Recognition assembly of inspection robot for coal conveying trestle of thermal power plant

Abstract

The invention discloses a thermal power plant coal handling trestle inspection robot identification assembly, and particularly relates to the field of coal handling inspection, which comprises the following steps: establishing a three-dimensional coordinate system of a coal handling trestle, collecting specified basic parameters through each sensor, and presetting a fault feature standard parameter set; constructing an environment influence model for quantifying the interference degree of the environment on image acquisition based on the basic parameters; and calculating a fault risk value by combining the fault feature matching deviation value and the environmental influence coefficient, and comparing the fault risk value with a fault early warning threshold to realize fault early warning. Storing inspection total data, constructing an overall optimization model, and adjusting a threshold value according to the overall optimization model; according to the invention, through multi-parameter acquisition, in combination with complex calculation of the environmental influence coefficient and the cleaning demand degree, the cleaning injection pressure and the scraper rotating speed are dynamically adjusted, precise adaptation of lens cleaning is realized, the problem of incomplete cleaning or excessive cleaning caused by the existing cleaning strategy stiffness is solved, the definition of image acquisition is guaranteed, and the cleaning efficiency is improved. And reliable data support is provided for fault identification.

Description

Technical Field

[0001] This invention relates to the field of coal conveying inspection technology, and more specifically, to a robot identification component for inspecting coal conveying trestle in a thermal power plant. Background Technology

[0002] Coal conveyor bridges in thermal power plants are the core hubs for coal transportation. Currently, the main method for inspecting coal conveyor bridges is to use inspection robots equipped with walking mechanisms, inspection cameras, and simple cleaning components. These robots move along preset tracks to monitor the surface condition of the coal conveyor belts, bridge steel structures, and auxiliary equipment, replacing the traditional manual inspection mode.

[0003] The advantages of existing inspection robots lie in their ability to avoid the high-risk environments such as high altitudes, dust, and noise faced by manual inspections, thereby increasing inspection frequency and expanding coverage. However, they also have significant technical shortcomings: First, the dust concentration inside the coal conveyor bridge fluctuates greatly, and coal ash easily adheres to the surface of the inspection camera lens. Existing cleaning components can only perform cleaning operations at a fixed frequency or in a single mode, and cannot dynamically adjust the cleaning intensity and duration according to the degree of lens contamination and the concentration of ambient dust, resulting in poor cleaning effects or excessive cleaning and wasted resources. Second, the inspection path relies on a preset program and cannot adjust the inspection focus and movement speed according to the operating status of the coal conveyor belt and the distribution of potential equipment malfunctions, resulting in superficial inspections of key areas and redundant inspections of non-key areas. Third, fault identification relies solely on visual image comparison without combining environmental parameters and equipment operating parameters for multi-dimensional analysis, resulting in low identification accuracy and a lack of a scientific fault risk assessment mechanism. It can only passively identify faults after they occur and cannot provide early warnings. Fourth, there is a lack of a closed-loop optimization mechanism for inspection data, making it difficult to improve inspection efficiency and accuracy in the long term.

[0004] To address the shortcomings of existing technologies, such as rigid cleaning strategies, fixed inspection paths, low recognition accuracy, lack of fault warning and continuous optimization mechanisms, there is an urgent need for a recognition component for a coal conveyor bridge inspection robot in thermal power plants. This component should solve the pain points of existing technologies through multi-parameter real-time acquisition, dynamic cleaning strategy formulation, adaptive path planning, multi-dimensional fault identification, and closed-loop optimization. Summary of the Invention

[0005] In order to overcome the above-mentioned defects of the prior art, embodiments of the present invention provide a recognition component for a coal conveying trestle inspection robot in a thermal power plant, which solves the problems mentioned in the background art through the following solutions.

[0006] To achieve the above objectives, the present invention provides the following technical solution: a recognition component for a coal conveying trestle inspection robot in a thermal power plant, comprising: Establish a three-dimensional coordinate system for the coal conveying trestle, collect specified basic parameters through various sensors, and preset a set of standard parameters for fault characteristics. An environmental impact model for quantifying the degree of interference in image acquisition is constructed based on fundamental parameters. The current light transmittance of the lens is collected and the thickness of coal ash adhesion is calculated. The environmental impact coefficient is combined to assess the lens cleaning requirement and compare it with the cleaning start threshold to determine whether cleaning is necessary. Construct a model of key inspection areas and adjust the inspection trajectory, while adjusting the movement speed of the inspection robot according to the speed model; Images are acquired according to the adjusted inspection speed and shooting angle, the image clarity is calculated, and it is determined whether image enhancement preprocessing is required. Extract real-time fault feature parameters from the image and calculate the fault feature matching deviation value by matching them with a preset fault feature standard parameter set. The fault risk value is calculated by combining the fault characteristic matching deviation value and the environmental impact coefficient, and then compared with the fault warning threshold to achieve fault warning. If cleaning is determined to be necessary, the cleaning jet pressure and scraper speed are calculated based on the cleaning demand and failure risk value, dynamic cleaning is performed, and the cleaning effect is verified. Store all inspection data, build an overall optimization model, and adjust the thresholds accordingly.

[0007] Preferably, the three-dimensional coordinate system has its origin at the center point of the starting end of the coal conveying trestle, with the length direction of the trestle as the X-axis, the width direction as the Y-axis, and the direction perpendicular to the plane of the trestle as the Z-axis; the basic parameters include the initial light transmittance of the camera lens. Standard working light transmittance d) the real-time shooting distance between the camera and the surface of the coal conveyor belt; and the camera shooting angle. Coal conveyor belt operating speed v, inspection robot moving base speed The required laser emission and reception time difference for calculating the dust concentration (C) in the trestle bridge environment, the image pixel resolution (P), and the thickness of coal ash adhesion on the lens surface is as follows: Where the laser propagation speed is c; the fault characteristic standard parameter set includes standard values ​​for crack length. Standard value of wear depth Standard value of deviation Fault warning threshold Y and clean start threshold .

[0008] Preferably, the environmental impact model is used to incorporate the following factors: dust concentration C within the trestle, coal conveyor belt speed v, real-time shooting distance d between the camera and the surface of the coal conveyor belt, and initial light transmittance of the camera lens. Standard working light transmittance Camera shooting angle Import the data into the system to calculate the environmental impact coefficient. .

[0009] Preferably, the thickness of the coal ash adhesion on the lens The method for assessing the lens cleaning requirement is as follows: calculate the lens cleaning requirement. ,in The current light transmittance of the lens, when Q> When it is determined that cleaning needs to be started, Q<= At that time, maintain the current inspection status.

[0010] Preferably, the inspection key area model is used to calculate the environmental impact coefficient. The operating speed v of the coal conveyor belt, the real-time shooting distance d between the camera and the surface of the coal conveyor belt, and the camera shooting angle. Importing this data yields the coefficients for key inspection areas. The speed model is used to calculate the adjusted moving speed of the inspection robot. Where L is the length of the trestle and W is the width of the trestle; the method of adjusting the inspection trajectory is as follows: when A>1.2, the inspection trajectory adopts an S-shaped turnaround path with a turnaround interval; when A<=1.2, a straight inspection path is adopted.

[0011] Preferably, the image acquisition is performed at fixed time intervals. Execution, the image clarity The image enhancement preprocessing method is as follows: when At that time, image enhancement processing is activated, adjusting contrast and brightness. If the fault is detected, proceed directly to the fault identification step.

[0012] Preferably, the real-time fault characteristic parameters include real-time crack length. Real-time wear depth Real-time deviation Fault feature matching deviation value The smaller the M value, the smaller the deviation between the real-time state and the standard state.

[0013] Preferably, the fault risk value The fault warning is as follows: when R>Y, the control center issues an audible and visual warning and records the fault location coordinates; when R<=Y, it issues a warning at intervals of... Recalculate R.

[0014] Preferably, the cleaning jet pressure Actual rotation speed of the rubber scraper ,in The base rotation speed of the rubber scraper. The duration of a single cleaning cycle; the dynamic cleaning process controls the annular cleaning spray assembly according to... Cleaning fluid is sprayed, and the drive motor rotates the rubber scraper at n-fold speed to clean; the cleaning effect is verified by collecting samples again after cleaning. And calculate Q until Q <= Stop cleaning.

[0015] Preferably, the full inspection data includes all basic parameters, calculated coefficients and values, fault records, and cleaning execution status; the overall optimization model is used to calculate the overall optimization coefficients. ,in This is the sum of all R values ​​from this inspection. This is the sum of all Q values ​​from this inspection. For all of this inspection The sum of values; the adjustment threshold includes the cleaning start threshold and fault warning threshold for the next inspection, specifically: when When the value is >0.8, the next inspection will be Down 10%, Y down 8%, when When <=0.8, the original parameter is maintained.

[0016] The technical effects and advantages of this invention are as follows: 1. This invention dynamically adjusts the cleaning spray pressure and scraper speed by collecting multiple parameters and combining complex calculations of environmental impact coefficients and cleaning requirements, thereby achieving precise adaptation for lens cleaning. This solves the problem of incomplete or over-cleaning caused by the rigidity of existing cleaning strategies, ensures the clarity of image acquisition, and provides reliable data support for fault identification. 2. This invention adjusts the inspection movement speed and trajectory by calculating the coefficient of key areas to achieve adaptive path planning. This solves the problems of insufficient inspection of key areas and redundancy in non-key areas caused by the fixed inspection path in the existing system. While ensuring the comprehensiveness of the inspection, it improves the inspection efficiency and reduces the energy consumption of the robot. Through image clarity quantification evaluation and preprocessing, combined with multi-dimensional fault feature matching calculation and fault risk value assessment, it breaks through the recognition limitations of single image comparison and solves the problems of low recognition accuracy and high false detection and missed detection rates in the existing technology. It achieves accurate fault identification and early warning, buys sufficient time for equipment maintenance, and reduces the losses caused by the expansion of faults. 3. This invention constructs a closed-loop optimization mechanism for inspection strategies by storing inspection data and calculating overall optimization coefficients, solving the problem of insufficient continuous optimization capabilities in existing technologies. As the number of inspections increases, the inspection parameters continuously adapt to actual working conditions, and the accuracy and efficiency of inspections gradually improve over long-term operation, extending the service life of the inspection robot and reducing equipment maintenance costs. Through the collaborative work of various modules and the logical connection of multiple steps, the cleaning strategy, path planning, fault identification, early warning, and optimization are organically integrated, solving the problem of functional fragmentation in existing technologies. This achieves intelligent, efficient, and precise inspection of coal conveyor bridges, ensuring the safe and stable operation of coal conveyor bridges and improving the overall reliability of thermal power plant production. Attached Figure Description

[0017] Figure 1 This is a schematic diagram of the overall structure of the present invention. Detailed Implementation

[0018] 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 skilled in the art without creative effort are within the scope of protection of the present invention.

[0019] A recognition component for an inspection robot on a coal conveying trestle in a thermal power plant includes an inspection robot body, an environmental perception module, an equipment status monitoring module, an image acquisition and processing module, a cleaning execution module, a path planning module, a fault identification and early warning module, and a control center. The environmental perception module includes a dust sensor and a transmittance meter; the equipment status monitoring module includes a speed sensor and an angle sensor; the image acquisition and processing module includes a high-definition camera; and the cleaning execution module includes a ring-shaped cleaning spray assembly, a drive motor, and a rubber scraper. Each module is communicatively connected to the control center. The system also includes an inspection box and a traveling track. The inspection box is equipped with a traveling mechanism, and an electric telescopic rod is fixedly installed inside the inspection box. The inspection camera body is fixedly installed at the bottom of the electric telescopic rod. This monorail liftable coal conveyor belt inspection system, through the combined use of cleaning components and cleaning structures, sprays cleaning fluid in a ring onto the inspection camera body, making the camera body fully covered with cleaning fluid. At the same time, the drive motor drives the rotating rod to rotate, which in turn drives the drive gear to rotate the external gear ring, which in turn drives the rubber scraper to rotate around the outside of the inspection camera body. The rubber scraper cleans the coal dust off the outside of the inspection camera body, thereby ensuring that the image captured by the inspection camera body is clear and improving the inspection effect.

[0020] As attached Figure 1 The identification component shown is for a coal conveyor bridge inspection robot in a thermal power plant, specifically including: Establish a three-dimensional coordinate system for the coal conveying trestle, collect specified basic parameters through various sensors, and preset a set of standard parameters for fault characteristics. In this embodiment, it should be specifically noted that: the three-dimensional coordinate system has its origin at the center point of the starting end of the coal conveying trestle, with the length direction of the trestle as the X-axis, the width direction as the Y-axis, and the direction perpendicular to the plane of the trestle as the Z-axis; the basic parameters include the initial light transmittance of the camera lens. Standard working light transmittance d) the real-time shooting distance between the camera and the surface of the coal conveyor belt; and the camera shooting angle. Coal conveyor belt operating speed v, inspection robot moving base speed The required laser emission and reception time difference for calculating the dust concentration (C) in the trestle bridge environment, the image pixel resolution (P), and the thickness of coal ash adhesion on the lens surface is as follows: Where the laser propagation speed is c; the fault characteristic standard parameter set includes standard values ​​for crack length. Standard value of wear depth Standard value of deviation Fault warning threshold Y and clean start threshold The establishment of the three-dimensional coordinate system follows the principle of "spatial uniqueness." With the center point of the starting end of the trestle as the origin and coordinate axes set along the key dimensions of the trestle, the system can accurately calibrate the position of the inspection robot and the coordinates of any faults. This provides a spatial benchmark for subsequent fault location and path planning, avoiding inefficient fault diagnosis due to ambiguous positioning. The collected basic parameters are all "necessary inputs" for subsequent core calculations: transmittance parameters relate to lens cleaning requirements and image quality assessment; distance and angle parameters affect image clarity and inspection coverage; speed parameters relate to path planning and fault risk assessment; and dust concentration quantifies environmental interference. All parameters are selected around the three core dimensions of "environment-equipment-image," with no redundant data collection. Preset fault characteristic standard parameter sets, warning thresholds, and cleaning start thresholds provide a clear benchmark for subsequent comparisons between "real-time status" and "standard status," avoiding misjudgments and incorrect cleaning timing due to a lack of unified judgment standards. The existence significance addresses the problem of ambiguous fault location caused by the lack of a unified spatial benchmark in existing technologies, enabling precise fault location tracing; it fills the gap in the fragmented parameter acquisition of existing technologies, providing complete data support for subsequent environmental assessment, path optimization, fault identification and other steps, ensuring closed-loop computational logic in each stage; and it establishes standardized judgment benchmarks to avoid the randomness of human subjective judgment, laying the foundation for subsequent automated decision-making (clean start-up, fault early warning).

[0021] An environmental impact model for quantifying the degree of interference in image acquisition is constructed based on fundamental parameters. In this embodiment, it should be specifically noted that: the environmental impact model is used to incorporate the dust concentration C within the trestle, the operating speed v of the coal conveyor belt, the real-time shooting distance d between the camera and the surface of the coal conveyor belt, and the initial light transmittance of the camera lens. Standard working light transmittance Camera shooting angle Import the data into the system to calculate the environmental impact coefficient. Among these factors, environmental interference is the core factor affecting image acquisition quality. However, a single environmental parameter (such as dust concentration C) cannot fully reflect the intensity of interference. The higher the dust concentration and the faster the conveyor belt runs (the more severe the dust), the stronger the interference. The farther the shooting distance and the greater the lens transmittance redundancy, the weaker the interference. The larger the shooting angle (larger oblique angle), the more obvious the light reflection interference. The formula scientifically integrates multi-dimensional environmental influencing factors through the combination of "positive correlation parameter × reciprocal of negative correlation parameter × logarithmic correction term" to achieve quantitative normalization of the degree of interference. The natural logarithmic function is used to correct the shooting angle because the influence of angle on interference increases non-linearly (interference changes slowly within a small angle range, and increases rapidly after a large angle). The logarithmic function can accurately fit this characteristic and avoid the interference assessment bias caused by linear calculation. The existence significance addresses the problem of existing technologies that "only qualitatively judge environmental interference without quantification," transforming abstract environmental impacts into specific calculable coefficients, providing a quantitative basis for subsequent path planning, image preprocessing, and cleaning strategies; it enables the linkage between environmental interference and subsequent operations, avoiding problems such as poor image quality and resource waste caused by existing technologies that "ignore environmental differences and adopt a uniform inspection / cleaning mode."

[0022] The current light transmittance of the lens is collected and the thickness of coal ash adhesion is calculated. The environmental impact coefficient is combined to assess the lens cleaning requirement and compare it with the cleaning start threshold to determine whether cleaning is necessary. In this embodiment, it is specifically necessary to explain the thickness of the coal ash adhesion on the lens. The method for assessing the lens cleaning requirement is as follows: calculate the lens cleaning requirement. ,in The current light transmittance of the lens, when Q> When it is determined that cleaning needs to be started, Q<= The current inspection status is maintained. Lens cleaning requirements are jointly determined by the "degree of contamination" and "environmental interference": relying solely on current transmittance (degree of contamination) without considering environmental interference may result in situations where "low contamination but high interference still produces blurry images" without cleaning; relying solely on environmental interference without considering actual contamination may result in situations where "high interference occurs but the lens is clean" but cleaning is done blindly. The thickness of coal ash adhesion is calculated using laser ranging time difference, directly reflecting the physical state of lens contamination, which is more comprehensive than simply relying on transmittance (avoiding misjudgments due to transmittance variations). A clear cleaning initiation judgment rule is established by comparing with a preset cleaning initiation threshold, avoiding manual intervention and ensuring automated and consistent cleaning decisions. Its significance lies in breaking through the rigid strategy of existing "fixed frequency / single mode cleaning" strategies, achieving "on-demand cleaning," and solving the problems of incomplete or over-cleaning; establishing a linkage mechanism of "contamination status - environmental interference - cleaning requirements" ensures that lens transmittance always meets image acquisition requirements, providing a clear image foundation for subsequent fault identification.

[0023] Construct a model of key inspection areas and adjust the inspection trajectory, while adjusting the movement speed of the inspection robot according to the speed model; In this embodiment, it should be specifically noted that the key inspection area model is used to calculate the environmental impact coefficient. The operating speed v of the coal conveyor belt, the real-time shooting distance d between the camera and the surface of the coal conveyor belt, and the camera shooting angle. Importing this data yields the coefficients for key inspection areas. The speed model is used to calculate the adjusted moving speed of the inspection robot. Where L is the length of the trestle and W is the width of the trestle; the method of adjusting the inspection trajectory is as follows: when A>1.2, the inspection trajectory adopts an S-shaped back-and-forth path with a back-and-forth interval; when A<=1.2, a straight inspection path is adopted. Among them, the calculation logic of the key inspection area coefficient closely follows the "risk orientation": the greater the environmental interference (higher image recognition difficulty) and the faster the belt running speed (higher probability of failure), the stronger the demand for key inspection; the farther the shooting distance (wider coverage) and the larger the shooting angle (larger coverage area of ​​a single image), the weaker the demand for key inspection. The formula accurately quantifies the priority of key areas; the adjustment of the inspection speed follows the principle of "slow inspection of key areas and fast inspection of non-key areas": the base speed is combined with the trestle size, key coefficient, and shooting distance, and calculated by formula to ensure that the movement speed of key areas is slow and the inspection is more detailed; the trajectory selection is linked to the A value: an "S" shaped back-and-forth path is adopted (to ensure no dead corners in coverage) and a straight path is adopted to balance inspection efficiency and coverage quality, avoiding the "fixed path leading to missed inspection of key areas or redundancy of non-key areas" in the existing technology. The existence significance addresses the problem of "mismatch between inspection paths and risks" in existing technologies, enabling inspection resources to be tilted towards high-risk and high-interference areas, thereby improving the targeting of inspections; it also optimizes inspection efficiency, reducing inspection time in non-key areas and lowering robot energy consumption while ensuring the quality of coverage in key areas.

[0024] Images are acquired according to the adjusted inspection speed and shooting angle, the image clarity is calculated, and it is determined whether image enhancement preprocessing is required. In this embodiment, it should be specifically noted that the image acquisition is performed at fixed time intervals. Execution, the image clarity The image enhancement preprocessing method is as follows: when At that time, image enhancement processing is activated, adjusting contrast and brightness. Upon reaching the fault identification step, the system directly proceeds to the fault identification stage. The image acquisition frequency is linked to the inspection speed: the faster the speed, the farther the coverage distance per unit time. Acquiring images at fixed time intervals ensures that the image overlap rate meets the requirements for stitching and recognition, avoiding missed shots due to a mismatch between speed and acquisition frequency. The calculation of image clarity integrates "hardware parameters - environmental parameters - lens status": pixel resolution (hardware upper limit), shooting distance (the farther the distance, the lower the clarity), dust concentration (environmental interference), and current transmittance (lens status). The formula comprehensively reflects image quality, avoiding misjudgments of clarity due to a single indicator (such as looking only at pixels). The preprocessing judgment threshold setting is based on engineering practice of "minimum image recognition clarity requirements": when S is below this threshold, fault features are easily masked by blurry noise, requiring enhanced processing; when it is above this threshold, the image quality meets the recognition requirements, requiring no additional processing, balancing processing efficiency and recognition accuracy. The existence significance ensures that the quality of the acquired images meets the requirements for fault identification, solving the core problem of "false or missed fault identification caused by image blurring" in existing technologies; it avoids the waste of computing resources caused by indiscriminate image preprocessing, realizes precise preprocessing of "processing only when necessary", and improves the efficiency of fault identification.

[0025] Extract real-time fault feature parameters from the image and calculate the fault feature matching deviation value by matching them with a preset fault feature standard parameter set. In this embodiment, it should be specifically noted that the real-time fault characteristic parameters include the real-time crack length. Real-time wear depth Real-time deviation Fault feature matching deviation value The smaller the M value, the smaller the deviation between the real-time state and the standard state. Three fault characteristic parameters—crack length, wear depth, and deviation—were selected to cover high-frequency fault types of the core equipment (belt conveyor and steel structure) of the coal conveyor bridge, ensuring no key characteristics were omitted. The fault characteristic matching deviation value M was calculated using the Euclidean distance formula combined with pixel resolution P and inspection speed correction: the Euclidean distance comprehensively reflects the combined deviation of multi-dimensional characteristics, divided by... This is because higher pixel resolution (clearer details) and slower inspection speed (more accurate feature acquisition) lead to more precise deviation assessment, avoiding the amplification or reduction of deviations caused by a single Euclidean distance calculation. Its existence significance breaks through the limitations of existing technologies that rely solely on visual qualitative comparison, quantifying fault characteristics into specific parameters to achieve precise quantitative comparison between "real-time state" and "standard state." This provides core input for subsequent fault risk assessment, avoiding subjectivity in fault judgment and improving the objectivity and accuracy of fault identification.

[0026] The fault risk value is calculated by combining the fault characteristic matching deviation value and the environmental impact coefficient, and then compared with the fault warning threshold to achieve fault warning. In this embodiment, it is specifically necessary to explain that the fault risk value The fault warning is as follows: when R>Y, the control center issues an audible and visual warning and records the fault location coordinates; when R<=Y, it issues a warning at intervals of... Recalculate R. The fault risk depends not only on the "current deviation M" but also on the "fault development speed" and "environmental interference": the faster the belt speed v (the faster the fault develops), the slower the inspection speed (the longer the continuous fault monitoring time), and the shorter the data collection interval (the more frequent the data updates), the higher the risk; the greater the environmental interference (the higher the uncertainty of fault identification, the greater the risk of delayed processing), the higher the risk. The formula, through multi-factor coupling, accurately quantifies the "urgency" and "potential loss" of the fault. The natural exponential function is used to amplify the impact of environmental interference on risk—under high environmental interference, the probability of fault misjudgment / missed judgment increases. If timely warning is not given, it may lead to the expansion of the fault. The exponential function can reflect this characteristic of "accelerated risk accumulation." The warning threshold Y is compared with the risk value R, combined with (dynamic re-inspection interval, linked with bridge size, inspection speed, and pixel resolution to ensure re-inspection efficiency), to achieve a graded warning mechanism of "immediate warning for high risk and dynamic monitoring for low risk." The existence significance addresses the shortcomings of existing technologies that "passively identify faults only after they occur," enabling early prediction and tiered warning of fault risks, thus providing buffer time for equipment maintenance; it avoids the proliferation of warnings or delays in critical faults caused by "one-size-fits-all warnings," thereby improving the effectiveness and relevance of warnings.

[0027] If cleaning is determined to be necessary, the cleaning jet pressure and scraper speed are calculated based on the cleaning demand and failure risk value, dynamic cleaning is performed, and the cleaning effect is verified. In this embodiment, it is specifically necessary to explain that the cleaning spray pressure Actual rotation speed of the rubber scraper ,in The base rotation speed of the rubber scraper. The duration of a single cleaning cycle; the dynamic cleaning process controls the annular cleaning spray assembly according to... Cleaning fluid is sprayed, and the drive motor rotates the rubber scraper at n-fold speed to clean; the cleaning effect is verified by collecting samples again after cleaning. And calculate Q until Q <= Stop cleaning. The cleaning intensity (spray pressure, scraper speed) needs to be linked to "cleaning demand Q" and "failure risk R": the larger Q is (the more severe the contamination), the higher the required spray pressure and the faster the scraper speed; the larger R is (the higher the failure risk), the faster the image clarity needs to be restored for accurate fault identification. Therefore, Q and R in the formula are both positively correlated factors. The spray pressure formula incorporates the rubber scraper's base speed, single cleaning duration, shooting distance, and shooting angle because the base speed determines the lower limit of cleaning, duration and distance affect the cleaning coverage, and angle affects the accuracy of the cleaning fluid spray direction. The scraper speed formula is introduced because there is a synergistic effect between spray pressure and scraper speed (the higher the pressure, the more the scraper speed needs to be increased to avoid cleaning fluid residue, and the rate of increase gradually slows down). After cleaning, Q is re-collected and verified, forming a closed loop of "cleaning-verification-re-cleaning" to ensure that the cleaning effect meets the requirements and avoids "incomplete cleaning" or "over-cleaning". The existence significance enables the "dynamic adaptation" of cleaning strategies, solving the problem of poor cleaning effect or waste of resources caused by the "fixed cleaning intensity" of existing technologies; it establishes a closed-loop verification mechanism for cleaning effect to ensure that the image quality after cleaning meets the requirements for fault identification and avoids misjudgment of faults due to lens contamination after cleaning.

[0028] Store all inspection data, build an overall optimization model, and adjust the thresholds accordingly.

[0029] In this embodiment, it should be specifically noted that: the total inspection data includes all basic parameters, calculated coefficients and values, fault records, and cleaning execution status; the overall optimization model is used to calculate the overall optimization coefficients. ,in This is the sum of all R values ​​from this inspection. This is the sum of all Q values ​​from this inspection. For all of this inspection The sum of values; the adjustment threshold includes the cleaning start threshold and fault warning threshold for the next inspection, specifically: when When the value is >0.8, the next inspection will be Down 10%, Y down 8%, when When the value is less than or equal to 0.8, the original parameters are maintained. The calculation of the overall optimization coefficient O integrates the "total risk of all faults in this inspection," the "total cleaning requirement," the "total inspection speed," and the "stack size," comprehensively reflecting the "overall risk level," "cleaning resource consumption," and "inspection efficiency" of this inspection, and quantifying the adaptability of this inspection strategy. A value of >0.8 indicates that there were many high-risk and high-cleanliness-requirement scenarios in this inspection, and the original threshold was too lenient, so it needs to be lowered to improve sensitivity. A threshold value <= 0.8 indicates that the threshold is suitable for the current operating conditions and requires no adjustment. The threshold adjustment range is based on the "sensitivity gradient optimization" principle in engineering practice, avoiding strategy oscillations caused by excessive one-time adjustments and ensuring the stability of optimization. Its significance lies in establishing a "closed-loop optimization mechanism" for the inspection strategy, addressing the shortcomings of existing technologies where "inspection parameters are fixed and cannot adapt to changes in operating conditions," and enabling dynamic iteration of the inspection strategy according to operating conditions (dust concentration, fault distribution, equipment status). With long-term operation, the accuracy, efficiency, and resource utilization of the inspection will continuously improve, extending the service life of the inspection robot and reducing overall operation and maintenance costs.

[0030] Secondly: The accompanying drawings of the embodiments disclosed in this invention only involve the structures involved in the embodiments disclosed in this invention. Other structures can refer to the general design. In the absence of conflict, the same embodiment and different embodiments of this invention can be combined with each other. In conclusion, the above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A coal-fired power plant coal trestle inspection robot identification component, characterized in that, The method comprises the following steps: Establishing a three-dimensional coordinate system of the coal conveying trestle, collecting specified basic parameters through various sensors, and presetting a set of standard parameters of fault characteristics; Based on the basic parameters, an environmental influence model is constructed to quantify the degree of interference of the environment on image acquisition; The current light transmittance of the lens is collected and the thickness of the coal ash attached is calculated, and the lens cleaning demand is evaluated in combination with the environmental influence coefficient, and compared with the cleaning start threshold to determine whether cleaning is needed; A key area model is constructed and the inspection trajectory is adjusted, and the moving speed of the inspection robot is adjusted according to the speed model; According to the adjusted inspection speed and shooting angle, the image is collected, the image clarity is calculated, and it is determined whether image enhancement preprocessing is needed; Real-time fault characteristic parameters in the image are extracted, matched with the preset set of standard parameters of fault characteristics, and the fault characteristic matching deviation value is calculated; The fault risk value is calculated in combination with the fault characteristic matching deviation value and the environmental influence coefficient, and the fault early warning is realized by comparing with the fault early warning threshold; If it is determined that cleaning is needed, the cleaning spray pressure and scraper rotating speed are calculated in combination with the cleaning demand and the fault risk value, dynamic cleaning is performed, and the cleaning effect is verified; The whole amount of data of the inspection is stored, an overall optimization model is constructed, and the threshold is adjusted accordingly.

2. The coal plant jetty inspection robot identification component according to claim 1, characterized in that: The three-dimensional coordinate system takes the center point of the starting end of the coal conveying trestle as the origin, the length direction of the trestle as the X axis, the width direction as the Y axis, and the direction perpendicular to the plane of the trestle as the Z axis; the basic parameters include the initial light transmittance of the camera lens , the standard working light transmittance , the real-time shooting distance d of the camera and the surface of the coal conveying belt, the shooting angle of the camera , the running speed v of the coal conveying belt, the moving basic speed of the inspection robot , the dust concentration C in the environment inside the trestle, the image pixel resolution P, and the time difference required for laser emission and reception for calculating the thickness of the attached coal ash on the surface of the lens , wherein the laser propagation speed is c; the fault feature standard parameter set contains the crack length standard value , the wear depth standard value , the deviation offset standard value , the fault warning threshold Y, and the cleaning starting threshold .

3. The coal conveyor belt inspection robot identification component of claim 1, wherein: The environmental influence model is used for calculating the environmental dust concentration C in the trestle, the running speed v of the coal conveying belt, the real-time shooting distance d of the camera and the surface of the coal conveying belt, the initial light transmittance of the camera lens , the standard working light transmittance , the shooting angle of the camera , and the environmental influence coefficient is calculated.

4. The coal plant jetty inspection robot identification component according to claim 2, characterized in that: The lens coal ash attachment thickness The lens cleaning requirement degree is calculated Wherein is the current light transmittance of the lens, when Q> determines that cleaning needs to be started, Q<= maintains the current inspection state.

5. The coal plant jetty inspection robot identification component according to claim 2, characterized in that: The inspection focus area model is used for calculating an environmental influence coefficient , a coal conveying belt running speed v, a real-time shooting distance d of a camera and a coal conveying belt surface, and a camera shooting angle The inspection focus area coefficient is obtained in the introduction ; the speed model is used for calculating a moving speed of the inspection robot after adjustment , wherein L is a length of the stack bridge, and W is a width of the stack bridge; and the adjustment of the inspection track is as follows: when A>1.2, the inspection track adopts an S-shaped turning path with a turning interval, and when A<=1.2, the inspection track adopts a straight line inspection path.

6. The coal plant jetty inspection robot identification component according to claim 2, characterized in that: The image acquisition is at fixed time intervals The image definition is executed The image enhancement preprocessing is performed in the following manner: when The image enhancement processing is started, the contrast and brightness are adjusted, when The fault recognition step is directly entered.

7. The coal conveyor belt inspection robot identification component of claim 2, wherein: The real-time fault characteristic parameters include real-time crack length , real-time wear depth , real-time deviation offset , fault characteristic matching deviation value , wherein the smaller the M value is, the smaller the deviation between the real-time state and the standard state is.

8. The coal conveying trestle inspection robot identification assembly of claim 7, wherein: The fault risk value The fault warning is as follows: when R>Y, the control center issues an audible and visual warning and records the fault location coordinates; when R<=Y, it issues a warning at intervals of... Recalculate R.

9. The coal conveying trestle inspection robot identification assembly of claim 8, wherein: The cleaning jet pressure Actual rotation speed of the rubber scraper ,in The base rotation speed of the rubber scraper. The duration of a single cleaning cycle; the dynamic cleaning process controls the annular cleaning spray assembly according to... Cleaning fluid is sprayed, and the drive motor rotates the rubber scraper at n-fold speed to clean; the cleaning effect is verified by collecting samples again after cleaning. And calculate Q until Q <= Stop cleaning.

10. The coal conveying trestle inspection robot identification assembly of claim 5, wherein: The inspection total data includes all basic parameters, various coefficients and values calculated, fault records, and cleaning execution situations; and the overall optimization model is used for calculating an overall optimization coefficient , wherein is a sum of all R values of this inspection, is a sum of all Q values of this inspection, is a sum of all values of this inspection; and the adjustment threshold includes a cleaning start threshold and a fault warning threshold of the next inspection, specifically: when > 0.8, the next inspection will be decreased by 10%, and Y is decreased by 8%; and when <= 0.8, the original parameters are maintained.

Images