Automatic fastening system and method for variable pitch bearing bolt

Abstract

The invention discloses an automatic fastening system and method for a variable pitch bearing bolt, and the method comprises the following steps: S1, collecting and preprocessing data, and constructing a stress distribution data set; s2, constructing a multi-layer pheromone field by using an improved ant foraging algorithm, and generating an initial fastening path in combination with a dynamic sensing function; s3, optimizing the initial fastening path through a pheromone updating rule, and generating a dynamically optimized fastening path; s4, based on a dynamic feedback mechanism, a bolt torque target is adjusted by constructing a thermal expansion compensation model and a friction correction model; s5, the fastening result is verified through finite element analysis, and a verification report is generated; s6, bolt nodes with torque deviation or abnormal stress are recognized in real time in combination with an anomaly detection mechanism, pheromone distribution is dynamically adjusted, and a fastening path is re-planned; and S7, generating a complete fastening task report and providing maintenance suggestions. A dynamic optimization and feedback control method is adopted, and variable pitch bearing bolt fastening path optimization and stress balance are intelligently achieved.

Description

Technical Field

[0001] The present invention relates to the technical fields of artificial intelligence and mechanical fastening, and particularly to a variable pitch bearing bolt automatic fastening system and method. Background Art

[0002] The fastening of variable pitch bearing bolts is one of the important processes in the fields of wind power generation, aerospace, and high-precision machinery manufacturing, and its quality is directly related to the operation stability and safety of equipment. In the prior art, bolt fastening is usually completed by manual operation or semi-automatic mechanical equipment. Due to relying on the experience and technical level of operators, this traditional fastening method is prone to problems such as uneven bolt torque and unbalanced stress distribution. Especially in the variable pitch bearings of wind turbines, bolts need to bear high-intensity dynamic loads. If the fastening quality does not meet the standard, it may lead to fatigue damage or even structural failure of the equipment.

[0003] Although the existing automatic fastening systems reduce human errors to a certain extent, there are still many technical defects. On the one hand, due to the usually complex bolt distribution and fixed fastening paths, the existing systems cannot dynamically adapt to environmental changes. For example, dynamic factors such as thermal expansion, elastic recovery, and friction change of materials will cause deviations between the actual torque and the target value during the fastening process, and the fixed path planning cannot optimize the path execution in real time, resulting in uneven stress distribution. On the other hand, traditional path planning mostly adopts fixed sequences or simple cross-sequences, lacking comprehensive optimization of global stress uniformity and regional stress distribution, and it is easy to have over-tightened or over-loosened bolts locally, thus affecting the performance of the entire system.

[0004] In addition, the current automatic fastening technologies usually lack an efficient feedback control mechanism. In industrial applications, uncontrollable factors such as environmental temperature, mechanical vibration, and equipment wear may cause changes in the real-time torque of bolts. However, the existing systems are mostly based on static optimization models and are difficult to timely perceive these dynamic changes and make corresponding adjustments. Especially in complex scenarios with the coupling effect of multiple bolts, the torque deviation of a single bolt may cause the stress redistribution of other bolts, further exacerbating the global stress non-uniformity. This situation not only reduces the fastening accuracy but also may lead to increased rework and maintenance costs.

[0005] At the same time, finite element analysis technology has been widely used in the field of bolt fastening to evaluate stress distribution and torque error. However, the finite element analysis in the prior art is mostly offline calculation and cannot be linked with the actual fastening process in real time. This limitation leads to a lag in the verification of fastening results, and abnormal bolts and uneven stress cannot be identified and corrected in time. In addition, traditional finite element analysis methods only focus on the stress conditions of single bolts or local areas, lacking a comprehensive assessment of the global stress distribution, and it is difficult to provide effective guidance for path planning and strategy adjustment.

[0006] For the detection and handling of abnormal bolts, the existing technologies usually make simple judgments by presetting thresholds, lacking the intelligent ability of abnormal classification and hierarchical processing. When there are torque deviations or abnormal stress distributions, the existing systems are difficult to quickly identify the types of abnormalities and their influence ranges, and cannot dynamically adjust the path planning according to the actual situation. Especially in high-frequency fastening tasks, the existing methods lack an efficient dynamic response mechanism, which is prone to abnormal accumulation, further affecting the performance and reliability of the entire system.

[0007] The existing ways of generating fastening task reports also have obvious deficiencies. Traditional reports are mostly summaries of result data, lacking a comprehensive analysis of torque distribution, stress uniformity, and abnormal handling, and it is difficult to provide effective references for subsequent maintenance. In addition, the generation process of reports usually relies on manual sorting and analysis, which is time-consuming and inefficient, and is not suitable for high-efficiency industrial scenarios.

[0008] Therefore, how to provide an automated pitch bearing bolt tightening system and method is an urgent problem to be solved by those skilled in the art. Summary of the Invention

[0009] An object of the present invention is to propose an automated pitch bearing bolt tightening system and method. The present invention adopts a dynamic optimization and feedback control method. By constructing a multi-layer pheromone field, a thermal expansion compensation model, and a friction correction model, it intelligently realizes the optimization of the tightening path of pitch bearing bolts, torque adjustment, and stress balance. Combining real-time data acquisition and finite element analysis, it ensures the tightening accuracy and the uniformity of stress distribution, and at the same time has an efficient abnormal detection and dynamic correction ability. The generated complete fastening task report provides decision-making support, adapts to complex industrial scenarios, significantly improves the high precision, high efficiency, and strong adaptability of the system, and meets the stringent requirements of the wind power generation and high-precision machinery fields.

[0010] According to the automated pitch bearing bolt tightening method of the embodiment of the present invention, the following steps are included:

[0011] S1. Collect the geometric distribution information, target torque range, and material property parameters of the bolt group, and perform preprocessing to construct a stress distribution data set;

[0012] S2. Based on the stress distribution data set, use an improved ant foraging algorithm to construct a multi-layer pheromone field, the multi-layer pheromone field includes a local torque layer, a regional stress layer, and a global balance layer, and generate an initial tightening path in combination with a dynamic perception function;

[0013] S3. Based on the bolt torque, displacement, and temperature data collected in real time, optimize the initial tightening path through the pheromone update rule, the pheromone includes local pheromone, regional pheromone, and global pheromone, and generate a dynamically optimized tightening path;

[0014] S4. Based on the dynamic feedback mechanism, the current torque, displacement and temperature parameters of the bolt are monitored in real time, and the bolt torque target is adjusted by constructing a thermal expansion compensation model and a friction correction model;

[0015] S5. Use finite element analysis to verify the fastening result, evaluate the single bolt torque error, regional stress uniformity and global stress distribution deviation, and generate a verification report;

[0016] S6. Combine the anomaly detection mechanism to identify the bolt nodes with torque deviation or stress anomaly in real time, dynamically adjust the pheromone distribution and re-plan the fastening path;

[0017] S7. Generate a complete fastening task report, including the fastening path, torque distribution, stress uniformity evaluation and anomaly correction record, and provide maintenance suggestions.

[0018] Optionally, the S2 specifically includes:

[0019] S21. Extract the initial torque deviation and stress distribution parameters of the bolt from the stress distribution dataset, and initialize the pheromone concentration of the local torque layer:

[0020]

[0021] Among them, τ local (i) represents the pheromone concentration of the local torque layer of the i-th bolt, M current (i) represents the current measured torque of the i-th bolt, M target (i) represents the target torque value of the i-th bolt;

[0022] S22. Based on the stress distribution data of the bolts in the area, calculate the average regional stress and the target stress deviation, and initialize the pheromone concentration of the regional stress layer:

[0023]

[0024] Among them, τ region (r) represents the pheromone concentration of area r, σ region (r) represents the average stress value of all bolts in area r, σ target represents the target regional stress value;

[0025] S23. According to the global bolt stress distribution, calculate the global stress uniformity deviation, and initialize the pheromone concentration of the global equilibrium layer:

[0026]

[0027] Among them, τ global represents the global pheromone concentration, Δσ global represents the maximum deviation value of the global stress distribution;

[0028] S24. Combine multi-layer pheromone fields and define the path selection probability P for an ant to move from the i-th bolt to the j-th bolt ij :

[0029]

[0030] where τ ij represents the pheromone concentration on the path from the i-th bolt to the j-th bolt, α, β, and γ represent weight factors, and τ ik represents the pheromone concentration on the path from the i-th bolt to the k-th bolt, allowed represents the set of next nodes, and τ local (k) represents the local torque layer pheromone concentration of the k-th bolt, and τ region (r k ) represents the pheromone concentration in region r k . η ij represents an index of path superiority when moving from the i-th bolt to the j-th bolt, and η ik represents an index of path superiority when moving from the i-th bolt to the k-th bolt:

[0031]

[0032] where d ij represents the distance from the i-th bolt to the j-th bolt, and σ ij represents the stress distribution value from the i-th bolt to the j-th bolt;

[0033] D(t) = α1·f temp (t) + β1·f elastic (t) + γ1·f friction (t);

[0034] where D(t) represents the dynamic perception function, and α1, β1, and γ1 represent weight coefficients;

[0035] f temp (t) = 1 + k temp ·(T(t) - T ref );

[0036] where f temp (t) represents the temperature influence function, k temp represents the thermal expansion coefficient of the material, T ref represents the reference temperature, and T(t) represents the current ambient temperature;

[0037] f elastic (t) = 1 + k elastic ·ΔM;

[0038] where felastic (t) represents the elastic recovery function, and k elastic represents the elastic recovery influence coefficient, and ΔM represents the moment deviation;

[0039] f friction (t) = 1 + k friction ·ΔF;

[0040] Among them, f friction (t) represents the friction change function, and k friction represents the friction change influence coefficient, and ΔF represents the friction force change value;

[0041] S25. During the path selection process, pheromone update is performed on the paths completed by ants in each iteration. The pheromone update rule is determined by the pheromone decay coefficient and the increment:

[0042] τ ij (t + 1) = (1 - ρ)·τ ij +Δτ ij ;

[0043]

[0044] Among them, τ ij (t + 1) represents the pheromone concentration from the i-th bolt to the j-th bolt after update, ρ represents the pheromone decay coefficient, and Δτ ij represents the pheromone increment, Q represents the pheromone enhancement constant, and L best represents the length of the current optimal path;

[0045] S26. Use the improved ant foraging algorithm to iterate multiple times until the path converges, and output the initial tightening path and bolt sequence optimized by combining the local moment layer, regional stress layer, and global equilibrium layer.

[0046] Optionally, the specific content of S3 includes:

[0047] S31. Real-time collect bolt torque, displacement, and temperature data through sensors, and generate a comprehensive dynamic deviation data set of torque deviation, displacement deviation, and temperature change;

[0048] S32. Dynamically adjust the local pheromone concentration according to the comprehensive dynamic deviation data set, and real-time update the pheromone value of a single bolt node, so that the local pheromone concentration is inversely proportional to the current torque deviation;

[0049] S33. Dynamically calculate the average value of bolt stress deviation in each region, update the regional pheromone concentration according to the real-time change of regional deviation, and optimize the path planning within the region;

[0050] S34. Dynamically adjust the global pheromone concentration based on the global deviation data, correct the global stress uniformity deviation, and optimize the global path planning;

[0051] S35. Combine the dynamic perception function during the path planning process to generate a dynamically optimized fastening path.

[0052] Optionally, the S4 specifically includes:

[0053] S41. Real-time collect the current torque, displacement, and temperature data of the bolt through a sensor, compare it with the reference state, and generate real-time deviation data, including torque deviation, displacement deviation, and temperature change value;

[0054] S42. Construct a thermal expansion compensation model, and calculate the influence value of thermal expansion on the target torque according to the thermal expansion coefficient of the bolt material and the temperature change value:

[0055] ΔM temp (i) = k temp ·ΔT(i)·M target (i);

[0056] Where, ΔM temp (i) represents the target torque compensation value generated by the thermal expansion effect of the i-th bolt, k temp represents the thermal expansion coefficient of the bolt material, ΔT(i) represents the temperature change value of the i-th bolt, and M target (i) represents the initial target torque value of the i-th bolt;

[0057] S43. Construct a friction correction model, and calculate the friction correction value according to the current displacement and the friction correction coefficient:

[0058] ΔM friction (i) = k friction ·Δd(i);

[0059] Where, ΔM friction (i) represents the friction correction value of the i-th bolt, k friction represents the friction correction coefficient, and Δd(i) represents the displacement deviation of the i-th bolt;

[0060] S44. Integrate the thermal expansion compensation model and the friction correction model to dynamically calculate the adjusted target torque value:

[0061] M adjusted (i) = M target (i) + ΔM temp (i) + ΔM friction (i);

[0062] Where, M adjusted (i) represents the adjusted target torque value of the i-th bolt;

[0063] S45. Compare the current torque value and the adjusted target torque value in real time, and calculate the current deviation ΔM final (i):

[0064] ΔM final (i) = M current (i) - M adjusted (i);

[0065] where, M current (i) represents the current torque value of the i-th bolt;

[0066] If |ΔM final (i)| > ∈, trigger the deviation correction mechanism, and optimize the global bolt tightening sequence through the adjusted path planning, where ∈ represents the deviation setting threshold.

[0067] Optionally, the S5 specifically includes:

[0068] S51. Collect the torque, displacement and temperature data of the bolts after tightening, and input the data into the finite element analysis model to simulate the actual stress state of the bolts;

[0069] S52. Calculate the current stress and torque states of each bolt based on the finite element model, evaluate the torque error of a single bolt, and mark the abnormal bolts that exceed the preset allowable range;

[0070] S53. Divide the bolt group into regions, calculate the stress distribution and uniformity within each region, and identify the regions where the stress distribution deviation is greater than the preset threshold deviation;

[0071] S54. Analyze the stress distribution of the bolts in the global range, calculate the global stress uniformity deviation, and evaluate whether the overall tightening effect meets the target requirements;

[0072] S55. Combine the bolt stress state with the analysis results of the regional and global stress uniformity to generate a verification report, including data on the torque error of a single bolt, regional stress uniformity, and global stress deviation.

[0073] Optionally, the S6 specifically includes:

[0074] S61. Collect the real-time torque, displacement and stress distribution data of the bolts during the tightening process, and compare them with the preset target parameters to identify the abnormal bolt nodes that deviate from the allowable range;

[0075] S62. Through the anomaly detection mechanism, classify the identified abnormal nodes, including single bolt torque deviation anomaly, regional stress distribution anomaly, and global stress uniformity anomaly;

[0076] S63. Dynamically adjust the pheromone distribution according to the type and distribution range of abnormal nodes, where local pheromone is used to compensate for a single abnormal node, regional pheromone is used to optimize the paths within the abnormal area, and global pheromone is used to rebalance the overall path planning;

[0077] S64. Real-time monitor the execution of the adjusted path, verify whether the abnormal nodes are effectively corrected. If there are still abnormalities, further optimize the pheromone distribution and repeat the path planning until the tightening path meets the preset target requirements.

[0078] The variable pitch bearing bolt automatic tightening system according to the embodiment of the present invention includes the following modules:

[0079] The data acquisition module is used to collect the geometric distribution information, target torque range and material property parameters of the bolt group, perform preprocessing, and construct a stress distribution data set;

[0080] The pheromone field construction module is used to construct pheromone fields for the local torque layer, regional stress layer and global equilibrium layer based on the stress distribution data set, and generate an initial tightening path;

[0081] The dynamic optimization module is used to dynamically update the pheromone distribution according to the torque, displacement and temperature data collected in real time, and optimize the tightening path;

[0082] The compensation and correction module is used to adjust the bolt target torque value based on the thermal expansion compensation model and the friction correction model;

[0083] The verification and analysis module is used to verify the tightening result based on finite element analysis, and evaluate the torque error, stress uniformity and global stress distribution;

[0084] The abnormal detection module is used to identify abnormal bolts, adjust the pheromone distribution and re-plan the tightening path;

[0085] The report generation module is used to generate a tightening task report and maintenance suggestions.

[0086] The beneficial effects of the present invention are:

[0087] First of all, the present invention adopts an improved ant foraging algorithm combined with a multi-layer pheromone field, which can dynamically sense environmental changes during the tightening process and optimize the tightening path. By constructing a local torque layer, a regional stress layer and a global equilibrium layer, the system realizes multi-objective comprehensive optimization from single-point accuracy to global uniformity, significantly improving the stress distribution uniformity and tightening efficiency of the bolt group.

[0088] Secondly, by collecting the torque, displacement, and temperature data of bolts in real time and combining with the thermal expansion compensation model and friction correction model, the present invention realizes the dynamic adjustment of the target torque, adapting to the influence of temperature changes and friction losses on the fastening accuracy in the industrial environment. The introduction of the dynamic perception function enables the system to adjust and optimize the strategy in real time under complex industrial scenarios, thus ensuring the high precision and high stability of the fastening process. Compared with the traditional fixed path planning method, the path optimization of the present invention has higher flexibility and adaptability, and can significantly reduce the stress unevenness problem caused by changes in material properties.

[0089] In addition, the present invention uses the finite element analysis model to verify the fastening results in real time, enabling the system to dynamically evaluate the torque error of a single bolt, the regional stress uniformity, and the global stress distribution deviation during the fastening process. Through the linkage with the fastening process, the finite element analysis not only provides immediate verification results but also provides a scientific basis for subsequent path planning and strategy optimization, avoiding the lag of traditional offline analysis. The anomaly detection mechanism further enhances the intelligence level of the system, which can identify and classify the bolt nodes with torque deviation and abnormal stress distribution in real time, and quickly correct the anomalies by dynamically adjusting the pheromone distribution and path planning, ensuring the reliability and stability of the system in complex tasks.

[0090] Finally, by generating a complete fastening task report, including the fastening path, torque distribution, stress uniformity evaluation, and anomaly correction records, the present invention provides important data support for subsequent maintenance and optimization. The report not only contains the execution results but also can provide maintenance suggestions based on verification analysis, providing a decision-making basis for the long-term stable operation and preventive maintenance of the equipment. This automated and intelligent fastening solution significantly improves the efficiency and quality of bolt fastening tasks in industrial scenarios and has broad application prospects and practical value. Description of the Drawings

[0091] The drawings are used to provide a further understanding of the present invention and constitute a part of the specification. They are used together with the embodiments of the present invention to explain the present invention and do not constitute a limitation to the present invention. In the drawings:

[0092] Figure 1 is the overall flowchart of the automated fastening method for the bolts of the pitch bearing proposed by the present invention;

[0093] Figure 2 is the structural schematic diagram of the automated fastening system for the bolts of the pitch bearing proposed by the present invention. Detailed Embodiments

[0094] Now, the present invention will be further described in detail with reference to the drawings. These drawings are all simplified schematic diagrams, only illustrating the basic structure of the present invention in a schematic way, so they only show the components related to the present invention.

[0095] Reference Figure 1 , the automated tightening method for pitch bearing bolts includes the following steps:

[0096] S1. Collect the geometric distribution information, target torque range, and material property parameters of the bolt group, perform preprocessing, and construct a stress distribution data set;

[0097] S2. Based on the stress distribution data set, use an improved ant foraging algorithm to construct a multi-layer pheromone field. The multi-layer pheromone field includes a local torque layer, a regional stress layer, and a global equilibrium layer, and generate an initial tightening path in combination with a dynamic perception function;

[0098] S3. Based on the bolt torque, displacement, and temperature data collected in real time, optimize the initial tightening path through pheromone update rules. The pheromone includes local pheromone, regional pheromone, and global pheromone, and generate a dynamically optimized tightening path;

[0099] S4. Based on a dynamic feedback mechanism, monitor the current torque, displacement, and temperature parameters of the bolts in real time, and adjust the bolt torque target by constructing a thermal expansion compensation model and a friction correction model;

[0100] S5. Use finite element analysis to verify the tightening result, evaluate the single bolt torque error, regional stress uniformity, and global stress distribution deviation, and generate a verification report;

[0101] S6. Combine an anomaly detection mechanism to identify bolt nodes with torque deviation or stress anomaly in real time, dynamically adjust the pheromone distribution, and re-plan the tightening path;

[0102] S7. Generate a complete tightening task report, including the tightening path, torque distribution, stress uniformity evaluation, and anomaly correction record, and provide maintenance suggestions.

[0103] In this embodiment, the S2 specifically includes:

[0104] S21. Extract the initial torque deviation and stress distribution parameters of the bolts from the stress distribution data set, and initialize the pheromone concentration of the local torque layer:

[0105]

[0106] Among them, τ local (i) represents the pheromone concentration of the local torque layer of the i-th bolt, M current (i) represents the current measured torque of the i-th bolt, M target (i) represents the target torque value of the i-th bolt;

[0107] S22. Based on the stress distribution data of the bolts within the region, calculate the regional stress mean and the target stress deviation, and initialize the pheromone concentration of the regional stress layer:

[0108]

[0109] Among them, τ region (r) represents the pheromone concentration of region r, and σ region (r) represents the average stress value of all bolts in region r, and σ target represents the stress value of the target region;

[0110] S23. Calculate the global stress uniformity deviation based on the global bolt stress distribution, and initialize the pheromone concentration of the global equilibrium layer:

[0111]

[0112] Among them, τ global represents the global pheromone concentration, and Δσ global represents the maximum deviation value of the global stress distribution;

[0113] S24. Combine the multi-layer pheromone field and define the path selection probability P of the ant from the i-th bolt to the j-th bolt ij :

[0114]

[0115] Among them, τ ij represents the pheromone concentration on the path from the i-th bolt to the j-th bolt, α, β, and γ represent weight factors, and τ ik represents the pheromone concentration on the path from the i-th bolt to the k-th bolt, allowed represents the set of next nodes, and τ local (k) represents the pheromone concentration of the local moment layer of the k-th bolt, and τ region (r k ) represents the pheromone concentration of region r k , η ij represents the index of path superiority when moving from the i-th bolt to the j-th bolt, and η ik represents the index of path superiority when moving from the i-th bolt to the k-th bolt:

[0116]

[0117] Among them, d ij represents the distance from the i-th bolt to the j-th bolt, and σ ij represents the stress distribution value from the i-th bolt to the j-th bolt;

[0118] D(t) = α1·f temp (t) + β1·f elastic (t) + γ1·f friction (t);

[0119] Among them, D(t) represents the dynamic perception function, and α1, β1, and γ1 represent weight coefficients;

[0120] f temp (t) = 1 + k temp ·(T(t) - T ref );

[0121] Among them, f temp (t) represents the temperature influence function, k temp represents the thermal expansion coefficient of the material, T ref represents the reference temperature, and T(t) represents the current ambient temperature;

[0122] f elastic (t) = 1 + k elastic ·ΔM;

[0123] Among them, f elastic (t) represents the elastic recovery function, k elastic represents the elastic recovery influence coefficient, and ΔM represents the torque deviation;

[0124] f friction (t) = 1 + k friction ·ΔF;

[0125] Among them, f friction (t) represents the friction change function, k friction represents the friction change influence coefficient, and ΔF represents the friction force change value;

[0126] S25. During the path selection process, pheromone update is performed on the paths completed by ants in each iteration. The pheromone update rule is determined by the pheromone decay coefficient and the increment:

[0127] τ ij (t + 1) = (1 - ρ)·τ ij + Δτ ij ;

[0128]

[0129] Among them, τ ij (t + 1) represents the pheromone concentration from the i-th bolt to the j-th bolt after update, ρ represents the pheromone decay coefficient, Δτ ij represents the pheromone increment, Q represents the pheromone enhancement constant, and L best represents the length of the current optimal path;

[0130] S26. Use the improved ant foraging algorithm to iterate multiple times until the path converges, and output the initial tightening path and bolt sequence optimized by combining the local torque layer, regional stress layer, and global equilibrium layer.

[0131] In this embodiment, S3 specifically includes:

[0132] S31. Collect bolt torque, displacement, and temperature data in real time through sensors, and generate a comprehensive dynamic deviation data set of torque deviation, displacement deviation, and temperature change;

[0133] S32. Dynamically adjust the local pheromone concentration according to the comprehensive dynamic deviation data set, and update the pheromone value of a single bolt node in real time, so that the local pheromone concentration is inversely proportional to the current torque deviation;

[0134] S33. Dynamically calculate the average value of bolt stress deviation in each area, update the area pheromone concentration according to the real-time change of area deviation, and optimize the path planning within the area;

[0135] S34. Dynamically adjust the global pheromone concentration based on the global deviation data, correct the global stress uniformity deviation, and optimize the global path planning;

[0136] S35. Combine the dynamic perception function during the path planning process to generate a dynamically optimized tightening path.

[0137] In this embodiment, S4 specifically includes:

[0138] S41. Collect the current torque, displacement, and temperature data of the bolt in real time through sensors, compare with the reference state, and generate real-time deviation data, including torque deviation, displacement deviation, and temperature change value;

[0139] S42. Construct a thermal expansion compensation model, and calculate the influence value of thermal expansion on the target torque according to the thermal expansion coefficient of the bolt material and the temperature change value:

[0140] ΔM temp (i) = k temp ·ΔT(i)·M target (i);

[0141] Where, ΔM temp (i) represents the target torque compensation value generated by the i-th bolt due to the thermal expansion effect, k temp represents the thermal expansion coefficient of the bolt material, ΔT(i) represents the temperature change value of the i-th bolt, and M target (i) represents the initial target torque value of the i-th bolt;

[0142] S43. Construct a friction correction model, and calculate the friction correction value according to the current displacement and the friction correction coefficient:

[0143] ΔM friction (i) = k friction ·Δd(i);

[0144] Among them, ΔM friction (i) represents the friction correction value of the i-th bolt, and k friction represents the friction correction coefficient, and Δd(i) represents the displacement deviation of the i-th bolt;

[0145] S44. Integrate the thermal expansion compensation model and the friction correction model to dynamically calculate the adjusted target torque value:

[0146] M adjusted (i) = M target (i) + ΔM temp (i) + ΔM friction (i);

[0147] Among them, M adjusted (i) represents the adjusted target torque value of the i-th bolt;

[0148] S45. Compare the current torque value with the adjusted target torque value in real time, and calculate the current deviation ΔM final (i):

[0149] ΔM final (i) = M current (i) - M adjusted (i);

[0150] Among them, M current (i) represents the current torque value of the i-th bolt;

[0151] If |ΔM final (i)| > ∈, trigger the deviation correction mechanism, and optimize the global bolt tightening sequence through adjusted path planning, where ∈ represents the deviation setting threshold.

[0152] In this embodiment, the S5 specifically includes:

[0153] S51. Collect the bolt torque, displacement, and temperature data after tightening, and input the data into the finite element analysis model to simulate the actual stress state of the bolts;

[0154] S52. Calculate the current stress and torque states of each bolt based on the finite element model, evaluate the torque error of a single bolt, and mark the abnormal bolts that exceed the preset allowable range;

[0155] S53. Divide the bolt group into regions, calculate the stress distribution and uniformity of the bolts in each region, and identify the regions where the stress distribution deviation is greater than the preset threshold deviation;

[0156] S54. Analyze the stress distribution of the bolts within the global range, calculate the global stress uniformity deviation, and evaluate whether the overall tightening effect meets the target requirements;

[0157] S55. Generate a verification report by combining the force states of the fastening bolts, the regional and global stress uniformity analysis results, including data on the torque error of a single bolt, regional stress uniformity, and global stress deviation.

[0158] In this embodiment, the S6 specifically includes:

[0159] S61. Collect the real-time torque, displacement, and stress distribution data of the bolts during the fastening process, compare them with the preset target parameters, and identify abnormal bolt nodes that deviate from the allowable range.

[0160] S62. Classify the identified abnormal nodes through an anomaly detection mechanism, including abnormal single-bolt torque deviation, abnormal regional stress distribution, and abnormal global stress uniformity.

[0161] S63. Dynamically adjust the pheromone distribution according to the type and distribution range of the abnormal nodes, where local pheromones are used to compensate for individual abnormal nodes, regional pheromones are used to optimize the paths within the abnormal regions, and global pheromones are used to rebalance the overall path planning.

[0162] S64. Real-time monitor the execution of the adjusted path, verify whether the abnormal nodes are effectively corrected. If there are still abnormalities, further optimize the pheromone distribution and repeat the path planning until the fastening path meets the preset target requirements.

[0163] Reference Figure 2 , the variable pitch bearing bolt automatic fastening system includes the following modules:

[0164] The data acquisition module is used to collect the geometric distribution information of the bolt group, the target torque range, and the material property parameters, perform preprocessing, and construct a stress distribution data set.

[0165] The pheromone field construction module is used to construct pheromone fields for the local torque layer, regional stress layer, and global equilibrium layer based on the stress distribution data set, and generate an initial fastening path.

[0166] The dynamic optimization module is used to dynamically update the pheromone distribution according to the torque, displacement, and temperature data collected in real time, and optimize the fastening path.

[0167] The compensation and correction module is used to adjust the bolt target torque value based on the thermal expansion compensation model and the friction correction model.

[0168] The verification and analysis module is used to verify the fastening results based on finite element analysis, and evaluate the torque error, stress uniformity, and global stress distribution.

[0169] The anomaly detection module is used to identify abnormal bolts, adjust the pheromone distribution, and re-plan the fastening path.

[0170] Report generation module for generating tightening task reports and maintenance recommendations.

[0171] Embodiment 1:

[0172] In order to verify the feasibility of the present invention in implementation, the present invention is applied to the task of tightening the bolts of the variable pitch bearings of large wind turbines in a certain wind farm. The wind farm is located in a coastal area and is affected by temperature changes, sea breeze humidity and mechanical vibrations all year round. The environment is complex and the tightening operation is difficult. The task requires bolt tightening of the variable pitch bearing with a diameter of 3 meters, involving 72 high-strength bolts. The target torque range of each bolt is 850-900Nm, and the global stress uniformity deviation is required to be no more than ±5%. Traditional manual tightening methods usually take more than 12 hours under similar conditions, and the tightening accuracy is difficult to meet the above requirements.

[0173] In this embodiment, the system of the present invention first obtains the geometric distribution information and material characteristic parameters (elastic modulus is 210 GPa, thermal expansion coefficient is 1.2×10 -5 / ℃, friction coefficient of 0.15) and target torque range. Combined with the actual temperature fluctuations in the coastal high humidity environment (temperature range of 15℃-35℃), the system constructed a real-time stress distribution dataset and initialized a multi-layer pheromone field. Through the pheromone field construction module, the system generated an initial tightening path containing a local torque layer, a regional stress layer, and a global equilibrium layer. The dynamic perception function takes into account the dynamic effects of temperature, friction changes, and elastic recovery in real time during the path planning process.

[0174] In the actual tightening process, the system combines the dynamic optimization module to analyze the collected real-time torque deviation, displacement changes and temperature fluctuations, and dynamically adjusts the target torque through the thermal expansion compensation model and the friction correction model. For example, in a certain tightening, the temperature gradually rose from the initial 20°C to 32°C, causing the actual torque deviation of the bolt to reach 12%, which could not be corrected in time by traditional methods. However, the compensation model of this system can accurately calculate the impact of thermal expansion on the torque and control the deviation within ±3%. In addition, in response to the fluctuation of the friction coefficient, the system adjusted the target torque in real time through the friction correction model, so that the final torque distribution is more uniform.

[0175] Through the finite element analysis module, the system verifies the stress distribution of bolts after tightening in real time. Data analysis shows that the torque error of a single bolt is controlled within ±2%, and after regional stress uniformity optimization, the uniformity deviation does not exceed ±4%, and the global stress distribution deviation decreases by 35%. Combined with the anomaly detection module, the system automatically identifies two abnormal bolt nodes during the path optimization process, and quickly corrects the anomaly through dynamic adjustment of the pheromone field and path replanning.

[0176] The system of the present invention shows a significant improvement in efficiency in this scenario. The complete bolt group tightening task is shortened from 12 hours by the traditional method to 6 hours, and the rework rate is reduced by 40%. The task report generated after tightening details the final torque of each bolt, the stress uniformity assessment, and the abnormal correction process, providing accurate reference data for subsequent maintenance.

[0177] Table 1 Comparison Table of Pitch Bearing Bolt Tightening Verification Results between Traditional Method and the Method of the Present Invention

[0178] Project Traditional method Method of the present invention Improvement effect Single bolt torque error range ±10％ ±2％ Reduce by 80% Regional stress uniformity deviation ±8％ ±4％ Reduce by 50% Global stress distribution deviation ±12％ ±6％ Reduce by 50% Number of abnormal bolt nodes 6 pieces 2 pieces Reduce by 67% Total fastening time (hours) 12 6 Shorten by 50% Rework rate 15％ 9％ Reduce by 40% Task report generation time (minutes) 45 10 Shorten by 77%

[0179] This embodiment verifies the feasibility and superiority of the present invention in pitch bearing bolt tightening in a complex industrial environment. By applying the present invention to the pitch bearing bolt group tightening task of a coastal wind farm, it fully demonstrates the excellent performance of dynamic optimization and feedback control under complex working conditions. The present invention effectively solves the problems of static path planning, difficult real-time adjustment of torque deviation, poor stress distribution uniformity, and low abnormal correction efficiency in the traditional method.

[0180] First, the present invention uses an improved ant foraging algorithm combined with a multi-layer pheromone field to generate a dynamically optimized initial tightening path, which not only significantly improves the path planning efficiency but also realizes the organic combination of single-point, regional, and global stress optimization. Second, by real-time collecting torque, displacement, and temperature data and using a thermal expansion compensation model and a friction correction model, the system can adapt to environmental temperature fluctuations and friction coefficient changes, effectively control torque deviation, and ensure the high precision and stability of bolt tightening.

[0181] The introduction of the finite element analysis module enables the full verification of the tightening results. The implementation results show that the present invention has significant advantages in single-bolt torque error control, regional stress uniformity optimization, and global stress distribution improvement. The torque error of a single bolt is reduced from ±10% of the traditional method to ±2%, the regional stress deviation is reduced by 50%, and the global stress distribution deviation is improved by 50%, which fully demonstrates the technological breakthrough of the present invention in precision control. In addition, the abnormal detection module can real-time identify and classify abnormal bolt nodes and quickly correct them through path re-planning, further improving the reliability and efficiency of the system.

[0182] From the perspective of task efficiency, the present invention shortens the overall tightening time from 12 hours of the traditional method to 6 hours, doubling the efficiency. The rework rate is reduced from 15% to 9%, and the task report generation time is shortened from 45 minutes to 10 minutes. These data indicate that the present invention not only significantly improves work efficiency but also greatly reduces the time cost and the need for manual intervention, providing strong support for efficient operation in actual industrial scenarios.

[0183] This embodiment fully demonstrates the superiority of the present invention in the tightening of bolts for pitch bearings, not only improving the tightening accuracy and efficiency, but also significantly enhancing the adaptability and reliability of the system through intelligent and dynamic control strategies, providing a practical solution for industrial scenarios such as wind power generation and high-precision machining.

[0184] The above is only a preferred specific embodiment of the present invention, but the protection scope of the present invention is not limited thereto. Any person skilled in the art within the technical scope disclosed by the present invention, according to the technical solution and inventive concept of the present invention, making equivalent substitutions or changes should be covered within the protection scope of the present invention.

Claims

1. An automated tightening method for pitch bearing bolts, characterized in that, It includes the following steps: S1. Collect the geometric distribution information, target torque range, and material property parameters of the bolt group, perform preprocessing, and construct a stress distribution data set; S2. Based on the stress distribution data set, use an improved ant foraging algorithm to construct a multi-layer pheromone field. The multi-layer pheromone field includes a local torque layer, a regional stress layer, and a global equilibrium layer, and generate an initial tightening path in combination with a dynamic perception function; S3. Based on the bolt torque, displacement, and temperature data collected in real time, optimize the initial tightening path through pheromone update rules. The pheromones include local pheromones, regional pheromones, and global pheromones, and generate a dynamically optimized tightening path; S4. Based on a dynamic feedback mechanism, monitor the current torque, displacement, and temperature parameters of the bolts in real time, and adjust the bolt torque target by constructing a thermal expansion compensation model and a friction correction model; S5. Use finite element analysis to verify the tightening result, evaluate the single bolt torque error, regional stress uniformity, and global stress distribution deviation, and generate a verification report; S6. Combine the anomaly detection mechanism to identify bolt nodes with torque deviation or stress anomaly in real time, dynamically adjust the pheromone distribution, and re-plan the tightening path; S7. Generate a complete tightening task report, including the tightening path, torque distribution, stress uniformity evaluation, and anomaly correction record, and provide maintenance suggestions.

2. The automated tightening method for pitch bearing bolts according to claim 1, wherein The specific content of S2 includes: S21. Extract the initial torque deviation and stress distribution parameters of the bolts from the stress distribution data set, and initialize the pheromone concentration of the local torque layer: Among them, τ local (i) represents the pheromone concentration of the local torque layer of the i-th bolt, M current (i) represents the currently measured torque of the i-th bolt, M target (i) represents the target torque value of the i-th bolt; S22. Based on the stress distribution data of the bolts within the region, calculate the regional stress mean and the target stress deviation, and initialize the pheromone concentration of the regional stress layer: Among them, τ region (r) represents the pheromone concentration of region r, σ region (r) represents the average stress value of all bolts within region r, σ target represents the stress value of the target region; S23. According to the global bolt stress distribution, calculate the global stress uniformity deviation, and initialize the pheromone concentration of the global equilibrium layer: Among them, τ global represents the global pheromone concentration, and Δσ global represents the maximum deviation value of the global stress distribution; S24. Combine multiple pheromone fields and define the path selection probability P for an ant to move from the i-th bolt to the j-th bolt ij : Among them, τ ij represents the pheromone concentration on the path from the i-th bolt to the j-th bolt, α, β, and γ represent weight factors, and τ ik represents the pheromone concentration on the path from the i-th bolt to the k-th bolt, allowed represents the set of next nodes, and τ local (k) represents the local moment layer pheromone concentration of the k-th bolt, and τ region (r k ) represents the pheromone concentration of region r k , η ij represents the index of path superiority when moving from the i-th bolt to the j-th bolt, and η ik represents the index of path superiority when the i-th bolt moves to the k-th bolt: where d ij represents the distance from the i-th bolt to the j-th bolt, and σ ij represents the stress distribution value from the i-th bolt to the j-th bolt; D(t) = α1·f temp (t) + β1·f elastic (t) + γ1·f friction (t); Among them, D(t) represents the dynamic perception function, and α1, β1, and γ1 represent weight coefficients; f temp f(t) = 1 + k temp ·(T(t) - T ref ); Among them, f temp (t) represents the temperature influence function, k temp represents the thermal expansion coefficient of the material, T ref represents the reference temperature, and T(t) represents the current ambient temperature; f elastic (t) = 1 + k elastic ·ΔM; Among them, f elastic (t) represents the elastic recovery function, k elastic represents the elastic recovery influence coefficient, and ΔM represents the moment deviation; f friction (t) = 1 + k friction ·ΔF; Among them, f friction (t) represents the friction change function, k friction represents the friction change influence coefficient, and ΔF represents the friction force change value; S25. During the path selection process, perform pheromone update for the path completed by the ants in each iteration. The pheromone update rule is determined by the pheromone decay coefficient and increment: τ ij (t + 1) = (1 - ρ)·τ ij + Δτ ij ; Among them, τ ij (t + 1) represents the pheromone concentration from the i-th bolt to the j-th bolt after update, ρ represents the pheromone decay coefficient, and Δτ ij represents the pheromone increment, Q represents the pheromone enhancement constant, and L best represents the length of the current optimal path; S26. Use the improved ant foraging algorithm to iterate multiple times until the path converges, and output the initial tightening path and bolt sequence optimized by combining the local torque layer, regional stress layer, and global equilibrium layer.

3. The automated tightening method for pitch bearing bolts according to claim 1, wherein, The specific content of S3 includes: S31. Collect bolt torque, displacement, and temperature data in real time through sensors, and generate a comprehensive dynamic deviation data set of torque deviation, displacement deviation, and temperature change; S32. According to the comprehensive dynamic deviation data set, dynamically adjust the local pheromone concentration, and update the pheromone value of a single bolt node in real time, so that the local pheromone concentration is inversely proportional to the current torque deviation; S33. Dynamically calculate the mean value of the bolt stress deviation within each region, and update the regional pheromone concentration according to the real-time change of the regional deviation to optimize the path planning within the region; S34. Based on the global deviation data, dynamically adjust the global pheromone concentration, correct the global stress uniformity deviation, and optimize the global path planning; S35. Combine the dynamic perception function during the path planning process to generate a dynamically optimized tightening path.

4. The automated tightening method for the pitch bearing bolts according to claim 1, characterized in that The specific content of S4 includes: S41. Collect the current torque, displacement, and temperature data of the bolt in real time through sensors, compare with the reference state, and generate real-time deviation data, including torque deviation, displacement deviation, and temperature change value; S42. Construct a thermal expansion compensation model, and calculate the influence value of thermal expansion on the target torque according to the thermal expansion coefficient of the bolt material and the temperature change value: ΔM temp (i) = k temp ·ΔT(i)·M target (i); Among them, ΔM temp (i) represents the target torque compensation value generated by the i-th bolt due to the thermal expansion effect, and k temp represents the thermal expansion coefficient of the bolt material, ΔT(i) represents the temperature change value of the i-th bolt, and M target (i) represents the initial target torque value of the i-th bolt; S43. Construct a friction correction model, and calculate the friction correction value according to the current displacement and the friction correction coefficient: ΔM friction (i) = k friction ·Δd(i); Among them, ΔM friction (i) represents the friction correction value of the i-th bolt, k friction represents the friction correction coefficient, and Δd(i) represents the displacement deviation of the i-th bolt; S44. Integrate the thermal expansion compensation model and the friction correction model to dynamically calculate the adjusted target torque value: M adjusted (i) = M target (i) + ΔM temp (i) + ΔM friction (i); Among them, M adjusted (i) represents the target torque value after adjustment of the i-th bolt; S45. Compare the current torque value and the adjusted target torque value in real time, and calculate the current deviation ΔM final (i): ΔM final (i) = M current (i) - M adjusted (i); Among them, M current (i) represents the current torque value of the i-th bolt; If |ΔM final (i)| > ∈, the deviation correction mechanism is triggered to optimize the global bolt tightening sequence by adjusting the path planning, where ∈ represents the deviation setting threshold.

5. The automated tightening method for the pitch bearing bolts according to claim 1, characterized in that, The specific steps of S5 are as follows: S51. Collect the torque, displacement, and temperature data of the bolt after tightening, and input the data into the finite element analysis model to simulate the actual stress state of the bolt; S52. Calculate the current stress and torque state of each bolt based on the finite element model, evaluate the torque error of a single bolt, and mark the abnormal bolts that exceed the preset allowable range; S53. Divide the bolt group into regions, calculate the stress distribution and uniformity of the bolts in each region, and identify the regions where the stress distribution deviation is greater than the preset threshold deviation; S54. Analyze the stress distribution of the bolts in the global range, calculate the global stress uniformity deviation, and evaluate whether the overall tightening effect meets the target requirements; S55. Combine the bolt stress state with the analysis results of regional and global stress uniformity to generate a verification report, including data on single-bolt torque error, regional stress uniformity, and global stress deviation.

6. The automated tightening method for the pitch bearing bolts according to claim 1, characterized in that The specific steps of S6 are as follows: S61. Collect the real-time torque, displacement, and stress distribution data of the bolt during the tightening process, compare with the preset target parameters, and identify the abnormal bolt nodes that deviate from the allowable range; S62. Through the anomaly detection mechanism, classify the identified abnormal nodes, including single-bolt torque deviation anomaly, regional stress distribution anomaly, and global stress uniformity anomaly; S63. Dynamically adjust the pheromone distribution according to the type and distribution range of the abnormal nodes, where local pheromone is used to compensate for a single abnormal node, regional pheromone is used to optimize the path within the abnormal region, and global pheromone is used to rebalance the overall path planning; S64. Real-time monitor the execution of the adjusted path, verify whether the abnormal nodes are effectively corrected, and if there are still anomalies, further optimize the pheromone distribution and repeat the path planning until the tightening path meets the preset target requirements.

7. Pitch bearing bolt automatic tightening system, which executes the pitch bearing bolt automatic tightening method described in any one of claims 1 to 6, characterized in that It includes the following modules: Data acquisition module, which is used to collect the geometric distribution information, target torque range, and material property parameters of the bolt group, perform preprocessing, and construct a stress distribution data set; Pheromone field construction module, which is used to construct pheromone fields for the local torque layer, regional stress layer, and global equilibrium layer based on the stress distribution data set, and generate an initial tightening path; Dynamic optimization module, which is used to dynamically update the pheromone distribution according to the torque, displacement, and temperature data collected in real time, and optimize the tightening path; Compensation and correction module, which is used to adjust the bolt target torque value based on the thermal expansion compensation model and the friction correction model; Verification and analysis module, which is used to verify the tightening result based on finite element analysis, and evaluate the torque error, stress uniformity, and global stress distribution; Anomaly detection module, which is used to identify abnormal bolts, adjust the pheromone distribution and re-plan the tightening path; Report generation module, which is used to generate tightening task reports and maintenance suggestions.

Images