Structural simulation optimization method and system of wind power generation tower drum cleaning robot

Abstract

The invention relates to the field of robot structure simulation optimization, in particular to a structure simulation optimization method and system for a wind power generation tower cleaning robot, and the system comprises a parameter acquisition module, a stress decomposition module, a working condition evaluation module, a risk judgment module and a deformation screening module. On one hand, the safety accident risk caused by structure instability can be effectively reduced; on the other hand, scientificity and accuracy of analysis results are ensured, important theoretical basis and data support are provided for subsequent structure improvement, material optimization and control strategy design, and therefore efficient, safe and reliable operation of the cleaning robot is achieved.

Description

TECHNICAL FIELD

[0001] The application belongs to the field of robot structure simulation optimization, and relates to a wind power tower cleaning robot structure simulation optimization method and system. BACKGROUND

[0002] The operation safety and stability analysis of the cleaning robot under complex working conditions is a key link to ensure its reliable operation, and the simulation analysis technology plays a crucial role therein. Through multi-state simulation analysis of the cleaning robot, dynamic parameters such as the motion trajectory, speed and acceleration of the mass center and brush disc under different motion states can be obtained, thereby providing accurate basis for subsequent force and stability evaluation.

[0003] At present, there are still some aspects to be optimized in the simulation analysis process of the cleaning robot, mainly including the following aspects: 1. The existing simulation analysis is mainly focused on the overall motion trend, and lacks fine modeling and dynamic response analysis of the motion parameters of the mass center and brush disc under different operating states, resulting in deviation of the trajectory, speed and acceleration data, and making it difficult to accurately reflect the real force characteristics of the robot under complex paths and high-risk working conditions.

[0004] 2. The existing research does not comprehensively consider the force conditions during the operation of the robot, and fails to fully combine the mutual coupling relationship of gravity, friction, wind resistance and other forces, resulting in inaccurate force curve drawing, and failing to effectively identify the key components with large force and potential dangerous working conditions, thereby affecting the safety evaluation and structure optimization design of the cleaning robot under different working postures. SUMMARY

[0005] In view of the problems existing in the prior art, the application provides a wind power tower cleaning robot structure simulation optimization method and system, which is used to solve the above technical problems.

[0006] In order to achieve the above purpose and other purposes, the technical scheme adopted by the application is as follows: The first aspect of the application provides a wind power tower cleaning robot structure simulation optimization method, which comprises the following steps: Step S1: constructing an assembly model of the cleaning robot by a three-dimensional modeling software, and performing multi-body dynamics simulation on the cleaning robot to obtain motion trajectory, speed and acceleration parameter data of the mass center and brush disc of the cleaning robot under multiple motion states; performing data cleaning and normalization processing on the motion trajectory, speed and acceleration parameter data to obtain standardized motion parameter data; Step S2: based on the standardized motion parameter data, multi-physical field force analysis of the cleaning robot in each motion state is performed to obtain comprehensive force data including gravity, friction, wind resistance; the force component decomposition is performed on the comprehensive force data to obtain main force component data; Step S3: according to the main force component data, a time sequence force curve is drawn, and key components and dangerous working condition data whose force exceeds a preset force threshold are identified through a curve peak value; the dangerous working condition data is subjected to working condition severity evaluation to obtain high-risk working condition data; Step S4: for the high-risk working condition data, transient force analysis of the cleaning robot in the downward cleaning and return upward two states is performed, and instability sliding and overturning risk simulation is performed based on wind resistance and dirt accumulation factors to obtain instability risk data; the instability risk data is subjected to risk level division to obtain risk level data; Step S5: based on the risk level data, stress deformation cloud atlas analysis is combined to determine the occurrence position of the maximum stress and deformation in the mainframe assembly and the roller support assembly of the cleaning robot.

[0007] The second aspect of the application provides a structure simulation optimization system of a wind power tower cleaning robot, which comprises: The parameter acquisition module: the assembly model of the cleaning robot is constructed through a three-dimensional modeling software, and the multi-body dynamics simulation of the cleaning robot is performed to obtain the center of mass and the motion trajectory, speed and acceleration parameter data of the brush disc of the cleaning robot in multiple motion states; the motion trajectory, speed and acceleration parameter data are subjected to data cleaning and normalization processing to obtain standardized motion parameter data; The force decomposition module: based on the standardized motion parameter data, multi-physical field force analysis of the cleaning robot in each motion state is performed to obtain comprehensive force data including gravity, friction, wind resistance; the force component decomposition is performed on the comprehensive force data to obtain main force component data; The working condition evaluation module: according to the main force component data, a time sequence force curve is drawn, and key components and dangerous working condition data whose force exceeds a preset force threshold are identified through a curve peak value; the dangerous working condition data is subjected to working condition severity evaluation to obtain high-risk working condition data; The risk discrimination module: for the high-risk working condition data, transient force analysis of the cleaning robot in the downward cleaning and return upward two states is performed, and instability sliding and overturning risk simulation is performed based on wind resistance and dirt accumulation factors to obtain instability risk data; the instability risk data is subjected to risk level division to obtain risk level data; The deformation screening module: based on the risk level data, stress deformation cloud atlas analysis is combined to determine the occurrence position of the maximum stress and deformation in the mainframe assembly and the roller support assembly of the cleaning robot.

[0008] As described above, the wind power tower cleaning robot structure simulation optimization method and system provided by the application has at least the following beneficial effects: The wind power tower cleaning robot structure simulation optimization method and system provided by the application can accurately capture the mass center and brush disc motion trajectory, speed and acceleration parameter data of the cleaning robot in different operating states, and further obtain the real dynamic behavior characteristics, by constructing a three-dimensional assembly model of the cleaning robot and performing multi-body dynamics simulation. By cleaning and normalizing the simulation data, errors and noises can be effectively removed, ensuring the reliability and consistency of the motion parameters, and providing standardized input data for subsequent stress analysis. BRIEF DESCRIPTION OF DRAWINGS

[0009] In order to more clearly illustrate the technical solutions of the embodiments of the application, the following will briefly introduce the drawings needed to be used in the description of the embodiments. Obviously, the drawings in the following description are only some embodiments of the application, and for those skilled in the art, other drawings can also be obtained without creative labor on the basis of these drawings.

[0010] Figure 1 The schematic diagram for connecting each step of the method of the application.

[0011] Figure 2 The schematic diagram for connecting each module of the system of the application. DETAILED DESCRIPTION

[0012] The following will combine the above content of the application, which is only an example and description of the concept of the application. Those skilled in the art can make various modifications or supplements to the described embodiments or use similar ways to replace, as long as they do not deviate from the concept of the application or exceed the scope defined by the claims, which shall belong to the protection scope of the application.

[0013] Embodiment 1 Please refer to Figure 1 The wind power tower cleaning robot structure simulation optimization method comprises the following steps: Step S1: constructing an assembly model of the cleaning robot by a three-dimensional modeling software, and performing multi-body dynamics simulation on the cleaning robot to obtain motion trajectory, speed and acceleration parameter data of the mass center and brush disc of the cleaning robot in various motion states; performing data cleaning and normalization processing on the motion trajectory, speed and acceleration parameter data to obtain standardized motion parameter data; Exemplarily, step S1 comprises the following sub-steps: Step S11: establishing an assembly model containing driving components and brush discs based on the assembly structure of the cleaning robot in the three-dimensional modeling software, and setting motion constraint conditions of the motor drive; Step S12: Apply linear reciprocating motion and spiral motion to the assembled model under two typical working conditions, perform multi-body dynamics simulation, and obtain the initial parameter set of the center of mass displacement trajectory, brush disc motion speed and axial acceleration; Step S13: Based on the brush disc material deformation threshold, eliminate abnormal data nodes with axial acceleration exceeding the limit in the initial parameter set, and obtain the compliant motion parameter set; Step S14: Perform unit dimension unification on the compliant motion parameter set, convert the displacement trajectory into a normalized curvature coefficient, and convert the speed into a time-normalized ratio, denoted as unit time displacement ratio, to generate standardized motion parameter data.

[0014] In one specific embodiment, when constructing the kinematics model of the cleaning robot, first, the three-dimensional geometric models of the driving motor, transmission shaft and brush disc are imported into the three-dimensional modeling software based on the assembly drawing. According to the elastic modulus and Poisson's ratio parameters of the brush disc material in the material library, the six-degree-of-freedom constraint conditions at the connection between the transmission shaft and the brush disc are set, and the axial degree-of-freedom is locked based on the brush disc working pressure threshold. The kinematics joint constraints of the motor drive system are set, the joint rotation torque threshold is set by calculating the relationship between the maximum rotation torque of the motor shaft and the transmission ratio, and the displacement upper and lower limits of the linear reciprocating motion are determined according to the brush disc working stroke range. In the multi-body dynamics simulation execution stage, the kinematics parameter set of the linear reciprocating motion and the spiral motion is loaded respectively, wherein the acceleration change rate is set based on the trapezoidal acceleration control strategy of the servo motor under linear motion, and the lateral displacement component of the brush disc is calculated according to the preset spiral angle under spiral motion. The dynamic solution of the assembled model is performed by using the explicit numerical integration algorithm, and the time-domain waveform data of the center of mass coordinate displacement, the brush disc rotation angular velocity and the axial acceleration are synchronously collected at discrete time steps. The axial acceleration is set with an abnormal data threshold according to the fatigue strength limit of the brush disc material, the extreme value of the time series data sequence in the initial parameter set is traversed, and when the acceleration peak value of three consecutive sampling points exceeds the threshold corresponding to the material yield stress, the abnormal flag is triggered. The abnormal interval data is smoothed and corrected by using the cubic spline interpolation algorithm. In the dimension unification stage, when converting the Cartesian coordinate system data of the center of mass displacement trajectory into the curvature parameter, the tangent angle deviation of adjacent points is calculated by using the point-by-point coordinate difference method, and the normalized curvature coefficient is constructed based on the ratio of the trajectory arc length to the chord length; the speed parameter is dimensionless by calculating the percentage ratio of the actual linear speed to the maximum design speed, and the acceleration data is also standardized by root mean square to eliminate the unit system difference.

[0015] For example, step S12 specifically includes: Step S121: According to the load distribution condition of the brush disc contact surface, set the acceleration threshold of linear reciprocating motion and the curvature radius variation range of spiral motion, and generate two sets of motion working condition parameters; Step S122: Define the torque input threshold and frequency response bandwidth of the driving components under two motion conditions based on the motor rated power upper limit of the assembly model; Step S123: Load two sets of motion condition parameters into the multi-body dynamics simulation module, perform continuous motion simulation under the constraint of motor torque threshold, and output the original simulation data package containing time-stamped centroid space coordinates, brush disc rotation speed pulse waveform, and axial acceleration peak-to-valley value; Step S124: Analyze the motion phase of the original simulation data package, extract the centroid X-axis displacement extreme points under linear motion and the Z-axis angular velocity change gradient under spiral motion as the feature parameter set; Step S125: Input the extracted feature parameter set into the kinematics verification module, trigger condition parameter feedback correction when the axial acceleration peak-to-valley value exceeds the material deformation threshold, and generate the initial parameter set that meets the physical constraints.

[0016] In one specific embodiment, based on the contact pressure distribution model of the brush disc and the cleaning surface, the maximum acceleration threshold of linear reciprocating motion is first set as the critical value at which the material does not undergo plastic deformation. By traversing the contact stress distribution curves under different friction coefficients, the acceleration upper limit is determined as two-thirds of the yield strength of the brush disc material. For the spiral motion condition, the dynamic variation interval of the spiral curve radius is derived according to the coverage requirement of the cleaning path, and the minimum separation distance of adjacent spiral trajectories is calculated using the geometric envelope method, thereby generating a curvature radius parameter set proportional to the brush disc diameter. According to the peak power characteristic curve of the motor, the dynamic response upper limit of the torque input is set in the driving system simulation module, and the allowable fluctuation range of the driving frequency is determined by calculating the inertia matching degree of the motor rotor inertia and the load torque. After importing the assembly model into the multi-body dynamics solver, the Runge-Kutta algorithm is used to simultaneously advance the simulation process of the two conditions, and the six-degree-of-freedom coordinates of the centroid position are collected at each time step. The high-frequency component of the axial acceleration is separated from the angular velocity signal through a differential amplification algorithm. In the motion phase analysis stage, the sliding window extreme value detection method is used for the centroid displacement data of the linear motion condition to extract the maximum positive deviation point and the reverse turning point in the X-axis direction as characteristic markers. For the angular velocity data of the spiral motion condition, a cubic spline interpolation curve is constructed to calculate the change gradient value of the Z-axis angular displacement per unit time. During the verification of the feature parameter set, a mapping relationship between the axial acceleration peak-to-valley value and the material elastic modulus is established. When the acceleration oscillation amplitude exceeds the safety threshold corresponding to the material relaxation modulus, the reverse propagation correction algorithm is triggered: the driving torque parameters for that period are recalculated, the weighted average method is used for smooth transition processing of the abnormal interval motion condition parameters, and the energy non-conservation error caused by data mutation is eliminated through an iterative compensation algorithm.

[0017] Step S2: Based on the standardized motion parameter data, multi-physical field force analysis of the cleaning robot in each motion state is performed to obtain comprehensive force data including gravity, friction, and air resistance; the comprehensive force data is decomposed into force component data; For example, step S2 includes the following steps: Based on the standardized motion parameter data, a force analysis model of the cleaning robot in a multi-physical field coupling environment is established, and the attitude parameters, center of mass position, and velocity direction of the robot in different motion states are input to obtain initial parameters of the force analysis model; It should be noted that the logic for obtaining the initial parameters of the force analysis model is as follows: Based on the standardized motion parameter data, the center of mass position change, attitude angle, and velocity direction of the cleaning robot in different motion states are segmented and extracted to obtain motion attitude basic data; According to the motion attitude basic data, the geometric constraint relationship of the cleaning robot in three-dimensional space is structurally mapped to construct a three-dimensional motion constraint model including a driving unit, a brush disc assembly, and a connecting mechanism to obtain three-dimensional constraint structure data; The three-dimensional constraint structure data and the standardized motion parameter data are matched and input to associate and calibrate the velocity and acceleration of the robot in different attitudes to obtain motion state matching parameters; Based on the motion state matching parameters, the action conditions of the robot in the multi-physical field are physically bounded to obtain multi-physical field coupling boundary data; the multi-physical field includes a gravity field, an air resistance field, and a contact friction field; The multi-physical field coupling boundary data is input into a dynamics modeling module to model and calculate the force characteristics in different motion states to obtain preliminary force analysis model data; specifically including the following steps: Based on the multi-physical field coupling boundary data, the force distribution of the cleaning robot in the gravity field, the friction field, and the air resistance field is divided into sub-fields to obtain sub-field force data under the independent action of each physical field; Each sub-field force data is analyzed by sub-field superposition, and the weight coefficients are set according to the actual influence degree of different physical fields in the robot running process, and the weighted sum of each sub-field force data is obtained to obtain force superposition weight data; Based on the force superposition weight data, the mechanical equilibrium relationship of the robot in different motion states is calculated to verify the stability of the force synthesis and motion state matching in the model to obtain preliminary force analysis model data.

[0018] The preliminary force analysis model data is verified for parameter convergence and stability adjustment to remove unreasonable constraints or abnormal stress points to obtain the final initial parameters of the force analysis model.

[0019] According to the initial parameters of the force analysis model, the gravity, friction and wind resistance of the robot in different states are dynamically solved to obtain the variation trend of each force in the time sequence, and preliminary comprehensive force data is obtained; It should be added that the logic for obtaining preliminary comprehensive force data is: Based on the initial parameters of the force analysis model, the attitude angle, mass center position and speed direction of the cleaning robot in different motion states are input in time sequence to obtain motion state input data; According to the motion state input data, the force distribution of the cleaning robot in the gravitational field is calculated, and the gravity in different attitudes is decomposed into vectors to obtain gravity component data; Based on the gravity component data, the friction characteristics of the contact area between the robot and the ground and the water surface are analyzed, and the contact area and the material friction coefficient are combined to obtain friction change data; Combine the friction change data with the robot speed direction and surface morphology information to dynamically estimate the air resistance and obtain wind resistance change data; Align and synchronize the time sequence of the gravity component data, friction change data and wind resistance change data to ensure the correspondence of the three types of forces under the same time reference, and obtain force time matching data; Based on the force time matching data, the forces of each type are added and summed at each time, the total force change sequence under different states is obtained, and the change trend characteristics are extracted, and finally the preliminary comprehensive force data is obtained.

[0020] It should be added that based on the force time matching data, the forces of each type are added and summed at each time, the total force change sequence under different states is obtained, and the change trend characteristics are extracted, and finally the preliminary comprehensive force data is obtained, which also includes: Calibrate the time reference of the force time matching data, synchronize the gravity component data, friction change data and wind resistance change data according to the uniform time interval, ensure that each force data has a corresponding relationship at the same time node, and obtain force time synchronization data; Based on the force time synchronization data, the gravity, friction and wind resistance at each time node are vectorally added and summed to calculate the magnitude and direction of the resultant force at each time, and time sequence force superposition data is obtained; Trend feature extraction is performed on the time sequence force superposition data, the change amplitude of the resultant force in the continuous time period is smoothed, the dominant trend and periodic fluctuation characteristics of the force change are identified, and finally the preliminary comprehensive force data is obtained.

[0021] The preliminary comprehensive stress data is subjected to vector superposition and time domain smoothing processing to remove numerical noise and transient interference generated in the calculation, and stable comprehensive stress data is obtained; the specific operation logic is as follows: The direction component of the preliminary comprehensive stress data is extracted, the gravity, friction and wind resistance at each time are subjected to vector decomposition according to the coordinate axis direction, and the stress direction component data is obtained; based on the stress direction component data, the vectors of each component at the same time point are superposed, and the vectors of different stress components in the spatial coordinate system are synthesized, and stress superposition data is obtained; the stress superposition data is subjected to time domain sampling analysis, and the fluctuation section and abnormal peak value of the stress changing with time are identified, and stress fluctuation section data is obtained; based on the stress fluctuation section data, the mutation points and transient interference signals in a short time are smoothed and corrected, and the high-frequency numerical noise is eliminated by using moving average or weighted average, and the stress data after smoothing is obtained; the stress data after smoothing is subjected to trend consistency verification, and whether the stress direction components after superposition remain consistent in physical dimension and time sequence continuity is checked, and stable comprehensive stress data is obtained.

[0022] Based on the comprehensive stress data, the stress directionality of the cleaning robot is decomposed, the comprehensive stress is divided into components according to the robot coordinate system, and stress component data is obtained; including: The stable comprehensive stress data is subjected to coordinate system initialization, the corresponding relationship between the body coordinate system and the ground reference coordinate system is established according to the structural characteristics of the cleaning robot, and coordinate system matching data is obtained; Based on the coordinate system matching data, the coordinate system of each stress vector in the comprehensive stress data is converted, the external global stress information is mapped to the body coordinate system of the cleaning robot, and stress direction mapping data is obtained; The stress direction mapping data is subjected to direction angle analysis, the distribution proportion of each stress in the X, Y and Z three-axis directions of the robot body coordinate system is calculated, and stress direction distribution data is obtained; Based on the stress direction distribution data, the components of each direction stress are extracted, and the total stress is decomposed according to the direction proportion, and preliminary stress component data is obtained; The preliminary stress component data is subjected to consistency verification, and whether the difference between the total vector after synthesizing each component and the comprehensive stress data is within the allowable error range is checked, and if the difference exceeds the threshold value, the proportion is adjusted, and the corrected stress component data is obtained; Based on the corrected stress component data, the distribution balance of the stress direction is analyzed, and whether the stress state of the cleaning robot in each direction is stable is judged, and finally the stress component data is obtained.

[0023] It needs to be added that the consistency test is carried out on the preliminary force component data, and the difference between the total vector after the synthesis of each component and the comprehensive force data is checked whether it is within the allowable error range. If the difference exceeds the threshold, the proportion is adjusted, and the corrected force component data is obtained; it also includes: The vector synthesis test is carried out on the preliminary force component data, and the force components in X, Y, Z directions are vector superimposed to calculate the total force size and direction after synthesis, and the force synthesis verification data is obtained; Based on the force synthesis verification data, the difference between the synthesized total force and the comprehensive force data is compared and analyzed, the relative error rate of the two is calculated, and compared with the preset allowable error threshold, the force error analysis data is obtained; When there are components in the force error analysis data that exceed the threshold, the proportion of the corresponding direction component is corrected according to the error direction and amplitude, the component values are adjusted under the premise of keeping the overall vector direction unchanged, and finally the corrected force component data is obtained.

[0024] The primary and secondary feature analysis is carried out on the force component data, the average intensity and change rate of different direction components are compared, and the force item that plays a leading role in the robot motion is selected to obtain the main force component data; it specifically includes the following steps: The direction classification is carried out on the force component data, and the direction components are grouped according to the X, Y, Z three axes of the body coordinate system to obtain the force direction grouping data; Based on the force direction grouping data, the force values of each direction component in the time sequence are statistically analyzed, and the average intensity value of each direction component is calculated to obtain the force average intensity data; According to the force average intensity data, the fluctuation amplitude of the force value of each direction component with time is calculated by difference, and the force change rate data is obtained; The force average intensity data and the force change rate data are compared and analyzed, and the contribution of each direction component in the overall motion process is calculated by weighted evaluation method to obtain the force contribution degree data; Based on the force contribution degree data, the direction components are sorted and selected, and the main force item with contribution degree greater than the preset contribution degree threshold and change rate in the stable interval is extracted to obtain the preliminary main force data; The consistency review is carried out on the preliminary main force data, and the difference between the superposition result and the original force component data in time and direction is verified whether it is within the allowable error range. If it exceeds, the proportion is adjusted, and finally the main force component data is obtained.

[0025] The force average intensity data and the force change rate data are compared and analyzed, and the contribution of each direction component in the overall motion process is calculated by weighted evaluation method to obtain the force contribution degree data, which also specifically includes the following steps: The stress intensity data is normalized, and the average intensity values of different direction components are standardized according to the maximum intensity ratio, so that the direction components in the same numerical interval are comparable, and stress intensity normalized data is obtained; Based on the stress intensity normalized data, the stress rate of change data is analyzed by weight distribution, the weight coefficients of each direction component are determined according to the stability and fluctuation amplitude of the change rate, and the change rate weight distribution data is obtained; The stress intensity normalized data and the change rate weight distribution data are superimposed by weight, the contribution degree of each direction component in the overall stress is calculated by weighted average, and the stress contribution degree data is obtained.

[0026] Step S3: According to the main stress component data, a time sequence stress curve is drawn, and the key components and dangerous working condition data whose stress exceeds the preset stress threshold are identified through the peak value of the curve; The dangerous working condition data is evaluated, and high-risk working condition data is obtained. Illustratively, step S3 includes: Step S31: Based on the main stress component data, the stress change of the cleaning robot at different time nodes is time-sequentially arranged, and the stress components in each direction are collected in time sequence, to obtain time-sequentially arranged stress data; Step S32: According to the time-sequentially arranged stress data, the stress components in each direction are curve-drawn and smoothed to generate a time-sequentially stress curve, and the local peak points appearing in the curve are extracted to obtain stress peak value data; Step S33: Based on the stress peak value data, the robot structure state in the time period where the peak value is located is matched and analyzed to identify the stress concentration area and the corresponding components, and key stress component data is obtained; Step S34: According to the key stress component data, the working environment parameters at each peak time are comprehensively associated and analyzed, wherein the working environment parameters include motion speed, attitude angle and contact friction condition, and dangerous working condition data is obtained; Step S35: The dangerous working condition data is evaluated, and the risk level of the working condition is calculated according to the stress amplitude, duration and repeated appearance frequency, and finally high-risk working condition data is obtained.

[0027] Illustratively, step S31 includes the following steps: Step S311: The time stamp of the main stress component data is extracted, the sampling time corresponding to each stress component is identified, and the stress data in each direction is matched with the corresponding time information to obtain stress time matching data; Step S312: Based on the stress time matching data, the stress components in different directions (X, Y, Z) are time-sequentially sorted, and the component values at different time are arranged in time sequence to obtain time sequence sorting data; Step S313: integrity check on time series sorting data, check whether there is missing or abnormal mutation in the stress data in the continuous time period, and interpolate and correct the abnormal points to obtain the corrected stress time series data; specifically including the following steps: The continuity of the time series sorting data is detected, the stress change amplitude of adjacent time nodes is calculated, and the time section with mutation or missing value is identified to obtain stress anomaly identification data; based on the stress anomaly identification data, the interpolation repair is performed on the missing section, the stress value is completed by linear interpolation or smooth mean value according to the stress change trend of the adjacent time nodes, and the stress interpolation repair data is obtained; the stress interpolation repair data is subjected to smoothing consistency check, the average gradient of stress change in the continuous time period is calculated by using sliding window method, the high frequency fluctuation caused by interpolation is removed, and finally the corrected stress time series data is obtained. Based on the corrected stress time series data, the time series relationship of each direction component is uniformly collected to ensure that all stress components have corresponding values at the same time node, and finally the time series stress arrangement data is obtained.

[0028] Illustratively, step S32 includes the following steps: Step S321: direction grouping processing is performed on the time series stress arrangement data, the stress components of each direction (X, Y, Z) are extracted in time sequence respectively, the directional stress sequence is formed, and the stress direction sequence data is obtained; Step S322: based on the stress direction sequence data, the curve drawing and smoothing filtering processing is performed on each direction stress sequence, the time sliding average is adopted to reduce the influence of mutation points, the time series stress curve with good continuity is generated, and the smooth stress curve data is obtained; Step S323: according to the smooth stress curve data, local extreme value analysis is performed on the curve change trend, the peak value point information is extracted by identifying the change interval in which the stress rises to the extreme value and then falls in the curve, and finally the stress peak value data is obtained.

[0029] Illustratively, step S33 includes the following steps: Step S331: time period extraction is performed on the stress peak value data, the start and end time range of the peak value is determined, and the complete stress record of the corresponding time period is intercepted from the time series stress curve to obtain the peak time period stress data; Step S332: based on the peak time period stress data, the motion posture parameters and structure state information of the cleaning robot in the time period are matched, including the attitude angle, the connection node displacement and the contact support state, and the structure state matching data is obtained; Step S333: According to the structure state matching data, the stress distribution of each structural unit in the peak time period is compared and analyzed, the stress concentration area is identified, and the key structural unit with a stress value significantly higher than the average level is determined, and the stress concentration area data is obtained; wherein the refinement steps are as follows: Step S3331: The structure state matching data is regionally divided, the overall structure of the robot is spatially partitioned according to joints, connecting rods and shell units, and structure partition data is obtained; wherein the structure partition data includes but is not limited to partition unique identifier, spatial coordinate set and adjacency relationship set; Step S3332: Based on the structure partition data, the stress values of the nodes in each partition are statistically calculated, the average stress and the maximum stress of each partition are extracted, and partition stress characteristic data is obtained; wherein the partition stress characteristic data includes but is not limited to partition average stress value, partition maximum stress value and peak position coordinate; Step S3333: According to the partition stress characteristic data, the stress distribution of all partitions is compared and analyzed, the ratio of the maximum stress of each partition to the overall average stress is calculated, and when the ratio exceeds the preset threshold, it is marked as a stress abnormal area, and stress abnormal partition data is obtained; wherein the stress abnormal partition data includes but is not limited to abnormal partition identifier, stress ratio and threshold determination result; Step S3334: Based on the stress abnormal partition data, adjacent stress abnormal areas are aggregated and analyzed, a region group with a stress continuity greater than a preset continuity threshold and a spatial distance less than a preset distance threshold is identified, a final stress concentration area is determined, and stress concentration area data is output, including but not limited to stress concentration area group ID, continuity score, region aggregated stress peak, spatial center coordinate and boundary range.

[0030] Step S334: Based on the stress concentration area data, the key stress components are identified, the stress area is matched with the specific components according to the robot structure topology relationship, and finally the key stress component data is obtained.

[0031] For example, step S34 includes the following steps: Step S341: The peak time of the key stress component data is extracted, the time node corresponding to each key component when the stress peak appears is identified, and the stress and state data at the corresponding time from the time series stress curve are extracted, which is recorded as peak time matching data; Step S342: Based on the peak time matching data, the motion speed, attitude angle and contact friction condition in this period are synchronously collected and associated analyzed, a one-to-one correspondence between the key stress component and the working environment parameter is established, and working condition association data is obtained; Step S343: According to the working condition associated data, the stress performance of each key stress component under different environmental parameters is coupled and evaluated, the abnormal stress condition under the conditions of excessive speed, attitude angle deviation or friction coefficient mutation is judged, the potential dangerous working condition interval is identified, and finally the dangerous working condition data is obtained.

[0032] Step S343 includes the following steps: Step S3431: Parameter threshold setting is performed on the working condition associated data, the speed upper threshold value, the attitude angle deviation threshold value and the friction coefficient fluctuation threshold value are determined according to the rated working range and the structural safety factor of the key stress component, and the working condition threshold reference data is obtained. Step S3432: Based on the working condition threshold reference data, the working condition associated data is compared and analyzed at each time point, the time point exceeding any threshold value is identified, and the corresponding stress data and environmental parameters are extracted, and the working condition abnormal judgment data is obtained. Step S3433: According to the working condition abnormal judgment data, interval aggregation analysis is performed on the continuous or adjacent abnormal time points, the abnormal section with continuous time or consistent stress change trend is divided into a potential dangerous working condition interval, and the working condition interval aggregation data is obtained. Step S3434: Based on the working condition interval aggregation data, the stress intensity, duration and occurrence frequency of each dangerous interval are comprehensively evaluated, the high-risk interval with larger stress amplitude and longer duration is selected, and finally the dangerous working condition data is output.

[0033] For example, step S35 includes the following steps: Step S351: Parameter extraction is performed on the dangerous working condition data, the stress peak value, stress duration and working condition repeated occurrence frequency of each dangerous working condition are obtained, and the working condition parameter extraction data is obtained. Step S352: Based on the working condition parameter extraction data, the stress amplitude and duration of different dangerous working conditions are weighted and analyzed, the working condition intensity index is calculated according to the comprehensive degree of stress peak value and duration in the working condition, and the working condition intensity evaluation data is obtained. Step S353: According to the working condition intensity evaluation data, the repeated occurrence frequency of each working condition is associated and corrected, when a working condition repeatedly occurs in multiple working cycles, the risk weight of the working condition is improved, and the working condition risk weight data is obtained. Step S354: Based on the working condition risk weight data, the risk level of each dangerous working condition is divided, the working condition level distribution is determined according to the comprehensive value of intensity index and risk weight, and the high-level result is marked as a high-risk working condition, and finally the high-risk working condition data is obtained.

[0034] Step S354 includes the following steps: Step S3541: Threshold setting is performed on the working condition risk weight data. According to the distribution range of the working condition intensity index and the historical risk statistical result, the threshold boundary of the low-risk, medium-risk and high-risk grades is determined, and risk grade threshold data is obtained. Step S3542: Based on the risk grade threshold data, the intensity index and the risk weight of each dangerous working condition are comprehensively calculated, the working condition comprehensive risk value is calculated according to the weighted average principle, and working condition comprehensive risk data is obtained. Step S3543: According to the working condition comprehensive risk data, the risk value and the threshold boundary of each dangerous working condition are compared in the interval, the risk grade interval to which it belongs is determined, and risk grade distribution data is obtained. Step S3544: Based on the risk grade distribution data, the working conditions in the high-grade interval are screened and confirmed, the working conditions with high risk grade are identified and output, and finally high-risk working condition data is obtained.

[0035] In the embodiment of the present application, the system analyzes the collected main force component data in the time dimension, extracts the sampling time stamp corresponding to each group of force vector, and establishes an accurate mapping relationship between the force value and the time information. Subsequently, according to the logical order of time, the force components in X, Y and Z three-axis directions are sorted in full sequence. In this process, the system will deeply check the integrity of the data, identify the breakpoints or abnormal mutations in the data stream by calculating the change gradient of the force value of adjacent time nodes; for the identified missing sections, the system uses the force change trend of the adjacent time points before and after to complete the missing force value by using linear interpolation or smooth mean calculation method, and generates force interpolation repair data. Then, the sliding window algorithm is used to check the smooth consistency of the repaired data, and the average gradient in the window is calculated to eliminate the high-frequency fluctuations caused by interpolation or sensor noise, so as to output the corrected force time series data, and finally collect it as time series force arrangement data; the system decouples the time series force arrangement data in the spatial dimension, and extracts the independent force sequence of X, Y and Z three directions respectively. For each sequence, a time moving average filtering algorithm is used to process, and the arithmetic mean value in a certain step is calculated to suppress random noise, and a smooth and continuous time series force curve is generated. On this basis, the system performs extreme value analysis on the geometric shape of the curve, accurately captures the local force peak value by identifying the inflection point of the curve slope from positive to negative or the point whose value is significantly higher than the adjacent baseline, extracts the time and amplitude information of these peak points, and generates force peak value data. Further, the system extracts the complete force record in the preset time window before and after the time when the peak value appears, and synchronously calls the motion posture parameters, connection node displacement and contact support state and other structure information of the robot in this time period. In order to accurately identify the specific physical position of the force concentration, the system performs detailed spatial region analysis: the whole structure of the robot is divided into several spatial partitions with unique identifiers, including joints, connecting rods and shell units, and an adjacency set is constructed based on the topological relationship. The system calculates the average force value and the maximum force value of each partition. Subsequently, by comparative analysis, the ratio of the maximum force value of each partition to the average force value of the whole structure is calculated, and when the ratio exceeds the preset abnormal judgment threshold, the partition is marked as a force abnormal area. Further, the system uses region growing or clustering analysis logic to aggregate the abnormal partitions adjacent in space and with force continuity score higher than the preset standard, and identifies the force concentration area group. Finally, combined with the structure topology graph of the robot, the high force group is mapped to the specific physical components, and the key force component data is output.

[0036] The system extracts the specific time when the key force component appears force peak value, and correlates the working environment parameters at that time, including the motion speed of the cleaning robot, the attitude tilt angle and the friction condition of the contact surface. By establishing a coupling model of component force and environmental parameters, the system sets a safety threshold benchmark based on the rated working range of the component. The system compares the real-time parameters with the benchmark threshold at each time, and determines that the working condition is abnormal once the motion speed is out of limit, the attitude angle is too large or the friction coefficient is suddenly changed. Then, the system aggregates the continuous or adjacent abnormal points on the time axis, and combines the time period with consistent force trend and continuously abnormal environmental parameters into a potential dangerous working condition interval. The system comprehensively considers the force intensity and duration of these intervals, and selects those intervals with large force amplitude and long duration time, and confirms them as dangerous working condition data.

[0037] Finally, the core feature parameters are extracted from the dangerous working condition data, including the force peak value, the abnormal force duration and the frequency of repeated occurrence of the working condition in the historical period. Based on these parameters, the system constructs a working condition intensity evaluation model: using weighted summation, different weight coefficients are given to the force peak value and the duration time respectively, and the intensity index reflecting the severity of a single working condition is calculated. At the same time, a frequency correction factor is introduced, and for the same working condition that repeatedly occurs in multiple working periods, the risk weight is increased. Then, according to the pre-set low, medium and high risk level thresholds, the system compares the calculated working condition comprehensive risk value (weighted result of intensity index and risk weight) with the threshold boundary. Finally, the system selects the working condition falling in the high risk level interval, marks and outputs it as high-risk working condition data.

[0038] Step S4: For the high-risk working condition data, transient force analysis of the cleaning robot in the downward cleaning and return upward states is performed, and instability sliding and overturning risk simulation is performed based on wind resistance and dirt accumulation factors to obtain instability risk data; risk level data is obtained by dividing the risk level of the instability risk data; Illustratively, step S4 includes: Step S41: Based on the high-risk working condition data, the force state of the cleaning robot in the downward cleaning and return upward states is divided, and the key force parameters in each state are extracted, including the gravity component, the adsorption force, the driving force and the resistance, to obtain state force decomposition data; Step S42: According to the state force decomposition data, the time series calculation of transient force change is performed to obtain the change trend and peak response of the force in a short time, and transient force analysis data is obtained; Step S43: Based on the transient force analysis data, the wind resistance coefficient and the surface dirt accumulation coefficient in the external environmental factors are combined to simulate the disturbance of the balance state of the robot under different force conditions, calculate the sliding trend and the gravity center offset of the robot on the inclined surface, and obtain the instability risk simulation data; Step S44: According to the instability risk simulation data, the sliding and overturning risks are comprehensively judged, the risk level is divided in combination with the corresponding relationship between the instability probability and the overturning angle, and finally the risk level data is obtained.

[0039] For example, step S41 includes the following steps: Step S411: State recognition analysis is performed on the high-risk working condition data, and according to the cleaning robot motion direction and attitude angle parameters, the running state is divided into downward cleaning state and return upward state, and state division data is obtained; Step S412: Based on the state division data, the force composition under each state is analyzed, and the gravity direction component of the robot on the working surface, the adsorption force generated by the adsorption device, the driving force generated by the driving device, and the resistance generated by the surface friction are extracted, and preliminary force parameter data is obtained; Step S413: According to the preliminary force parameter data, the vector decomposition and direction normalization of each direction component are performed to ensure that the gravity component, the adsorption force, the driving force and the resistance have consistent direction definition in the unified coordinate system, and standardized force component data is obtained; Step S414: Based on the standardized force component data, the force distribution under the two working states is compared and analyzed to identify the main force change trend and the difference area, and finally state force decomposition data is obtained.

[0040] For example, step S42 includes the following steps: Step S421: Time series division is performed on the state force decomposition data, and the force data is segmented according to the sampling time interval, the start and end time of the force change in each segment is determined, and time series segmentation data is obtained; Step S422: Based on the time series segmentation data, the continuous difference calculation is performed on the force values of each segment, the force change rate in a short time is extracted, the rising and falling trend of the force is obtained, and force change trend data is obtained; Step S423: According to the force change trend data, the local peak value of the force fluctuation in each time period is identified, the maximum force point and its corresponding time in the change process are extracted, and transient peak value response data is obtained; Step S424: Based on the transient peak value response data, the fluctuation amplitude, peak density and change rate of the overall time series force are comprehensively evaluated, the short-time dynamic force characteristic description is formed, and finally the transient force analysis data is obtained.

[0041] Exemplarily, step S43 comprises the following steps: Step S431: external factor matching is performed on the transient force analysis data, environmental parameters corresponding to the force period are extracted, including the wind resistance coefficient and the surface dirt accumulation coefficient, a force-environment correspondence relationship is formed, and environment matching data is obtained; Step S432: based on the environment matching data, the overall balance moment of the cleaning robot under different force states is calculated, the effects of gravity, adsorption force and wind resistance are comprehensively considered, the moment balance deviation under each state is evaluated, and balance deviation data is obtained; Step S433: according to the balance deviation data, the force disturbance of the robot on the inclined surface is dynamically simulated, the sliding trend and attitude deflection response caused by external disturbance are analyzed, and sliding trend simulation data is obtained; Step S434: based on the sliding trend simulation data, the overall center of gravity trajectory of the robot is calculated, the deflection amount and direction change of the center of gravity in the disturbance process are obtained, and finally instability risk simulation data is obtained.

[0042] Exemplarily, step S44 comprises the following steps: Step S441: instability risk simulation data is extracted, the sliding trend and sliding distance change of the robot under different force states are analyzed, the sliding occurrence probability per unit time is calculated, and sliding risk data is obtained; Step S442: based on the sliding risk data, the overturning angle change of the robot in the disturbance process is statistically analyzed, the corresponding relationship between the inclination angle and the center of gravity offset at each time is determined, and overturning angle feature data is obtained; Step S443: according to the overturning angle feature data and the sliding risk data, the coupling effect of the two risks is comprehensively judged, the instability probability distribution is calculated, and comprehensive instability probability data is obtained; Step S444: based on the comprehensive instability probability data, combined with the grade interval threshold of the overturning angle, different risk levels are divided, the comprehensive probability value is matched with the angle deflection range, and finally risk grade data is obtained.

[0043] In this embodiment of the invention, based on the motion direction vector and the body posture angle, the system accurately identifies whether the robot is in a "downward cleaning" or "returning upward" working state, and constructs a corresponding force model accordingly. In this model, the system performs projection decomposition on the gravity vector, calculating its components in the normal and tangential directions of the slope; simultaneously, it combines the vacuum feedback of the adsorption device and the torque parameters of the drive motor to quantify the adsorption force and driving force; and calculates the basic frictional resistance based on the material properties of the contact surface. All mechanical components are mapped to a unified local coordinate system of the body for directional normalization, thereby generating standardized state force decomposition data, and performing transient time-series analysis on the above force data. The system segments the continuous force data stream according to the sampling frequency, uses a differential algorithm to calculate the force increment between adjacent time slices, and derives the rate of change of force over time, i.e., the trend of the first derivative of the force. By scanning these rate of change curves, the system locks the local extreme points of force fluctuations and identifies the impact load or sudden peak value that appears at a specific instant. By comprehensively evaluating the amplitude of fluctuations, the density of peaks, and the steepness of changes, the system generates transient force analysis data. Based on this, the system retrieves the drag coefficient (simulating wind load during high-altitude operations) and surface dirt accumulation coefficient (simulating the attenuation of the friction coefficient) matched to the current working conditions, and inputs these parameters as perturbation variables into the mechanical equilibrium equation. The system calculates the moment balance state of the robot under the combined action of gravity, adhesion force, wind load, and dynamic friction, focusing on evaluating the deviation between the overturning moment and the anti-overturning moment around the fuselage contact fulcrum. Simultaneously, it simulates the robot's sliding trend along the inclined plane when dirt reduces friction, and uses a kinematic integral algorithm to track the spatial trajectory offset of the fuselage's center of gravity during the perturbation process, thereby quantifying the instability risk simulation data.

[0044] Subsequently, the system statistically analyzes the frequency and magnitude of slippage during the simulation, calculating the probability of slippage per unit time. Simultaneously, it analyzes the change in overturning angle caused by center of gravity shift, establishing a mapping relationship between the tilt angle and the instability critical value. The system employs multi-dimensional weighted logic to couple the slippage probability with the overturning angle characteristics, calculating a comprehensive instability probability value. Finally, based on a preset risk level threshold matrix, this probability value is mapped to the corresponding high, medium, and low risk ranges, outputting the final risk level data.

[0045] Step S5: Based on the risk level data and combined with stress-deformation cloud map analysis, determine the locations of maximum stress and deformation in the main frame assembly and roller support assembly of the cleaning robot.

[0046] For example, step S5 includes the following steps: Step S51: Perform working condition matching on the risk level data. Based on the load conditions corresponding to high-risk, medium-risk and low-risk levels, select representative working conditions and import the finite element analysis results to obtain working condition simulation matching data. Step S52: Based on the working condition simulation matching data, the stress nephogram and the deformation nephogram of the mainframe assembly and the roller support assembly are regionally divided, the main stress region and the deformation concentrated region are identified, and structure region identification data is obtained; Step S53: According to the structure region identification data, the stress value and the displacement amount of each region are peak value searched, the maximum stress point and the maximum deformation point and their position coordinates in each component are extracted, and peak value extraction data is obtained; Step S54: Based on the peak value extraction data, the stress cause of the peak value region is analyzed in combination with the risk level data, the stress and deformation characteristics caused by load concentration, structure weakness or boundary constraint are judged, and finally maximum stress and deformation position identification data is obtained.

[0047] Step S52 includes the following steps: Step S521: The stress nephogram and the deformation nephogram in the working condition simulation matching data are imported and preprocessed, the coordinate system and the scale parameter are unified, the model resolution difference is eliminated, and cloud atlas standardization data is obtained; Step S522: Based on the cloud atlas standardization data, the grid nodes of the mainframe assembly and the roller support assembly are regionally marked, and the initial regional division is performed according to the structure connection characteristics and the geometric boundary, and initial regional division data is obtained; Step S523: According to the initial regional division data, the stress value and the displacement value of each regional node are statistically analyzed, the average stress, the maximum stress and the average deformation amount indicators are extracted, and regional stress and deformation characteristic data is obtained; Step S524: Based on the regional stress and deformation characteristic data, the stress gradient and the deformation gradient of adjacent regions are compared, the stress concentration area and the deformation concentration area are identified, and concentration characteristic identification data is obtained; Step S525: According to the concentration characteristic identification data, the main stress region and the deformation concentrated region are identified and numbered, the structure region identification result is generated, and finally the structure region identification data is obtained.

[0048] Step S53 includes the following steps: Step S531: The stress distribution and the deformation distribution in the structure region identification data are read and layered processed, the regional data of the mainframe assembly and the roller support assembly are respectively imported into the analysis module, and regional layered data is obtained; Step S532: Based on the regional layered data, the node stress value and the node displacement amount of each region are numerically scanned, and abnormal nodes and calculation errors are screened out, and effective node data is obtained; Step S533: According to the effective node data, the stress value and the displacement amount of each region are locally extreme value detected, the stress peak point and the deformation peak point of each region are determined, and regional peak value identification data is obtained; Step S534: Based on the regional peak value identification data, a global comparison is performed on each regional peak point, the global maximum points of stress value and deformation amount in the mainframe assembly and the roller support assembly are screened out, and the corresponding spatial position coordinates are recorded to obtain global peak value data; Step S535: According to the global peak value data, the numerical accuracy of the maximum stress point and the maximum deformation point is verified and consistency comparison is performed to ensure that the peak value extraction is accurate and reliable, and finally peak value extraction data is obtained.

[0049] Step S54 includes the following steps: Step S541: The peak value extraction data is regionally positioned, the maximum stress point and the maximum deformation point are respectively mapped to the structure coordinates of the mainframe assembly and the roller support assembly, and peak value region positioning data is obtained; Step S542: Based on the peak value region positioning data, the local stress distribution of the peak value region is refined and analyzed, the stress gradient, displacement gradient and direction vector of the nodes around the peak value region are extracted, and local stress distribution data is obtained; Step S543: According to the local stress distribution data, in combination with the load working condition information in the risk level data, the load concentration degree of the peak value region is determined, the stress concentration characteristics caused by load concentration are identified, and load concentration determination data is obtained; Step S544: Based on the load concentration determination data, the structure geometric characteristics and constraint boundary conditions of the peak value region are compared, and whether there is stress and deformation abnormality caused by structural section mutation or constraint rigidity difference is analyzed, and structure constraint analysis data is obtained; Step S545: According to the structure constraint analysis data, the load concentration and structure constraint factors are comprehensively judged, the main stress cause of the peak value region is determined, and the appearance position of the maximum stress and the maximum deformation is output, and finally maximum stress and deformation position identification data is obtained.

[0050] Step S543 includes the following steps: Step S5431: Load information extraction is performed on the local stress distribution data, according to the load working condition input in the risk level data, the load distribution value and direction component of each node in the peak value region are extracted, and load distribution extraction data is obtained; Step S5432: Based on the load distribution extraction data, the load intensity in the peak value region is compared in space, the load difference and change gradient between adjacent nodes are calculated, and load change gradient data is obtained; Step S5433: According to the load change gradient data, the load concentration trend is identified, the mutation region appearing in the load distribution is determined by the concentration coefficient, and the comparison analysis with the stress gradient characteristics is performed, and load concentration trend data is obtained; Step S5434: Based on the load concentration trend data, classify the load concentration level in the peak region, determine whether load concentration is the main factor causing the peak stress, and finally obtain the load concentration judgment data.

[0051] Example 2 like Figure 2 As shown, a structural simulation and optimization system for a wind turbine tower cleaning robot includes: Parameter acquisition module: Constructs an assembly model of the cleaning robot using 3D modeling software, and performs multibody dynamics simulation on the cleaning robot to obtain the motion trajectory, velocity, and acceleration parameter data of the cleaning robot's center of mass and brush plate under various motion states; performs data cleaning and normalization on the motion trajectory, velocity, and acceleration parameter data to obtain standardized motion parameter data; Force decomposition module: Based on standardized motion parameter data, multi-physics force analysis is performed on the cleaning robot under various motion states to obtain comprehensive force data including gravity, friction, and wind resistance; the comprehensive force data is decomposed into force components to obtain the main force component data; Working condition assessment module: Based on the main stress component data, it plots time-series stress curves, identifies key components and dangerous working condition data whose stress exceeds the preset stress threshold by the peak value of the curve; it assesses the severity of dangerous working condition data to obtain high-risk working condition data; Risk assessment module: For high-risk working condition data, transient force analysis is performed on the cleaning robot in two states: downward cleaning and return upward. Based on wind resistance and dirt accumulation factors, instability, slippage and overturning risks are simulated to obtain instability risk data. The instability risk data is classified into risk levels to obtain risk level data. Deformation screening module: Based on risk level data and stress-deformation cloud map analysis, the module determines the locations of maximum stress and deformation in the main frame assembly and roller support assembly of the cleaning robot.

[0052] It should be noted that the interval and threshold sizes are set for ease of comparison. The size of the threshold depends on the amount of sample data and the base number set by those skilled in the art for each set of sample data, as long as it does not affect the proportional relationship between the parameter and the quantized value. Furthermore, the above formulas are all dimensionless calculations, and the formulas are derived from software simulations using a large amount of collected data to obtain the most recent real-world results. The preset parameters in the formulas are set by those skilled in the art according to the actual situation.

[0053] It should be understood that in the various embodiments of this application, the order of the above-mentioned processes does not imply the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of this application.

[0054] It should be understood that determining B from A does not mean that B is determined only from A, but also B can be determined from A and / or other information.

[0055] The above description is merely illustrative of the application and is not intended to limit the scope of the application, which is defined by the appended claims. As such, the application is not limited to that precisely set forth above and can vary other than as described in the claims and equivalents thereof.

[0056] Finally, the above description is merely the preferred embodiment of the application, and is not intended to limit the application. Any modification, equivalent replacement and improvement made within the principle and technical scope of the application shall be included in the protection scope of the application.

Claims

1. A method for structural simulation optimization of a wind power tower-cleaning robot, characterized in that, The method comprises: Step S1: constructing an assembly model of the cleaning robot by a three-dimensional modeling software, and performing multi-body dynamics simulation on the cleaning robot to obtain a center of mass of the cleaning robot and motion trajectory, velocity, and acceleration parameter data of the brush disc under multiple motion states; performing data cleaning and normalization processing on the motion trajectory, velocity, and acceleration parameter data to obtain standardized motion parameter data; Step S2: based on the standardized motion parameter data, performing multi-physical field force analysis of the cleaning robot under each motion state to obtain comprehensive force data including gravity, friction, and wind resistance; performing force component decomposition on the comprehensive force data to obtain main force component data; Step S3: according to the main force component data, drawing a time sequence force curve, identifying key components and dangerous working condition data whose force exceeds a preset force threshold through a curve peak value; performing working condition severity evaluation on the dangerous working condition data to obtain high-risk working condition data; Step S4: for the high-risk working condition data, performing transient force analysis of the cleaning robot under two states of downward cleaning and returning upward, and performing instability sliding and overturning risk simulation based on wind resistance and dirt accumulation factors to obtain instability risk data; performing risk level division on the instability risk data to obtain risk level data; Step S5: based on the risk level data, combining stress deformation cloud map analysis to determine the occurrence positions of the maximum stress and deformation in the mainframe assembly and the roller support assembly of the cleaning robot.

2. The structural simulation optimization method of a wind power tower washing robot according to claim 1, wherein, Step S1 comprises the following sub-steps: Step S11: establishing an assembly model containing driving components and a brush disc based on the assembly structure of the cleaning robot in the three-dimensional modeling software, and setting motion constraint conditions of motor driving; Step S12: applying two typical working conditions of linear reciprocating motion and spiral motion to the assembly model, performing multi-body dynamics simulation to obtain an initial parameter set of center of mass displacement trajectory, brush disc motion velocity, and axial acceleration; Step S13: based on a brush disc material deformation threshold, eliminating abnormal data nodes with axial acceleration exceeding the limit in the initial parameter set to obtain a compliant motion parameter set; Step S14: performing unit dimension unification on the compliant motion parameter set, converting the displacement trajectory into a normalized curvature coefficient, and converting the velocity into a time normalized ratio, denoted as a unit time displacement ratio, to generate standardized motion parameter data.

3. The structural simulation optimization method of a wind power tower washing robot according to claim 2, characterized in that, Step S12 specifically comprises: Step S121: setting an acceleration threshold of linear reciprocating motion and a curvature radius variation range of spiral motion according to the load distribution conditions of the brush disc contact surface to generate two sets of motion working condition parameters; Step S122: defining torque input threshold and frequency response bandwidth of the driving components under the two motion working conditions based on the upper limit of motor rated power of the assembly model; Step S123: loading the two sets of motion working condition parameters to the multi-body dynamics simulation module, performing continuous motion simulation under the constraint of motor torque threshold, and outputting an original simulation data package containing time stamp center of mass space coordinate sequence, brush disc rotation speed pulse waveform, and axial acceleration peak and valley value; Step S124: performing motion phase analysis on the original simulation data package, and extracting the center of mass X-axis displacement extreme point under linear working condition and the Z-axis angular velocity variation gradient under spiral working condition as a feature parameter set; Step S125: input the extracted feature parameter set into a kinematics verification module, trigger working condition parameter feedback correction when the axial acceleration peak-to-valley value exceeds the material deformation threshold, and generate an initial parameter set that meets the physical constraints.

4. The structural simulation optimization method of a wind power tower washing robot according to claim 1, wherein Step S2 includes: Based on the standardized motion parameter data, a force analysis model of the cleaning robot in a multi-physical field coupled environment is established, the attitude parameters, mass center positions and speed directions of the robot in different motion states are input, and initial parameters of the force analysis model are obtained; According to the initial parameters of the force analysis model, the gravity, friction and wind resistance of the robot in different states are dynamically solved, the change trend of each force in the time sequence is obtained, and preliminary comprehensive force data is obtained; The preliminary comprehensive force data is subjected to vector superposition and time domain smoothing processing to remove numerical noise and transient interference generated in the calculation, and stable comprehensive force data is obtained; Based on the comprehensive force data, the overall force directionality of the cleaning robot is decomposed, the comprehensive force is divided into components according to the robot coordinate system, and force component data is obtained; The force component data is subjected to principal and secondary feature analysis, the average intensity and change rate of different direction components are compared, and the force term that plays a leading role in the robot motion is selected, and main force component data is obtained.

5. The structural simulation optimization method of a wind power tower washing robot according to claim 4, wherein, The logic for obtaining the preliminary comprehensive force data is as follows: Based on the initial parameters of the force analysis model, the attitude angle, mass center position and speed direction of the cleaning robot in different motion states are input in time sequence, and motion state input data is obtained; According to the motion state input data, the force distribution of the cleaning robot in the gravity field is calculated, the gravity action in different attitudes is decomposed into vectors, and gravity component data is obtained; Based on the gravity component data, the friction characteristics of the contact area between the robot and the ground and the water surface are analyzed, and the friction force change data is obtained by combining the contact area and the material friction coefficient; The air resistance is dynamically estimated by combining the friction force change data with the robot speed direction and surface morphology information, and wind resistance change data is obtained; The gravity component data, friction force change data and wind resistance change data are subjected to time sequence alignment and synchronous analysis to ensure the correspondence of the three types of forces under the same time reference, and force time matching data is obtained; Based on the force time matching data, the forces of each type are added up at each time, the total force change sequence under different states is obtained, and the change trend characteristics are extracted, and finally the preliminary comprehensive force data is obtained.

6. The structural simulation optimization method of a wind power tower washing robot according to claim 4, wherein The steps for obtaining the main force component data are as follows: The force component data is classified by direction, the data of each direction component is grouped according to the X, Y and Z axes of the body coordinate system, and force direction grouping data is obtained; Based on the force direction grouping data, the force values of each direction component in the time sequence are statistically analyzed, the average intensity value of each direction component is calculated, and force average intensity data is obtained; According to the force average intensity data, the fluctuation amplitude of the force value of each direction component with time is calculated by difference, and force change rate data is obtained; The force contribution degree data is obtained by comparing and analyzing the force average intensity data and the force change rate data, and calculating the contribution degree of each direction component in the overall motion process through a weighted evaluation method; Based on the force contribution degree data, the main force items with a contribution degree greater than a preset contribution degree threshold and a change rate within a stable interval are extracted by sorting and screening each direction component, to obtain preliminary main force data; The consistency of the preliminary main force data is reviewed, and whether the difference between the superposition result and the original force component data in time and direction is within the allowable error range is verified. If it exceeds, it is proportionally corrected, and finally the main force component data is obtained.

7. The method of structural simulation optimization of a wind power tower washing robot according to claim 1, wherein, Step S3 includes: Step S31: Based on the main force component data, the force changes of the cleaning robot at different time nodes are time-sequentially arranged, and the force components in each direction are collected in time sequence to obtain time-sequential force arrangement data; Step S32: According to the time-sequential force arrangement data, the force components in each direction are curve-drawn and smoothed to generate a time-sequential force curve, and local peak points appearing in the curve are extracted to obtain force peak value data; Step S33: Based on the force peak value data, the robot structure state of the time period where the peak value is located is matched and analyzed to identify the force concentration area and the corresponding components, and key force component data is obtained; Step S34: According to the key force component data, the working environment parameters at each peak time are comprehensively associated and analyzed, wherein the working environment parameters include motion speed, attitude angle and contact friction condition, and dangerous working condition data is obtained; Step S35: The working condition severity of the dangerous working condition data is evaluated, and the risk level of the working condition is calculated according to parameters such as force amplitude, duration and repeated occurrence frequency, and finally high-risk working condition data is obtained.

8. The method of structural simulation optimization of a wind power tower washing robot according to claim 1, wherein, Step S4 includes: Step S41: Based on the high-risk working condition data, the force state of the cleaning robot in the downward cleaning and return upward states is divided, and the key force parameters in each state are extracted, including gravity component, adsorption force, driving force and resistance, to obtain state force decomposition data; Step S42: According to the state force decomposition data, the transient force change is time-sequentially calculated to obtain the change trend and peak response of the force in a short time, and transient force analysis data is obtained; Step S43: Based on the transient force analysis data, the balance state of the robot under different force conditions is simulated by combining the wind resistance coefficient and the surface dirt accumulation coefficient in the external environment factors, the sliding trend and the center of gravity offset of the robot on the inclined plane are calculated, and instability risk simulation data is obtained; Step S44: According to the instability risk simulation data, the sliding and overturning risks are comprehensively judged, and the risk level is divided according to the corresponding relationship between the instability probability and the overturning angle, and finally risk level data is obtained.

9. The method of structural simulation optimization of a wind power tower washing robot according to claim 1, wherein, Step S5 includes: Step S51: The risk level data is inputted and processed, and a cleaning robot finite element analysis model is established according to the load conditions and constraint boundaries corresponding to different risk levels, and simulation model data is obtained; Step S52: Based on the simulation model data, static force calculation is performed on the robot structure to obtain stress distribution and displacement response of the main frame assembly and the roller support assembly under steady-state load, and static simulation result data is obtained; Step S53: According to the static simulation result data, load time sequence input and transient solution are performed on the dynamic characteristics of the robot under high-risk working conditions, and dynamic simulation result data is obtained; Step S54: Based on the static simulation result data and the dynamic simulation result data, stress and deformation cloud maps are superimposed and analyzed, the concentrated parts of the maximum stress and the maximum deformation are identified, and the causes are comprehensively judged in combination with the risk level data, and finally stress and deformation analysis data are obtained. 10.A structural simulation optimization system of a wind power tower cleaning robot, based on the structural simulation optimization method of a wind power tower cleaning robot according to any one of claims 1-9, characterized in that, The system comprises: A parameter acquisition module: a three-dimensional modeling software is used to construct an assembly model of the cleaning robot, and multi-body dynamics simulation is performed on the cleaning robot to obtain the center of mass of the cleaning robot and the motion trajectory, speed and acceleration parameter data of the brush disc under various motion states; the motion trajectory, speed and acceleration parameter data are cleaned and normalized to obtain standardized motion parameter data; A stress decomposition module: based on the standardized motion parameter data, multi-physical field stress analysis of the cleaning robot under each motion state is performed to obtain comprehensive stress data including gravity, friction, wind resistance; the stress component decomposition is performed on the comprehensive stress data to obtain main stress component data; A working condition evaluation module: according to the main stress component data, a time sequence stress curve is drawn, and key components and dangerous working condition data whose stress exceeds the preset stress threshold are identified through the curve peak value; the working condition severity evaluation is performed on the dangerous working condition data to obtain high-risk working condition data; A risk discrimination module: for the high-risk working condition data, transient stress analysis of the cleaning robot under the two states of downward cleaning and return upward is performed, and instability sliding and overturning risk simulation is performed based on wind resistance and dirt accumulation factors to obtain instability risk data; the instability risk data is divided into risk levels to obtain risk level data; A deformation screening module: based on the risk level data, the stress and deformation cloud map analysis is combined to determine the positions of the maximum stress and deformation in the main frame assembly and the roller support assembly of the cleaning robot.

Images