A structural simulation optimization method and system for a wind power tower cylinder cleaning robot

Abstract

The present application relates to the field of robot structure simulation optimization, and relates to a wind power tower cylinder cleaning robot structure simulation optimization method and system, which comprises a parameter acquisition module, a stress decomposition module, a working condition evaluation module, a risk discrimination module and a deformation screening module, can effectively reduce the risk of safety accidents caused by structural instability on the one hand; on the other hand, the scientificity and accuracy of the analysis results are ensured, important theoretical basis and data support are provided for subsequent structure improvement, material optimization and control strategy design, so as to realize efficient, safe and reliable operation of the cleaning robot.

Description

Technical Field

[0001] This invention belongs to the field of robot structure simulation and optimization, and relates to a method and system for structural simulation and optimization of a wind power tower cleaning robot. Background Technology

[0002] Analyzing the operational safety and stability of cleaning robots under complex working conditions is a crucial step in ensuring their reliable operation, and simulation analysis technology plays a vital role in this process. By conducting multi-state simulation analysis on the cleaning robot, dynamic parameters such as the motion trajectory, velocity, and acceleration of the center of mass and brush disc under different motion states can be obtained, providing accurate basis for subsequent stress and stability assessments.

[0003] Currently, there are still some aspects in the simulation analysis of cleaning robots that need to be optimized, mainly in the following areas:

[0004] 1. Existing simulation analyses mostly focus on the overall motion trend, but lack detailed modeling and dynamic response analysis of the motion parameters of the center of mass and brush disk under different operating conditions. This results in deviations in trajectory, velocity, and acceleration data, making it difficult to accurately reflect the true force characteristics of the robot under complex paths and high-risk working conditions.

[0005] 2. Existing research does not take into account the forces acting on the robot during operation in a comprehensive manner. It fails to fully integrate the coupling relationship between various forces such as gravity, friction, and wind resistance, resulting in inaccurate force curve plotting and an inability to effectively identify critical components and potentially dangerous working conditions under high stress. This affects the safety assessment and structural optimization design of the cleaning robot under different working postures. Summary of the Invention

[0006] In view of the problems existing in the prior art, the present invention provides a structural simulation optimization method and system for a wind power tower cleaning robot to solve the above-mentioned technical problems.

[0007] To achieve the above and other objectives, the technical solution adopted by the present invention is as follows:

[0008] The first aspect of this invention provides a structural simulation optimization method for a wind power tower cleaning robot, the method comprising the following steps:

[0009] Step S1: Construct an assembly model of the cleaning robot using 3D modeling software, and perform 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; perform data cleaning and normalization on the motion trajectory, velocity, and acceleration parameter data to obtain standardized motion parameter data;

[0010] Step S2: Based on standardized motion parameter data, perform multi-physics force analysis on the cleaning robot under various motion states to obtain comprehensive force data including gravity, friction, and wind resistance; decompose the comprehensive force data into force components to obtain the main force component data;

[0011] Step S3: Based on the main force component data, plot the time-series force curve, and identify the key components and dangerous working conditions whose forces exceed the preset force threshold by the peak values ​​of the curves; assess the severity of the dangerous working conditions to obtain high-risk working condition data.

[0012] Step S4: For high-risk working condition data, perform transient force analysis on the cleaning robot in two states: downward cleaning and return upward. Based on wind resistance and dirt accumulation factors, simulate the risks of instability, slippage, and overturning to obtain instability risk data. Classify the instability risk data into risk levels to obtain risk level data.

[0013] 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.

[0014] A second aspect of the present invention provides a structural simulation and optimization system for a wind power tower cleaning robot, the system comprising:

[0015] 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;

[0016] 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;

[0017] 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;

[0018] 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.

[0019] 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.

[0020] As described above, the structural simulation optimization method and system for a wind power tower cleaning robot provided by the present invention has at least the following beneficial effects:

[0021] This invention provides a structural simulation optimization method and system for a wind power tower cleaning robot. By constructing a three-dimensional assembly model of the cleaning robot and performing multibody dynamics simulation, it can accurately capture parameter data such as the center of mass and brush plate motion trajectory, velocity, and acceleration of the cleaning robot under different operating conditions, thereby obtaining its realistic dynamic behavior characteristics. By cleaning and normalizing the simulation data, errors and noise can be effectively eliminated, ensuring the reliability and consistency of motion parameters and providing standardized input data for subsequent force analysis. Attached Figure Description

[0022] To more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0023] Figure 1 This is a schematic diagram showing the connections between the steps of the method of the present invention.

[0024] Figure 2 This is a schematic diagram showing the connections of the various modules in the system of the present invention. Detailed Implementation

[0025] The following description, in conjunction with the implementation of this invention, is merely an example and illustration of the concept of this invention. Those skilled in the art can make various modifications or additions to the specific embodiments described, or use similar methods to replace them, as long as they do not deviate from the inventive concept or exceed the scope defined in these claims, all of which should fall within the protection scope of this invention.

[0026] Example 1

[0027] Please see Figure 1 As shown, a structural simulation optimization method for a wind power tower cleaning robot is proposed, which includes the following steps:

[0028] Step S1: Construct an assembly model of the cleaning robot using 3D modeling software, and perform 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; perform data cleaning and normalization on the motion trajectory, velocity, and acceleration parameter data to obtain standardized motion parameter data;

[0029] For example, step S1 includes the following sub-steps:

[0030] Step S11: Based on the assembly structure of the cleaning robot, establish an assembly model including the drive components and brush in the 3D modeling software, and set the motion constraints of the motor drive.

[0031] Step S12: Apply two typical working conditions, linear reciprocating motion and helical motion, to the assembly model, perform multibody dynamics simulation, and obtain the initial parameter set of the centroid displacement trajectory, brush motion velocity, and axial acceleration.

[0032] Step S13: Based on the material deformation threshold of the brush disk, remove abnormal data nodes with excessive axial acceleration in the initial parameter set to obtain a compliant motion parameter set;

[0033] Step S14: Unify the unit dimensions of the compliant motion parameter set, convert the displacement trajectory into normalized curvature coefficients, and convert the velocity into a time normalized ratio, which is recorded as the displacement ratio per unit time, to generate standardized motion parameter data.

[0034] In a specific embodiment, when constructing the kinematic model of the cleaning robot, the three-dimensional geometric models of the drive motor, transmission shaft, and brush disk are first imported into the 3D modeling software based on the assembly drawings. According to the elastic modulus and Poisson's ratio parameters of the brush disk material in the material library, six-degree-of-freedom constraints are set at the connection between the transmission shaft and the brush disk, with the axial degree of freedom locked based on the brush disk's working pressure threshold. Kinematic joint constraints are set for the motor drive system. The joint torque threshold is set by calculating the relationship between the maximum torque of the motor shaft and the transmission ratio, and the upper and lower limits of the linear reciprocating motion displacement are determined based on the brush disk's working stroke range. During the multibody dynamics simulation execution phase, kinematic parameter sets for the linear reciprocating and helical motion conditions are loaded respectively. In the linear motion condition, the acceleration change rate is set based on the servo motor trapezoidal acceleration and deceleration control strategy. In the helical motion condition, the lateral displacement component of the brush disk is calculated based on the preset helix angle. An explicit numerical integration algorithm is used to dynamically solve the assembly model, synchronously acquiring time-domain waveform data of the centroid coordinate displacement, brush disk rotational angular velocity around the axis, and axial acceleration at discrete time steps. The axial acceleration threshold is set based on the fatigue strength limit of the brush disk material. Anomalies are triggered by iterating through the time-series data of the initial parameter set and triggering an anomaly marker when the acceleration peak at three consecutive sampling points exceeds the threshold corresponding to the material's yield stress. A cubic spline interpolation algorithm is used to smooth and correct the data in the anomaly interval. In the dimension unification stage, when converting the Cartesian coordinate data of the centroid displacement trajectory to curvature parameters, the point-by-point coordinate difference method is used to calculate the tangential angle deviation between adjacent points, and a normalized curvature coefficient is constructed based on the ratio of the trajectory arc length to the chord length. The brush disk velocity parameters are dimensionlessly eliminated by calculating the percentage ratio of the actual linear velocity to the maximum design velocity, and root mean square standardization is applied to the acceleration data to eliminate unit system differences.

[0035] For example, step S12 specifically includes:

[0036] Step S121: Based on the load distribution conditions of the brush contact surface, set the acceleration threshold for linear reciprocating motion and the range of curvature radius variation for helical motion, and generate two sets of motion condition parameters.

[0037] Step S122: Based on the upper limit of the rated power of the motor in the assembly model, define the torque input threshold and frequency response bandwidth of the drive components under two motion conditions;

[0038] Step S123: Load two sets of motion parameters into the multibody dynamics simulation module, perform continuous motion simulation under the motor torque threshold constraint, and output the original simulation data package containing the centroid spatial coordinate sequence, brush speed pulse waveform and axial acceleration peak and valley values.

[0039] Step S124: Perform motion phase analysis on the original simulation data package and extract the extreme points of the centroid X-axis displacement under the linear condition and the gradient of the Z-axis angular velocity change under the helical condition as the feature parameter set;

[0040] Step S125: Input the extracted feature parameter set into the kinematic verification module. When the peak and valley values ​​of axial acceleration exceed the material deformation threshold, trigger the working condition parameter feedback correction to generate an initial parameter set that conforms to physical constraints.

[0041] In one specific embodiment, based on the contact pressure distribution model between the brush and the cleaning surface, the maximum acceleration threshold for 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 upper limit of acceleration is determined to be two-thirds of the yield strength of the brush material. For the helical motion condition, the dynamic variation range of the helical radius of curvature is derived according to the coverage requirements of the cleaning path. The minimum distinguishing distance between adjacent helical trajectories is calculated using the geometric envelope method, thereby generating a set of curvature radius parameters proportional to the brush diameter. Based on the peak power characteristic curve of the motor, the dynamic response upper limit of the torque input is set in the drive system simulation module. By calculating the inertia matching degree between the motor rotor inertia and the load torque, the allowable fluctuation range of the drive frequency is determined. After importing the assembly model into the multibody dynamics solver, the Runge-Kutta algorithm is used to simultaneously advance the simulation process of the two working conditions. The six-degree-of-freedom coordinates of the center of mass position are collected at each time step, and the high-frequency component of axial acceleration is separated from the angular velocity signal using a differential amplification algorithm. During the motion phase analysis stage, a sliding window extreme value detection method is used for the centroid displacement data of the straight-line condition to extract the maximum positive offset point and the reverse inflection point in the X-axis direction as feature markers. For the angular velocity data of the helical condition, a cubic spline interpolation curve is constructed to calculate the gradient value of the change in Z-axis angular displacement per unit time. During the feature parameter set verification, a mapping relationship between the peak and valley values ​​of axial acceleration and the elastic modulus of the material is established. When the acceleration oscillation amplitude is detected to exceed the safety threshold corresponding to the material relaxation modulus, the backpropagation correction algorithm is triggered: the driving torque parameters for that period are recalculated, a weighted average method is used to smooth the transition of the condition parameters in the abnormal interval, and an iterative compensation algorithm is used to eliminate the energy non-conservation error caused by data mutation.

[0042] Step S2: Based on standardized motion parameter data, perform multi-physics force analysis on the cleaning robot under various motion states to obtain comprehensive force data including gravity, friction, and wind resistance; decompose the comprehensive force data into force components to obtain the main force component data;

[0043] For example, step S2 includes the following steps:

[0044] Based on standardized motion parameter data, a force analysis model of the cleaning robot under multi-physics coupling environment is established. The robot's posture parameters, center of mass position and velocity direction under different motion states are input to obtain the initial parameters of the force analysis model.

[0045] It should be added that the logic for obtaining the initial parameters of the force analysis model is as follows:

[0046] Based on standardized motion parameter data, the changes in the center of mass position, posture angle and velocity direction of the cleaning robot under different motion states are segmented and extracted to obtain basic motion posture data;

[0047] Based on the basic motion posture data, the geometric constraint relationship of the cleaning robot in three-dimensional space is structurally mapped, and a three-dimensional motion constraint model including the drive unit, brush assembly and connecting mechanism is constructed to obtain the three-dimensional constraint structure data.

[0048] By matching the three-dimensional constraint structure data with the standardized motion parameter data, the velocity and acceleration of the robot in different postures are correlated and calibrated to obtain the motion state matching parameters.

[0049] Based on motion state matching parameters, physical boundaries are set for the robot's interaction conditions in a multi-physics field to obtain multi-physics field coupling boundary data; the multi-physics field includes gravity field, air resistance field and contact friction field.

[0050] The multiphysics coupled boundary data is input into the dynamic modeling module to perform overall modeling and parameter constraint calculations on the force characteristics under different motion states, thereby obtaining preliminary force analysis model data; specifically, the following steps are included:

[0051] Based on multi-physics field coupled boundary data, the force distribution of the cleaning robot in the gravity field, friction field and air resistance field is divided into subfields to obtain the subfield force data under the independent action of each physical field.

[0052] The force data of each subfield are analyzed by superposition. The weight coefficients are set according to the actual influence of different physical fields on the robot's operation. The force data of each subfield are weighted and summed to obtain the force superposition weight data.

[0053] Based on the superposition weighted force data, the mechanical equilibrium relationship of the robot under different motion states is constrained and calculated to verify the stability of the force composition and motion state matching in the model, and to obtain preliminary force analysis model data.

[0054] The parameter convergence and stability of the preliminary stress analysis model data are verified and adjusted. Unreasonable constraints or abnormal stress points are removed to obtain the initial parameters of the final stress analysis model.

[0055] Based on the initial parameters of the force analysis model, the gravity, friction and wind resistance of the robot under different states are dynamically solved to obtain the changing trend of each force in the time series and obtain preliminary comprehensive force data;

[0056] It should be added that the logic for obtaining the preliminary comprehensive force data is as follows:

[0057] Based on the initial parameters of the force analysis model, the posture angle, center of mass position and velocity direction of the cleaning robot under different motion states are input in time sequence to obtain motion state input data;

[0058] Based on the motion state input data, the force distribution of the cleaning robot in the gravity field is calculated, and the gravity effect under different postures is decomposed into vector components to obtain gravity component data.

[0059] Based on gravity component data, friction characteristics of the contact area between the robot and the ground and water surface are analyzed, and friction force variation data are obtained by combining the contact area and material friction coefficient.

[0060] By combining friction force change data with robot velocity direction and surface morphology information, airflow resistance is dynamically estimated to obtain wind resistance change data.

[0061] Time series alignment and synchronization analysis were performed on gravity component data, friction force change data and wind resistance change data to ensure that the correspondence of the three types of forces is consistent under the same time reference, and force time matching data was obtained.

[0062] Based on the force-time matching data, the forces of various types are superimposed and summed at each time step to obtain the total force change sequence under different states, and the trend characteristics of the change are extracted to finally obtain preliminary comprehensive force data.

[0063] It should be added that, based on the force-time matching data, various forces are superimposed and summed time-by-time to obtain the total force change sequence under different states, and their change trend characteristics are extracted. Finally, preliminary comprehensive force data is obtained, which also includes:

[0064] Time reference calibration is performed on the force-time matching data. Gravity component data, friction force change data and wind resistance change data are synchronized at a uniform time interval to ensure that each force data has a corresponding relationship at the same time node, thus obtaining force-time synchronization data.

[0065] Based on the time-synchronized force data, the gravity, friction and wind resistance at each time point are vector superimposed and summed to calculate the magnitude and direction of the resultant force at each moment, thus obtaining the time-series force superposition data.

[0066] Trend features are extracted from the superimposed temporal force data. By smoothing the magnitude of the resultant force change over a continuous time period, the dominant trend and periodic fluctuation characteristics of the force change are identified, and preliminary comprehensive force data are finally obtained.

[0067] Vector superposition and time-domain smoothing are performed on the preliminary comprehensive stress data to remove numerical noise and transient interference generated during the calculation, resulting in stable comprehensive stress data; the specific operation logic is as follows:

[0068] The directional components of the preliminary integrated force data are extracted. Gravity, friction, and wind resistance at each time point are vectorized according to the coordinate axis direction to obtain force directional component data. Based on the force directional component data, the components at the same time point are vector-superimposed, and different force components are vector-synthesized in the spatial coordinate system to obtain force superposition data. Temporal sampling analysis is performed on the force superposition data to identify fluctuation segments and abnormal peaks in force over time, obtaining force fluctuation segment data. Based on the force fluctuation segment data, abrupt changes and transient interference signals in a short period of time are smoothed and corrected. High-frequency numerical noise is eliminated by using moving average or weighted average methods to obtain smoothed force data. The trend consistency of the smoothed force data is verified to check whether each force directional component maintains the consistency of physical dimensions and temporal continuity after superposition, thus obtaining stable integrated force data.

[0069] Based on the comprehensive force data, the overall force situation of the cleaning robot is decomposed into force directions. The comprehensive force is divided into components according to the robot coordinate system to obtain force component data, including:

[0070] The coordinate system is initialized based on the stable comprehensive force data. The correspondence between the body coordinate system and the ground reference coordinate system is established according to the structural characteristics of the cleaning robot to obtain coordinate system matching data.

[0071] Based on coordinate system matching data, coordinate system transformation is performed on each force vector in the comprehensive force data to map the external global force information to the body coordinate system of the cleaning robot, thus obtaining force direction mapping data.

[0072] The force direction mapping data is analyzed by direction angle, and the distribution ratio of each force in the X, Y, and Z axes of the robot body coordinate system is calculated to obtain the force direction distribution data.

[0073] Based on the force direction distribution data, the force components in each direction are extracted, and the total force is decomposed according to the direction ratio to obtain preliminary force component data.

[0074] The initial force component data is checked for consistency. The difference between the total vector after the synthesis of each component and the comprehensive force data is checked to see if it is within the allowable error range. If the difference exceeds the threshold, the ratio is adjusted to obtain the corrected force component data.

[0075] The distribution uniformity of the force direction is analyzed based on the corrected force component data to determine whether the force state of the cleaning robot in each direction is stable, and finally the force component data is obtained.

[0076] It should be added that a consistency check is performed on the preliminary force component data, verifying whether the difference between the total vector after combining the components and the comprehensive force data is within the allowable error range. If the difference exceeds the threshold, a proportional adjustment is made to obtain the corrected force component data; this also includes:

[0077] Vector synthesis verification is performed on the preliminary force component data. The force components in the X, Y, and Z directions are vector superimposed, and the magnitude and direction of the synthesized total force are calculated to obtain the force synthesis verification data.

[0078] 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 between the two is calculated and compared with the preset allowable error threshold to obtain the force error analysis data.

[0079] When there are components in the force error analysis data that exceed the threshold, the corresponding directional components are proportionally corrected according to the error direction and magnitude. While keeping the overall vector direction unchanged, the values ​​of each component are adjusted to finally obtain the corrected force component data.

[0080] The force component data is analyzed for its primary and secondary characteristics. By comparing the average intensity and rate of change of components in different directions, the force terms that play a dominant role in the robot's motion are selected, thus obtaining the main force component data. Specifically, this includes the following steps:

[0081] The force component data is categorized by direction, and the force components are grouped according to the X, Y, and Z axes of the body coordinate system to obtain the force direction grouping data.

[0082] Based on the force direction grouping data, statistical analysis is performed on the force values ​​of each directional component in the time series, and the average intensity value of each directional component is calculated to obtain the average force intensity data.

[0083] Based on the average strength data, the fluctuation amplitude of the force value of each directional component over time is calculated by difference to obtain the force change rate data;

[0084] By comparing and analyzing the average force intensity data with the force change rate data, and calculating the contribution of each directional component in the overall motion process through a weighted evaluation method, the force contribution data is obtained.

[0085] Based on the force contribution data, the components in each direction are sorted and filtered to extract the main force items whose contribution is greater than the preset contribution threshold and whose rate of change is within the stable range, thus obtaining preliminary main force data.

[0086] The initial main force data is checked for consistency to verify whether the difference between the superposition result and the original force component data in time and direction is within the allowable error range. If it exceeds the range, a proportional correction is performed to finally obtain the main force component data.

[0087] The average force intensity data is compared and analyzed with the force change rate data. The contribution of each directional component in the overall motion process is calculated through a weighted evaluation method to obtain the force contribution data. The specific steps also include:

[0088] The average strength data under stress is normalized by standardizing the average strength values ​​of different directional components according to the maximum strength ratio to ensure that the components in each direction are comparable within the same numerical range, thus obtaining normalized strength data.

[0089] Based on the normalized stress intensity data, a weight allocation analysis is performed on the stress change rate data. According to the stability and fluctuation range of the change rate, the weight coefficients of each directional component are determined, and the change rate weight allocation data is obtained.

[0090] The normalized stress intensity data and the weighted data of the rate of change are weighted and superimposed, and the contribution of each directional component to the overall stress is calculated by weighted averaging to obtain the stress contribution data.

[0091] Step S3: Based on the main force component data, plot the time-series force curve, and identify the key components and dangerous working conditions whose forces exceed the preset force threshold by the peak values ​​of the curves; assess the severity of the dangerous working conditions to obtain high-risk working condition data.

[0092] For example, step S3 includes:

[0093] Step S31: Based on the main force component data, perform time-series processing on the force changes of the cleaning robot at different time points, and collect the force components in each direction in time order to obtain time-series force processing data.

[0094] Step S32: Based on the time-series force data, plot and smooth the force components in each direction to generate a time-series force curve, and extract the local peak points that appear in the curve to obtain the peak force data.

[0095] Step S33: Based on the peak force data, perform matching analysis on the robot's structural state during the time period of the peak force to identify the concentrated force area and corresponding components, and obtain the key force component data;

[0096] Step S34: Based on the data of key stress-bearing components, perform a comprehensive correlation analysis on the working environment parameters at each peak moment. The working environment parameters include motion speed, attitude angle and contact friction conditions to obtain hazardous working condition data.

[0097] Step S35: Assess the severity of hazardous working conditions by calculating the risk level of the working condition based on parameters such as stress amplitude, duration, and recurrence frequency, and finally obtain high-risk working condition data.

[0098] For example, step S31 includes the following steps:

[0099] Step S311: Extract timestamps from the main force component data, identify the sampling time corresponding to each force component, and match the force data in each direction with the corresponding time information to obtain force-time matching data;

[0100] Step S312: Based on the force-time matching data, sort the force components in different directions (X, Y, Z) by time series, and arrange the component values ​​at different times in chronological order to obtain the sorted time series data;

[0101] Step S313: Perform integrity verification on the time series sorted data, check whether there are missing or abnormal abrupt changes in the stress data within continuous time periods, and perform interpolation correction on outliers to obtain corrected stress time series data; specifically including the following steps:

[0102] Continuity detection is performed on the time-series sorted data, calculating the magnitude of force changes between adjacent time nodes and identifying time segments with abrupt changes or missing values, resulting in force anomaly identification data. Based on the force anomaly identification data, interpolation is performed to repair missing segments. Force values ​​are supplemented using linear interpolation or smoothed mean methods based on the force change trends of adjacent time nodes, resulting in force interpolation repair data. Smoothing consistency verification is performed on the force interpolation repair data. The average gradient of force changes within continuous time periods is calculated using a sliding window method to remove high-frequency fluctuations caused by interpolation, ultimately yielding corrected force time-series data. Based on the corrected force time-series data, the temporal relationships of each directional component are uniformly aggregated to ensure that all force components have corresponding values ​​at the same time node, ultimately yielding time-series force consolidation data.

[0103] For example, step S32 includes the following steps:

[0104] Step S321: Perform directional grouping processing on the time-series force data, extract the force components in each direction (X, Y, Z) in chronological order to form a directional force sequence, and obtain the force direction sequence data;

[0105] Step S322: Based on the force direction sequence data, perform curve plotting and smoothing filtering on the force sequence in each direction, and use time moving average to reduce the influence of abrupt change points to generate a time-series force curve with better continuity, thus obtaining smooth force curve data;

[0106] Step S323: Based on the smoothed force curve data, perform local extreme value analysis on the curve change trend, extract peak point information by identifying the change interval in the curve where the force rises to the extreme value and then falls, and finally obtain the peak force data.

[0107] For example, step S33 includes the following steps:

[0108] Step S331: Extract time periods from the peak force data, determine the start and end time range of the peak occurrence, and extract the complete force record of the corresponding time period from the time-series force curve to obtain the peak time period force data;

[0109] Step S332: Based on the force data during the peak time period, match the motion posture parameters and structural state information of the cleaning robot during that time period, including posture angle, displacement of connection nodes and contact support state, to obtain structural state matching data;

[0110] Step S333: Based on the structural state matching data, compare and analyze the force distribution of each structural unit during the peak time period, identify the stress concentration area, and determine the key structural units with force values ​​significantly higher than the average level, thus obtaining the stress concentration area data; this also includes the following detailed steps:

[0111] Step S3331: Perform region division processing on the structural state matching data, divide the overall robot structure into spatial partitions according to joints, connecting rods and shell units, and obtain structural partition data; wherein the structural partition data includes, but is not limited to, a unique partition identifier, a set of spatial coordinates and a set of adjacency relationships;

[0112] Step S3332: Based on the structural partition data, perform statistical calculations on the force values ​​of nodes within each partition, extract the average force and maximum force of each partition, and obtain partition force characteristic data; wherein the partition force characteristic data includes, but is not limited to, the partition average force value, the partition maximum force value, and the coordinates of the peak position;

[0113] Step S3333: Based on the force characteristic data of the partitions, compare and analyze the force distribution of all partitions, calculate the ratio of the maximum force of each partition to the overall average force, and mark the partition as a force abnormal area when the ratio exceeds the preset threshold, thus obtaining the force abnormal partition data; wherein the force abnormal partition data includes, but is not limited to, the abnormal partition identifier, the force ratio and the threshold determination result;

[0114] Step S3334: Based on the stress anomaly partition data, perform aggregate analysis on adjacent stress anomaly areas, identify the area groups where the stress continuity is greater than the preset continuity threshold and the spatial distance is less than the preset distance threshold, determine the final stress concentration area, and output the stress concentration area data, which includes but is not limited to the stress concentration area group ID, continuity score, regional aggregated stress peak, spatial center coordinates, and boundary range.

[0115] Step S334: Based on the data of the concentrated stress area, perform component mapping and identification on the structural unit with concentrated stress. Combined with the topological relationship of the robot structure, match the stress area with the specific component to finally obtain the data of the key stress component.

[0116] For example, step S34 includes the following steps:

[0117] Step S341: Extract peak moments from the data of key stress-bearing components, identify the time nodes corresponding to the peak stress of each key component, and extract the stress and state data of the corresponding moment from the time-series stress curve, which are recorded as peak moment matching data.

[0118] Step S342: Based on the peak moment matching data, synchronously collect and correlate the motion speed, attitude angle and contact friction conditions during the period, establish a one-to-one correspondence between key stress components and working environment parameters, and obtain working condition correlation data;

[0119] Step S343: Based on the working condition correlation data, perform coupled evaluation on the stress performance of each key stress component under different environmental parameters, determine the abnormal stress conditions under conditions of excessive speed, deviation of attitude angle or sudden change of friction coefficient, identify potential dangerous working condition ranges, and finally obtain dangerous working condition data.

[0120] Step S343 includes the following steps:

[0121] Step S3431: Set parameter thresholds for the working condition correlation data. Based on the rated working range of key stress-bearing components and the structural safety factor, determine the upper limit threshold of speed, the attitude angle deviation threshold, and the friction coefficient fluctuation threshold to obtain the working condition threshold benchmark data.

[0122] Step S3432: Based on the benchmark data of working condition thresholds, perform time-by-time comparative analysis on the working condition related data, identify the time points that exceed any threshold, and extract the corresponding stress data and environmental parameters to obtain the working condition anomaly judgment data.

[0123] Step S3433: Based on the abnormal working condition judgment data, perform interval aggregation analysis on continuous or adjacent abnormal time points, divide the abnormal segments that are continuous in time or have the same stress change trend into a potential dangerous working condition interval, and obtain the working condition interval aggregation data.

[0124] Step S3434: Based on the aggregated data of the working condition intervals, comprehensively evaluate the stress intensity, duration and frequency of occurrence of each dangerous interval, screen out the high-risk intervals with large stress amplitude and long duration, and finally output the dangerous working condition data.

[0125] For example, step S35 includes the following steps:

[0126] Step S351: Extract parameters from the hazardous working condition data to obtain basic parameters such as peak force, duration of force, and frequency of recurrence of each hazardous working condition, and obtain working condition parameter extraction data.

[0127] Step S352: Extract data based on working condition parameters, perform weighted analysis on the force amplitude and duration of different dangerous working conditions, calculate the working condition strength index based on the comprehensive degree of the peak force and duration of action in the working condition, and obtain working condition strength assessment data.

[0128] Step S353: Based on the working condition intensity assessment data, the recurrence frequency of each working condition is correlated and corrected. When a certain working condition repeatedly occurs in multiple working cycles, its risk weight is increased to obtain working condition risk weight data.

[0129] Step S354: Based on the working condition risk weight data, classify the risk level of each hazardous working condition, determine the working condition level distribution according to the comprehensive value of the intensity index and risk weight, and mark the high-level results as high-risk working conditions, and finally obtain the high-risk working condition data.

[0130] Step S354 includes the following steps:

[0131] Step S3541: Set thresholds for the working condition risk weight data. Based on the distribution range of the working condition intensity index and the historical risk statistics, determine the threshold boundaries for low-risk, medium-risk, and high-risk levels to obtain the risk level threshold data.

[0132] Step S3542: Based on the risk level threshold data, the intensity index and risk weight of each hazardous working condition are comprehensively calculated, and the comprehensive risk value of the working condition is obtained according to the weighted average principle to obtain the comprehensive risk data of the working condition.

[0133] Step S3543: Based on the comprehensive risk data of the working conditions, compare the risk value of each hazardous working condition with the threshold boundary to determine its risk level range and obtain the risk level distribution data.

[0134] Step S3544: Based on the risk level distribution data, filter and confirm the working conditions in the high-level range, mark and output the working conditions with a high risk level, and finally obtain the high-risk working condition data.

[0135] In this embodiment of the invention, the system analyzes the collected main force component data in the time dimension, extracts the sampling timestamps corresponding to each set of force vectors, and establishes a precise mapping relationship between force values ​​and time information. Subsequently, according to the logical order of time, the force components in the X, Y, and Z axes are sorted in full sequence. During this process, the system performs in-depth verification of the data integrity, identifying breakpoints or abnormal abrupt changes in the data stream by calculating the gradient of force values ​​at adjacent time nodes. For the identified missing segments, the system uses the force change trends of adjacent time points and employs linear interpolation or smoothing mean calculation methods to fill in the missing force values, generating force interpolation repair data. Next, the sliding window algorithm is used to perform smoothing consistency verification on the repaired data, calculating the average gradient within the window to eliminate high-frequency fluctuations caused by interpolation or sensor noise, thereby outputting the corrected force time series data, which is finally aggregated into time-series force processing data. The system decouples the time-series force processing data according to the spatial dimension, extracting independent force sequences in the X, Y, and Z directions respectively. For each sequence, a time-moving average filtering algorithm is used to suppress random noise by calculating the arithmetic mean within a specific step size, generating a smooth and continuous temporal force curve. Based on this, the system performs extremum analysis on the curve's geometry, accurately capturing local force peaks by identifying inflection points where the curve's slope changes from positive to negative, or points whose values ​​are significantly higher than the neighborhood baseline. The system then extracts the time and amplitude information of these peak points to generate force peak data. Furthermore, based on the time of the peak occurrence, the system extracts complete force records within a preset time window before and after it, and simultaneously retrieves structural information such as the robot's motion posture parameters, connection node displacements, and contact support states during that time period. To accurately identify the specific physical location of force concentration, the system performs refined spatial region analysis: dividing the robot's overall structure into several spatial partitions with unique identifiers, including joints, links, and shell units, and constructing adjacency sets based on topological relationships. The system statistically analyzes the force situation of nodes within each partition, calculating the average and maximum force values ​​for each partition. Subsequently, through comparative analysis, the ratio of the maximum stress value of each partition to the average stress value of the overall structure is calculated. When this ratio exceeds a preset anomaly detection threshold, the partition is marked as a stress anomaly area. Further, the system employs region growing or clustering analysis logic to aggregate spatially adjacent anomaly partitions with stress continuity scores higher than a preset standard, identifying groups of concentrated stress areas. Finally, combined with the robot's structural topology map, these high-stress groups are mapped to specific physical components, thereby outputting key stress component data.

[0136] The system extracts the specific moments when critical stress components experience peak stress and correlates these moments with environmental parameters, including the robot's speed, tilt angle, and friction conditions at the contact surfaces. By establishing a coupled model of component stress and environmental parameters, the system sets safety threshold benchmarks based on the component's rated operating range. The system compares real-time parameters with these benchmark thresholds hourly. Any instances of excessive speed, excessive tilt angle, or abrupt changes in friction coefficient are identified as abnormal operating conditions. Subsequently, the system aggregates consecutive or adjacent abnormal points on the timeline, merging time periods with consistent stress trends and persistently abnormal environmental parameters into a single potentially hazardous operating condition interval. The system comprehensively considers the stress intensity and duration within these intervals, selecting those with large stress amplitudes and long durations as hazardous operating condition data.

[0137] Finally, core characteristic parameters are extracted from the hazardous working condition data, including the peak stress magnitude, the duration of abnormal stress, and the frequency of recurrence of the working condition in historical cycles. Based on these parameters, the system constructs a working condition intensity assessment model: using a weighted summation method, different weight coefficients are assigned to the peak stress and duration to calculate an intensity index reflecting the severity of a single working condition. Simultaneously, a frequency correction factor is introduced to increase the risk weight of similar working conditions that repeatedly occur in multiple work cycles. Next, the system compares the calculated comprehensive risk value of the working condition (the weighted result of the intensity index and risk weight) with the threshold boundaries according to pre-set low, medium, and high risk level thresholds. Finally, the system filters out working conditions falling within the high-risk level range, marks them, and outputs them as high-risk working condition data.

[0138] Step S4: For high-risk working condition data, perform transient force analysis on the cleaning robot in two states: downward cleaning and return upward. Based on wind resistance and dirt accumulation factors, simulate the risks of instability, slippage, and overturning to obtain instability risk data. Classify the instability risk data into risk levels to obtain risk level data.

[0139] For example, step S4 includes:

[0140] Step S41: Based on high-risk working condition data, the force state of the cleaning robot in the two states of downward cleaning and return upward is divided, and the key force parameters in each state are extracted, including the component of gravity, adsorption force, driving force and resistance, to obtain state force decomposition data.

[0141] Step S42: Based on the state-force decomposition data, perform time-series calculations on the transient force changes to obtain the force change trend and peak response in a short time, and obtain transient force analysis data;

[0142] Step S43: Based on transient force analysis data, combined with the drag coefficient and surface dirt accumulation coefficient in the external environment, the equilibrium state of the robot under different force conditions is disturbed and simulated. The sliding trend and center of gravity offset of the robot on the inclined surface are calculated to obtain the instability risk simulation data.

[0143] Step S44: Based on the instability risk simulation data, make a comprehensive judgment on the risks of slippage and overturning, and combine the correspondence between the instability probability and the overturning angle to classify the risk level, and finally obtain the risk level data.

[0144] For example, step S41 includes the following steps:

[0145] Step S411: Perform state identification analysis on the high-risk working condition data. Based on the movement direction and attitude angle parameters of the cleaning robot, divide its operating state into downward cleaning state and return upward state to obtain state division data.

[0146] Step S412: Based on the state division data, analyze the force composition in each state, extract 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 surface friction to obtain preliminary force parameter data.

[0147] Step S413: Based on the preliminary force parameter data, perform vector decomposition and direction normalization of the force components in each direction to ensure that the gravitational force, adsorption force, driving force and resistance have consistent direction definitions in a unified coordinate system, and obtain standardized force component data.

[0148] Step S414: Based on the standardized force component data, compare and analyze the force distribution under the two working conditions, identify the main force change trends and difference areas, and finally obtain the state force decomposition data.

[0149] For example, step S42 includes the following steps:

[0150] Step S421: Divide the state-force decomposition data into time series segments, segment the force data according to the sampling time interval, determine the start and end time of the force change in each segment, and obtain time series segmented data;

[0151] Step S422: Based on the time-series segmented data, perform continuous difference calculation on the force values ​​of each segment, extract the force change rate in a short time, obtain the force increase and decrease trend, and obtain the force change trend data;

[0152] Step S423: Based on the force change trend data, identify the local peak values ​​of the force fluctuations in each time period, extract the maximum force point and its corresponding time during the change process, and obtain the transient peak response data;

[0153] Step S424: Based on the transient peak response data, comprehensively evaluate the fluctuation amplitude, peak density and rate of change of the overall time-series force to form a short-time dynamic force characteristic description, and finally obtain transient force analysis data.

[0154] For example, step S43 includes the following steps:

[0155] Step S431: Perform external factor matching on transient stress analysis data, extract environmental parameters corresponding to the stress period, including wind resistance coefficient and surface dirt accumulation coefficient, form a stress-environment correspondence, and obtain environmental matching data;

[0156] Step S432: Based on environmental matching data, calculate the overall balance torque of the cleaning robot under different force states, and evaluate the torque balance deviation under each state by taking into account the effects of gravity, adsorption force and wind resistance, and obtain balance deviation data.

[0157] Step S433: Based on the balance deviation data, perform dynamic simulation of the force disturbance on the robot on the inclined surface, analyze the slip trend and attitude deviation response caused by external disturbance, and obtain slip trend simulation data.

[0158] Step S434: Based on the slip trend simulation data, calculate the overall center of gravity trajectory of the robot, obtain the offset and directional change of the center of gravity during the disturbance process, and finally obtain the instability risk simulation data.

[0159] For example, step S44 includes the following steps:

[0160] Step S441: Extract slip risk from the instability risk simulation data, analyze the slip trend and sliding distance changes of the robot under different force states, calculate the probability of slippage occurring per unit time, and obtain slip risk data;

[0161] Step S442: Based on the slip risk data, perform statistical analysis on the overturning angle change of the robot during the disturbance process, determine the correspondence between the tilt angle and the center of gravity shift at each moment, and obtain the overturning angle characteristic data.

[0162] Step S443: Based on the overturning angle characteristic data and the slip risk data, make a comprehensive judgment on the coupling effect of the two risks, calculate its instability probability distribution, and obtain comprehensive instability probability data;

[0163] Step S444: Based on the comprehensive instability probability data and combined with the level range threshold of the overturning angle, different risk levels are divided, and the comprehensive probability value is matched with the corresponding angle offset range to finally obtain the risk level data.

[0164] 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.

[0165] 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.

[0166] 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.

[0167] For example, step S5 includes the following steps:

[0168] 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.

[0169] Step S52: Based on the working condition simulation matching data, divide the stress cloud map and deformation cloud map of the main frame component and the roller support component into regions, identify the main stress area and deformation concentration area, and obtain structural region identification data.

[0170] Step S53: Based on the structural region identification data, perform peak search on the stress values ​​and displacements of each region, extract the maximum stress point and maximum deformation point and their location coordinates within each component, and obtain peak extraction data;

[0171] Step S54: Based on the peak extraction data, the stress causes of the peak area are analyzed in combination with the risk level data to determine the stress and deformation characteristics caused by load concentration, structural weakness or boundary constraints, and finally obtain the identification data of the maximum stress and deformation location.

[0172] Step S52 includes the following steps:

[0173] Step S521: Import and preprocess the stress cloud map and deformation cloud map in the working condition simulation matching data, unify the coordinate system and scale parameters, eliminate model resolution differences, and obtain standardized cloud map data.

[0174] Step S522: Based on the cloud map standardized data, the grid nodes of the main frame component and the roller support component are partitioned and marked, and the preliminary region is divided according to the structural connection characteristics and geometric boundaries to obtain the preliminary region division data;

[0175] Step S523: Based on the preliminary regional division data, perform statistical analysis on the stress and displacement values ​​of nodes in each region, extract the average stress, maximum stress and average deformation indexes, and obtain regional stress and deformation characteristic data.

[0176] Step S524: Based on the regional stress and deformation characteristic data, compare the stress gradient and deformation gradient of adjacent regions to identify stress concentration areas and deformation concentration areas, and obtain concentration feature identification data;

[0177] Step S525: Based on the concentrated feature identification data, identify and number the main stress areas and deformation concentration areas, generate structural area identification results, and finally obtain structural area identification data.

[0178] Step S53 includes the following steps:

[0179] Step S531: Read and process the stress distribution and deformation distribution in the structural region identification data, and import the region data of the main frame assembly and the roller support assembly into the analysis module to obtain the region layering data;

[0180] Step S532: Based on the regional layered data, perform numerical scanning of the nodal stress values ​​and nodal displacements in each region, filter out abnormal nodes and calculation errors, and obtain valid node data;

[0181] Step S533: Based on the effective node data, perform local extreme value detection on the stress value and displacement of each region to determine the stress peak point and deformation peak point of each region, and obtain the region peak identification data;

[0182] Step S534: Based on the regional peak identification data, perform a global comparison of the peak points in each region, filter out the global maximum points of stress value and deformation in the main frame component and roller support component, and record the corresponding spatial coordinates to obtain global peak data;

[0183] Step S535: Based on the global peak data, verify and compare the numerical accuracy and consistency of the maximum stress point and the maximum deformation point to ensure that the peak extraction is accurate and reliable, and finally obtain the peak extraction data.

[0184] Step S54 includes the following steps:

[0185] Step S541: Perform region localization on the peak extraction data, and map the maximum stress point and the maximum deformation point to the structural coordinates of the main frame assembly and the roller support assembly respectively to obtain the peak region localization data;

[0186] Step S542: Based on the peak region location data, perform a detailed analysis of the local stress distribution in the peak region, extract the stress gradient, displacement gradient and direction vector of the surrounding nodes, and obtain the local force distribution data.

[0187] Step S543: Based on the local stress distribution data and the load condition information in the risk level data, determine the degree of load concentration in the peak area, identify the stress concentration characteristics caused by load concentration, and obtain load concentration determination data.

[0188] Step S544: Based on the load concentration judgment data, compare the structural geometric features and constraint boundary conditions of the peak region to analyze whether there are stress and deformation anomalies caused by abrupt changes in the structural cross section or differences in constraint rigidity, and obtain structural constraint analysis data;

[0189] Step S545: Based on the structural constraint analysis data, comprehensively judge the load concentration and structural constraint factors, determine the main stress causes in the peak region, and output the location of the maximum stress and maximum deformation, finally obtaining the identification data of the location of the maximum stress and deformation.

[0190] Step S543 includes the following steps:

[0191] Step S5431: Extract load information from the local stress distribution data. Based on the load condition input in the risk level data, extract the load distribution value and direction component of each node in the peak area to obtain the load distribution extraction data.

[0192] Step S5432: Extract data based on load distribution, perform spatial comparison of load intensity in the peak region, calculate the load difference and change gradient between adjacent nodes, and obtain load change gradient data;

[0193] Step S5433: Based on the load change gradient data, identify the load concentration trend, determine the concentration coefficient of the abrupt change areas in the load distribution, and compare and analyze it with the stress gradient characteristics to obtain the load concentration trend data.

[0194] 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.

[0195] Example 2

[0196] like Figure 2 As shown, a structural simulation and optimization system for a wind turbine tower cleaning robot includes:

[0197] 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;

[0198] 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;

[0199] 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;

[0200] 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.

[0201] 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.

[0202] 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.

[0203] 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.

[0204] It should be understood that determining B based on A does not mean determining B solely based on A; it also means determining B based on A and / or other information.

[0205] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

[0206] 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 structural simulation optimization method for a wind power tower cleaning robot, characterized in that, The method includes: Step S1: Construct an assembly model of the cleaning robot using 3D modeling software, and perform 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; perform data cleaning and normalization on the motion trajectory, velocity, and acceleration parameter data to obtain standardized motion parameter data; Step S1 includes the following sub-steps: Step S11: Based on the assembly structure of the cleaning robot, establish an assembly model including the drive components and brush in the 3D modeling software, and set the motion constraints of the motor drive. Step S12: Apply two typical working conditions, linear reciprocating motion and helical motion, to the assembly model, perform multibody dynamics simulation, and obtain the initial parameter set of the centroid displacement trajectory, brush motion velocity, and axial acceleration. Step S13: Based on the material deformation threshold of the brush disk, remove abnormal data nodes with excessive axial acceleration in the initial parameter set to obtain a compliant motion parameter set; Step S14: Unify the unit dimensions of the compliant motion parameter set, convert the displacement trajectory into normalized curvature coefficients, convert the velocity into a time normalized ratio, record it as the displacement ratio per unit time, and generate standardized motion parameter data. Step S12 specifically includes: Step S121: Based on the load distribution conditions of the brush contact surface, set the acceleration threshold for linear reciprocating motion and the range of curvature radius variation for helical motion, and generate two sets of motion condition parameters. Step S122: Based on the upper limit of the rated power of the motor in the assembly model, define the torque input threshold and frequency response bandwidth of the drive components under two motion conditions; Step S123: Load two sets of motion parameters into the multibody dynamics simulation module, perform continuous motion simulation under the motor torque threshold constraint, and output the original simulation data package containing the centroid spatial coordinate sequence, brush speed pulse waveform and axial acceleration peak and valley values. Step S124: Perform motion phase analysis on the original simulation data package and extract the extreme points of the centroid X-axis displacement under the linear condition and the gradient of the Z-axis angular velocity change under the helical condition as the feature parameter set; Step S125: Input the extracted feature parameter set into the kinematic verification module. When the peak and valley values ​​of axial acceleration exceed the material deformation threshold, trigger the working condition parameter feedback correction to generate an initial parameter set that conforms to physical constraints. Step S2: Based on standardized motion parameter data, perform multi-physics force analysis on the cleaning robot under various motion states to obtain comprehensive force data including gravity, friction, and wind resistance; decompose the comprehensive force data into force components to obtain the main force component data; Step S3: Based on the main force component data, plot the time-series force curve, and identify the key components and dangerous working conditions whose forces exceed the preset force threshold by the peak values ​​of the curves; assess the severity of the dangerous working conditions to obtain high-risk working condition data. Step S4: For high-risk working condition data, perform transient force analysis on the cleaning robot in two states: downward cleaning and return upward. Based on wind resistance and dirt accumulation factors, simulate the risks of instability, slippage, and overturning to obtain instability risk data. Classify the instability risk data into risk levels to obtain risk level data. 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.

2. The structural simulation optimization method for a wind power tower cleaning robot according to claim 1, characterized in that, Step S2 includes: Based on standardized motion parameter data, a force analysis model of the cleaning robot under multi-physics coupling environment is established. The robot's posture parameters, center of mass position and velocity direction under different motion states are input to obtain the initial parameters of the force analysis model. Based on the initial parameters of the force analysis model, the gravity, friction and wind resistance of the robot under different states are dynamically solved to obtain the changing trend of each force in the time series and obtain preliminary comprehensive force data; Vector superposition and time-domain smoothing are performed on the preliminary comprehensive stress data to remove numerical noise and transient interference generated in the calculation, and stable comprehensive stress data are obtained. Based on the comprehensive force data, the overall force situation of the cleaning robot is decomposed into force direction, and the comprehensive force is divided into components according to the robot coordinate system to obtain force component data; By performing a primary and secondary feature analysis on the force component data and comparing the average intensity and rate of change of components in different directions, the force terms that play a dominant role in the robot's motion are selected, and the main force component data are obtained.

3. The structural simulation optimization method for a wind power tower cleaning robot according to claim 2, characterized in that, The logic for obtaining preliminary comprehensive stress data is as follows: Based on the initial parameters of the force analysis model, the posture angle, center of mass position and velocity direction of the cleaning robot under different motion states are input in time sequence to obtain motion state input data; Based on the motion state input data, the force distribution of the cleaning robot in the gravity field is calculated, and the gravity effect under different postures is decomposed into vector components to obtain gravity component data. Based on gravity component data, friction characteristics of the contact area between the robot and the ground and water surface are analyzed, and friction force variation data are obtained by combining the contact area and material friction coefficient. By combining friction force change data with robot velocity direction and surface morphology information, airflow resistance is dynamically estimated to obtain wind resistance change data. Time series alignment and synchronization analysis were performed on gravity component data, friction force change data and wind resistance change data to ensure that the correspondence of the three types of forces is consistent under the same time reference, and force time matching data was obtained. Based on the force-time matching data, the forces of various types are superimposed and summed at each time step to obtain the total force change sequence under different states, and the trend characteristics of the change are extracted to finally obtain preliminary comprehensive force data.

4. The structural simulation optimization method for a wind power tower cleaning robot according to claim 2, characterized in that, The steps for obtaining the main force component data are as follows: The force component data is categorized by direction, and the force components are grouped according to the X, Y, and Z axes of the body coordinate system to obtain the force direction grouping data. Based on the force direction grouping data, statistical analysis is performed on the force values ​​of each directional component in the time series, and the average intensity value of each directional component is calculated to obtain the average force intensity data. Based on the average strength data, the fluctuation amplitude of the force value of each directional component over time is calculated by difference to obtain the force change rate data; By comparing and analyzing the average force intensity data with the force change rate data, and calculating the contribution of each directional component in the overall motion process through a weighted evaluation method, the force contribution data is obtained. Based on the force contribution data, the components in each direction are sorted and filtered to extract the main force items whose contribution is greater than the preset contribution threshold and whose rate of change is stable, thus obtaining preliminary main force data. The initial main force data is checked for consistency to verify whether the difference between the superposition result and the original force component data in time and direction is within the allowable error range. If it exceeds the range, a proportional correction is performed to finally obtain the main force component data.

5. The structural simulation optimization method for a wind power tower cleaning robot according to claim 1, characterized in that, Step S3 includes: Step S31: Based on the main force component data, perform time-series processing on the force changes of the cleaning robot at different time points, and collect the force components in each direction in time order to obtain time-series force processing data. Step S32: Based on the time-series force data, plot and smooth the force components in each direction to generate a time-series force curve, and extract the local peak points that appear in the curve to obtain the peak force data. Step S33: Based on the peak force data, perform matching analysis on the robot's structural state during the time period of the peak force to identify the concentrated force area and corresponding components, and obtain the key force component data; Step S34: Based on the data of key stress-bearing components, perform a comprehensive correlation analysis on the working environment parameters at each peak moment. The working environment parameters include motion speed, attitude angle and contact friction conditions to obtain hazardous working condition data. Step S35: Assess the severity of hazardous working conditions by calculating the risk level of the working condition based on parameters such as stress amplitude, duration, and recurrence frequency, and finally obtain high-risk working condition data.

6. The structural simulation optimization method for a wind power tower cleaning robot according to claim 1, characterized in that, Step S4 includes: Step S41: Based on high-risk working condition data, the force state of the cleaning robot in the two states of downward cleaning and return upward is divided, and the key force parameters in each state are extracted, including the component of gravity, adsorption force, driving force and resistance, to obtain state force decomposition data. Step S42: Based on the state-force decomposition data, perform time-series calculations on the transient force changes to obtain the force change trend and peak response in a short time, and obtain transient force analysis data; Step S43: Based on transient force analysis data, combined with the drag coefficient and surface dirt accumulation coefficient in the external environment, the equilibrium state of the robot under different force conditions is disturbed and simulated. The sliding trend and center of gravity offset of the robot on the inclined surface are calculated to obtain the instability risk simulation data. Step S44: Based on the instability risk simulation data, make a comprehensive judgment on the risks of slippage and overturning, and combine the correspondence between the instability probability and the overturning angle to classify the risk level, and finally obtain the risk level data.

7. The structural simulation optimization method for a wind power tower cleaning robot according to claim 1, characterized in that, Step S5 includes: Step S51: Process the risk level data for working condition input, and establish a finite element analysis model of the cleaning robot based on the load conditions and constraint boundaries corresponding to different risk levels to obtain simulation model data; Step S52: Based on the simulation model data, perform static stress calculation on the robot structure, obtain the stress distribution and displacement response of the main frame component and roller support component under steady-state load, and obtain static simulation result data; Step S53: Based on the static simulation results, load timing input and transient solution are performed on the dynamic characteristics of the robot under high-risk working conditions to obtain dynamic simulation results. Step S54: Based on the static simulation results and dynamic simulation results, perform overlay analysis on the stress and deformation cloud maps, identify the concentrated areas of maximum stress and maximum deformation, and combine the risk level data to make a comprehensive judgment on their causes, and finally obtain stress and deformation analysis data.

8. A structural simulation optimization system for a wind turbine tower cleaning robot, implemented based on the structural simulation optimization method for a wind turbine tower cleaning robot according to any one of claims 1-7, characterized in that, The system 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.

Images