input conditions using numerical methods to obtain the temperature prediction value of each discrete node at each preset time point within a future set time period.

[0111] In step S604, "numerical method" specifically refers to numerical calculation techniques used to solve transient heat conduction equations. This includes three commonly used discretization methods in engineering thermal analysis: the finite difference method, the finite element method, and the finite volume method. The finite difference method discretizes by approximating the derivative through the difference quotient; the finite element method discretizes based on variational principles and interpolation functions; and the finite volume method discretizes by controlling the volume integral. Solving the transient heat conduction equation involves calculating the temperature change over time at each discrete node using numerical methods.

[0112] In this embodiment, the finite difference method is used to iteratively solve the discretized heat conduction equation to obtain the temperature prediction values ​​of each node at different time steps.

[0113] Step S605: Organize the temperature prediction values ​​of each discrete node at the same preset time point into a temperature prediction value set according to spatial location.

[0114] In this embodiment of the application, the temperature values ​​of each node at the same time are arranged according to the physical location relationship of the circuit to form a temperature field snapshot containing spatial information.

[0115] Here is a specific example: When a certain inspection robot performs a task in area G, it first extracts two key parameters from the model input vector [1650, 8.2, 1180, 27, 0.87]: airflow velocity of 8.2 m / s and light intensity of 1180 W / m². Based on the principles of heat transfer, the airflow velocity is converted into the convective heat transfer coefficient, specifically using the formula h = 0.664·(ρvL / μ)^(1 / 2)·Pr^(1 / 3)·(k / L), where ρ is the air density (0.9 kg / m³), v is the airflow velocity (8.2 m / s), L is the characteristic length (0.15 m), and μ is the aerodynamic viscosity (2.0 × 10⁻⁶). -5 Pascal-second, Pr is the Prandtl number taken as 0.7, k is the thermal conductivity of air taken as 0.03 W / m Kelvin, and the calculated h = 30.5 W / m Kelvin is obtained. Simultaneously, the light intensity is converted to radiative heat flux density q = εσT^4, where ε is the emissivity of radiation (0.87), σ is the Stefan-Boltzmann constant (5.67 × 10^-8 W / m Kelvin to the power of 4), and T is the ambient temperature taken as 303 Kelvin, and the calculated q = 412 W / m. The circuit board geometry is discretized into 45 nodes, with node spacing ranging from 4 to 6 mm depending on the component layout. The transient heat conduction equation is solved using the finite difference method. Where Q is the heat source term including the aforementioned radiative heat flux density, the calculated temperature of node 18 is 61.5℃ at 15 seconds and 63.8℃ at 25 seconds; the temperature values ​​of each node at the same moment are arranged according to the actual position of the circuit board, such as the set of predicted temperature values ​​at 15 seconds as [59.2,60.1,...,61.5,...62.3]℃, a total of 45 data points, forming complete temperature field distribution data.

[0116] In this embodiment, accurate prediction of the circuit temperature field of the inspection robot is achieved through precise environmental parameter conversion, reasonable spatial discretization, and reliable numerical solution, providing a high-quality data foundation for subsequent condition monitoring.

[0117] To improve the accuracy of circuit temperature monitoring in inspection robots, refer to Figure 7 In some embodiments, step S103: dynamically calibrating the temperature evolution data based on the virtual model includes: Step S701: Obtain the actual measurement value collected by the temperature sensor bound in the virtual model at the target timestamp.

[0118] In step S701, the actual measured value refers to the real temperature data collected by the physical temperature sensor bound to the virtual model at a specific moment, reflecting the actual thermal state of the circuit.

[0119] In this embodiment of the application, the measured temperature values ​​of each sensor node at a specified time point are obtained in real time through the digital twin data channel, and these values ​​are established in a one-to-one correspondence with the virtual model nodes.

[0120] Step S702: Extract the calculated temperature value of the corresponding spatial location under the target timestamp from the temperature evolution data stream.

[0121] In step S702, the calculated temperature value refers to the temperature result predicted by the heat conduction model at the same time point and spatial location.

[0122] In this embodiment of the application, predicted values ​​that spatiotemporally match the measured data are retrieved from the temperature evolution data stream to ensure that the comparison data have the same spatiotemporal reference.

[0123] Step S703: Calculate the numerical deviation between the actual measured value and the calculated temperature value.

[0124] In step S703, the numerical deviation is the absolute value of the temperature difference between the measured value and the predicted value.

[0125] In this embodiment of the application, the difference between the two is calculated by simple arithmetic operations, and this value reflects the prediction error of the model under the current working conditions.

[0126] Step S704: When the numerical deviation exceeds the set tolerance, dynamically calibrate the temperature evolution data to obtain calibrated temperature evolution data.

[0127] In this embodiment, when the deviation exceeds the allowable range, a weighted compensation algorithm is used to progressively adjust the subsequent predicted values, making the predicted curve gradually approach the measured curve. For example, when the inspection robot flies to the 85th second, the actual measured value is 62.3℃ obtained through temperature sensor No. 12 bound in the virtual model. This sensor is located at the heatsink of the main control chip. Simultaneously, the calculated temperature value at the same time point (85 seconds) corresponding to the main control chip location is extracted from the temperature evolution data stream; this value is 58.7℃, calculated by the heat conduction model based on prior environmental parameters. The numerical deviation ΔT = 62.3℃ - 58.7℃ = 3.6℃ is calculated. The deviation compensation formula Tcalibrated = Toriginal + α·ΔT is used to dynamically calibrate the subsequent data stream, where the compensation coefficient α is 0.85, determined through historical data regression analysis. Starting from the 85th second, the calculated value of the main control chip position in the temperature evolution data stream is uniformly increased by 3.1℃, so that the temperature value at this position is calibrated from 59.2℃ to 62.3℃ in the 86th second and from 59.8℃ to 62.9℃ in the 87th second, realizing real-time correction of the data stream.

[0128] Here is a specific example: When the inspection robot performs its task in area B, based on the system architecture of embodiment 1, the actual temperature sensor corresponding to node G-12 bound in the virtual model measures a temperature of 64.8℃ at time 120 seconds. Simultaneously, the model prediction value for the same time point at this node, extracted from the temperature evolution data stream, is 61.2℃. The calculated numerical deviation ΔT = 64.8 - 61.2 = 3.6℃ is then calculated. Since this deviation exceeds the set tolerance threshold of 2.0℃, a dynamic calibration mechanism is triggered. The subsequent predicted values ​​are corrected using the compensation formula T_calibration = T_original + α·ΔT, where α = 0. .82 is the weighting coefficient optimized based on historical calibration results; the original predicted value of node G-12 at 125 seconds (62.5℃) is corrected to 62.5 + 0.82 × 3.6 = 65.45℃, and the original predicted value at 130 seconds (63.8℃) is corrected to 63.8 + 0.82 × 3.6 = 66.75℃; at the same time, this weighting coefficient is fed back to the dynamic optimization algorithm for boundary parameter adjustment, forming a closed-loop optimization system; the calibrated temperature evolution data is used for subsequent state monitoring map generation, making the calculation of the temperature change slope of node G-12 during 250-255 seconds more accurate.

[0129] In this embodiment, by comparing data in real time and performing dynamic compensation calibration, the systematic deviation between model prediction and actual situation is effectively eliminated, the reliability of temperature monitoring results is improved, and more accurate data support is provided for the circuit health management of the inspection robot.

[0130] Figure 8A schematic diagram of a smart inspection control system for an inspection robot based on edge computing, provided in an embodiment of this application, is shown below. Figure 8 As shown, the system includes: The acquisition module 21 is used to acquire multi-dimensional parameters of the inspection robot during flight.

[0131] The input module 22 is used to input the multidimensional parameters into the pre-trained heat conduction model and output the temperature evolution data of the inspection robot circuit. The boundary parameters of the heat conduction model are obtained by multiple rounds of iterative adjustment through a dynamic optimization algorithm.

[0132] The construction module 23 is used to construct a virtual model corresponding to the circuit of the inspection robot in the digital twin framework, and to dynamically calibrate the temperature evolution data based on the virtual model.

[0133] The generation module 24 is used to generate a state monitoring map of the inspection robot circuit based on the calibrated temperature evolution data and the virtual model.

[0134] Figure 8 The aforementioned intelligent inspection control system for inspection robots based on edge computing can execute... Figure 1 The implementation principle and technical effects of the edge computing-based intelligent inspection control method for inspection robots described in the illustrated embodiment will not be repeated here. The specific methods by which each module and unit of the edge computing-based intelligent inspection control system for inspection robots in the above embodiments have been described in detail in the embodiments related to this method, and will not be elaborated upon here.

[0135] In one possible design, Figure 8 The edge computing-based intelligent inspection control system for inspection robots shown in the embodiment can be implemented as a computing device, such as... Figure 9 As shown, the computing device may include a storage component 31 and a processing component 32; The storage component 31 stores one or more computer instructions, wherein the one or more computer instructions are invoked and executed by the processing component 32.

[0136] The processing component 32 is used to perform the above. Figure 1 The embodiment describes an intelligent inspection control method for an inspection robot based on edge computing.

[0137] The processing component 32 may include one or more processors to execute computer instructions to complete all or part of the steps in the above-described method. Alternatively, the processing component may be implemented as one or more application-specific integrated circuits (ASICs), digital signal processors (DSPs), digital signal processing devices (DSPDs), programmable logic devices (PLDs), field-programmable gate arrays (FPGAs), controllers, microcontrollers, microprocessors, or other electronic components to execute the above-described method.

[0138] Storage component 31 is configured to store various types of data to support operations at the terminal. The storage component can be implemented from any type of volatile or non-volatile storage device or a combination thereof, such as Static Random-Access Memory (SRAM), Electrically Erasable Programmable Read Only Memory (EEPROM), Erasable Programmable Read Only Memory (EPROM), Programmable Read Only Memory (PROM), Read Only Memory (ROM), magnetic storage, flash memory, magnetic disk, or optical disk.

[0139] Of course, computing devices may also include other components, such as input / output interfaces, display components, communication components, etc.

[0140] Input / output interfaces provide interfaces between processing components and peripheral interface modules, which can be output devices, input devices, etc.

[0141] The communication components are configured to facilitate wired or wireless communication between computing devices and other devices.

[0142] The computing device can be a physical device or an elastic computing host provided by a cloud computing platform. In this case, the computing device can refer to a cloud server, and the aforementioned processing components, storage components, etc., can be basic server resources rented or purchased from the cloud computing platform.

[0143] This application also provides a computer storage medium storing a computer program, which, when executed by a computer, can perform the above-described functions. Figure 1 The embodiment shown is an intelligent inspection control method for an inspection robot based on edge computing.

[0144] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the specific working processes of the systems and units described above can be referred to the corresponding processes in the foregoing method embodiments, and will not be repeated here.

[0145] The system embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs. Those skilled in the art can understand and implement this without any creative effort.

[0146] Through the above description of the embodiments, those skilled in the art can clearly understand that each embodiment can be implemented by means of software plus necessary general-purpose hardware platforms, and of course, it can also be implemented by hardware. Based on this understanding, the above technical solutions, in essence 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 ROM / RAM, magnetic disk, optical disk, etc., and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute the methods described in the various embodiments or some parts of the embodiments.

[0147] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of this application, and are not intended to limit them. Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of this application.

Claims

1. A method for intelligent inspection control of an inspection robot based on edge computing, characterized in that, include: Acquire multi-dimensional parameters of the inspection robot during its flight; The multidimensional parameters are input into a pre-trained heat conduction model, and the temperature evolution data of the inspection robot circuit is output. The boundary parameters of the heat conduction model are obtained by multiple rounds of iterative adjustment through a dynamic optimization algorithm. A virtual model corresponding to the circuitry of the inspection robot is constructed within a digital twin framework, and the temperature evolution data is dynamically calibrated based on the virtual model. Based on the calibrated temperature evolution data and the virtual model, a state monitoring map of the inspection robot circuit is generated; The process of generating a state monitoring map of the inspection robot circuit based on the calibrated temperature evolution data and the virtual model includes: The calibrated temperature evolution data is mapped to the three-dimensional spatial topology mesh of the virtual model according to spatial coordinates; Extract the temperature change trajectory data corresponding to the monitoring period from each grid cell in the three-dimensional spatial topology grid; Based on the temperature change trajectory data, calculate the instantaneous slope of each grid cell; The grid cells with instantaneous slopes exceeding a set threshold are combined into a grid cell set; Visually label the set of grid cells in the three-dimensional spatial topological grid of the virtual model to form a status monitoring map; The step of visually marking the set of grid cells in the three-dimensional spatial topological grid of the virtual model to form a state monitoring map includes: Calculate the temperature change of each grid cell in the grid cell set during the monitoring period; Based on the temperature change, and in conjunction with preset risk mapping rules, a risk level code is generated; In the three-dimensional spatial topological grid, the spatial coordinates of each grid cell are bound to the corresponding risk level code to generate data tags; A condition monitoring map is generated based on the data labels of all grid cells.

2. The method according to claim 1, characterized in that, The generation of the state monitoring map based on the data tags of all grid cells includes: Convert the spatial coordinates in the data markers of each grid cell into planar coordinates in a preset projected coordinate system; Based on the risk level code in the data markers, select symbols of the corresponding type from a preset visualization symbol library; In the projected coordinate system, the display position of each type of symbol is determined according to the planar coordinates, and the display color of each symbol is adjusted according to the risk level code; Each symbol is integrated into the circuit diagram template of the inspection robot according to its corresponding display position, display color and type to form a status monitoring diagram.

3. The method according to claim 1, characterized in that, The step of inputting the multidimensional parameters into a pre-trained heat conduction model and outputting temperature evolution data of the inspection robot circuit includes: Obtain a boundary parameter sequence, wherein each boundary parameter in the boundary parameter sequence is generated during the parameter adjustment process by a dynamic optimization algorithm; The multidimensional parameters are combined with the latest parameter value in the boundary parameter sequence to form the model input vector; The heat conduction model is used to calculate based on the model input vector to generate a set of predicted temperature values ​​for each location point of the inspection robot circuit within a future set time period. All the predicted temperature values ​​are organized in chronological order to form temperature evolution data.

4. The method according to claim 3, characterized in that, The step of calculating based on the model input vector using the heat conduction model to generate a set of predicted temperature values ​​for each location point of the inspection robot circuit within a future set time period includes: Extract airflow velocity parameters and light intensity parameters from the model input vector; Configure the input conditions for the heat conduction model based on the airflow velocity parameters and the light intensity parameters; In the heat conduction model, the geometric structure of the inspection robot circuit is spatially discretized to obtain multiple discrete nodes; The transient heat conduction equations of each discrete node under the input conditions are solved by numerical methods to obtain the temperature prediction values ​​of each discrete node at each preset time point within a future set time period. The temperature prediction values ​​of each discrete node at the same preset time point are organized into a temperature prediction value set according to their spatial location.

5. The method according to claim 1, characterized in that, The dynamic calibration of the temperature evolution data based on the virtual model includes: Obtain the actual measurement value collected by the temperature sensor bound to the virtual model at the target timestamp; Extract the calculated temperature value of the corresponding spatial location at the target timestamp from the temperature evolution data stream; Calculate the numerical deviation between the actual measured value and the calculated temperature value; When the numerical deviation exceeds the set tolerance, the temperature evolution data is dynamically calibrated to obtain calibrated temperature evolution data.

6. An intelligent inspection control system for inspection robots based on edge computing, characterized in that, include: The acquisition module is used to acquire multi-dimensional parameters of the inspection robot during flight. The input module is used to input the multidimensional parameters into the pre-trained heat conduction model and output the temperature evolution data of the inspection robot circuit. The boundary parameters of the heat conduction model are obtained by multiple rounds of iterative adjustment through a dynamic optimization algorithm. A construction module is used to build a virtual model corresponding to the circuitry of the inspection robot within a digital twin framework, and to dynamically calibrate the temperature evolution data based on the virtual model; The generation module is used to generate a state monitoring map of the inspection robot circuit based on the calibrated temperature evolution data and the virtual model. The edge computing-based intelligent inspection control system for inspection robots is used to execute the edge computing-based intelligent inspection control method for inspection robots as described in any one of claims 1-5.

7. A computing device, characterized in that, It includes a processing component and a storage component; the storage component stores one or more computer instructions; the one or more computer instructions are invoked and executed by the processing component to implement the intelligent inspection control method for an inspection robot based on edge computing as described in any one of claims 1 to 5.

8. A computer storage medium, characterized in that, The system contains a computer program that, when executed by a computer, implements an intelligent inspection control method for an inspection robot based on edge computing as described in any one of claims 1 to 5.

Images