A robotic operation method for dismantling and assembling busbars in high-voltage substations

Abstract

This invention discloses a method for the operation of a robot for dismantling and assembling a busbar in a high-voltage substation, relating to the field of power construction technology. The method includes: acquiring parameter data outside the busbar; constructing a positioning model, which, based on the parameter data, outputs a work plan for the robot, including the work location, work path, and work time; constructing a detection model, which, based on the parameter data, acquires environmental data outside the busbar affecting the robot's operation; acquiring attribute data of the robot; and constructing an influence discrimination model. This invention uses a feature tree diagram to efficiently match attribute data and environmental data, forming a precise mapping relationship. This ensures that the robot can accurately identify and understand the work environment during operation, improving data processing efficiency and enhancing the robot's adaptability to the work environment.

Description

Technical Field

[0001] This invention relates to the field of power construction technology, specifically to a method for dismantling and assembling a robot for the main busbar in a high-voltage substation. Background Technology

[0002] With the rapid development of modern industry and the improvement of people's living standards, power users are constantly increasing their requirements for power supply reliability. In order to reduce power outage time or even ensure uninterrupted power supply, power supply companies in various regions have accelerated the application and promotion of live-line working robots in substations. Substations are places in the power system that transform voltage and current, receive electrical energy, and distribute electrical energy. Therefore, the robot operation method for dismantling and assembling busbars in high-voltage substations is crucial for the application of robots in substations.

[0003] A search revealed Chinese invention patent publication number "CN113708290 A", which discloses a "control method, device and robot terminal for threading a guide wire in a live-line working robot". This application obtains a state image before threading, identifies the pose information of the end of the guide wire from the state image, uses the pose of the end of the guide wire as a standard, adjusts the pose of the wire clamp arm to adjust the alignment of the wire clamp tool with the end of the guide wire, and allows the wire clamp arm equipped with the wire clamp tool to actively complete the threading action, avoiding the situation where the pose of the guide wire is deviated due to arm movement, thereby improving the threading success rate.

[0004] In addition, Chinese invention patent with publication number "CN113964720A" discloses "a power distribution network live-line operation robot system and a method for disconnecting diversion lines". This application, through a unique robot structure design, enables the robot's main control to control the movement of each joint of the robotic arm through overall task planning. The autonomous control uses the identification information collected and processed by the vision device combined with a deep learning vision detection and control method to identify the object to be operated and its position, and control the robotic arm to complete the operation, achieving short-distance high-precision displacement and positioning of the robotic arm.

[0005] Finally, Chinese invention patent CN116787466A disclosed "A method for identifying a current-carrying cable based on an insulating sleeve, a storage medium, and a robot." This application involves installing an insulating sleeve on the head of the cable during installation. The insulating sleeve has fixing holes for mounting the cable head. During installation, the head of the current-carrying cable is inserted into the fixing holes on the insulating sleeve for fixation. Multiple feature code areas composed of strong and weak reflection areas are set on the insulating sleeve. Radar point cloud data is obtained by scanning the current environment using a lidar scanner. The signal strength of the radar point cloud data is matched, as the weak reflection areas absorb most of the laser signal. LiDAR, with its strong reflective areas reflecting most of the signal, allows the LiDAR to scan the feature code area on the insulating sleeve. This means that when the LiDAR scans the feature code area, it can determine the data block corresponding to the signal feature code based on the reflection intensity of the strong and weak reflective areas. The obtained data block can then be analyzed to obtain the spatial information of the insulating sleeve. Since the insulating sleeve is placed on the cable drain line, the spatial location of the drain line can be determined based on the obtained spatial information of the insulating sleeve, allowing for identification of the drain line. Placing the easily identifiable insulating sleeve on the cable drain line makes it easier for the LiDAR to identify the cable drain line, facilitating subsequent live-line work.

[0006] However, the methods disclosed above and similar methods, when actually used, focus on the operation and recognition technology of the robotic arm, but lack comprehensive collection and fusion processing of environmental data of the work area. As a result, when the robot is working in a complex environment, changes in environmental parameters will affect the efficiency and safety of the operation. Summary of the Invention

[0007] The purpose of this invention is to provide a robotic operation method for dismantling and assembling busbars in high-voltage substations, so as to solve the problems mentioned in the background art.

[0008] To achieve the above objectives, the present invention provides the following technical solution:

[0009] A method for robotic operation of dismantling and assembling busbars in high-voltage substations includes:

[0010] Obtain parameter data outside the main drain line;

[0011] A positioning model is constructed, which outputs a work plan for the assembly / disassembly robot based on parameter data. The work plan includes: work location, work path, and work time.

[0012] Construct a detection model, and based on the detection model and parameter data, obtain environmental data outside the main drain line that affects the operation of the disassembly and assembly robot;

[0013] Obtain the attribute data of the disassembly and assembly robot;

[0014] An influence discrimination model is constructed, which outputs discrimination results based on attribute data and environmental data. The discrimination results are used to show the factors affecting the operation of the assembly and disassembly robot.

[0015] A command control model is constructed. Based on the judgment results, the command control model outputs dynamically adjusted commands, which then apply to the work plan.

[0016] A monitoring and feedback model is constructed to track the operation status of the assembly and disassembly robot in real time. Based on the comparison between the real-time operation status of the assembly and disassembly robot and the preset target, adjustment data is generated. The adjustment data acts on the command control model and affects the dynamic adjustment command.

[0017] The influence discrimination model includes:

[0018] The feature data of attribute data and environmental data are obtained separately. The feature data that matches the attribute data and environmental data are selected and a mapping relationship is constructed. Based on the mapping relationship, the attribute data and environmental data are integrated into several feature-based fusion datasets.

[0019] Acquire features within the assembly / disassembly robot's work plan and assign high weights to fused data that match the features within the assembly / disassembly robot's work plan;

[0020] A preset discrimination threshold is set manually based on the historical work plan and safety regulations of the disassembly and assembly robot. The discrimination threshold is compared with the data values ​​in the high-weighted fused data, and the discrimination result is output.

[0021] As a further preferred embodiment of this technical solution, the feature data within the attribute data includes: the length of the robotic arm of the assembly / disassembly robot, the rotation range of the robotic arm, the type of tools assembled by the assembly / disassembly robot, and the specifications of the assembly / disassembly tools assembled by the assembly / disassembly robot.

[0022] The environmental data features include: temperature, humidity, wind speed, light intensity, electromagnetic interference level, and material and specifications of the main drain line in the work area.

[0023] As a further preferred embodiment of this technical solution, the method for constructing the mapping relationship includes:

[0024] Construct a feature tree diagram. The feature tree diagram is constructed with several branch nodes based on the feature data types in the attribute data and environmental data. Each branch node has several child nodes, and the number of child nodes represents the specific feature data values.

[0025] The robot acquires real-time operational data. Based on this data, it retrieves feature data values ​​that match the attribute data and environmental data in a feature tree diagram. The child nodes corresponding to the matching feature data values ​​are then connected to form a mapping path, which represents the mapping relationship.

[0026] As a further preferred embodiment of this technical solution, the method for comparing the discrimination threshold and the data values ​​within the high-weighted fusion dataset includes:

[0027] Based on the feature retrieval and discrimination threshold of the high-weight fused dataset, the data values ​​of the matching features within the dataset are determined.

[0028] If matching data values ​​are found, the data values ​​are compared directly.

[0029] If no matching feature data value is found, then data values ​​with similar features are retrieved, interpolated, and approximate values ​​are obtained before comparison.

[0030] When the data values ​​in the high-weighted fusion dataset exceed the discrimination threshold, the discrimination result indicates that there is a risk in the disassembly and assembly robot operation;

[0031] When the data value in the high-weighted fusion dataset is lower than or equal to the discrimination threshold, the discrimination result is that the disassembly and assembly robot operation is safe.

[0032] As a further preferred embodiment of this technical solution, the method for distinguishing similar feature data values ​​includes:

[0033] Features of the attribute data and environmental data are obtained separately, and a higher-level feature summary table is constructed. The higher-level feature summary table integrates and summarizes the obtained features based on physical, environmental and temporal aspects to form feature classifications at different levels.

[0034] As a further preferred embodiment of this technical solution, the method for obtaining parameter data includes:

[0035] Several parameter acquisition nodes, based on LiDAR technology, are constructed and evenly distributed around the main pipe drainage line. A data encryption and cleaning method is set up to clean redundant data in real time during data collection and output the cleaned data. A data shielding enhancement model is constructed to acquire multiple sets of cleaned data in adjacent time periods in real time, and obtain parameter data by extracting either the median or the average value. An intermediate storage node is created to acquire and back up the parameter data. A feedback rule is constructed between the intermediate storage node and the data shielding enhancement model. The feedback rule presets a critical threshold, which is dynamically set based on historical parameter data. When the parameter data exceeds the critical threshold, the feedback rule will trigger the data replacement of the intermediate storage node with the latest acquired parameter data.

[0036] As a further preferred embodiment of this technical solution, the parameter data includes: the length, diameter, curvature, position coordinates, distance from surrounding objects, and current and voltage of the main drain line;

[0037] The positioning model includes:

[0038] A three-dimensional spatial model is constructed. The three-dimensional spatial model is based on the main pipe drainage line and a three-dimensional coordinate system is established. The parameter data of the main pipe drainage line is converted into coordinate points and vectors in the three-dimensional space. Based on the coordinate points and vectors, a three-dimensional spatial model of the main pipe drainage line is constructed. In addition, the three-dimensional spatial model is centered on the main pipe drainage line and simulates the spatial environment around the main pipe drainage line. Based on the position coordinates in the parameter data and the distance to the surrounding objects, the simulation space is divided into different work areas. Each work area corresponds to the possible work position and work path of the disassembly and assembly robot.

[0039] The operation location and path of the disassembly and assembly robot in each operation area are obtained and combined with the length of the main pipe guide line. The operation time of the disassembly and assembly robot in each operation area is calculated. The operation time, operation location and operation path are summarized and integrated to form the operation plan of the disassembly and assembly robot.

[0040] As a further preferred embodiment of this technical solution, the detection model includes:

[0041] An environmental perception system is constructed based on the cross-combination of several different sensing units. It is used to collect environmental data of the work area in real time, obtain parameter data features, label sensing units based on feature types, construct a fusion algorithm, assign high weight to labeled sensing units, assign equal weight to the remaining sensing units within different sensing units, and fuse the data of labeled sensing units and the data of the remaining units to obtain environmental data.

[0042] The formula for the fusion algorithm is as follows: Where d m This represents the environmental data value collected by the sensor unit m, which is marked as having high weight, and α represents the weighting coefficient assigned to the high-weight sensor unit m. d i This represents the sum of environmental data values ​​collected by all other sensing units (n-1 units) except for the high-weight sensing unit m. This is a normalization factor used to evenly distribute the sum of data from other sensing units across each sensing unit.

[0043] As a further preferred embodiment of this technical solution, the instruction control model includes:

[0044] The system acquires the features of the discrimination results, including the risk level of the operation, the safe area of ​​the operation, and the priority of the operation. It then constructs an instruction generation system, which dynamically generates control instructions based on the features of the discrimination results, the operation plan and capabilities of the disassembly and assembly robot, and an adaptive control strategy based on machine learning.

[0045] The work plan includes: work area selection, work tool selection, work path planning, and work time allocation.

[0046] As a further preferred embodiment of this technical solution, the monitoring feedback model includes:

[0047] Acquire real-time operation data of the assembly / disassembly robot, extract features from the real-time operation data, construct an operation status database based on the features in the operation data, and use the operation status database to update the operation data of the assembly / disassembly robot in real time.

[0048] By linking the operational status database and timeline, extracting operational data parameters from multiple adjacent time periods, analyzing the magnitude of change in operational data parameters, and predicting the operational status of the disassembly and assembly robot in future time periods based on the magnitude of change, the operational status parameter values ​​for future time periods are set as preset targets.

[0049] Compare preset targets with real-time running data, and generate adjustment data based on the deviation.

[0050] Compared with the prior art, the beneficial effects of the present invention are:

[0051] The robot operation method for dismantling and assembling the busbar in a high-voltage substation uses a feature tree diagram to efficiently match attribute data and environmental data, forming a precise mapping relationship. This ensures that the robot can accurately identify and understand the working environment during operation, improves data processing efficiency, and enhances the robot's adaptability to the working environment.

[0052] Secondly, by comparing data values ​​within a high-weighted fusion dataset with a discrimination threshold, this invention can accurately and in real time determine the risk level of the assembly and disassembly robot operation, enabling the robot to fully assess the working environment before operation, thereby avoiding potential safety risks.

[0053] Furthermore, by constructing parameter acquisition nodes and a data masking enhancement model, this invention can acquire and clean parameter data in real time, ensuring the accuracy and reliability of the data. At the same time, by constructing a three-dimensional spatial model and a detection model, the robot can better understand the working environment and improve the accuracy of the operation.

[0054] Finally, this invention achieves comprehensive monitoring and dynamic control of the assembly and disassembly robot's operation through an instruction control model and a monitoring feedback model. By acquiring the robot's operation data in real time, predicting its future operating status, and generating adjustment data based on deviations, the robot can make adaptive adjustments during the operation, ensuring the smooth progress of the operation.

[0055] It should also be noted that by constructing an influence discrimination model, this invention enables the dismantling and assembly robot to comprehensively consider multiple influencing factors when making operational risk assessments, thereby obtaining more comprehensive and accurate judgment results. This improves the safety and stability of the dismantling and assembly robot in the dismantling and assembly of the busbar lead-in line in high-voltage substations, and also provides more reliable technical support for the operation and maintenance management of substations. Attached Figure Description

[0056] Figure 1 This is a flowchart illustrating the operational logic of the method of the present invention.

[0057] Figure 2 This is a flowchart illustrating the operational logic of the impact discrimination model of this invention.

[0058] Figure 3 This is a diagram illustrating the method for constructing the mapping relationship in this invention;

[0059] Figure 4 This is a diagram illustrating the method for obtaining parameter data in this invention;

[0060] Figure 5 The operational logic diagram for the invention positioning model;

[0061] Figure 6 This is a flowchart illustrating the operational logic of the detection model of this invention.

[0062] Figure 7 This is a flowchart illustrating the operational logic of the monitoring and feedback model of this invention. Detailed Implementation

[0063] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the scope of protection of the present invention.

[0064] The application of this invention in the scenario of dismantling and assembling busbars in high-voltage substations requires clarification. Specifically, in this scenario, the robot operation method for dismantling and assembling busbars in high-voltage substations proposed in this application necessitates ensuring the stability and safety of the robot system before its operation. Firstly, a comprehensive system check must be conducted to ensure that all robot components, such as the robotic arm, sensors, and control system, are in good working order. Secondly, a detailed survey of the work area is required to clarify the location, height, angle, and other information of the busbars, providing accurate data support for the robot's operation.

[0065] like Figures 1-7 As shown, this solution includes: acquiring parameter data outside the main pipe drainage line; constructing a positioning model; the positioning model outputs a work plan for the assembly / disassembly robot based on the parameter data; the work plan includes: work location, work path, and work time; constructing a detection model; the detection model acquires environmental data outside the main pipe drainage line that affects the assembly / disassembly robot's operation based on the parameter data; acquiring attribute data of the assembly / disassembly robot; constructing an influence discrimination model; the influence discrimination model outputs discrimination results based on the attribute data and environmental data; the discrimination results are used to display the factors affecting the assembly / disassembly robot's operation; constructing an instruction control model; the instruction control model outputs dynamic adjustment instructions based on the discrimination results; the dynamic adjustment instructions act on the work plan; and constructing a monitoring feedback model; the monitoring feedback model is used to track the assembly / disassembly robot's operation status in real time, and generates adjustment data based on the comparison between the assembly / disassembly robot's real-time operation status and the preset target; the adjustment data acts on the instruction control model and influences the dynamic adjustment instructions.

[0066] It should be noted that, in this embodiment, the parameter data includes: the length, diameter, curvature, position coordinates, distance from surrounding objects, and current and voltage of the main drain line; the work plan includes: work area selection, work tool selection, work path planning, and work time allocation; the characteristic data within the attribute data includes: the length of the robotic arm of the assembly / disassembly robot, the range of rotation of the robotic arm, the type of tools equipped on the assembly / disassembly robot, and the specifications of the assembly / disassembly tools equipped on the assembly / disassembly robot; and the characteristic data of the environmental data includes: the temperature, humidity, wind force, light intensity, electromagnetic interference level, and material and specifications of the main drain line in the work area.

[0067] It should be added that during the actual operation of the assembly / disassembly robot, due to the complex and ever-changing working environment, the real-time acquired data may contain noise or interference. In order to ensure the accuracy and stability of the acquired data, this invention further introduces a data cleaning and verification mechanism. Specifically, after acquiring the real-time operating data of the assembly / disassembly robot, the monitoring feedback model first cleans the data to remove outliers and noise, ensuring the accuracy and reliability of the data. Then, through the data verification mechanism, the cleaned data is verified to ensure that it meets the preset data range and format requirements.

[0068] As a preferred embodiment, refer to Figure 2 As can be seen, in this application, the influence discrimination model includes: acquiring feature data of attribute data and environmental data respectively; selecting feature data that matches attribute data and environmental data and constructing a mapping relationship; integrating attribute data and environmental data into several feature-based fusion datasets based on the mapping relationship; acquiring features within the disassembly and assembly robot's work plan; assigning high weights to fusion data that matches the features within the disassembly and assembly robot's work plan; setting a discrimination threshold based on the disassembly and assembly robot's historical work plan and safety regulations; comparing the discrimination threshold with the data values ​​within the high-weight fusion dataset; and outputting the discrimination result.

[0069] It should be noted that in this embodiment, the mapping relationship is the correspondence between characteristic data such as the length of the robotic arm, the range of rotation of the robotic arm, and the type and specifications of tools assembled by the disassembly and assembly robot in the attribute data, and characteristic data such as the temperature, humidity, wind force, light intensity, and electromagnetic interference level of the working area in the environmental data. This mapping relationship is used to help the disassembly and assembly robot better understand the working environment, thereby making more accurate judgment results. In addition, the judgment threshold setting of the influence judgment model is not fixed in the actual operation process, but will be dynamically adjusted according to the real-time feedback data and historical operation data of the disassembly and assembly robot during the operation to adapt to the needs of different working environments and tasks. For example, when the disassembly and assembly robot is performing disassembly and assembly operations, if it finds that the temperature or humidity of the working area exceeds the preset judgment threshold, the influence judgment model will immediately respond, output the corresponding judgment result, and send a dynamic adjustment command to the disassembly and assembly robot through the command control model, requiring the robot to adjust the operation plan, such as changing the operation path or adjusting the operation time, to ensure the safety and stability of the operation.

[0070] Furthermore, it should be added that, in actual operation, the work plan output by the positioning model in this implementation may include any one or more of the following: work location, work path, and work time. When only one feature exists, the fusion dataset that matches the feature in the assembly / disassembly robot work plan is given a high weight. When multiple features exist, the weights of the fusion dataset that matches each feature in the assembly / disassembly robot work plan are equally distributed.

[0071] As a preferred embodiment, refer to Figure 3 As can be seen, in this application, the method for constructing the mapping relationship includes: constructing a feature tree diagram, wherein the feature tree diagram constructs several branch nodes based on the feature data types in the attribute data and environmental data, and each branch node has several child nodes, the number of child nodes representing specific feature data values; obtaining real-time operation data of the assembly / disassembly robot; based on the real-time operation data, retrieving feature data values ​​that match the attribute data and environmental data in the feature tree diagram; connecting the child nodes corresponding to the matching feature data values ​​to form a mapping path, and the mapping path is the mapping relationship.

[0072] It should be noted that in actual operation, this implementation method also needs to verify and optimize the mapping relationship to ensure its accuracy and effectiveness. Specifically, when the assembly / disassembly robot generates new operating data during actual operation, this new data will be used to verify the mapping relationship. If the verification result shows that the mapping relationship can accurately reflect the relationship between the assembly / disassembly robot's operating environment and attributes, the mapping relationship is considered valid and can continue to be used. However, if the verification result shows that there is a deviation or error in the mapping relationship, the mapping relationship needs to be optimized. The optimization methods include: adjusting the branch nodes and child nodes of the feature tree diagram based on the new operating data, adding or deleting certain feature data values, and adjusting the mapping path, etc.

[0073] In a preferred embodiment, the method for comparing the discrimination threshold and the data values ​​within the high-weight fusion dataset in this application includes: retrieving data values ​​of matching features within the discrimination threshold based on features within the high-weight fusion dataset; if matching feature data values ​​are found, direct data value comparison is performed; if no matching feature data values ​​are found, data values ​​of similar features are retrieved, interpolation calculation is performed, and an approximate value of the feature is obtained before comparison; when the data values ​​within the high-weight fusion dataset exceed the discrimination threshold, the discrimination result is that the disassembly and assembly robot operation is risky; when the data values ​​within the high-weight fusion dataset are lower than or equal to the discrimination threshold, the discrimination result is that the disassembly and assembly robot operation is safe.

[0074] As a preferred embodiment, in this application, the method for distinguishing similar feature data values ​​includes: acquiring the features of feature data within attribute data and environmental data respectively, constructing a higher-level feature summary table, and integrating and summarizing the acquired features based on physical, environmental and temporal aspects to form feature classifications at different levels.

[0075] It should be noted that, in this embodiment, the construction of the superordinate feature summary table aims to provide a systematic framework to help the assembly / disassembly robot better understand and analyze the complexity of the working environment. By integrating and summarizing physical, environmental, and temporal features, the superordinate feature summary table can form a multi-level feature classification system, thereby achieving a more comprehensive and detailed description of the working environment. In actual operation, the assembly / disassembly robot retrieves the corresponding feature classification from the superordinate feature summary table based on the real-time acquired operating data, and performs more accurate operational risk assessment and judgment based on these feature classifications.

[0076] As a preferred embodiment, refer to Figure 4 As can be seen, the method for obtaining parameter data in this application includes: constructing several parameter acquisition nodes based on LiDAR technology, with the nodes evenly distributed around the main pipe drainage line; setting a data encryption and cleaning method; cleaning redundant data in real time during data collection and outputting the cleaned data; constructing a data shielding enhancement model; acquiring multiple sets of cleaned data in adjacent time periods in real time, and obtaining parameter data by extracting either the median or the average value; creating an intermediate storage node; acquiring and backing up the parameter data; and constructing a feedback rule between the intermediate storage node and the data shielding enhancement model. The feedback rule presets a critical threshold, which is dynamically set based on historical parameter data. When the parameter data exceeds the critical threshold, the feedback rule will trigger data replacement at the intermediate storage node, replacing it with the latest acquired parameter data.

[0077] It should be noted that, in this embodiment, when the method of extracting median values ​​is used to obtain parameter data, if the data fluctuation is small and the overall trend is stable, then the median value effectively represents the overall state of the current parameter data, reducing the impact of extreme values ​​or noisy data. At the same time, the method of obtaining median values ​​can also reduce the system's demand for data processing speed, thereby improving the stability and reliability of the system to a certain extent. When the method of extracting average values ​​is used to obtain parameter data, the average value can more comprehensively reflect the overall state of the data. Especially when the data fluctuation is large or there are extreme values, the average value can weaken the impact of extreme values ​​on the overall data through the "averaging" effect, providing more robust and accurate parameter data.

[0078] As a preferred embodiment, refer to Figure 5It can be seen that the positioning model includes: constructing a three-dimensional spatial model. The three-dimensional spatial model is based on the main pipe drainage line, establishing a three-dimensional coordinate system, converting the parameter data of the main pipe drainage line into coordinate points and vectors in the three-dimensional space, and constructing a three-dimensional spatial model of the main pipe drainage line based on the coordinate points and vectors. In addition, the three-dimensional spatial model is centered on the main pipe drainage line and simulates the spatial environment around the main pipe drainage line. Based on the position coordinates in the parameter data and the distance to surrounding objects, the simulated space is divided into different work areas. Each work area corresponds to the possible work position and work path of the disassembly and assembly robot. The work position and work path of the disassembly and assembly robot in each work area are obtained and combined with the length of the main pipe drainage line to calculate the work time of the disassembly and assembly robot in each work area. The work time, work position and work path are summarized and integrated to form the work plan of the disassembly and assembly robot.

[0079] It should be added that, based on the content of this implementation method and Figure 5 As can be seen, the integrated calculation involves obtaining the working position and path of the assembly / disassembly robot in each work area and combining them with the length of the main pipe guide line. The calculation also involves determining the working time of the assembly / disassembly robot in each work area, summarizing the working time, working position, and working path, and integrating them to form a work plan for the assembly / disassembly robot. It should be further explained that the integrated calculation method uses computer vision and machine learning algorithms to perform in-depth analysis of the coordinate points, vectors, and work areas in the 3D spatial model. First, based on the working position and path of the assembly / disassembly robot in each work area, combined with the length and shape of the main pipe guide line, the required movement distance and work complexity of the robot are assessed. Then, using historical work data and time series analysis, the time required for the assembly / disassembly robot to complete the work in each work area is predicted. Finally, the predicted working time is combined with the actual working position and working path to generate a detailed work plan.

[0080] As a preferred embodiment, refer to Figure 6 As can be seen, in this embodiment, the detection model includes: constructing an environmental perception system, which is constructed based on the cross-combination of several different sensing units, for real-time acquisition of environmental data in the work area, obtaining parameter data features, labeling sensing units based on feature types, constructing a fusion algorithm, assigning high weights to the labeled sensing units, assigning equal weights to the remaining sensing units within different sensing units, and fusing the data of the labeled sensing units and the data of the remaining units to obtain environmental data.

[0081] It should be noted that the fusion algorithm is as follows: Where d m This represents the environmental data value collected by the sensor unit m, which is marked as having high weight, and α represents the weighting coefficient assigned to the high-weight sensor unit m. di This represents the sum of environmental data values ​​collected by all other sensing units (n-1 units) except for the high-weight sensing unit m. This is a normalization factor used to evenly distribute the sum of data from other sensing units across each sensing unit.

[0082] It should be noted that in actual operation, the fusion algorithm, for example, when the high-weight sensing unit is a voltage sensing unit, collects environmental data values ​​of voltage, for example, 100kV (this is just an example; the actual value may vary depending on the environment and sensing unit). Then, the total data of 100kV from the other sensing units is evenly distributed among the remaining n-1 sensing units. Assuming there are 3 sensing units (n=4), there are 2 sensing units besides the high-weight sensing unit, so each sensing unit receives 100kV / 2 = 50kV. Finally, these values ​​are substituted into the running formula of the fusion algorithm to calculate the fused environmental data values.

[0083] R = 5 * 215 kV + 2 * (50 kV / 2) = 1075 kV + 50 kV = 1125 kV (Note: The unit kV here is only for example; other units or dimensionless values ​​may be used in actual calculations).

[0084] In a preferred embodiment, the instruction control model includes: acquiring discrimination result features, which include the operation risk level, the operation safety zone, and the operation priority order; constructing an instruction generation system; and dynamically generating control instructions based on the discrimination result features, combined with the operation plan and operation capabilities of the assembly / disassembly robot, and based on an adaptive control strategy using machine learning.

[0085] As a preferred embodiment, refer to Figure 7 As can be seen, in this embodiment, the monitoring feedback model includes: acquiring real-time operation data of the assembly / disassembly robot, extracting features from the real-time operation data, constructing an operation status database based on the features in the operation data, using the operation status database to update the operation data of the assembly / disassembly robot in real time, associating the operation status database with the time axis, extracting operation data parameters in multiple adjacent time periods, analyzing the change range of the operation data parameters, predicting the operation status of the assembly / disassembly robot in future time periods based on the change range, setting the operation status parameter values ​​of future time periods as preset targets, comparing the preset targets with the real-time operation data, and generating adjustment data based on the deviation.

[0086] It should be noted that in this embodiment, the comparison of the preset target and real-time operating data, and the generation of adjustment data based on the deviation, are achieved using a deviation analysis algorithm in the prior art. This algorithm is widely used in data monitoring and predictive adjustment. It calculates the deviation value between the preset target and real-time data, and generates corresponding adjustment data based on this deviation value. This adjustment data can be used to guide the operation of the assembly / disassembly robot to ensure that it can perform its tasks according to the preset target.

[0087] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended embodiments and their equivalents.

Claims

1. A high-voltage substation pipe busbar dismounting robot operation method, characterized by, The method comprises the following steps: acquiring parameter data outside the pipe mother drainage line; constructing a positioning model based on the parameter data, the positioning model outputting a work plan of the disassembly robot, the work plan comprising a work position, a work path, and a work time; constructing a detection model based on the parameter data to acquire environmental data affecting the work of the disassembly robot; acquiring attribute data of the disassembly robot; constructing an influence discrimination model based on the attribute data and the environmental data, the influence discrimination model outputting a discrimination result, the discrimination result being used to show factors affecting the work of the disassembly robot; constructing an instruction regulation model based on the discrimination result, the instruction regulation model outputting a dynamic adjustment instruction, the dynamic adjustment instruction acting on the work plan; constructing a monitoring feedback model, the monitoring feedback model being used to track the work state of the disassembly robot in real time, and generating adjustment data based on the comparison between the real-time work state of the disassembly robot and a preset target, the adjustment data acting on the instruction regulation model and affecting the dynamic adjustment instruction; the influence discrimination model comprises: respectively acquiring feature data of the attribute data and the environmental data, selecting feature data matched with the attribute data and the environmental data and constructing a mapping relationship, and integrating the attribute data and the environmental data into a plurality of feature-based fusion data sets based on the mapping relationship; acquiring features in the work plan of the disassembly robot, and giving high weights to the fusion data matched with the features in the work plan of the disassembly robot; presetting a discrimination threshold, the discrimination threshold being set based on the historical work plan of the disassembly robot and safety specifications, comparing the discrimination threshold and the data values in the high-weight fusion data, and outputting a discrimination result; the comparison method of the discrimination threshold and the data values in the high-weight fusion data set comprises: searching for data values of matching features in the discrimination threshold based on the features in the high-weight fusion data set; if the data values of the matching features are searched, the data value comparison is directly performed; if the data values of the matching features are not searched, the data values of similar features are searched for interpolation calculation, and the approximate values are acquired before the comparison is performed; when the data values in the high-weight fusion data set exceed the discrimination threshold, the discrimination result is that the work of the disassembly robot is at risk; when the data values in the high-weight fusion data set are lower than or equal to the discrimination threshold, the discrimination result is that the work of the disassembly robot is safe. the discrimination method of the similar feature data values comprises: respectively acquiring features of the feature data in the attribute data and the environmental data, and constructing a superordinate feature induction table, the superordinate feature induction table integrating and inducing the acquired features based on physics, environment, and time to form different levels of feature classification.

2. The method of claim 1, wherein the method further comprises: determining a location of the robot relative to the busbar; and determining a location of the busbar relative to the robot. the feature data in the attribute data comprises a length of a mechanical arm of the disassembly robot, a rotation range of the mechanical arm, a type of a tool assembled by the disassembly robot, and a specification of a disassembly tool assembled by the disassembly robot; the feature data of the environmental data comprises a temperature, a humidity, a wind power, an illumination intensity, an electromagnetic interference degree of a work area, and a material and a specification of the pipe mother drainage line.

3. The method of claim 2, wherein the method further comprises: determining a position of the robot relative to the busbar; and determining a position of the busbar relative to the robot. the construction method of the mapping relationship comprises: The characteristic tree diagram is constructed by taking attribute data and characteristic data in environment data as branch nodes, and each branch node is provided with a plurality of sub-nodes, and the number of the sub-nodes represents a specific characteristic data value; Real-time operation data of the disassembly robot is acquired, and based on the real-time operation data, characteristic data values matching the attribute data and the environment data are searched in the characteristic tree diagram, the sub-nodes corresponding to the matched characteristic data values are connected, a mapping path is formed, and the mapping path is the mapping relationship.

4. The method for robotic operation of dismantling and assembling a high-voltage substation busbar as described in claim 1, characterized in that: The parameter data acquisition method comprises: A plurality of parameter acquisition nodes are constructed, the parameter acquisition nodes are based on laser radar technology, the plurality of parameter acquisition nodes are uniformly distributed around the pipe drain wire, a data encryption and cleaning method is set, redundant data is cleaned in real time during data collection, and cleaned data is output, a data shielding enhancement model is constructed, a plurality of sets of cleaned data in adjacent time periods are acquired in real time, and any one of an intermediate value or an average value is extracted to obtain parameter data, a middle storage node is created, parameter data is acquired and backed up, a feedback rule between the middle storage node and the data shielding enhancement model is constructed, the feedback rule is provided with a critical threshold, the critical threshold is dynamically set based on historical parameter data as a standard, and when the parameter data exceeds the critical threshold, the feedback rule triggers data replacement of the middle storage node, and the latest acquired parameter data is replaced.

5. A robotic operation method for dismantling and assembling a high-voltage substation busbar as described in claim 4, characterized in that: The parameter data comprises a length, a diameter, a bending degree, a position coordinate, a distance from the pipe drain wire to surrounding objects, and a current voltage of the pipe drain wire. The positioning model comprises: A three-dimensional space modeling is constructed, the three-dimensional space modeling takes the pipe drain wire as a reference to establish a three-dimensional coordinate system, converts the parameter data of the pipe drain wire into coordinate points and vectors in the three-dimensional space, constructs a three-dimensional space model of the pipe drain wire based on the coordinate points and vectors, and in addition, the three-dimensional space modeling takes the pipe drain wire as the center to simulate a space environment around the pipe drain wire, divides the simulated space into different work areas based on the position coordinate and the distance from the pipe drain wire to surrounding objects in the parameter data, and each work area corresponds to a work position and a work path that the disassembly robot can perform. The work position and the work path of the disassembly robot in each work area are acquired and combined with the length of the pipe drain wire, the work time of the disassembly robot in each work area is calculated, the work time, the work position, and the work path are summarized, and a work plan of the disassembly robot is integrated and formed.

6. A robotic operation method for dismantling and assembling a high-voltage substation busbar as described in claim 4, characterized in that: The detection model comprises: An environment perception system is constructed, the environment perception system is constructed based on cross combination of a plurality of different sensing units, is used for real-time collection of work area environment data, acquires parameter data characteristics, labels the sensing units based on the characteristic types, constructs a fusion algorithm, gives a high weight to the labeled sensing units, gives an equal weight to the remaining sensing units in different sensing units, fuses data of the labeled sensing units and data of the remaining sensing units, and obtains environment data. The operation formula of the fusion algorithm is: where dm represents the environmental data value collected by the sensing unit m marked as high weight, a represents the weight coefficient given to the high weight sensing unit m, represents the sum of the environmental data values collected by all other sensing units, i.e., n−1, except for the high weight sensing unit m, is a normalization factor for equally distributing the sum of data of other sensing units to each sensing unit.

7. A robotic operation method for dismantling and assembling a high-voltage substation busbar as described in claim 1, characterized in that: The instruction regulation model comprises: The discrimination result features include a work risk level, a work safety area, and a work priority order, an instruction generation system is constructed, the instruction generation system is based on the discrimination result features, combines a work plan and a work capacity of the disassembly and assembly robot, and dynamically generates a regulation and control instruction based on a self-adaptive regulation and control strategy of machine learning; The work plan includes work area selection, work tool selection, work path planning, and work time allocation.

8. The method of claim 1, wherein the method further comprises: determining a location of the robot relative to the high voltage substation; and determining a location of the robot relative to the high voltage substation pipe busbar. The monitoring feedback model includes: Real-time running data of the disassembly and assembly robot is acquired, and features in the real-time running data are extracted, a running state database is constructed based on the features in the running data, and the running state database is used to update the running data of the disassembly and assembly robot in real time; The running state database and a time axis are associated, running data parameters in multiple groups of adjacent time periods are intercepted, a change amplitude of the running data parameters is analyzed, and a running state of the disassembly and assembly robot in a future time period is predicted based on the change amplitude, a running state parameter value of the future time period is set as a preset target; The preset target and the real-time running data are compared, and adjustment data is generated based on a deviation.

Images