An ai-based segmented blade connection safety monitoring method and system

Abstract

This invention relates to the field of wind power equipment assembly monitoring technology, and provides an AI-based method and system for safety monitoring of segmented blade connections. The method includes: analyzing the locking torque curve generated by the hydraulic locking device performing a locking operation when the hydraulic pressure is within a reasonable fluctuation range in samples from various temperature ranges; identifying the synchronization of the pulse oscillation response to hydraulic pressure fluctuations in the locking torque curve; and if the proportion of the synchronization response decreases from a certain temperature range, and the decreasing trend intensifies as the temperature decreases, then that temperature range is set as a designated temperature range. This invention introduces a temperature-sensitive locking torque response law identification mechanism, starting from the synchronization relationship between hydraulic pressure fluctuations and locking torque pulse oscillations, to dynamically evaluate the assembly quality of the segmented blade flange connection interface.

Description

Technical Field

[0001] This invention belongs to the field of wind power equipment assembly monitoring technology, and particularly relates to an AI-based method and system for monitoring the safety of segmented blade connections. Background Technology

[0002] During on-site assembly, segmented blades of wind turbines commonly employ flange connections, with hydraulic locking devices applying preload to the bolts to ensure reliable connections between blade segments. Existing assembly quality evaluation systems typically generate assembly scores based on indicators such as the achievement of the locking device's target torque, process execution records, and fit testing. They generally treat normal fluctuations in hydraulic pressure during the locking process as a routine phenomenon within the system, and as long as the fluctuations are within acceptable limits, they do not further analyze their impact on torque transmission behavior.

[0003] However, extensive assembly practice has shown that the flange connection interface exhibits significant temperature sensitivity in response to hydraulic pressure disturbances. Under normal temperature conditions, reasonable fluctuations in hydraulic pressure typically produce corresponding pulse oscillations on the locking torque curve, indicating that the torque is adequately transmitted to the flange interface. However, under low temperature conditions, due to the increased contact stiffness and decreased micro-adhesion at the blade connection interface, the same pressure fluctuations often fail to produce normal response oscillations on the torque curve, exhibiting weakened, delayed, or even absent responses.

[0004] If this "synchronous decrease in response" phenomenon is not identified during the assembly stage, it may lead to potential risks such as preload decay, uneven bolt load, and fretting wear on the flange surface, thereby affecting the overall operational safety of the machine. However, existing technologies do not provide an analytical mechanism for quantifying the relationship between hydraulic pressure fluctuations and locking torque response, nor do they automatically identify temperature-induced insufficient response problems based on historical assembly data, and they lack the ability to dynamically correct assembly scores. Summary of the Invention

[0005] The purpose of this invention is to provide an AI-based method and system for monitoring the safety of segmented blade connections, aiming to solve the problems mentioned in the background art.

[0006] This invention is implemented as follows: an AI-based segmented blade connection safety monitoring method, the method comprising:

[0007] When assembling the flange connection interface of the target segmented blade, obtain the assembly database and the initial assembly score of the current assembly process;

[0008] Several samples were selected from the assembly database that were consistent with the background of this assembly and the locking operation of the hydraulic locking device, but with different ambient temperatures. The samples were then grouped according to a preset temperature range.

[0009] Analyze the locking torque curve generated by the hydraulic locking device when the hydraulic pressure is within a reasonable fluctuation range in the samples of each temperature range, and identify the response synchronization of the pulse oscillation in the locking torque curve relative to the hydraulic pressure fluctuation. If the proportion of the response synchronization decreases from a certain temperature range and the decreasing trend intensifies as the temperature decreases, then set that temperature range as the specified temperature range.

[0010] Determine whether the current ambient temperature is within or below the specified temperature range. If so, set the sample corresponding to the current ambient temperature as the target sample and select a reference sample from the samples above the specified temperature range.

[0011] The new locking torque curve deviation features compared to the reference sample are extracted from the target sample, and a correction factor is generated based on the deviation features to correct the initial assembly score value.

[0012] As a further limitation of the technical solution of the present invention, the phrase "consistent with the background of the current assembly condition" means that the sample and the current assembly process are consistent in terms of blade type, flange connection interface structure, assembly sequence, locking position setting, and flange connection interface assembly posture conditions.

[0013] As a further limitation of the technical solution of the present invention, the fact that the plurality of samples are consistent with the locking operation of the hydraulic locking device in the current assembly process means that the plurality of samples are consistent with the current assembly process in terms of the operating parameter range of the hydraulic locking device, the target value of the applied locking torque, and the control strategy of the locking action.

[0014] As a further limitation of the technical solution of the present invention, the hydraulic pressure being within a reasonable fluctuation range means that during the locking operation, the instantaneous deviation of the output hydraulic pressure of the hydraulic locking device is within the allowable deviation range of the preset pressure target value, and the amplitude and rate of change of the deviation do not exceed the pressure disturbance threshold allowed by the hydraulic system design.

[0015] As a further limitation of the technical solution of this invention, the locking torque curve generated by the hydraulic locking device performing the locking operation when the hydraulic pressure is within a reasonable fluctuation range is analyzed in the samples of each temperature range, and the response synchronization of the pulse oscillation in the locking torque curve relative to the hydraulic pressure fluctuation is identified. If the proportion of the response synchronization decreases from a certain temperature range, and the decreasing trend intensifies as the temperature decreases, then the step of setting this temperature range as a specified temperature range includes:

[0016] The samples corresponding to each temperature range are analyzed sequentially, and the locking torque curve formed by the hydraulic locking device performing the locking operation is obtained under the premise that the hydraulic pressure in the sample is within a reasonable fluctuation range.

[0017] Identify the synchronization of the pulse oscillation in the locking torque curve with respect to the hydraulic pressure fluctuation. The synchronization refers to the fact that when a fluctuation in hydraulic pressure is detected, the locking torque curve exhibits a torque oscillation that is positively correlated with the amplitude of the hydraulic pressure fluctuation within the corresponding time window, and the ratio of the amplitude of the torque oscillation to the amplitude of the hydraulic pressure fluctuation reaches a preset response ratio threshold used to characterize the normal response level.

[0018] If it is found in a number of samples that, starting from a certain temperature range, the proportion of responses to synchronization in samples above that temperature range remains basically constant, while the proportion of responses to synchronization in samples at and below that temperature range gradually decreases and the decreasing trend intensifies as the temperature decreases, then that temperature range is set as the designated temperature range.

[0019] As a further limitation of the technical solution of this embodiment of the invention, the step of extracting the new locking torque curve deviation features from the target sample compared with the reference sample, and generating a correction factor based on the deviation features to correct the initial assembly score value includes:

[0020] Extract the locking torque curves corresponding to the reference sample and the target sample, and identify the torque data points of the target sample that do not meet the response synchronization condition within the time window of the corresponding pressure fluctuation event. Determine each torque data point that does not meet the response synchronization condition as a locking torque deviation feature.

[0021] For each deviation feature, the stress torque values ​​of the reference sample and the target sample within the same pressure fluctuation event time window are obtained, and the absolute value of the deviation of the torque value of the target sample from the torque value of the reference sample is calculated. A correction factor is generated based on the comprehensive result of the absolute values ​​of all deviations.

[0022] The initial assembly score is corrected by a correction factor to obtain the corrected assembly score. The corrected assembly score is then used for the assembly safety assessment, locking process quality judgment, or risk warning strategy implementation of the current flange connection interface.

[0023] As a further limitation of the technical solution of this embodiment of the invention, when correcting the initial assembly score value, a preset correction formula is used, and the correction formula is defined as follows:

[0024] ;

[0025] in, This refers to the revised assembly score. This refers to the initial assembly score. This represents the total number of deviations from the characteristic. This refers to the first [item] in the target sample. The torque value corresponding to each deviation feature. This refers to the first in the reference sample. The torque value corresponding to each deviation feature. This refers to the average of the absolute values ​​of all deviations. This refers to a preset correction strength coefficient, and satisfies... Greater than 0.

[0026] As a further limitation of the technical solution of the present invention, in the process of determining whether the current ambient temperature is within or below the specified temperature range, if the current ambient temperature is higher than the specified temperature range, the initial assembly score value is directly used as the assembly score value of this assembly process.

[0027] As a further limitation of the technical solution of this invention, an AI-based segmented blade connection safety monitoring system is provided, the system comprising:

[0028] The data acquisition module is used to acquire the assembly database and the initial assembly score of the current assembly process when assembling the target segmented blade at the flange connection interface.

[0029] The sample screening module is used to select several samples from the assembly database that are consistent with the background of this assembly and the locking operation of the hydraulic locking device but have different ambient temperatures, and to group the samples according to the preset temperature range.

[0030] The response analysis module is used to analyze the locking torque curve generated by the hydraulic locker performing the locking operation when the hydraulic pressure is within a reasonable fluctuation range in the samples of each temperature range, and to identify the response synchronization of the pulse oscillation in the locking torque curve relative to the hydraulic pressure fluctuation. If the proportion of the response synchronization decreases from a certain temperature range, and the decreasing trend intensifies as the temperature decreases, then that temperature range is set as the specified temperature range.

[0031] The temperature determination module is used to determine whether the current ambient temperature is within or below the specified temperature range. If so, the sample corresponding to the current ambient temperature is set as the target sample, and a reference sample is selected from the samples above the specified temperature range.

[0032] The scoring correction module is used to extract the new locking torque curve deviation features from the reference sample from the target sample, and generate a correction factor based on the deviation features to correct the initial assembly score value.

[0033] As a further limitation of the technical solution of the present invention, the phrase "consistent with the background of the current assembly condition" means that the sample and the current assembly process are consistent in terms of blade type, flange connection interface structure, assembly sequence, locking position setting, and flange connection interface assembly posture conditions.

[0034] Compared with the prior art, the present invention has the following beneficial effects:

[0035] This invention introduces a temperature-sensitive locking torque response pattern identification mechanism to dynamically evaluate the assembly quality of segmented blade flange connection interfaces by examining the response synchronization relationship between hydraulic pressure fluctuations and locking torque pulse oscillations. By constructing a historical assembly sample system spanning multiple temperature ranges, this invention can automatically identify latent characteristics of weakened mechanical response at the blade connection interface under low-temperature conditions and accordingly determine physically meaningful specified temperature ranges, enabling proactive assessment of potential assembly risks.

[0036] Furthermore, the deviation feature extraction and correction factor generation method based on the average deviation ratio proposed in this invention enable quantitative correction of the initial assembly score value according to the actual response under the target working condition, significantly improving the sensitivity and reliability of the scoring system. Through the above technical mechanism, this invention not only solves the problem of insufficient identification of temperature-induced response in the prior art, but also achieves a systematic improvement in the safety, locking quality, and risk warning capabilities of the assembly process, possessing broad engineering applicability and promotional value. Attached Figure Description

[0037] Figure 1 A flowchart of the method provided in the embodiments of the present invention;

[0038] Figure 2 This is a flowchart illustrating the method for determining a specified temperature range provided in an embodiment of the present invention;

[0039] Figure 3 This is a flowchart illustrating the process of correcting the initial assembly score value in the method provided by the embodiments of the present invention;

[0040] Figure 4 The application architecture diagram of the system provided in the embodiments of the present invention. Detailed Implementation

[0041] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the invention.

[0042] Figure 1 A flowchart of the method provided by an embodiment of the present invention is shown.

[0043] Specifically, an AI-based segmented blade connection safety monitoring method includes the following steps:

[0044] Step S100: When assembling the flange connection interface of the target segmented blade, obtain the assembly database and the initial assembly score of the current assembly process.

[0045] In this embodiment of the invention, when assembling the flange connection interface of the target segmented blade, it is necessary to first obtain the assembly database and the initial assembly score of the current assembly process. Segmented blades are commonly used in large wind power generation equipment or aviation equipment. Due to size, transportation, and installation limitations, the overall blade is divided into multiple segments, which are connected on-site via blade root flanges or transition flanges. The flange connection interface typically uses a multi-point bolt distribution method to achieve structural fixation. However, to ensure flange surface fit, preload distribution, and connection reliability, actual assembly often relies on hydraulic locking devices to perform the locking operation. Hydraulic locking devices have the characteristics of stable output torque, accurate torque feedback, and the ability to achieve multi-point synchronous control, and therefore have become a commonly used professional assembly equipment in the flange connection process of large structures.

[0046] Assembly databases typically originate from historical assembly records, debugging data, experimental data, and standardized process data accumulated by manufacturers over long-term assembly processes. This data can be automatically collected by the assembly monitoring system or manually entered after assembly. The database generally includes the following types of data: real-time hydraulic pressure curves and locking torque curves of the hydraulic locking device during the locking process, records of pressure fluctuation events, ambient temperature, assembly sequence and locking step information, target flange interface fit test results, and post-assembly quality inspection data. Data acquisition can be accomplished using pressure sensors, torque sensors, temperature and humidity acquisition modules integrated into the hydraulic locking device, flange fit measurement equipment, and assembly monitoring systems. These acquisition methods and equipment are all mature and readily available assembly monitoring configurations in the current technology.

[0047] The initial assembly score is used to conduct a basic evaluation of the assembly quality of the current flange connection interface. It can come from existing assembly quality evaluation systems, such as assembly score tables generated based on assembly process specifications, assembly torque completion, fit test results, and locking step completion, or from standard evaluation indicators commonly used in the wind power industry, such as "assembly quality grade judgment value" and "bolt locking completion score value". These scoring systems are all existing technologies that are maturely applied in industrial sites and can be directly used as the basic input parameters for the subsequent correction mechanism of this invention.

[0048] Furthermore, the AI-based segmented blade connection safety monitoring method also includes the following steps:

[0049] Step S200: Select several samples from the assembly database that are consistent with the current assembly working conditions and the locking operation of the hydraulic locking device but have different ambient temperatures, and group the samples according to the preset temperature range.

[0050] The phrase "consistent with the current assembly conditions" means that the sample and the current assembly process are consistent in terms of blade type, flange connection interface structure, assembly sequence, locking position setting, and flange connection interface assembly posture conditions.

[0051] The statement that the locking operation of the hydraulic locking device in the current assembly process is consistent with the current assembly process means that the samples are consistent with the current assembly process in terms of the operating parameter range of the hydraulic locking device, the target value of the applied locking torque, and the control strategy of the locking action.

[0052] In this embodiment of the invention, the core objective of step S200 is to select historical assembly samples from the existing assembly database that can be used for temperature sensitivity analysis, so as to identify the differences in the response characteristics of the flange connection interface under different temperature conditions, thereby determining whether the current assembly environment has potential adverse conditions.

[0053] The basis of this invention stems from the fact that, in the prior art, when a hydraulic locking device performs a locking operation, the hydraulic system experiences a certain degree of hydraulic pressure fluctuation. This is one of the normal operating states of a hydraulic system, typically caused by factors such as changes in oil flow resistance, valve opening and closing, and local pressure drop fluctuations. In traditional assembly quality evaluation systems, as long as the amplitude of the hydraulic pressure fluctuation falls within the specified allowable deviation range, the fluctuation is considered normal, and the evaluation system does not pay extra attention to it.

[0054] However, after analyzing a large amount of actual blade assembly data, those skilled in the art have found that when temperature conditions are favorable, i.e., when the material properties of the flange connection interface are within the design range, the blade flange contact area exhibits good elastic response capabilities. When the hydraulic pressure fluctuates instantaneously within a reasonable range, the locking torque curve generated by the hydraulic locking device performing the locking operation will show a pulse oscillation corresponding to the pressure fluctuation. This is a normal mechanical feedback behavior of the flange connection interface. Specifically, due to the micro-elastic compression of the flange contact interface, the compliance of the fitting process, and the stability of the coupling contact conditions, when a pressure disturbance occurs on the hydraulic side, the torque end will reflect the disturbance in an approximately synchronous and linearly proportional manner, thus forming a clearly identifiable pulse oscillation characteristic.

[0055] However, when the ambient temperature is low, the local stiffness of the interface materials (such as composite material layers, adhesive layers, metal inserts, etc.) of the blade flange connection increases significantly while the interface compliance decreases, resulting in a weakened or even delayed torque response to pressure disturbances. Under these conditions, even if the hydraulic pressure remains within the normal fluctuation range, the corresponding pulse oscillations in the locking torque curve may exhibit reduced amplitude, decreased proportion, or weakened synchronicity. This decrease in feedback capability reflects the adverse effects of low-temperature conditions on the installation process, such as insufficient interface adhesion, deterioration of local contact conditions, or insufficient fretting response. Therefore, to avoid such hidden quality risks, this invention proposes to perform a reduction correction on the initial assembly score value when the current temperature falls within the unfavorable range, in order to accurately reflect potential risks.

[0056] To achieve the above distinction, the screening of samples at different temperatures needs to ensure that "other influencing factors besides temperature are consistent." Therefore, this invention requires screening from the assembly database:

[0057] Samples consistent with the background of this assembly process include: blade type, flange connection interface structure, assembly sequence, locking position setting, and flange connection interface assembly posture conditions.

[0058] The sample that is consistent with this hydraulic locking operation has the same operating parameter range, target value of applied locking torque, and locking action control strategy.

[0059] The above screening criteria ensure that the only significant difference among samples is "ambient temperature," thus making the subsequent identification of response differences truly reflect the temperature factor, rather than differences caused by other operating conditions.

[0060] In addition, in practical applications, sample screening can further include, but is not limited to, the following auxiliary conditions: consistent type of assembly equipment; consistent batches of fasteners used or of the same quality grade; wind conditions (wind speed, humidity) in the assembly area controlled within the same range; and consistent installation team or construction process records to reduce human interference.

[0061] To facilitate temperature difference analysis, this invention divides the screened samples into preset temperature ranges. These temperature ranges can be based on temperature segmentation principles commonly used in the wind power industry, such as using 0℃ as a typical threshold for changes in structural material performance, or they can be divided using a fixed step size, such as dividing into ranges of 2℃ or 4℃. Each temperature range preferably corresponds to a representative sample, which is selected from all samples within that range based on evaluation criteria such as the completeness of quality records, data continuity, and sensor accuracy, thereby improving the representativeness of the analysis results.

[0062] It should be noted that the implementation of this invention relies on a sufficiently large and reliable historical assembly database. Only with a sufficient sample size, comprehensive temperature coverage, and standardized recorded parameters can the data-driven approach be used to identify the trend of temperature's influence on response synchronization. Therefore, this invention requires that the assembly database be constructed based on large-scale, multi-batch, and multi-climate assembly records generated from many years of actual wind power field applications to ensure the reliability and engineering feasibility of the method.

[0063] Furthermore, the AI-based segmented blade connection safety monitoring method also includes the following steps:

[0064] Step S300: Analyze the locking torque curve generated by the hydraulic locking device when the hydraulic pressure is within a reasonable fluctuation range in the samples of each temperature range, and identify the response synchronization of the pulse oscillation in the locking torque curve relative to the hydraulic pressure fluctuation. If the proportion of the response synchronization decreases from a certain temperature range, and the decreasing trend intensifies as the temperature decreases, then set that temperature range as the specified temperature range.

[0065] The hydraulic pressure being within a reasonable fluctuation range means that during the locking operation, the instantaneous deviation of the output hydraulic pressure of the hydraulic locking device is within the allowable deviation range of the preset pressure target value, and the amplitude and rate of change of the deviation do not exceed the pressure disturbance threshold allowed by the hydraulic system design.

[0066] Specifically, Figure 2 A flowchart for determining a specified temperature range is shown.

[0067] The process involves analyzing the locking torque curve generated by the hydraulic locking device during locking operations when the hydraulic pressure is within a reasonable fluctuation range in samples from each temperature range. It also involves identifying the synchronization of pulse oscillations in the locking torque curve relative to hydraulic pressure fluctuations. If the proportion of synchronization decreases from a certain temperature range, and this decreasing trend intensifies as the temperature decreases, then that temperature range is designated as a specific temperature range. The specific steps include:

[0068] Step S301: Analyze the samples corresponding to each temperature range in sequence, and on the premise that the hydraulic pressure in the sample is within a reasonable fluctuation range, obtain the locking torque curve formed by the hydraulic locking device performing the locking operation;

[0069] Step S302: Identify the synchronization of the pulse oscillation in the locking torque curve with respect to the hydraulic pressure fluctuation. The synchronization refers to the fact that when the hydraulic pressure fluctuation is detected, the locking torque curve shows a torque oscillation that is positively correlated with the amplitude of the hydraulic pressure fluctuation within the corresponding time window, and the ratio of the amplitude of the torque oscillation to the amplitude of the hydraulic pressure fluctuation reaches a preset response ratio threshold used to characterize the normal response level.

[0070] Step S303: If it is found in a number of samples that, starting from a certain temperature range, the proportion of response synchronization of samples above that temperature range remains basically constant, while the proportion of response synchronization of samples at and below that temperature range gradually decreases and the decreasing trend intensifies as the temperature decreases, then that temperature range is set as the specified temperature range.

[0071] In this embodiment of the invention, the core objective of step S300 is to: based on the sample data across temperature ranges selected in step S200, analyze the response characteristics of the hydraulic locking device during the locking operation to identify the differences in the response capability of the flange connection interface of the target segmented blade under different temperature conditions, and thus find the temperature range that leads to insufficient response feedback. In other words, this step aims to determine a critical temperature range, i.e., a "specified temperature range," that can distinguish between the "normal response temperature range" and the "low temperature range where the torque response begins to degrade," as the basis for subsequent scoring correction.

[0072] This analysis is based on the following engineering fact: When a hydraulic locking device performs a locking operation, its hydraulic pressure exhibits instantaneous fluctuations that are a superposition of random and systematic factors. As long as these fluctuations fall within the reasonable fluctuation range allowed by the system design, they are considered normal. Ideally, the flange connection interface should maintain good compliance, and its stress state should reflect changes in hydraulic pressure in real time. Therefore, the locking torque curve will show pulse oscillations that are almost synchronous with the pressure fluctuations. However, when the ambient temperature drops to a certain level, the local stiffness of the flange connection interface increases and the contact compliance decreases, making it unable to respond promptly or fully to disturbances from the hydraulic end, resulting in a decrease in the synchronicity of the torque oscillations. These phenomena can all be identified using the method of this invention.

[0073] In step S301, before analyzing the samples for each temperature range, it is necessary to first confirm that the hydraulic pressure in the samples is within a "reasonable fluctuation range." This is because pressure disturbances can only be considered normal dynamic inputs of the system when hydraulic pressure fluctuations are within the design allowable deviation range. Only then can the oscillations in the locking torque curve truly reflect the response characteristics of the flange connection interface. If the pressure fluctuations are too large or too rapid, exceeding the reasonable range, it will cause distortion in the torque curve, making it unusable for temperature response analysis. Therefore, only by extracting the torque curve under the premise of reasonable pressure fluctuations can the stability, comparability, and reference value of the analysis be guaranteed.

[0074] Step S302 is one of the core innovations of this invention. In this step, the invention digitizes, quantifies, and adaptively defines the "response synchronization status," systematically transforming the traditional experience-based response identification process into a calculable and thresholdable standard criterion. The pressure fluctuation amplitude of the hydraulic system may vary with factors such as the operating stage, locking position, and oil temperature, and the torque oscillation amplitude fed back from the flange connection interface will also change accordingly. Therefore, this invention does not use "whether oscillation occurs" as the response criterion, but proposes a general, robust, and adaptable standard—by comparing the ratio of the torque oscillation amplitude to the pressure fluctuation amplitude, it determines whether a preset response ratio threshold is reached; if so, it is determined that "response synchronization has been achieved."

[0075] This proportional criterion has three major advantages: it shields the influence of pressure disturbance amplitude differences under different locking conditions, making the judgment standard consistent across different samples; it can adapt to amplitude changes caused by differences in equipment performance, contact conditions, oil temperature, etc., thus not relying on absolute torque values; and it enables the system to automatically determine the boundary between normal and weak responses in a data-driven manner, making it highly practical for engineering applications.

[0076] The preset response ratio threshold can be obtained through large-sample statistics of historical normal temperature samples, or it can be set according to industry standards or the dynamic response index of the equipment manufacturer. Its value is usually in the empirical range of 0.55 to 0.85, which is used to characterize the minimum proportional relationship between torque oscillation and pressure fluctuation at the connection interface under normal response conditions.

[0077] In terms of implementation details, step S302 can employ signal processing methods such as window segmentation sampling, peak detection, time series fitting, and oscillation peak-to-amplitude ratio calculation to automatically extract the amplitude characteristics of pulse oscillations from the locking torque curve; simultaneously, it combines the pressure curve to extract the amplitude characteristics of corresponding pressure fluctuations. Preferably, the amplitudes are all characterized using relative change ratios (e.g., the increase ratio of the peak value relative to the reference value) to eliminate the influence of absolute values ​​under different operating conditions, making the response characteristics between samples in different temperature ranges more comparable. Subsequently, by comparing whether the ratio of the torque oscillation amplitude to the pressure fluctuation amplitude reaches a preset response ratio threshold, it is finally determined whether the response synchronization condition is met within the time window.

[0078] Step S303 is used to determine the specified temperature range based on statistical results across temperature ranges. In a large sample, it can be observed that above a certain temperature range, the flange connection interface exhibits sufficient compliance, and the proportion of synchronous response remains stable, unchanged with temperature variations, forming a normal response segment. However, when the temperature drops to a certain critical range, the stiffness of the flange connection interface increases significantly, leading to a decrease in the proportion of synchronous response, and this downward trend intensifies as the temperature continues to decrease. This indicates that the dynamic response capability of the flange connection interface degrades thereafter. Based on this objective trend, this invention identifies this critical temperature range as the "specified temperature range".

[0079] The setting of the specified temperature range plays a crucial role, as it directly corresponds to the core research point proposed in step S200, namely, identifying the response differences of the flange connection interface under different temperature conditions, and using this as the basis for judgment, making targeted corrections to the initial assembly score value under low temperature conditions, so that the final score can reflect the hidden risks caused by temperature.

[0080] Furthermore, the AI-based segmented blade connection safety monitoring method also includes the following steps:

[0081] Step S400: Determine whether the current ambient temperature is within or below the specified temperature range. If so, set the sample corresponding to the current ambient temperature as the target sample and select a reference sample from the samples above the specified temperature range.

[0082] In the process of determining whether the current ambient temperature is within or below the specified temperature range, if the current ambient temperature is higher than the specified temperature range, the initial assembly score value is directly used as the assembly score value for this assembly process.

[0083] Step S500: Extract the new locking torque curve deviation features from the target sample compared to the reference sample, and generate a correction factor based on the deviation features to correct the initial assembly score value.

[0084] Specifically, Figure 3A flowchart for correcting the initial assembly score value is shown.

[0085] The process of extracting the new locking torque curve deviation features from the reference sample from the target sample, and generating a correction factor based on the deviation features to correct the initial assembly score, specifically includes the following steps:

[0086] Step S501: Extract the locking torque curves corresponding to the reference sample and the target sample, and identify the torque data points of the target sample that do not meet the response synchronization condition within the time window of the corresponding pressure fluctuation event, and determine each torque data point that does not meet the response synchronization condition as the locking torque deviation feature.

[0087] Step S502: For each deviation feature, obtain the stress torque values ​​of the reference sample and the target sample within the time window of the same pressure fluctuation event, and calculate the absolute value of the deviation of the torque value of the target sample from the torque value of the reference sample. Generate a correction factor based on the comprehensive result of the absolute values ​​of all deviations.

[0088] Step S503: The initial assembly score value is corrected by a correction factor to obtain the corrected assembly score value. The corrected assembly score value is then used for the assembly safety assessment, locking process quality judgment, or risk warning strategy implementation of the current flange connection interface.

[0089] When correcting the initial assembly score, a preset correction formula is used, which is defined as follows:

[0090] ;

[0091] in, This refers to the revised assembly score. This refers to the initial assembly score. This represents the total number of deviations from the characteristic. This refers to the first [item] in the target sample. The torque value corresponding to each deviation feature. This refers to the first in the reference sample. The torque value corresponding to each deviation feature. This refers to the average of the absolute values ​​of all deviations. This refers to a preset correction strength coefficient, and satisfies... Greater than 0.

[0092] In this embodiment of the invention, the target sample is the sample corresponding to the current actual assembly environment temperature, while the reference sample is preferably selected from samples above the specified temperature range. Since samples in the high-temperature range are generally stable, their response to hydraulic pressure fluctuations during the locking operation of the hydraulic locking device is more consistent, and therefore can be considered a "standard reference under normal operating conditions." In this case, the differences between samples within the high-temperature range are small, and the sample with the most stable quality and the smallest curve disturbance within that temperature range can usually be selected as the reference sample to obtain a more representative comparison benchmark.

[0093] When the current ambient temperature is determined to be higher than the specified temperature range, the flange connection interface at higher temperatures usually has good assembly response characteristics, and the locking torque responds normally to pressure fluctuations. Therefore, there is no need to perform subsequent deviation feature extraction and correction factor generation processes. Instead, the initial assembly score value is directly used as the assembly score value for this assembly process.

[0094] In step S500, based on the aforementioned law that temperature causes a decrease in response synchronization, this invention proposes a novel and engineering-significant assembly quality correction mechanism by analyzing the new deviation characteristics of the target sample's locking torque curve compared to the reference sample. This process extracts the weak response behavior of the locking torque under low-temperature conditions, uses the deviation amplitude of the locking torque as a quantitative basis, and corrects the initial assembly score through mathematical calculations, making the final score more consistent with the actual assembly quality performance under adverse low-temperature conditions.

[0095] In step S501, the locking torque curves corresponding to the reference sample and the target sample during the locking operation are first extracted, and torque data points in the target sample that do not meet the response synchronization condition within the time window of the corresponding pressure fluctuation event are identified. Specifically, in each pressure fluctuation window, if the locking torque of the target sample fails to generate a pulse oscillation of sufficient amplitude, or the ratio of the oscillation amplitude to the pressure fluctuation amplitude does not reach the preset response ratio threshold, the torque data points in that time window are determined as locking torque deviation characteristics. These deviation characteristics typically correspond to increased material rigidity, decreased lubrication, or enhanced contact interface friction caused by low temperature, thus preventing the locking torque from responding to pressure disturbances in a timely or sufficient manner.

[0096] In step S502, for each deviation feature, the stress torque values ​​of the reference sample and the target sample within the same pressure fluctuation event time window are obtained. Typically, to avoid the influence of instantaneous noise, the average value of the torque values ​​within the window can be selected as the representative torque data. Then, the deviation magnitude is calculated by comparing the target sample torque value with the reference sample torque value using the method of "taking the absolute value of the relative deviation ratio". The absolute deviation magnitudes corresponding to multiple deviation features can be averaged, weighted averaged, or used other statistical synthesis methods to generate the final correction factor.

[0097] The significance of this correction process lies in the fact that the reference sample represents the ideal response at normal temperatures, while the weak response of the target sample at low temperatures directly causes the torque curve to deviate from the reference baseline. By calculating the absolute values ​​of these deviations as the source of correction factors, the adverse effects of low-temperature conditions on actual assembly quality can be accurately quantified. Compared to traditional correction methods that rely on human experience, this method is more digital, repeatable, and physically grounded.

[0098] In step S503, the initial assembly score is finally corrected using a correction factor. The correction method can be multiplying the initial score by a reduction coefficient, or using other mathematical correction methods, such as subtracting the correction factor multiplied by the weight value from the initial score, or using a nonlinear attenuation model. The addition, subtraction, multiplication, and division methods provided by this invention constitute an intuitive and effective implementation method, facilitating understanding and deployment by engineers. The corrected assembly score will be used for the assembly safety assessment of the current flange connection interface, the quality judgment of the locking process, or the implementation of risk warning strategies to reflect the true assembly reliability under low-temperature conditions.

[0099] To facilitate the demonstration of the specific implementation of the technical solution of this invention, the overall implementation process of this invention is described below with reference to a digital example. Assume that a wind turbine is assembled with flange connections in an environment of -10℃, and the initial assembly score obtained after testing is 90. Through the processing of steps S200 and S300, the specified temperature range is determined to be 0~5℃, and it is identified that the current ambient temperature is lower than this range, therefore, score correction is required.

[0100] Subsequently, in step S501, typical windows that fail to meet the response synchronization requirement within the time window of the corresponding pressure fluctuation event are identified from the target sample, and the average locking torque values ​​of the target sample and the reference sample are obtained in these windows respectively. For example, assuming three typical pressure fluctuation events that fail to meet the response synchronization requirement are identified in the target sample, their corresponding average locking torques are as follows:

[0101] (1) First non-compliance window: reference sample 100 N·m, target sample 90 N·m;

[0102] (2) Second non-compliance window: reference sample 120 N·m, target sample 100 N·m;

[0103] (3) Third non-compliant window: reference sample 110 N·m, target sample 95 N·m.

[0104] According to step S502, the absolute values ​​of the "relative deviation ratios" for the above three deviation characteristics are calculated: First window: (90–100)÷100=–0.10, i.e., absolute value 0.10. Second window: (100–120)÷120=–0.1667, i.e., absolute value approximately 0.167. Third window: (95–110)÷110=–0.1364, i.e., absolute value approximately 0.136. Averaging the three, the average deviation ratio is approximately 0.134.

[0105] Let the correction strength coefficient α = 0.2, then the correction factor is: 0.2 × 0.134 = 0.0402. The final corrected assembly score is: score = 90 × (1 – 0.0402) ≈ 86.38.

[0106] This correction process can accurately quantify the degree of insufficient response synchronization under low-temperature conditions, making the final score more consistent with the actual assembly quality performance.

[0107] The overall technical solution of this invention achieves intelligent and safe assessment of the assembly quality of segmented blade flange connections through temperature range division, response synchronization identification, deviation feature extraction, and scoring correction calculation. The weakening mechanism of locking torque response discovered in this invention under low-temperature conditions is an important engineering principle. Utilizing this mechanism through digital modeling can significantly improve the accuracy of assembly safety assessment, representing a technological improvement with outstanding substantive features and significant progress. This invention has broad application prospects in wind turbine manufacturing, operation and maintenance, low-temperature assembly monitoring of large equipment, and health assessment of connection structures under extreme environments.

[0108] Furthermore, Figure 4 An application architecture diagram of the system provided in an embodiment of the present invention is shown.

[0109] In another preferred embodiment of the present invention, an AI-based segmented blade connection safety monitoring system includes:

[0110] The data acquisition module 100 is used to acquire the assembly database and the initial assembly score of the current assembly process when assembling the flange connection interface of the target segmented blade.

[0111] Furthermore, the AI-based segmented blade connection safety monitoring system also includes:

[0112] The sample screening module 200 is used to select several samples from the assembly database that are consistent with the background of the current assembly and the locking operation of the hydraulic locking device but have different ambient temperatures, and to group the samples according to a preset temperature range.

[0113] The phrase "consistent with the current assembly conditions" means that the sample and the current assembly process are consistent in terms of blade type, flange connection interface structure, assembly sequence, locking position setting, and flange connection interface assembly posture conditions.

[0114] Furthermore, the AI-based segmented blade connection safety monitoring system also includes:

[0115] The response analysis module 300 is used to analyze the locking torque curve generated by the hydraulic locker performing the locking operation when the hydraulic pressure is within a reasonable fluctuation range in the samples of each temperature range, and to identify the response synchronization of the pulse oscillation in the locking torque curve relative to the hydraulic pressure fluctuation. If the proportion of the response synchronization decreases from a certain temperature range, and the decreasing trend intensifies as the temperature decreases, then that temperature range is set as the specified temperature range.

[0116] Furthermore, the AI-based segmented blade connection safety monitoring system also includes:

[0117] The temperature determination module 400 is used to determine whether the current ambient temperature is within or below a specified temperature range. If so, the sample corresponding to the current ambient temperature is set as the target sample, and a reference sample is selected from the samples above the specified temperature range.

[0118] Furthermore, the AI-based segmented blade connection safety monitoring system also includes:

[0119] The scoring correction module 500 is used to extract the new locking torque curve deviation features from the target sample compared with the reference sample, and generate a correction factor based on the deviation features to correct the initial assembly score value.

[0120] It should be understood that although the steps in the flowcharts of the various embodiments of the present invention are shown sequentially according to the arrows, these steps are not necessarily executed in the order indicated by the arrows. Unless explicitly stated herein, there is no strict order restriction on the execution of these steps, and they can be executed in other orders. Moreover, at least some steps in the various embodiments may include multiple sub-steps or multiple stages. These sub-steps or stages are not necessarily completed at the same time, but can be executed at different times. The execution order of these sub-steps or stages is not necessarily sequential, but can be performed alternately or in turn with other steps or at least a portion of the sub-steps or stages of other steps.

[0121] Those skilled in the art will understand that all or part of the processes in the methods of the above embodiments can be implemented by a computer program instructing related hardware. The program can be stored in a non-volatile computer-readable storage medium, and when executed, it can include the processes of the embodiments of the above methods. Any references to memory, storage, databases, or other media used in the embodiments provided in this application can include non-volatile and / or volatile memory. Non-volatile memory can include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM), or flash memory. Volatile memory can include random access memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in various forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), dual data rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), synchronous link DRAM (SLDRAM), Rambus direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM), etc.

[0122] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0123] The embodiments described above are merely illustrative of several implementations of the present invention, and while the descriptions are specific and detailed, they should not be construed as limiting the scope of the present invention. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of the present invention, and these modifications and improvements all fall within the scope of protection of the present invention. Therefore, the scope of protection of this patent should be determined by the appended claims.

[0124] 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, and improvements made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A segmented blade connection safety monitoring method based on AI, characterized in that, The method includes: When assembling the flange connection interface of the target segmented blade, obtain the assembly database and the initial assembly score of the current assembly process; Several samples were selected from the assembly database that were consistent with the background of this assembly and the locking operation of the hydraulic locking device, but with different ambient temperatures. The samples were then grouped according to a preset temperature range. Analyze the locking torque curve generated by the hydraulic locking device when the hydraulic pressure is within a reasonable fluctuation range in the samples of each temperature range, and identify the response synchronization of the pulse oscillation in the locking torque curve relative to the hydraulic pressure fluctuation. If the proportion of the response synchronization decreases from a certain temperature range and the decreasing trend intensifies as the temperature decreases, then set that temperature range as the specified temperature range. Determine whether the current ambient temperature is within or below the specified temperature range. If so, set the sample corresponding to the current ambient temperature as the target sample and select a reference sample from the samples above the specified temperature range. The new locking torque curve deviation features compared to the reference sample are extracted from the target sample, and a correction factor is generated based on the deviation features to correct the initial assembly score value.

2. The AI-based segmented blade connection safety monitoring method according to claim 1, characterized in that, The phrase "consistent with the current assembly conditions" means that the sample and the current assembly process are consistent in terms of blade type, flange connection interface structure, assembly sequence, locking position setting, and flange connection interface assembly posture conditions.

3. The AI-based segmented blade connection safety monitoring method according to claim 1, characterized in that, The statement that the locking operation of the hydraulic locking device in the current assembly process is consistent with the current assembly process means that the samples are consistent with the current assembly process in terms of the operating parameter range of the hydraulic locking device, the target value of the applied locking torque, and the control strategy of the locking action.

4. The AI-based segmented blade connection safety monitoring method according to claim 1, characterized in that, The hydraulic pressure being within a reasonable fluctuation range means that during the locking operation, the instantaneous deviation of the output hydraulic pressure of the hydraulic locking device is within the allowable deviation range of the preset pressure target value, and the amplitude and rate of change of the deviation do not exceed the pressure disturbance threshold allowed by the hydraulic system design.

5. The AI-based segmented blade connection safety monitoring method according to claim 4, characterized in that, Analyzing the locking torque curves generated by the hydraulic locking device during locking operations when the hydraulic pressure is within a reasonable fluctuation range in the samples of each temperature range, and identifying the synchronization of the pulse oscillations in the locking torque curve with respect to the hydraulic pressure fluctuations, if the proportion of the synchronization decreases from a certain temperature range, and the decreasing trend intensifies as the temperature decreases, then setting that temperature range as a specified temperature range includes the following steps: The samples corresponding to each temperature range are analyzed sequentially, and the locking torque curve formed by the hydraulic locking device performing the locking operation is obtained under the premise that the hydraulic pressure in the sample is within a reasonable fluctuation range. Identify the synchronization of the pulse oscillation in the locking torque curve with respect to the hydraulic pressure fluctuation. The synchronization refers to the fact that when a fluctuation in hydraulic pressure is detected, the locking torque curve exhibits a torque oscillation that is positively correlated with the amplitude of the hydraulic pressure fluctuation within the corresponding time window, and the ratio of the amplitude of the torque oscillation to the amplitude of the hydraulic pressure fluctuation reaches a preset response ratio threshold used to characterize the normal response level. If it is found in a number of samples that, starting from a certain temperature range, the proportion of responses to synchronization in samples above that temperature range remains basically constant, while the proportion of responses to synchronization in samples at and below that temperature range gradually decreases and the decreasing trend intensifies as the temperature decreases, then that temperature range is set as the designated temperature range.

6. The AI-based segmented blade connection safety monitoring method according to claim 5, characterized in that, The steps of extracting the new locking torque curve deviation features compared to the reference sample from the target sample, and generating a correction factor based on the deviation features to correct the initial assembly score include: Extract the locking torque curves corresponding to the reference sample and the target sample, and identify the torque data points of the target sample that do not meet the response synchronization condition within the time window of the corresponding pressure fluctuation event. Determine each torque data point that does not meet the response synchronization condition as a locking torque deviation feature. For each deviation feature, the stress torque values ​​of the reference sample and the target sample within the same pressure fluctuation event time window are obtained, and the absolute value of the deviation of the torque value of the target sample from the torque value of the reference sample is calculated. A correction factor is generated based on the comprehensive result of the absolute values ​​of all deviations. The initial assembly score is corrected by a correction factor to obtain the corrected assembly score. The corrected assembly score is then used for the assembly safety assessment, locking process quality judgment, or risk warning strategy implementation of the current flange connection interface.

7. The AI-based segmented blade connection safety monitoring method according to claim 6, characterized in that, When correcting the initial assembly score, a preset correction formula is used, which is defined as follows: ； in, This refers to the revised assembly score. This refers to the initial assembly score. This represents the total number of deviations from the characteristic. This refers to the first [item] in the target sample. The torque value corresponding to each deviation feature. This refers to the first in the reference sample. The torque value corresponding to each deviation feature. This refers to the average of the absolute values ​​of all deviations. This refers to a preset correction strength coefficient, and satisfies... Greater than 0.

8. The AI-based segmented blade connection safety monitoring method according to claim 1, characterized in that, In the process of determining whether the current ambient temperature is within or below the specified temperature range, if the current ambient temperature is higher than the specified temperature range, the initial assembly score value is directly used as the assembly score value for this assembly process.

9. An AI-based segmented blade connection safety monitoring system, characterized in that, The system includes: The data acquisition module is used to acquire the assembly database and the initial assembly score of the current assembly process when assembling the target segmented blade at the flange connection interface. The sample screening module is used to select several samples from the assembly database that are consistent with the background of this assembly and the locking operation of the hydraulic locking device but have different ambient temperatures, and to group the samples according to the preset temperature range. The response analysis module is used to analyze the locking torque curve generated by the hydraulic locker performing the locking operation when the hydraulic pressure is within a reasonable fluctuation range in the samples of each temperature range, and to identify the response synchronization of the pulse oscillation in the locking torque curve relative to the hydraulic pressure fluctuation. If the proportion of the response synchronization decreases from a certain temperature range, and the decreasing trend intensifies as the temperature decreases, then that temperature range is set as the specified temperature range. The temperature determination module is used to determine whether the current ambient temperature is within or below the specified temperature range. If so, the sample corresponding to the current ambient temperature is set as the target sample, and a reference sample is selected from the samples above the specified temperature range. The scoring correction module is used to extract the new locking torque curve deviation features from the reference sample from the target sample, and generate a correction factor based on the deviation features to correct the initial assembly score value.

10. The AI-based segmented blade connection safety monitoring system according to claim 9, characterized in that, The phrase "consistent with the current assembly conditions" means that the sample and the current assembly process are consistent in terms of blade type, flange connection interface structure, assembly sequence, locking position setting, and flange connection interface assembly posture conditions.

Images