Quality monitoring system and method for the production of insulating side-expanded bushings

Abstract

This invention discloses a quality monitoring system and method for the production of insulated side-expanded sleeves, relating to the field of pipe quality monitoring. The method includes: S1: acquiring ultrasonic wall thickness, infrared temperature, and capacitance signals, as well as process parameters; S2: normalizing the signals to generate enhanced signals and calculating local gradient anomaly indices; S3: calculating the flaring stress index based on thickness deviation and process parameters; S4: calculating coupling non-uniformity based on the local gradient anomaly indices; S5: calculating a comprehensive quality factor based on the local gradient anomaly indices, flaring stress index, coupling non-uniformity, and process parameter deviations; S6: generating anomaly identifiers based on the dynamic deviation rate of the comprehensive quality factor and an adaptive threshold; S7: mapping the temporal characteristics corresponding to the anomaly identifiers to spatial coordinates and outputting defect distribution information; S8: generating process parameter adjustment amounts based on the defect distribution information and feeding them back to the control system.

Description

Technical Field

[0001] This invention relates to the field of pipe quality monitoring, specifically to a quality monitoring system and method for the production of insulating side-expanding sleeves. Background Technology

[0002] Insulating side-expanding bushings are widely used in power equipment, chemical pipelines, and high-voltage transmission systems. Their forming quality directly affects the sealing performance and service life of the product. Current production processes typically rely on single sensors or manual sampling to monitor bushing quality, which suffers from insufficient monitoring dimensions, low accuracy in defect location, and lagging process adjustments. Especially in the flaring forming stage, factors such as wall thickness deviation, temperature fluctuations, and abnormal dielectric properties work together, making it difficult for traditional methods to achieve real-time, comprehensive, and adaptive monitoring of the production process.

[0003] In recent years, the development of multimodal sensing and intelligent data processing technologies has provided new ideas for quality monitoring in bushing production. However, existing solutions still have shortcomings in multi-source signal fusion, defect spatial mapping, and process closed-loop optimization, making it difficult to effectively solve the quality fluctuation problem caused by multi-factor coupling. Therefore, there is an urgent need to propose a quality monitoring method and system that can fuse multimodal signals, achieve accurate defect localization, and possess closed-loop process optimization capabilities to improve the stability and finished product quality of insulating side-expanded bushing production. Summary of the Invention

[0004] Based on the shortcomings of the prior art described above, the purpose of this invention is to provide a quality monitoring system and method for the production of insulating side expansion bushings, so as to solve the above-mentioned technical problems.

[0005] To achieve the above objectives, the present invention provides the following technical solution: a quality monitoring method for the production of insulating side expansion bushings, comprising:

[0006] S1: Collect multimodal sensing signals and process parameters during the production process. The multimodal sensing signals include ultrasonic wall thickness signals, infrared temperature signals, and capacitance signals.

[0007] S2: Normalize the multimodal sensing signals to generate enhanced normalized signals, and calculate the local gradient anomaly index of each mode based on the enhanced normalized signals;

[0008] S3: Calculate the flaring stress index based on the deviation between the ultrasonic wall thickness signal and the reference wall thickness value, combined with process parameters;

[0009] S4: Based on the local gradient anomaly index, perform coupling analysis on different modal sensing signals and calculate the multimodal coupling nonuniformity.

[0010] S5: Calculate the comprehensive quality factor based on the local gradient anomaly index, the flaring stress index, and the multimodal coupling nonuniformity, combined with the process parameter deviation.

[0011] S6: Based on the dynamic deviation rate of the comprehensive quality factor and the adaptive threshold, perform anomaly detection on the production status and generate anomaly labels;

[0012] S7: Map the multi-source temporal features corresponding to the anomaly identifier to the spatial coordinates of the casing, and output the spatial distribution information of the defect;

[0013] S8: Generate process parameter adjustment amounts based on the spatial distribution information of defects, and feed these adjustment amounts back to the production control system for closed-loop optimization control of the production process.

[0014] The present invention is further configured such that S2 includes:

[0015] The ultrasonic wall thickness signal, infrared temperature signal, and capacitance signal are independently normalized.

[0016] The signal sequences within the preset historical time window of each sensing mode are subjected to fractional power weighting, and the current signal normalization result of each sensing mode is fused with the corresponding historical signal weighting result to generate an enhanced normalized signal.

[0017] Based on the enhanced normalized signals of each sensing mode, the local gradient anomaly index of each sensing mode is generated by calculating the ratio of signal energy change to signal change intensity within a preset sliding time window.

[0018] The present invention is further configured such that S3 includes:

[0019] Based on the enhanced normalized ultrasonic wall thickness signal and the corresponding reference wall thickness value at each measurement location, the wall thickness deviation between the current ultrasonic wall thickness signal and the reference wall thickness value is calculated, and the wall thickness deviation is nonlinearly amplified.

[0020] The process difference between real-time process parameters and local equivalent process parameters at each measurement location is obtained, and the process difference is processed by nonlinear adjustment.

[0021] The local stress components at each measurement location are obtained by normalizing the nonlinearly amplified wall thickness deviation and the nonlinearly adjusted process difference.

[0022] The local stress components at each measurement location are aggregated to generate the flared stress index.

[0023] The present invention is further configured such that S4 includes:

[0024] Based on the local gradient anomaly index corresponding to each sensing mode, calculate the anomaly difference between the local gradient anomaly indices of any two different sensing modes, and perform nonlinear adjustment processing on the anomaly difference.

[0025] The local gradient anomaly indices of two different sensing modes are nonlinearly weighted and combined, and the anomaly difference quantity adjusted by nonlinearity is used to normalize the weighted combination result to obtain the interactive coupling components corresponding to the two sensing modes.

[0026] The interactive coupling components of all sensing mode pairs are aggregated to generate multimodal coupling nonuniformity.

[0027] The present invention is further configured such that S5 includes:

[0028] The local gradient anomaly indices of each sensing mode are summed, and the summation result is subjected to a first power transformation.

[0029] The flaring stress index is subjected to a second power transformation, and the results of the first power transformation and the second power transformation are fused to generate a signal fusion factor.

[0030] A third power transformation and offset processing are performed on the multimodal coupling nonuniformity to generate a coupling influence factor.

[0031] Calculate the absolute deviation between the real-time process parameters and the optimal process parameters, and perform exponential mapping on the absolute deviation to generate a process deviation penalty term;

[0032] The signal fusion factor, coupling influence factor, and process deviation penalty term are multiplied together to generate a comprehensive quality factor.

[0033] The present invention is further configured such that S6 includes:

[0034] The dynamic deviation rate is calculated based on the continuous change of the comprehensive quality factor within the preset historical time window and the continuous change of the real-time process parameters within the preset historical time window.

[0035] Calculate the adaptive threshold based on the current comprehensive quality factor, dynamic deviation rate, and multimodal coupling inhomogeneity.

[0036] The current comprehensive quality factor is compared with the adaptive threshold. If the current comprehensive quality factor is less than the adaptive dynamic threshold, an anomaly flag is generated.

[0037] The present invention is further configured such that S7 includes:

[0038] Obtain the local gradient anomaly index and multimodal coupling nonuniformity of each sensing mode at the time point corresponding to the anomaly identifier, as well as the enhanced normalized signal and corresponding reference signal of each sensing mode in the sleeve space coordinates;

[0039] Calculate the signal deviation between the enhanced normalized signal of each sensing mode and the corresponding reference signal at each spatial coordinate, and perform nonlinear adjustment processing on the signal deviation.

[0040] The local gradient anomaly index is nonlinearly amplified, and the nonlinearly adjusted signal deviation is used to normalize the nonlinearly amplified local gradient anomaly index to obtain the independent contribution of each sensing mode in each spatial coordinate.

[0041] The independent contributions of all sensing modes are summed to obtain the spatial contribution aggregate value; the spatial contribution aggregate value is then fused with the multimodal coupling inhomogeneity to generate the spatial mapping index.

[0042] Based on the magnitude and distribution of the spatial mapping index on each spatial coordinate, the spatial location and severity of the defect are determined, and the spatial distribution information of the defect is output.

[0043] The present invention is further configured such that S8 includes:

[0044] Based on the spatial mapping index corresponding to the spatial distribution information of defects, the spatial mapping index on all spatial coordinates is summed to generate the cumulative defect severity.

[0045] The current process parameters are nonlinearly transformed to generate process state adjustment factors.

[0046] The cumulative defect severity is weighted and modulated using a process state adjustment factor to generate process parameter adjustment amounts.

[0047] The process parameter adjustment is added to the current process parameter to generate an optimized process parameter setting value;

[0048] The optimized process parameter setpoints are fed back to the production control system to achieve closed-loop optimization control.

[0049] The present invention is further configured such that the multimodal sensing signal is acquired by a distributed sensor array, the sensor array being arranged along the axial and circumferential directions of the sleeve.

[0050] This invention also provides a quality monitoring system for the production of insulating side expansion bushings, the system comprising:

[0051] Signal acquisition module: used to acquire multimodal sensing signals and process parameters during the production process. The multimodal sensing signals include ultrasonic wall thickness signals, infrared temperature signals, and capacitance signals.

[0052] Normalization anomaly module: used to normalize multimodal sensing signals to generate enhanced normalized signals, and calculate local gradient anomaly indices for each mode based on the enhanced normalized signals;

[0053] Flaring stress assessment module: used to calculate the flaring stress index based on the deviation between the ultrasonic wall thickness signal and the reference wall thickness value, combined with process parameters;

[0054] Coupling Analysis Module: Used to perform coupling analysis on different modal sensing signals based on local gradient anomaly indices and calculate multimodal coupling nonuniformity;

[0055] The comprehensive quality assessment module is used to calculate the comprehensive quality factor based on local gradient anomaly indices, flaring stress index, and multimodal coupling nonuniformity, combined with process parameter deviations.

[0056] Anomaly detection module: Used to detect anomalies in production status and generate anomaly identifiers based on the dynamic deviation rate of the comprehensive quality factor and adaptive threshold;

[0057] Defect localization module: used to map the multi-source temporal features corresponding to the anomaly identifier to the spatial coordinates of the casing, and output the spatial distribution information of the defect;

[0058] Closed-loop process control module: Used to generate process parameter adjustment amounts based on the spatial distribution information of defects, and feed the process parameter adjustment amounts back to the production control system to perform closed-loop optimization control of the production process.

[0059] This invention provides a quality monitoring system and method for the production of insulating side-expanded bushings. The method comprises: S1: acquiring multimodal sensing signals and process parameters during the production process, including ultrasonic wall thickness signals, infrared temperature signals, and capacitance signals; S2: normalizing the multimodal sensing signals to generate enhanced normalized signals, and calculating local gradient anomaly indices for each mode based on the enhanced normalized signals; S3: calculating the flaring stress index based on the deviation between the ultrasonic wall thickness signal and a reference wall thickness value, combined with process parameters; and S4: performing coupled analysis of the different modal sensing signals based on the local gradient anomaly indices to calculate... S5: Calculate the comprehensive quality factor based on local gradient anomaly index, flaring stress index, and multimodal coupling nonuniformity, combined with process parameter deviation; S6: Detect anomalies in production status and generate anomaly identifiers based on the dynamic deviation rate and adaptive threshold of the comprehensive quality factor; S7: Map the multi-source temporal features corresponding to the anomaly identifiers to the spatial coordinates of the casing, outputting the spatial distribution information of defects; S8: Generate process parameter adjustment amounts based on the spatial distribution information of defects, and feed these adjustment amounts back to the production control system for closed-loop optimization control of the production process. The beneficial effects include:

[0060] 1. By synchronously acquiring multimodal sensing data such as ultrasonic wall thickness signal, infrared temperature signal and capacitance signal, and combining real-time process parameters, local anomaly indicators are extracted using enhanced normalization and fractional weighting mechanism. By calculating the flaring stress index and multimodal coupling nonuniformity, a comprehensive quality factor is constructed, realizing unified quantification and comprehensive evaluation of multi-source information in the production process, improving the completeness of data perception, the sensitivity of anomaly characteristics and the accuracy of stress state characterization.

[0061] 2. By constructing dynamic deviation rate and adaptive threshold, combined with comprehensive quality factor, real-time quality anomaly detection is carried out, and the mapping relationship between anomaly time sequence characteristics and casing spatial coordinates is established to generate spatial mapping index, thereby realizing accurate spatial positioning and severity assessment of defects. This breaks through the limitation of traditional methods that can only give overall anomaly signals and cannot realize spatial distribution representation, and improves the accuracy of anomaly detection and spatial diagnostic capability.

[0062] 3. By feeding back the spatial distribution information of defects to the process control link, and generating process parameter adjustment amounts based on the cumulative amount of defect severity and process state adjustment factors, closed-loop adaptive optimization control of the production process is realized. This enables rapid intervention on significant defects while maintaining system stability, thereby improving the intelligence level of the production process and the overall quality stability.

[0063] The above description is only an overview of the technical solution of this application. In order to better understand the technical means of this application and to implement it in accordance with the contents of the specification, and to make the above and other objects, features and advantages of this application more obvious and understandable, specific embodiments of this application are given below. Attached Figure Description

[0064] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort. In the drawings:

[0065] Figure 1 A flowchart illustrating a quality monitoring method for the production of insulating side expansion sleeves, as shown in an exemplary embodiment of the present invention;

[0066] Figure 2 This is a schematic diagram of a quality monitoring system for the production of insulating side expansion sleeves, as shown in an exemplary embodiment of the present invention. Detailed Implementation

[0067] The embodiments of the present invention will be described below with reference to the accompanying drawings and preferred embodiments. Those skilled in the art can easily understand other advantages and effects of the present invention from the content disclosed in this specification. The present invention can also be implemented or applied through other different specific embodiments, and various details in this specification can also be modified or changed based on different viewpoints and applications without departing from the spirit of the present invention. It should be understood that the preferred embodiments are only for illustrating the present invention and not for limiting the scope of protection of the present invention.

[0068] It should be noted that the illustrations provided in the following embodiments are only schematic representations of the basic concept of the present invention. Therefore, the drawings only show the components related to the present invention and are not drawn according to the actual number, shape and size of the components in the actual implementation. In the actual implementation, the form, quantity and proportion of each component can be arbitrarily changed, and the layout of the components may also be more complex.

[0069] In the following description, numerous details are explored to provide a more thorough explanation of embodiments of the invention. However, it will be apparent to those skilled in the art that embodiments of the invention may be practiced without these specific details. In other embodiments, well-known structures and devices are shown in block diagram form rather than in detail to avoid obscuring embodiments of the invention.

[0070] Example 1

[0071] Quality monitoring methods for the production of insulating side expansion bushings, such as Figure 1 As shown, it includes:

[0072] S1: Collect multimodal sensing signals and process parameters during the production process. The multimodal sensing signals include ultrasonic wall thickness signals, infrared temperature signals, and capacitance signals.

[0073] S2: Normalize the multimodal sensing signals to generate enhanced normalized signals, and calculate the local gradient anomaly index of each mode based on the enhanced normalized signals;

[0074] S3: Calculate the flaring stress index based on the deviation between the ultrasonic wall thickness signal and the reference wall thickness value, combined with process parameters;

[0075] S4: Based on the local gradient anomaly index, perform coupling analysis on different modal sensing signals and calculate the multimodal coupling nonuniformity.

[0076] S5: Calculate the comprehensive quality factor based on the local gradient anomaly index, the flaring stress index, and the multimodal coupling nonuniformity, combined with the process parameter deviation.

[0077] S6: Based on the dynamic deviation rate of the comprehensive quality factor and the adaptive threshold, perform anomaly detection on the production status and generate anomaly labels;

[0078] S7: Map the multi-source temporal features corresponding to the anomaly identifier to the spatial coordinates of the casing, and output the spatial distribution information of the defect;

[0079] S8: Generate process parameter adjustment amounts based on the spatial distribution information of defects, and feed these adjustment amounts back to the production control system for closed-loop optimization control of the production process.

[0080] The present invention is further configured such that S2 includes:

[0081] The ultrasonic wall thickness signal, infrared temperature signal, and capacitance signal are independently normalized.

[0082] The signal sequences within the preset historical time window of each sensing mode are subjected to fractional power weighting, and the current signal normalization result of each sensing mode is fused with the corresponding historical signal weighting result to generate an enhanced normalized signal.

[0083] Based on the enhanced normalized signals of each sensing mode, the ratio of signal energy change to signal intensity change is calculated within a preset sliding time window to generate local gradient anomaly indices for each sensing mode. Specifically, during the casing production process, ultrasonic wall thickness signals, infrared temperature signals, and capacitance signals are synchronously acquired through a distributed sensor array. Each sensing signal corresponds to multiple measurement positions on the casing, achieving spatial coverage of the casing surface and wall thickness. The ultrasonic wall thickness signal is detected by ultrasonic sensors arranged along the casing axis at each measurement position, with each sensor acquiring a signal corresponding to a specific spatial coordinate position on the casing. The infrared temperature signal is acquired by infrared sensors arranged along the casing circumference at each measurement position to ensure that temperature changes at different positions are monitored and quantified in real time during production. The capacitance signal is acquired by sensors arranged at key points on the casing. Capacitive sensors in the forming area sense the local process status. Each acquired signal corresponds to a specific spatial coordinate position of the sleeve, reflecting local dielectric properties and wall thickness changes. Each mode of sensing signal is independently normalized to unify the numerical scale of each sensing mode and eliminate dimensional differences between different physical quantities. The sensing modes include ultrasonic sensing, infrared temperature sensing, and capacitive sensing, used to acquire the sleeve's wall thickness signal, surface temperature signal, and capacitance signal reflecting local dielectric properties and wall thickness changes, respectively. Within the historical time window of each sensing mode, the sensing signal sequence is subjected to fractional-order power-weighted processing to enhance the characteristics of short-term abnormal responses. The current normalized signal of each sensing mode is fused with its corresponding historical weighted processing result to generate an enhanced normalized signal. The calculation logic for the enhanced normalized signal is as follows: ,in, For modality At any moment The enhanced normalized signal is used to unify the numerical scale of various sensing modes, while amplifying short-time anomalous responses. , Indicates ultrasonic wall thickness signal, Indicates infrared temperature signal, Indicates a capacitance signal; and Modal At any moment and The original signal; and For modality Minimum and maximum values ​​during the data collection process; This is an enhancement coefficient used to control the intensity of the influence of historical signals, with a value range of [0.1, 0.5]. The preset historical time window length; For historical time indexing; is a fractional power exponent used to control the attenuation characteristics of historical signals, with a value range of [0.5, 0.8]. This is a normalization process used to map the signal to the [0,1] range and eliminate amplitude differences between modes; For historical weighted enhancement, fractional-power accumulation of historical signals is used to amplify local anomalies while preserving trend information; based on enhanced normalized signals. At every moment The local gradient anomaly index is calculated within a preset sliding window. This index describes the nonlinear changes and energy shifts of the signal within the local window. The calculation logic for the local gradient anomaly index is as follows: ,in, For modality At any moment The local gradient anomaly index is used to quantify the fluctuation characteristics of a signal at the current moment and within its adjacent time window, highlighting local anomalies or abrupt changes. The preset sliding time window is half the length; For sliding time index; For modality At any moment Enhanced normalized signal; The adjustment parameter is used to control the sensitivity of the local gradient anomaly index calculation, and its value range is [1.2, 1.8]. This represents the change in signal energy, used to reflect the change in signal energy between the current moment and nearby moments. It represents the intensity of signal changes, used to quantify the absolute difference between signals at nearby times and the current time, and characterizes the intensity of local fluctuations of the signal within a short time scale.

[0084] The present invention is further configured such that S3 includes:

[0085] Based on the enhanced normalized ultrasonic wall thickness signal and the corresponding reference wall thickness value at each measurement location, the wall thickness deviation between the current ultrasonic wall thickness signal and the reference wall thickness value is calculated, and the wall thickness deviation is nonlinearly amplified.

[0086] The process difference between real-time process parameters and local equivalent process parameters at each measurement location is obtained, and the process difference is processed by nonlinear adjustment.

[0087] The local stress components at each measurement location are obtained by normalizing the nonlinearly amplified wall thickness deviation and the nonlinearly adjusted process difference.

[0088] The local stress components at each measurement location are aggregated to generate a flaring stress index. Specifically, during the sleeve production process, for each measurement location, the enhanced normalized ultrasonic wall thickness signal is acquired and compared with the corresponding reference wall thickness value to calculate the wall thickness deviation. This deviation is then nonlinearly amplified to enhance the response of local stress to changes in wall thickness. Local process parameters at the measurement location are collected and compared with real-time process parameters to calculate the process difference. This difference is then nonlinearly adjusted to reflect the sensitivity of local stress to process fluctuations. The nonlinearly amplified wall thickness deviation and the nonlinearly adjusted process difference are normalized to obtain the local stress component at each measurement location. This local stress component reflects the stress concentration at each measurement location during the flaring process. The local stress components at all measurement locations are aggregated to generate a flaring stress index for the sleeve. This index quantifies the local stress distribution during the flaring process and reflects the severity of local stress anomalies. The calculation logic for the flaring stress index is as follows: ,in, For a moment The flaring stress index; For a moment Spatial coordinates The enhanced normalized ultrasonic wall thickness signal at the location is the position. Enhanced normalized ultrasonic wall thickness signal at the location is used to reflect position. The actual wall thickness; This is the set of all measured locations; Spatial coordinates Reference wall thickness value at the location; Indicates time Spatial coordinates The wall thickness deviation at a certain point is used to quantify the difference between the current wall thickness and the reference wall thickness, reflecting the thickness change caused by local stress. The parameter is used to adjust the wall thickness deviation nonlinearly, amplify or compress it, and enhance the influence of the local stress concentration area. The value range is [1.5,2]. These are global real-time process parameters, such as flaring pressure, flaring speed, and flaring process temperature setpoints, used to provide a global reference for the current production process status. For a moment Spatial coordinates The estimated values ​​of local process parameters at a given location can be derived by interpolation based on sensor data; Indicates time Spatial coordinates The process difference at a point represents the amount by which local process parameters deviate from global process parameters; The adjustment parameter is used to control the influence of process variation on the flaring stress index, and its value range is [0.8, 1.2].

[0089] The present invention is further configured such that S4 includes:

[0090] Based on the local gradient anomaly index corresponding to each sensing mode, calculate the anomaly difference between the local gradient anomaly indices of any two different sensing modes, and perform nonlinear adjustment processing on the anomaly difference.

[0091] The local gradient anomaly indices of two different sensing modes are nonlinearly weighted and combined, and the anomaly difference quantity adjusted by nonlinearity is used to normalize the weighted combination result to obtain the interactive coupling components corresponding to the two sensing modes.

[0092] The interactive coupling components of all sensing modal pairs are aggregated to generate multimodal coupling nonuniformity. Specifically, for any two different sensing modalities, the local gradient anomaly indices are calculated to determine their anomaly differences. A nonlinear adjustment function is used to enhance the sensitivity of these differences, capturing the cooperative anomaly characteristics between modalities and strengthening the representation of defects that are difficult to identify with a single modality. The local gradient anomaly indices of the two sensing modalities are then nonlinearly weighted and combined. The weighted combination result is normalized using the nonlinearly adjusted anomaly difference to generate the interactive coupling components between the two modalities, revealing the cooperative response characteristics of the multimodal sensing signals on the local gradient. The interactive coupling components of all sensing modal pairs are aggregated to generate the multimodal coupling nonuniformity of the sleeve at that time point. This multimodal coupling nonuniformity reflects the strength of the cooperative effect of anomaly signals between different sensing modalities. The calculation logic for multimodal coupling nonuniformity is as follows: ,in, For a moment The multimodal coupling nonuniformity is used to reflect the time-varying multimodal coupling nonuniformity. The interaction coupling strength of local gradient anomaly indices between different modal signals; This means summing all modes pairwise, calculating each pair only once to avoid repetition. and ,and ; and Modal and At any moment Local gradient anomaly indicators; and These are weight parameters, which act on... and , is used to control the degree of influence of different modalities in the interaction, and its value range is [1,2]; The parameter is used to adjust the sensitivity of multimodal coupling inhomogeneity to the abnormal differences between modes, and its value range is [1.3, 1.7]. This represents a nonlinear weighted combination of local gradient anomaly indices for two sensing modes. This nonlinear processing can enhance the influence of modes with larger anomaly changes on the multimodal coupling inhomogeneity, while suppressing the contribution of modes with smaller anomaly changes.

[0093] The present invention is further configured such that S5 includes:

[0094] The local gradient anomaly indices of each sensing mode are summed, and the summation result is subjected to a first power transformation.

[0095] The flaring stress index is subjected to a second power transformation, and the results of the first power transformation and the second power transformation are fused to generate a signal fusion factor.

[0096] A third power transformation and offset processing are performed on the multimodal coupling nonuniformity to generate a coupling influence factor.

[0097] Calculate the absolute deviation between the real-time process parameters and the optimal process parameters, and perform exponential mapping on the absolute deviation to generate a process deviation penalty term;

[0098] The signal fusion factor, coupling influence factor, and process deviation penalty term are multiplied to generate a comprehensive quality factor. Specifically, the local gradient anomaly indices corresponding to the ultrasonic sensing mode, infrared temperature sensing mode, and capacitive sensing mode are summed, and the summation result is subjected to a first power transformation to quantify the cumulative effect of local anomalies in each mode. The flaring stress index is subjected to a second power transformation and fused with the result of the first power transformation to obtain a signal fusion factor, which reflects the comprehensive local stress effect caused by wall thickness deviation and process parameter fluctuations. For multimodal coupling nonuniformity, a third power transformation and offset processing are used to generate a coupling factor. A comprehensive quality factor is generated by multiplying the signal fusion factor, coupling influence factor, and process deviation penalty term to quantify the contribution of intermodal synergy to overall quality. The absolute deviation between real-time and optimal process parameters is processed through exponential mapping to generate a process deviation penalty term, reflecting the impact of process condition deviations on bushing quality. A comprehensive quality factor is generated by multiplying the signal fusion factor, coupling influence factor, and process deviation penalty term. This comprehensive quality factor is used to comprehensively evaluate the cumulative effect of local anomalies, intermodal coupling effects, and quality risks caused by process deviations under current production conditions, providing a unified indicator for subsequent dynamic deviation rate calculation and anomaly judgment. The calculation logic of the comprehensive quality factor is as follows: ,in, For a moment The comprehensive quality factor is used to measure the performance of the casing at any given time. The overall quality status reflects the severity of local defects or process deviations in the casing; , and For power-order parameters; The contribution of the local gradient anomaly index for each mode to the overall quality factor is used to adjust the value range of [2,3]. Used to adjust the intensity of the influence of the flaring stress index on the comprehensive quality factor, with a value range of [1.5, 2.5]. This is used to adjust the contribution of multimodal coupling nonuniformity to the coupling influence factor, with a value range of [1.2, 1.8]. The weighting parameter controls the proportion by which the multimodal coupling inhomogeneity amplifies or compresses the overall quality factor, and its value ranges from -0.3 to -0.1. The adjustment parameter is used to adjust the impact of the deviation between the real-time process parameters and the optimal process parameters on the decay of the comprehensive quality factor, and its value range is [0.5, 2]. These are the optimal process parameter values; As a signal fusion factor, it aggregates local gradient anomaly index and flaring stress index to reflect the contribution of single-mode and process stress anomalies to the overall quality of the bushing. As a coupling influence factor, it is used to characterize the interactive coupling effect between different sensing modes, and to nonlinearly amplify or compress the influence of cooperative changes on the comprehensive quality factor. This is a process deviation penalty term used to quantify the negative impact of process parameters deviating from optimal values ​​on casing quality.

[0099] The present invention is further configured such that S6 includes:

[0100] The dynamic deviation rate is calculated based on the continuous change of the comprehensive quality factor within the preset historical time window and the continuous change of the real-time process parameters within the preset historical time window.

[0101] Calculate the adaptive threshold based on the current comprehensive quality factor, dynamic deviation rate, and multimodal coupling inhomogeneity.

[0102] The current comprehensive quality factor is compared with an adaptive threshold. If the current comprehensive quality factor is less than the adaptive dynamic threshold, an anomaly flag is generated. Specifically, during the production process, the comprehensive quality factor at each time point is continuously recorded, and the continuous change is calculated within a preset historical time window to capture the fluctuation of the comprehensive quality factor over time. Real-time process parameters within the preset historical time window are collected, and the continuous change is calculated to quantify process fluctuations. Combining the changes in the comprehensive quality factor and the fluctuations in process parameters, the dynamic deviation rate of the comprehensive quality factor is calculated to reflect the degree of anomaly of the current state relative to the historical trend. The calculation logic of the dynamic deviation rate is as follows: ,in, For a moment The dynamic offset rate is used to quantify the degree of dynamic change of the comprehensive quality factor relative to the fluctuation of process parameters within a preset historical time window. and They are time points and The overall quality factor; and To adjust the parameters; Used to control the sensitivity of changes in the comprehensive quality factor in the calculation of dynamic deviation rate, with a value range of [1.2, 1.6]; Used to amplify or suppress the impact of process parameter fluctuations on dynamic deviation rate, with a value range of [1, 1.4]. and They are time points and Process parameters; Used to measure the intensity of changes in historical composite quality factors; This is used to measure the overall fluctuation of process parameters; considering the current comprehensive quality factor, dynamic offset rate, and multimodal coupling non-uniformity, an adaptive threshold is calculated. This adaptive threshold can dynamically adapt to changes in the production process, avoiding false alarms or missed alarms caused by a fixed threshold; the calculation logic of the adaptive threshold is as follows: ,in, For a moment An adaptive threshold is used to determine whether the current comprehensive quality factor deviates from the normal range, thereby generating an anomaly flag; and To adjust the index; Used to adjust the integrated part The nonlinear amplification or compression effect of the adaptive threshold ranges from [1.1, 1.3]. For adjustment The overall amplification or compression effect of the adaptive threshold ranges from [-1, -0.6]. The current comprehensive quality factor is compared with the adaptive threshold for anomaly detection. If the current comprehensive quality factor is less than the adaptive threshold, then: When an anomaly is detected, an anomaly flag is generated.

[0103] The present invention is further configured such that S7 includes:

[0104] Obtain the local gradient anomaly index and multimodal coupling nonuniformity of each sensing mode at the time point corresponding to the anomaly identifier, as well as the enhanced normalized signal and corresponding reference signal of each sensing mode in the sleeve space coordinates;

[0105] Calculate the signal deviation between the enhanced normalized signal of each sensing mode and the corresponding reference signal at each spatial coordinate, and perform nonlinear adjustment processing on the signal deviation.

[0106] The local gradient anomaly index is nonlinearly amplified, and the nonlinearly adjusted signal deviation is used to normalize the nonlinearly amplified local gradient anomaly index to obtain the independent contribution of each sensing mode in each spatial coordinate.

[0107] The independent contributions of all sensing modes are summed to obtain the aggregated spatial contribution value.

[0108] The spatial contribution aggregation value is fused with the multimodal coupling inhomogeneity to generate a spatial mapping index. Based on the magnitude and distribution of the spatial mapping index across spatial coordinates, the spatial location and severity of the defect are determined, and the spatial distribution information of the defect is output. Specifically, the time point marker information obtained from anomaly detection is mapped to the physical spatial coordinates of the casing to achieve defect location and spatial distribution assessment. By combining the enhanced normalized signals of each sensing mode, the local gradient anomaly index, and the multimodal coupling inhomogeneity, the independent contribution of each spatial point to the anomaly is quantified, and the multimodal interaction effect is fused to generate a spatial mapping index, providing a visual basis for defect location. For the time point corresponding to the anomaly marker, the enhanced normalized signal and corresponding reference signal of each sensing mode at each spatial coordinate are obtained, along with the local gradient anomaly index and multimodal coupling inhomogeneity at the same time point. For each spatial coordinate and sensing mode, the signal deviation between the enhanced normalized signal and the reference signal is calculated, and the signal deviation is nonlinearly adjusted to enhance the local anomaly response. The nonlinearly adjusted signal deviation is denoted as: ,in, For modality At any moment Spatial coordinates Enhanced normalized signal at the location; For modality In spatial coordinates Reference signal at the location; To adjust the parameters, the contribution of the signal deviation to the spatial mapping index is amplified or compressed, with a value range of [1.2, 1.6]. The signal deviation after nonlinear adjustment is used to normalize the amplified local gradient anomaly index to obtain the independent contribution of each sensing mode in each spatial coordinate. The independent contributions of all sensing modes are accumulated to obtain the spatial contribution aggregation value. This spatial contribution aggregation value is used to reflect the overall anomaly contribution degree of different sensing modes in each spatial coordinate. The calculation logic of the spatial contribution aggregation value is as follows: ,in, For modality At any moment Spatial coordinates Spatial contribution aggregation value at the location, Independent contribution; The parameter is used to adjust the influence of the local gradient anomaly index on the spatial mapping exponent, with a value range of [1.6, 2]. The spatial contribution aggregate value is... Coupling inhomogeneity with multimodal modes The data is then fused to generate a spatial mapping index, which is used to characterize the abnormal signal intensity and defect distribution at various spatial coordinates of the bushing. The calculation logic of the spatial mapping index is as follows: ,in, The spatial mapping index reflects the position of the bushing at time [time]. Spatial coordinates The overall signal anomaly intensity at the location; is a weighting parameter used to control the influence of multimodal coupling inhomogeneity on the spatial mapping index, with a value range of [0.8, 1.2]. The spatial mapping index is adjusted accordingly. By analyzing the magnitude and distribution of the values, the location and severity of defects can be determined, thus providing a reference for adjusting production process parameters.

[0109] The present invention is further configured such that S8 includes:

[0110] Based on the spatial mapping index corresponding to the spatial distribution information of defects, the spatial mapping index on all spatial coordinates is summed to generate the cumulative defect severity.

[0111] The current process parameters are nonlinearly transformed to generate process state adjustment factors.

[0112] The cumulative defect severity is weighted and modulated using a process state adjustment factor to generate process parameter adjustment amounts.

[0113] The process parameter adjustment is added to the current process parameter to generate an optimized process parameter setting value;

[0114] The optimized process parameter setpoints are fed back to the production control system to achieve closed-loop optimization control. Specifically, in the casing production process, the production process parameters are dynamically adjusted by utilizing the spatial mapping index corresponding to the spatial distribution information of defects, thereby achieving optimized control of the production process. The spatial mapping index for each spatial coordinate... After squaring, the values ​​are summed to obtain the cumulative defect severity of the entire casing at the current time point. This cumulative defect severity is used to quantify the overall defect level of the entire casing at the current moment, highlighting the contribution of high-anomaly locations to the total defect severity. The calculation logic for the cumulative defect severity is as follows: ,in, For a moment The cumulative severity of defects; for current process parameters A nonlinear transformation is performed to generate a process state adjustment factor. This factor considers the influence of the current process parameter state on the adjustment amount. Closed-loop control using this factor can intervene in defects while avoiding instability due to over-adjustment. The calculation logic of the process state adjustment factor is as follows: ,in, For a moment Process state adjustment factor; and This is the adjustment coefficient; Used to control the degree of influence of process status on adjustment amount, with a value range of [0.5, 0.9]; Used to adjust the direction of process parameter adjustments, with a value range of [-1.2, -0.8]; based on the cumulative defect severity. and process state adjustment factor The process parameter adjustment amount is calculated for closed-loop optimization of the production process. The calculation logic for the process parameter adjustment amount is as follows: ,in, For a moment The amount of process parameter adjustment; This is an adjustment coefficient used to control the overall magnitude of the process parameter adjustment, with a value range of [0.01, 0.1]. The obtained process parameter adjustment amount... Compared with current process parameters The parameters are superimposed to generate optimized process parameter setpoints that can be directly applied to the production control system. The calculation logic for the optimized process parameter setpoints is as follows: ,in, For a moment The optimized process parameter settings; the obtained optimized process parameter settings Feedback is sent to the production control system to achieve dynamic control of the casing production process. By adjusting process parameters in real time, the defect rate can be reduced and the overall quality of the casing can be improved.

[0115] The invention is further configured such that the multimodal sensing signals are acquired by a distributed sensor array, which is arranged along the axial and circumferential directions of the casing. Specifically, during the casing production process, the multimodal sensing signals are acquired through a distributed sensor array, with the sensors uniformly arranged along the axial and circumferential directions of the casing, and each sensor acquiring a signal corresponding to a specific spatial coordinate of the casing. Furthermore, the signals of each mode are synchronized through high-precision timestamps; through this sensor arrangement, the dynamic changes of the casing in the axial and circumferential directions can be monitored simultaneously, realizing comprehensive and multimodal monitoring of the flaring process and the overall quality status of the casing.

[0116] Example 2

[0117] Please see Figure 2 The exemplary quality monitoring system for the production of insulating side-expanding bushings includes:

[0118] Signal acquisition module: used to acquire multimodal sensing signals and process parameters during the production process. The multimodal sensing signals include ultrasonic wall thickness signals, infrared temperature signals, and capacitance signals.

[0119] Normalization anomaly module: used to normalize multimodal sensing signals to generate enhanced normalized signals, and calculate local gradient anomaly indices for each mode based on the enhanced normalized signals;

[0120] Flaring stress assessment module: used to calculate the flaring stress index based on the deviation between the ultrasonic wall thickness signal and the reference wall thickness value, combined with process parameters;

[0121] Coupling Analysis Module: Used to perform coupling analysis on different modal sensing signals based on local gradient anomaly indices and calculate multimodal coupling nonuniformity;

[0122] The comprehensive quality assessment module is used to calculate the comprehensive quality factor based on local gradient anomaly indices, flaring stress index, and multimodal coupling nonuniformity, combined with process parameter deviations.

[0123] Anomaly detection module: Used to detect anomalies in production status and generate anomaly identifiers based on the dynamic deviation rate of the comprehensive quality factor and adaptive threshold;

[0124] Defect localization module: used to map the multi-source temporal features corresponding to the anomaly identifier to the spatial coordinates of the casing, and output the spatial distribution information of the defect;

[0125] Closed-loop process control module: Used to generate process parameter adjustment amounts based on the spatial distribution information of defects, and feed the process parameter adjustment amounts back to the production control system to perform closed-loop optimization control of the production process.

[0126] It should be noted that the quality monitoring system for the production of insulating side-expanding bushings provided in the above embodiments and the quality monitoring method for the production of insulating side-expanding bushings provided in the above embodiments belong to the same concept. The specific operation methods of each module and unit have been described in detail in the method embodiments and will not be repeated here. In practical applications, the quality monitoring system for the production of insulating side-expanding bushings provided in the above embodiments can be assigned to different functional modules as needed, that is, the internal structure of the system can be divided into different functional modules to complete all or part of the functions described above. This is not a limitation here.

[0127] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

Claims

1. A quality monitoring method for the production of insulating side expansion bushings, characterized in that, include: S1: Collect multimodal sensing signals and process parameters during the production process. The multimodal sensing signals include ultrasonic wall thickness signals, infrared temperature signals, and capacitance signals. S2: Normalize the multimodal sensing signals to generate enhanced normalized signals. Calculate the local gradient anomaly index of each sensing mode based on the enhanced normalized signals. The sensing modes include ultrasonic sensing mode, infrared temperature sensing mode, and capacitance sensing mode, which are used to collect the ultrasonic wall thickness signal, infrared temperature signal, and capacitance signal of the local dielectric properties and wall thickness changes of the sleeve, respectively. S3: Calculate the flaring stress index based on the deviation between the ultrasonic wall thickness signal and the reference wall thickness value, combined with process parameters; S4: Based on the local gradient anomaly index, perform coupling analysis on different modal sensing signals and calculate the multimodal coupling nonuniformity. S5: Calculate the comprehensive quality factor based on the local gradient anomaly index, the flaring stress index, and the multimodal coupling nonuniformity, combined with the process parameter deviation. The process parameter deviation is the absolute deviation between the collected real-time process parameters and the optimal process parameters. S6: Based on the dynamic deviation rate of the comprehensive quality factor and the adaptive threshold, perform anomaly detection on the production status and generate anomaly labels; S7: Map the multi-source temporal features corresponding to the anomaly identifier to the spatial coordinates of the casing, and output the spatial distribution information of the defect; S8: Generate process parameter adjustment amounts based on the spatial distribution information of defects, and feed these adjustment amounts back to the production control system for closed-loop optimization control of the production process.

2. The quality monitoring method for the production of insulating side expansion bushings according to claim 1, characterized in that, S2 includes: The ultrasonic wall thickness signal, infrared temperature signal, and capacitance signal are independently normalized. The signal sequences within the preset historical time window of each sensing mode are subjected to fractional power weighting, and the current signal normalization result of each sensing mode is fused with the corresponding historical signal weighting result to generate an enhanced normalized signal. Based on the enhanced normalized signals of each sensing mode, the local gradient anomaly index of each sensing mode is generated by calculating the ratio of signal energy change to signal change intensity within a preset sliding time window.

3. The quality monitoring method for the production of insulating side expansion bushings according to claim 1, characterized in that, S3 includes: Based on the enhanced normalized ultrasonic wall thickness signal and the corresponding reference wall thickness value at each measurement location, the wall thickness deviation between the current ultrasonic wall thickness signal and the reference wall thickness value is calculated, and the wall thickness deviation is nonlinearly amplified. The process difference between real-time process parameters and local equivalent process parameters at each measurement location is obtained, and the process difference is processed by nonlinear adjustment. The local stress components at each measurement location are obtained by normalizing the nonlinearly amplified wall thickness deviation and the nonlinearly adjusted process difference. The local stress components at each measurement location are aggregated to generate the flared stress index.

4. The quality monitoring method for the production of insulating side expansion bushings according to claim 1, characterized in that, S4 includes: Based on the local gradient anomaly index corresponding to each sensing mode, calculate the anomaly difference between the local gradient anomaly indices of any two different sensing modes, and perform nonlinear adjustment processing on the anomaly difference. The local gradient anomaly indices of two different sensing modes are nonlinearly weighted and combined, and the anomaly difference quantity adjusted by nonlinearity is used to normalize the weighted combination result to obtain the interactive coupling components corresponding to the two sensing modes. The interactive coupling components of all sensing mode pairs are aggregated to generate multimodal coupling nonuniformity.

5. The quality monitoring method for the production of insulating side expansion bushings according to claim 1, characterized in that, S5 includes: The local gradient anomaly indices of each sensing mode are summed, and the summation result is subjected to a first power transformation. The flaring stress index is subjected to a second power transformation, and the results of the first power transformation and the second power transformation are fused to generate a signal fusion factor. A third power transformation and offset processing are performed on the multimodal coupling nonuniformity to generate a coupling influence factor. Calculate the absolute deviation between the real-time process parameters and the optimal process parameters, and perform exponential mapping on the absolute deviation to generate a process deviation penalty term; The signal fusion factor, coupling influence factor, and process deviation penalty term are multiplied together to generate a comprehensive quality factor.

6. The quality monitoring method for the production of insulating side expansion bushings according to claim 1, characterized in that, S6 includes: The dynamic deviation rate is calculated based on the continuous change of the comprehensive quality factor within the preset historical time window and the continuous change of the real-time process parameters within the preset historical time window. Calculate the adaptive threshold based on the current comprehensive quality factor, dynamic deviation rate, and multimodal coupling inhomogeneity. The current comprehensive quality factor is compared with the adaptive threshold. If the current comprehensive quality factor is less than the adaptive dynamic threshold, an anomaly flag is generated.

7. The quality monitoring method for the production of insulating side expansion bushings according to claim 1, characterized in that, S7 includes: Obtain the local gradient anomaly index and multimodal coupling nonuniformity of each sensing mode at the time point corresponding to the anomaly identifier, as well as the enhanced normalized signal and corresponding reference signal of each sensing mode in the sleeve space coordinates; Calculate the signal deviation between the enhanced normalized signal of each sensing mode and the corresponding reference signal at each spatial coordinate, and perform nonlinear adjustment processing on the signal deviation. The local gradient anomaly index is nonlinearly amplified, and the nonlinearly adjusted signal deviation is used to normalize the nonlinearly amplified local gradient anomaly index to obtain the independent contribution of each sensing mode in each spatial coordinate. The independent contributions of all sensing modes are summed to obtain the aggregated spatial contribution value. The spatial contribution aggregation value is fused with the multimodal coupling inhomogeneity to generate a spatial mapping index; Based on the magnitude and distribution of the spatial mapping index on each spatial coordinate, the spatial location and severity of the defect are determined, and the spatial distribution information of the defect is output.

8. The quality monitoring method for the production of insulating side expansion bushings according to claim 1, characterized in that, S8 includes: Based on the spatial mapping index corresponding to the spatial distribution information of defects, the spatial mapping index on all spatial coordinates is summed to generate the cumulative defect severity. The current process parameters are nonlinearly transformed to generate process state adjustment factors. The cumulative defect severity is weighted and modulated using a process state adjustment factor to generate process parameter adjustment amounts. The process parameter adjustment is added to the current process parameter to generate an optimized process parameter setting value; The optimized process parameter setpoints are fed back to the production control system to achieve closed-loop optimization control.

9. The quality monitoring method for the production of insulating side expansion bushings according to claim 1, characterized in that, The multimodal sensing signals are acquired by a distributed sensor array, which is arranged along the axial and circumferential directions of the sleeve.

10. A quality monitoring system for the production of insulating side-expanded bushings, used to implement the quality monitoring method for the production of insulating side-expanded bushings as described in any one of claims 1-9, characterized in that, include: Signal acquisition module: used to acquire multimodal sensing signals and process parameters during the production process. The multimodal sensing signals include ultrasonic wall thickness signals, infrared temperature signals, and capacitance signals. Normalization anomaly module: Used to normalize multimodal sensing signals to generate enhanced normalized signals, and calculate local gradient anomaly indices for each sensing mode based on the enhanced normalized signals. The sensing modes include ultrasonic sensing mode, infrared temperature sensing mode and capacitance sensing mode, used to collect ultrasonic wall thickness signal, infrared temperature signal and capacitance signal of local dielectric properties and wall thickness changes of the sleeve respectively. Flaring stress assessment module: used to calculate the flaring stress index based on the deviation between the ultrasonic wall thickness signal and the reference wall thickness value, combined with process parameters; Coupling Analysis Module: Used to perform coupling analysis on different modal sensing signals based on local gradient anomaly indices and calculate multimodal coupling nonuniformity; The comprehensive quality assessment module is used to calculate the comprehensive quality factor based on the local gradient anomaly index, the flaring stress index, and the multimodal coupling non-uniformity, combined with the process parameter deviation. The process parameter deviation is the absolute deviation between the collected real-time process parameters and the optimal process parameters. Anomaly detection module: Used to detect anomalies in production status and generate anomaly identifiers based on the dynamic deviation rate of the comprehensive quality factor and adaptive threshold; Defect localization module: used to map the multi-source temporal features corresponding to the anomaly identifier to the spatial coordinates of the casing, and output the spatial distribution information of the defect; Closed-loop process control module: Used to generate process parameter adjustment amounts based on the spatial distribution information of defects, and feed the process parameter adjustment amounts back to the production control system to perform closed-loop optimization control of the production process.

Images