Self-correction method of 600MW unit air cooling energy-saving and efficiency-improving robot

Abstract

The invention discloses a self-correction method of a 600MW unit air-cooling energy-saving and efficiency-improving robot, and relates to the field of thermal power generation, comprising the following steps: the 600MW unit air-cooling energy-saving and efficiency-improving robot initializes a laser radar, an inertial measurement unit, an encoder, a high-pressure water pump and an electrically operated valve and calibrates a sensor; constructing a three-dimensional virtual track model by adopting a finite element analysis method by integrating laser radar scanning data, inertial measurement unit attitude data and material mechanical property parameters; predicting a future track deformation trend and generating a deformation compensation instruction by combining historical deformation data and material fatigue characteristics based on the track digital twinning model; and receiving a deformation compensation instruction, dynamically adjusting parameters of a proportional-integral-differential controller based on a reinforcement learning algorithm, and optimizing a control strategy of a driving motor. The cleaning effect is evaluated through multi-sensor fusion, fin damage is detected through the acoustic emission technology, and finally intelligent monitoring and one-button operation management of a centralized control center on the state of the robot are achieved.

Description

Technical Field

[0001] This invention relates to the field of thermal power generation, and in particular to a self-calibration method for an air-cooled energy-saving and efficiency-enhancing robot for a 600MW unit. Background Technology

[0002] In the field of thermal power generation, the air-cooling system of a 600MW unit is a key piece of equipment for energy conservation and consumption reduction. Its heat dissipation efficiency directly affects the unit's back pressure and economy. The accumulation of dirt on the surface of the air-cooling island tube bundle will lead to a decrease in heat exchange efficiency. In the traditional method of using a track-type cleaning robot for regular flushing and maintenance, the system collects track geometric parameters in real time and uses a PID control algorithm to correct the robot's trajectory online. The system adopts multi-sensor fusion technology, uses a laser ranging unit to measure changes in track spacing in real time, and combines encoder feedback to achieve synchronous control of the drive motor. This effectively solves the problem of track parallelism deviation caused by thermal deformation. A temperature compensation algorithm is introduced to correct the measured values ​​according to changes in ambient temperature, which significantly improves the measurement accuracy under conditions of large day-night temperature differences.

[0003] However, existing technologies still have room for improvement in predictive compensation and adaptive control of track deformation. Because air-cooled island tracks are exposed to the outdoor environment for extended periods, they are affected by a combination of factors, including wind load, temperature cycling, and material creep, resulting in significant time-varying and nonlinear characteristics in their deformation process. While existing fixed-parameter PID control methods can achieve real-time deviation correction, they struggle to address the control challenges posed by sudden deformation and gradual deformation accumulation. Especially under conditions of sudden wind speed changes or rapid temperature fluctuations, fixed control parameters cannot simultaneously meet the requirements of response speed and stability, potentially leading to overshoot or oscillations during the cleaning process. Summary of the Invention

[0004] In view of the aforementioned existing problems, the present invention is proposed.

[0005] Therefore, this invention provides a self-calibration method for a 600MW unit air-cooled energy-saving and efficiency-enhancing robot, which solves the overshoot or oscillation problem caused by the difficulty of adapting to complex operating conditions due to fixed control parameters in the prior art.

[0006] To solve the above-mentioned technical problems, the present invention provides the following technical solution: This invention provides a self-calibration method for a 600MW unit air-cooled energy-saving and efficiency-enhancing robot, which includes initializing and calibrating the LiDAR, inertial measurement unit, encoder, high-pressure water pump and electric valve of the 600MW unit air-cooled energy-saving and efficiency-enhancing robot, and constructing a three-dimensional virtual track model by integrating LiDAR scanning data, inertial measurement unit attitude data and material mechanical property parameters using the finite element analysis method. Based on the digital twin model of the track, combined with historical deformation data and material fatigue characteristics, the future deformation trend of the track is predicted and deformation compensation commands are generated. Upon receiving deformation compensation commands, the parameters of the proportional-integral-derivative controller are dynamically adjusted based on reinforcement learning algorithms to optimize the control strategy of the drive motor. Based on the topology diagram of the air-cooled island tube bundle layout and combined with the three-dimensional coordinate map of the track generated by LiDAR scanning, the cleaning coverage path is optimized and the travel path is set through the A* algorithm. The 600MW unit air-cooled energy-saving and efficiency-enhancing robot follows its path, with lidar and inertial measurement unit continuously monitoring track deformation and dynamically adjusting the output of the proportional-integral-derivative controller to control the magnetorheological damper to suppress vibration. The cleaning effect is evaluated by multi-sensor fusion using infrared thermal imagers, high-definition cameras, and ultrasonic thickness gauges, generating a comprehensive cleaning effect score. The control center stores the cleaning data and updates the track digital twin model. Based on acoustic emission sensor data, the risk of fin damage is analyzed, and areas requiring maintenance are marked. The central control center monitors the status of the 600MW unit's air-cooled energy-saving and efficiency-enhancing robot in real time through a human-machine interface, and performs self-calibration and cleaning operations.

[0007] As a preferred embodiment of the self-calibration method for the 600MW unit air-cooled energy-saving and efficiency-enhancing robot described in this invention, the 600MW unit air-cooled energy-saving and efficiency-enhancing robot initializes and calibrates sensors such as lidar, inertial measurement unit, encoder, high-pressure water pump, and electric valve, including the following steps. The control center sends an initialization command, which sequentially performs zero-position calibration of the reflective target of the lidar, six-axis static calibration of the inertial measurement unit, speed feedback verification of the encoder, no-load pressure test of the high-pressure water pump, and opening and closing response test of the electric valve. All calibration data is uploaded to the control center in real time via industrial Ethernet. In case of abnormality, an alarm is triggered and the startup process is paused.

[0008] As a preferred embodiment of the self-calibration method for the 600MW unit air-cooled energy-saving and efficiency-enhancing robot described in this invention, the method involves: constructing a three-dimensional virtual track model using finite element analysis by integrating lidar scanning data, inertial measurement unit attitude data, and material mechanical property parameters, including the following steps. By integrating the orbit point cloud data scanned by lidar, the orbit inclination data recorded by the inertial measurement unit, and the elastic modulus and fatigue limit parameters of the orbit material, a three-dimensional virtual orbit model is established using finite element analysis software, which includes mesh generation, boundary condition setting, and load application.

[0009] As a preferred embodiment of the self-calibration method for the 600MW unit air-cooled energy-saving and efficiency-enhancing robot described in this invention, the method includes: based on a digital twin model of the track, combined with historical deformation data and material fatigue characteristics, predicting future track deformation trends, and generating deformation compensation commands, comprising the following steps. Using deformation records under temperature and wind loads from the historical deformation database, the longitudinal bending and lateral offset of each track segment are calculated using a time series prediction algorithm. Deformation compensation commands with deformation position, direction, and compensation amount are generated and sent to the motion controller.

[0010] As a preferred embodiment of the self-calibration method for the 600MW unit air-cooled energy-saving and efficiency-enhancing robot described in this invention, the method includes the following steps: receiving deformation compensation commands, dynamically adjusting the parameters of the proportional-integral-derivative controller based on a reinforcement learning algorithm, and optimizing the control strategy of the drive motor. The deviation between the actual track deformation and the predicted value serves as the state input of the deep reinforcement learning algorithm, with the motor torque adjustment as the action output. The proportional-integral-derivative controller's proportional coefficient, integral coefficient, and derivative coefficient are optimized online through a reward function that balances energy consumption and trajectory tracking accuracy, and then transmitted to the drive motor in real time via the CAN bus.

[0011] As a preferred embodiment of the self-calibration method for the 600MW unit air-cooled energy-saving and efficiency-enhancing robot described in this invention, the method includes the following steps: based on the topology diagram of the air-cooled island tube bundle layout, combined with the three-dimensional coordinate map of the track generated by lidar scanning, the cleaning and coverage path is optimized and the travel path is set using the A* algorithm. A topology network is established based on the CAD drawings of the 560 tube bundles in the air-cooled island. A three-dimensional coordinate map generated by the lidar is overlaid, and the A* algorithm is used to calculate the global optimal path to avoid high-risk deformation areas. The path coordinate sequence is then sent to the steering mechanism.

[0012] As a preferred embodiment of the self-calibration method for the 600MW unit air-cooled energy-saving efficiency-enhancing robot described in this invention, the 600MW unit air-cooled energy-saving efficiency-enhancing robot follows a travel path, with lidar and inertial measurement unit continuously monitoring track deformation, dynamically adjusting the output of the proportional-integral-derivative controller, and controlling the magnetorheological damper to suppress vibration. This includes the following steps: The track deformation is scanned at a frequency of 100Hz by lidar, the vibration spectrum is monitored in real time by inertial measurement unit, the motor torque distribution is dynamically adjusted by motion controller, and the magnetorheological damper is controlled by current signal to suppress high-frequency vibration.

[0013] As a preferred embodiment of the self-calibration method for the 600MW unit air-cooled energy-saving and efficiency-enhancing robot described in this invention, the method includes the following steps: evaluating the cleaning effect through multi-sensor fusion using an infrared thermal imager, a high-definition camera, and an ultrasonic thickness gauge to generate a comprehensive cleaning effect score. An infrared thermal imager scans the temperature distribution of the fins, a high-definition camera identifies areas of residual dirt using a convolutional neural network, an ultrasonic thickness gauge detects the thickness of dirt at key points, and the entire process is combined to calculate a comprehensive score for the cleaning effect.

[0014] As a preferred embodiment of the self-calibration method for the 600MW unit air-cooled energy-saving and efficiency-enhancing robot described in this invention, the method includes the following steps: the control center stores cleaning data and updates the track digital twin model; analyzes fin damage risk based on acoustic emission sensor data; and marks areas requiring maintenance. The operational data is encrypted and stored in a blockchain database. Microcrack signals are identified based on the spectral characteristics of the 20-50kHz frequency band of the acoustic emission sensor. Damage risk areas are marked in the digital twin model and maintenance work orders are generated.

[0015] As a preferred embodiment of the self-calibration method for the 600MW unit air-cooled energy-saving and efficiency-enhancing robot described in this invention, the central control center monitors the status of the 600MW unit air-cooled energy-saving and efficiency-enhancing robot in real time through a human-machine interface. The self-calibration and cleaning operations include the following steps. The human-machine interface displays real-time deformation heat maps, cleaning scores, and equipment status, supports one-click start of the entire process, pushes abnormal alarm information to mobile terminals in real time, and enables closed-loop processing of maintenance work orders.

[0016] The beneficial effects of this invention are as follows: a three-dimensional virtual track model is constructed by fusing data from multiple sensors such as lidar and inertial measurement units, and predictive compensation and adaptive control of track deformation are achieved based on digital twin technology and reinforcement learning algorithms. This method first performs precise calibration of various sensors, then integrates laser scanning data, attitude information and material parameters to establish a finite element model, predicts track deformation trends and generates compensation commands through time series algorithms, dynamically optimizes PID control parameters using reinforcement learning, plans the optimal cleaning path using the A* algorithm, monitors and suppresses track vibration in real time, evaluates the cleaning effect through multi-sensor fusion, and detects fin damage using acoustic emission technology. Finally, it enables intelligent monitoring and one-click operation management of the robot status by the central control center. Attached Figure Description

[0017] To more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings used in the following description of the embodiments will be briefly introduced. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0018] Figure 1 A flowchart of the self-calibration method for an air-cooled energy-saving and efficiency-enhancing robot for a 600MW unit.

[0019] Figure 2 This is a flowchart of the sensor calibration process.

[0020] Figure 3 This is a flowchart for adaptive control and path planning. Detailed Implementation

[0021] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, the specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings.

[0022] Many specific details are set forth in the following description in order to provide a full understanding of the invention. However, the invention may also be practiced in other ways different from those described herein, and those skilled in the art can make similar extensions without departing from the spirit of the invention. Therefore, the invention is not limited to the specific embodiments disclosed below.

[0023] Secondly, the term "one embodiment" or "embodiment" as used herein refers to a specific feature, structure, or characteristic that may be included in at least one implementation of the present invention. The phrase "in one embodiment" appearing in different places in this specification does not necessarily refer to the same embodiment, nor is it a single or selective embodiment that is mutually exclusive with other embodiments.

[0024] Reference Figures 1-3 As one embodiment of the present invention, this embodiment provides a self-calibration method for a 600MW unit air-cooled energy-saving efficiency-enhancing robot, comprising the following steps: The S1, 600MW unit air-cooled energy-saving and efficiency-enhancing robot initializes and calibrates the lidar, inertial measurement unit, encoder, high-pressure water pump and electric valve.

[0025] S1.1 The control center sends an initialization command to perform zero-position calibration of the reflective target of the lidar, six-axis static calibration of the inertial measurement unit, speed feedback verification of the encoder, no-load pressure test of the high-pressure water pump, and opening and closing response test of the electric valve. All calibration data are uploaded to the control center in real time via industrial Ethernet. If an abnormality occurs, an alarm is triggered and the startup process is paused.

[0026] Furthermore, before the robot system starts, the control center sends initialization commands to all key sensing and execution units. First, a high-precision reflective target is used to perform zero-point calibration of the lidar to ensure the accuracy of the ranging reference. Then, a six-axis static calibration is performed on the inertial measurement unit on a horizontal reference platform to eliminate zero-point drift errors in attitude measurement. Next, the encoder is dynamically verified by comparing the actual rotational speed with the feedback signal to ensure the accuracy of motor control. At the same time, the no-load pressure characteristics of the high-pressure water pump are tested to verify its pressure output linearity, and the electric valve is tested for full-stroke opening and closing response to confirm its operational reliability. All data generated during the calibration process is transmitted to the control center in real time via industrial Ethernet. When any abnormal value exceeding the technical specifications is detected in any link, the system immediately triggers multi-level audible and visual alarms and suspends the startup process. S2. By integrating lidar scanning data, inertial measurement unit attitude data, and material mechanical property parameters, a three-dimensional virtual track model is constructed using the finite element analysis method.

[0027] S2.1 By integrating the orbit point cloud data scanned by lidar, the orbit inclination data recorded by the inertial measurement unit, and the elastic modulus and fatigue limit parameters of the orbit material, a three-dimensional virtual orbit model is established using finite element analysis software, which includes mesh generation, boundary condition setting, and load application.

[0028] Furthermore, in constructing the 3D virtual track model, the system first uses a high-precision lidar to perform a comprehensive scan of the air-cooled island track, acquiring dense point cloud spatial coordinate data. Simultaneously, it utilizes an inertial measurement unit to collect real-time dynamic tilt angle changes at various points on the track. Combining the elastic deformation characteristics and long-term fatigue performance parameters of the track steel, professional finite element analysis software is used for modeling. The modeling process begins with fine-grained mesh generation of the track structure to ensure mesh density in key areas. Then, reasonable boundary constraints are set based on the actual support conditions of the track, and loads are applied considering multiple physical field factors such as wind pressure loads, temperature stress, and mechanical vibration. Finally, a 3D virtual model that accurately reflects the true state of the track is generated. This model, by integrating multi-source measured data and material properties, achieves a precise digital representation of the mechanical behavior of the track structure.

[0029] S3. Based on the digital twin model of the track, combined with historical deformation data and material fatigue characteristics, predict the future deformation trend of the track and generate deformation compensation commands.

[0030] S3.1 Using the deformation records under temperature load and wind load in the historical deformation database, the longitudinal bending and lateral offset of each track segment are calculated by the time series prediction algorithm, and deformation compensation commands with deformation position, direction and compensation amount are generated and sent to the motion controller.

[0031] The expression for longitudinal bending is, ; in, For the shift operator, The first-order moving average coefficient, for, The change in temperature For constant terms, The wind load influence coefficient is... This is the temperature influence coefficient. For wind speed, These are the first-order autoregressive coefficients. It is the difference order.

[0032] The expression for the lateral offset is: ; in, Seasonal coefficient, It is a second-order MA term. This is the random error term.

[0033] S4. Receive deformation compensation instructions and dynamically adjust the parameters of the proportional-integral-derivative controller based on reinforcement learning algorithm to optimize the control strategy of the drive motor.

[0034] S4.1 The deviation between the actual deformation of the track and the predicted value is used as the state input of the deep reinforcement learning algorithm, and the motor torque adjustment is used as the action output. The proportional coefficient, integral coefficient and derivative coefficient of the proportional-integral-derivative controller are optimized online through a reward function that takes into account both energy consumption and trajectory tracking accuracy, and then sent to the drive motor in real time via the CAN bus.

[0035] Furthermore, during the adaptive control process, the system collects dynamic deviation data between the actual track deformation and the predicted values ​​from the digital twin model in real time, using this data as the environmental state input for the deep reinforcement learning algorithm. The control algorithm evaluates the current operating conditions through a deep neural network and outputs the optimal motor torque adjustment command. This adjustment comprehensively considers instantaneous response requirements and long-term operational stability, employing a multi-objective optimization reward mechanism. While ensuring trajectory tracking accuracy, it intelligently balances factors such as energy efficiency and mechanical wear, autonomously adjusting the parameter combinations of the proportional, integral, and derivative control dimensions. The optimized control parameters are transmitted in real time to each drive motor execution unit via a high-speed CAN bus communication protocol, achieving millisecond-level dynamic response. This AI-based adaptive control strategy effectively solves the adjustment lag problem of traditional fixed-parameter PID control under complex operating conditions, enabling the robot to intelligently cope with challenges such as sudden track deformation and gradual deformation accumulation.

[0036] S5. Based on the topology diagram of the air-cooled island tube bundle layout, combined with the three-dimensional coordinate map of the track generated by lidar scanning, the A* algorithm is used to optimize the cleaning coverage path and set the travel path.

[0037] S5.1. Establish a topology network based on the CAD drawings of the 560 tube bundles of the air-cooled island, overlay a three-dimensional coordinate map generated by the lidar, use the A* algorithm to calculate the global optimal path, avoid high-risk deformation areas, and send the path coordinate sequence to the steering mechanism.

[0038] Specifically, the expression is, ; in, For nodes The overall cost estimate, From the starting point to the node The actual cost, For the node Heuristic cost estimation to the target point.

[0039] Furthermore, in the path planning phase, the system first constructs a topological network based on the engineering design drawings of the 560 tube bundles in the air-cooled island, accurately reflecting the spatial connections between the tube bundles. Simultaneously, it integrates a real-time 3D point cloud map generated by high-precision LiDAR scanning to establish an environmental model that includes the track deformation state. Through an improved A* search algorithm, the system intelligently calculates the globally optimal path within the topological network. This algorithm comprehensively considers factors such as path length, deformation risk level, and mechanical steering constraints, automatically avoiding high-risk areas where deformation exceeds safety thresholds. The final optimized path coordinate sequence is transmitted in real-time to the steering control system via an industrial communication protocol, guiding the cleaning robot to precisely operate along the predetermined trajectory. This path planning method, which integrates prior design and real-time perception, ensures comprehensive coverage of the cleaning operation while effectively avoiding operational risks caused by track deformation.

[0040] The S6 600MW unit air-cooled energy-saving and efficiency-enhancing robot follows its path, with lidar and inertial measurement unit continuously monitoring track deformation and dynamically adjusting the output of the proportional-integral-derivative controller to control the magnetorheological damper to suppress vibration.

[0041] S6.1 The track deformation is scanned at a frequency of 100Hz by lidar, the vibration spectrum is monitored in real time by inertial measurement unit, the motion controller dynamically adjusts the motor torque distribution, and the magnetorheological damper is controlled by current signal to suppress high-frequency vibration.

[0042] Furthermore, during robot operation, a high-precision lidar continuously monitors the track's geometric deformation at a scanning frequency of 100 times per second, capturing microscopic displacement changes on the track surface in real time. Simultaneously, an inertial measurement unit samples the robot's multidimensional vibration spectrum at millisecond levels, accurately identifying the mechanical vibration characteristics of different frequency bands. Based on this real-time sensor data, the motion controller dynamically allocates the output torque of each drive motor through intelligent algorithms, while simultaneously sending precisely modulated current control signals to the magnetorheological damper, enabling its damping characteristics to adaptively suppress high-frequency mechanical vibrations. This multi-sensor collaborative closed-loop control system ensures the robot's operational stability under complex track conditions while effectively avoiding positioning errors and mechanical wear caused by vibration, significantly improving the accuracy of cleaning operations and extending equipment lifespan.

[0043] S7. The cleaning effect is evaluated by multi-sensor fusion using an infrared thermal imager, a high-definition camera, and an ultrasonic thickness gauge, and a comprehensive cleaning effect score is generated.

[0044] S7.1: An infrared thermal imager scans the temperature distribution of the fins, a high-definition camera identifies areas of residual dirt through a convolutional neural network, an ultrasonic thickness gauge detects the thickness of dirt at key points, and the system calculates a comprehensive score for the cleaning effect.

[0045] The comprehensive scoring expression for cleaning effectiveness is as follows: ; in, To determine the overall cleaning effect, To account for the temperature difference before and after cleaning, As a temperature reference value, This is the back pressure reference value. This represents the decrease in back pressure. For the area to be cleaned, This represents the total area of ​​the tube bundle.

[0046] Furthermore, during the cleaning effect evaluation phase, the system uses an infrared thermal imager to scan the entire temperature field of the air-cooled island fins, accurately capturing the temperature distribution characteristics of the heat exchange tube bundle. Simultaneously, a high-definition camera acquires images of the fin surface, and a deep learning convolutional neural network intelligently identifies areas of residual dirt and calculates the contamination coverage. At the same time, an ultrasonic thickness gauge precisely measures the thickness of dirt deposits in key areas. Finally, through a multi-source data fusion algorithm, an objective and quantitative comprehensive evaluation result of the cleaning effect is generated by integrating multiple dimensions such as temperature uniformity, visual cleanliness, and dirt thickness. This multi-modal sensing collaborative evaluation mechanism overcomes the limitations of single detection methods and achieves a comprehensive and scientific evaluation of cleaning quality, providing accurate data support for subsequent maintenance decisions.

[0047] S8. The control center stores cleaning data and updates the track digital twin model. Based on acoustic emission sensor data, it analyzes the risk of fin damage and marks areas requiring maintenance.

[0048] S8.1. Encrypt and store the operation data in the blockchain database, identify microcrack signals based on the spectral characteristics of the 20-50kHz frequency band of the acoustic emission sensor, mark the damage risk area in the digital twin model and generate maintenance work orders.

[0049] Furthermore, during the data management and equipment maintenance phase, the system processes all operational data using encryption algorithms and stores it in a distributed database based on blockchain technology, ensuring the immutability and traceability of data records. Simultaneously, it continuously monitors acoustic signals in the 20-50 kHz frequency band using acoustic emission sensors and employs pattern recognition algorithms to accurately extract the microcrack characteristic spectrum of the finned tube bundle. When a damage characteristic signal is detected, the system automatically marks the risk level at the corresponding location in the digital twin model and generates an intelligent work order containing specific coordinates, risk type, and maintenance priority, which is then pushed to the equipment management system. This technical solution, integrating blockchain data preservation and acoustic fault diagnosis, not only achieves full lifecycle management of equipment status but also provides a scientific basis for predictive maintenance, effectively avoiding heat exchange efficiency degradation and safety accidents caused by fin damage.

[0050] S9, the central control center monitors the status of the 600MW unit's air-cooled energy-saving and efficiency-enhancing robot in real time through the human-machine interface, and performs self-calibration and cleaning operations.

[0051] S9.1 Displays real-time deformation heat map, cleaning score and equipment status on the human-machine interface, supports one-click start of the entire process operation, pushes abnormal alarm information to mobile terminal in real time, and handles maintenance work orders in a closed loop.

[0052] Furthermore, in the intelligent monitoring phase of the central control center, the human-machine interface displays real-time key information such as track deformation heat maps, multi-dimensional cleaning effect scoring curves, and equipment operating status parameters. Operators can trigger fully automated cleaning operations with a single click. The system implements multi-level early warning management for abnormal operational events. Alarm information is instantly pushed to the mobile terminals of relevant personnel via a secure communication protocol. Simultaneously, it automatically generates intelligent maintenance work orders containing fault location, risk level, and handling solutions, and tracks and records the entire closed-loop management process from work order issuance to acceptance confirmation. This control system, integrating visual monitoring, intelligent early warning, and process-oriented management, significantly reduces the intensity of manual operations and achieves refined and standardized equipment maintenance, significantly improving the operational efficiency and reliability of the air-cooled system.

[0053] This embodiment also provides a computer device applicable to the self-calibration method of the air-cooled energy-saving and efficiency-enhancing robot for a 600MW unit, comprising: a memory and a processor; the memory is used to store computer-executable instructions, and the processor is used to execute the computer-executable instructions to implement the self-calibration method of the air-cooled energy-saving and efficiency-enhancing robot for a 600MW unit as proposed in the above embodiment.

[0054] The computer device can be a terminal, comprising a processor, memory, communication interface, display screen, and input devices connected via a system bus. The processor provides computing and control capabilities. The memory includes non-volatile storage media and internal memory. The non-volatile storage media stores the operating system and computer programs. The internal memory provides an environment for the operation of the operating system and computer programs stored in the non-volatile storage media. The communication interface is used for wired or wireless communication with external terminals; wireless communication can be achieved through Wi-Fi, carrier networks, NFC (Near Field Communication), or other technologies. The display screen can be an LCD screen or an e-ink screen. The input devices can be a touch layer covering the display screen, buttons, a trackball, or a touchpad on the computer device's casing, or an external keyboard, touchpad, or mouse.

[0055] This embodiment also provides a storage medium storing a computer program. When executed by a processor, the program implements the self-calibration method for implementing the 600MW unit air-cooled energy-saving efficiency-enhancing robot as proposed in the above embodiment. The storage medium can be implemented by 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 Red-Only Memory (PROM), Read-Only Memory (ROM), magnetic storage, flash memory, magnetic disk, or optical disk.

[0056] In summary, this invention constructs a three-dimensional virtual track model through the fusion of data from multiple sensors, including lidar and inertial measurement units. Based on digital twin technology and reinforcement learning algorithms, it achieves predictive compensation and adaptive control of track deformation. The method first performs precise calibration of various sensors, then integrates laser scanning data, attitude information, and material parameters to establish a finite element model. It predicts the track deformation trend and generates compensation commands through time series algorithms, dynamically optimizes PID control parameters using reinforcement learning, plans the optimal cleaning path using the A* algorithm, monitors and suppresses track vibration in real time, evaluates the cleaning effect through multi-sensor fusion, and uses acoustic emission technology to detect fin damage. Finally, it enables intelligent monitoring and one-click operation management of the robot's status by the central control center.

[0057] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention, and all such modifications or substitutions should be covered within the scope of the claims of the present invention.

Claims

1. A self-calibration method for a 600MW unit air-cooled energy-saving and efficiency-enhancing robot, characterized in that: Comprising, 600MW unit air cooling energy saving and efficiency increasing robot initializes and calibrates laser radar, inertial measurement unit, encoder, high-pressure water pump and electric valve, and constructs three-dimensional virtual track model by integrating laser radar scanning data, inertial measurement unit attitude data and material mechanics characteristic parameters through finite element analysis method; Based on the track digital twin model, combined with historical deformation data and material fatigue characteristics, the future track deformation trend is predicted, and the deformation compensation instruction is generated; Receive the deformation compensation instruction, dynamically adjust the parameters of the proportional-integral-derivative controller based on the reinforcement learning algorithm, optimize the control strategy of the drive motor, and set the marching path through the A* algorithm based on the topological structure diagram of the air cooling island tube bundle layout and the track three-dimensional coordinate map generated by laser radar scanning. 600MW unit air cooling energy saving and efficiency increasing robot marches according to the marching path, laser radar and inertial measurement unit continuously monitor the track deformation, dynamically adjust the proportional-integral-derivative controller output, control the magneto-rheological damper to suppress vibration; Through multi-sensor fusion of infrared thermal imager, high-definition camera and ultrasonic thickness gauge, the cleaning effect evaluation is carried out, the cleaning effect comprehensive score is generated, the cleaning data is stored in the control center and the track digital twin model is updated, the fin damage risk is analyzed based on acoustic emission sensor data, and the maintenance area is marked. The centralized control center monitors the state of 600MW unit air cooling energy saving and efficiency increasing robot in real time through human-machine interface, self-corrects and cleans the work.

2. The self-correction method of the air cooling energy-saving and efficiency-improving robot for a 600 MW unit as claimed in claim 1, characterized in that: 600MW unit air cooling energy saving and efficiency increasing robot initializes and calibrates laser radar, inertial measurement unit, encoder, high-pressure water pump and electric valve, including the following steps, The control center sends initialization instructions, and in turn calibrates the laser radar to the reflection target zero position, calibrates the inertial measurement unit to six-axis static calibration, calibrates the encoder to speed feedback verification, calibrates the high-pressure water pump to no-load pressure test, and calibrates the electric valve to open and close response test. All calibration data is uploaded to the control center in real time through industrial Ethernet, and the alarm is triggered and the start process is suspended in case of abnormality.

3. The self-correction method of the air cooling energy-saving and efficiency-improving robot for a 600 MW unit as claimed in claim 2, characterized in that: Through the integration of laser radar scanning data, inertial measurement unit attitude data and material mechanics characteristic parameters, a three-dimensional virtual track model is constructed by using finite element analysis method, including the following steps, Through the integration of laser radar scanning track point cloud data, inertial measurement unit recorded track inclination data and track material elastic modulus and fatigue limit parameters, a three-dimensional virtual track model is established by finite element analysis software through grid division, boundary condition setting and load application.

4. The self-correction method of the air cooling energy-saving and efficiency-improving robot for a 600 MW unit as claimed in claim 3, characterized in that: Based on the track digital twin model, combined with historical deformation data and material fatigue characteristics, the future track deformation trend is predicted, and the deformation compensation instruction is generated, including the following steps, Using the deformation records under the action of temperature load and wind load in the historical deformation database, the longitudinal bending and transverse offset of each section of the track are calculated by time series prediction algorithm, the deformation compensation instruction of deformation position, direction and compensation amount is generated and sent to the motion controller.

5. The self-correction method of the air cooling energy-saving and efficiency-improving robot for a 600 MW unit as claimed in claim 4, characterized in that: Receive the deformation compensation instruction, dynamically adjust the parameters of the proportional-integral-derivative controller based on the reinforcement learning algorithm, optimize the control strategy of the drive motor, including the following steps, The deviation of the actual deformation amount of the track from the predicted value is taken as the state input of the deep reinforcement learning algorithm, and the motor torque adjustment amount is taken as the action output. The proportional coefficient, integral coefficient and differential coefficient of the proportional-integral-differential controller are optimized online through a reward function considering the energy consumption and trajectory tracking accuracy, and are transmitted to the drive motor in real time through the CAN bus.

6. The self-correction method of the 600 MW unit air cooling energy-saving and efficiency-improving robot according to claim 5, characterized in that: Based on the topology diagram of the air-cooled island tube bundle layout, combined with the track three-dimensional coordinate map generated by laser radar scanning, the A* algorithm is used to optimize the cleaning coverage path to set the travel path, including the following steps, According to the layout CAD drawing of the air-cooled island 560 tube bundle, the topology network is established, the three-dimensional coordinate map generated by the laser radar is superimposed, the A* algorithm is used to calculate the global optimal path, and the path coordinate sequence is transmitted to the steering mechanism.

7. The self-correction method of the air cooling energy-saving and efficiency-improving robot for a 600 MW unit as claimed in claim 6, characterized in that: The 600MW unit air-cooled energy-saving robot follows the travel path, the laser radar and the inertial measurement unit continuously monitor the track deformation, dynamically adjusts the proportional-integral-differential controller output, controls the magneto-rheological damper to suppress vibration, including the following steps, The laser radar scans the track deformation at a frequency of 100Hz, the inertial measurement unit monitors the vibration frequency spectrum in real time, the motion controller dynamically adjusts the motor torque distribution, and the magneto-rheological damper is controlled through the current signal to suppress high-frequency vibration.

8. The self-correction method of the air cooling energy-saving and efficiency-improving robot for a 600 MW unit as claimed in claim 7, characterized in that: The cleaning effect evaluation is carried out through multi-sensor fusion of infrared thermal imager, high-definition camera and ultrasonic thickness gauge, and the cleaning effect comprehensive score is generated, including the following steps, The infrared thermal imager scans the fin temperature distribution, the high-definition camera identifies the dirt residue area through convolutional neural network, the ultrasonic thickness gauge detects the key point dirt thickness, and the cleaning effect comprehensive score is calculated through fusion.

9. The self-correction method of the air cooling energy-saving and efficiency-improving robot for a 600 MW unit as claimed in claim 8, characterized in that: The control center stores the cleaning data and updates the track digital twin model, analyzes the fin damage risk based on the acoustic emission sensor data, Including the following steps, The running data is encrypted and stored in the blockchain database, the micro-crack signal is identified based on the frequency spectrum characteristics of the acoustic emission sensor in the 20-50kHz frequency band, the damage risk area is marked in the digital twin model, and the maintenance work order is generated.

10. The self-correction method of the air cooling energy-saving and efficiency-improving robot for a 600 MW unit as claimed in claim 9, characterized in that: The control center monitors the state of the 600MW unit air-cooled energy-saving robot in real time through the human-machine interface, self-corrects and cleans the work, Including the following steps, The real-time deformation thermodynamic diagram, cleaning score and equipment state are displayed on the human-machine interface, one-key start of the whole process operation is supported, abnormal alarm information is pushed to the mobile terminal in real time, and the maintenance work order is closed loop processed.

Images