Intelligent defect detection and repair method and system for vacuum isothermal forging based on multi-modal sensing data

Abstract

The invention relates to the technical field of intelligent manufacturing, in particular to an intelligent defect detecting and repairing method and system for a vacuum isothermal forging based on multi-mode sensing data. The method comprises the following steps: respectively acquiring image data, heat distribution data and stress monitoring data of the vacuum isothermal forging through a visual sensor, a thermal sensor and a stress sensor, and preprocessing the data to generate a time-space aligned multi-modal data set; performing fusion processing on the multi-modal data set, extracting multi-dimensional features of defects through a deep learning network, and generating a defect feature sequence; optimizing a repair path planning model based on the defect feature sequence in combination with the stress distribution features, and generating a repair task sequence; and inputting the repairing task sequence into the intelligent repairing equipment to execute repairing operation. According to the method, the multi-mode sensing technology and the deep learning algorithm are combined, high-precision detection, classification and repair of the defects of the vacuum isothermal forging are achieved, the production efficiency and the product quality are remarkably improved, and the defective rate and the repair cost are reduced.

Description

Technical Field

[0001] The present invention relates to the field of intelligent manufacturing technology, and in particular to an intelligent defect detection and repair method and system for vacuum isothermal forgings based on multimodal sensing data. Background Art

[0002] Vacuum isothermal forgings are widely used in high-end manufacturing fields such as aviation, aerospace, and gas turbines. The production process has extremely high requirements for superplastic deformation of billets, process parameter control, and microstructure and performance control. However, in the production process of forgings, due to the complex process conditions and the great difficulty of operation under high temperature conditions, forgings are prone to complex defects such as cracks, uneven microstructure, and uneven heat distribution. If the above defects cannot be detected and effectively repaired in time, the quality and service performance of the forgings will be seriously affected.

[0003] Existing defect detection and repair technologies mainly rely on single-mode sensors (such as visual ultrasonic), which have low detection range and accuracy and are difficult to adapt to complex multi-type defects of forgings. At the same time, traditional repair methods lack intelligent means to judge the location and type of defects, and usually rely on manual experience to make decisions, resulting in low repair efficiency and poor accuracy.

[0004] With the development of multimodal sensing technology and artificial intelligence algorithms, the application of deep learning algorithms based on multimodal sensing data fusion in complex industrial scenarios has gradually become a trend. Multimodal data can comprehensively utilize information such as vision, thermal sensing and stress sensing to characterize the characteristics of defects from multiple dimensions, and deep learning algorithms can achieve accurate classification of defects and automatic optimization of repair strategies through data learning. However, there is currently a lack of a complete system that can integrate multimodal data acquisition, deep learning processing and intelligent repair operations for defect detection and repair of vacuum isothermal forgings. Therefore, proposing a system that can efficiently and accurately complete defect detection and intelligent repair has become the key to solving the above problems. Summary of the invention

[0005] The present invention provides an intelligent defect detection and repair method and system for vacuum isothermal forgings based on multimodal sensing data, so as to solve the problem of how to comprehensively analyze and process complex defects of vacuum isothermal forgings through multimodal sensing technology and deep learning algorithms, and realize high-precision detection, classification and intelligent repair of defects.

[0006] In order to solve the above technical problems, the present invention provides an intelligent defect detection and repair method for vacuum isothermal forgings based on multimodal sensing data, comprising: Acquire image data of vacuum isothermal forgings through a visual sensor, collect thermal distribution data through a thermal sensor, collect stress monitoring data through a stress sensor, pre-process the image data, thermal distribution data and stress monitoring data, and generate a multi-modal data set aligned in time and space; The multimodal data set is fused and processed, and multidimensional features of defects are extracted through a deep learning network to generate a defect feature sequence of defect location, category, and severity; Based on the defect feature sequence and combined with stress distribution characteristics, the repair path planning model is optimized to generate a repair task sequence of repair methods, repair parameters and path optimization solutions; The repair task sequence is input into the intelligent repair equipment, the vacuum isothermal forging is repaired, and the repair status is monitored in real time through sensors, and a repair quality report is generated and fed back to the repair path planning model for dynamic adjustment.

[0007] Furthermore, the step of acquiring image data through a visual sensor includes: The image data is subjected to noise filtering and feature enhancement processing to generate preprocessed image data.

[0008] Furthermore, the step of collecting heat distribution data by means of a thermal sensor comprises: The heat distribution data is subjected to denoising and normalization processing to generate pre-processed heat distribution data.

[0009] Furthermore, the step of collecting stress monitoring data through the stress sensor includes: The stress monitoring data is subjected to time alignment and feature extraction processing to generate stress gradient features consistent with the image data and the thermal distribution data.

[0010] Furthermore, the fusion processing step of the multimodal data set includes: A convolution operation is performed on the image data to extract image defect features, a gradient analysis is performed on the thermal distribution data to extract temperature anomaly features, and feature extraction is performed on the stress monitoring data to generate a multi-modal feature sequence of defects.

[0011] Furthermore, the step of extracting multi-dimensional features by the deep learning network includes: The weight of each modal feature is calculated through the attention mechanism, and the image features, thermal distribution features and stress features are weightedly fused to generate the defect location, category and severity.

[0012] Furthermore, the repair path planning model optimization step includes: The preliminary repair method and repair parameters are generated according to the defect feature sequence, the repair path is optimized based on the stress distribution characteristics, and the repair task sequence is generated.

[0013] Furthermore, the repair task sequence includes: Repair method, repair parameters and path optimization scheme, wherein the path optimization scheme includes the spatial position of the repair points and the repair sequence.

[0014] Furthermore, the steps of executing and providing feedback on the repair task include: Repair operations are performed through intelligent repair equipment according to the repair task sequence, real-time data of the repair status is collected, repair monitoring data is generated and fed back to the path planning model, and the repair parameters and paths are dynamically adjusted.

[0015] Furthermore, a vacuum isothermal forging intelligent defect detection and repair system based on multimodal sensing data comprises: A data acquisition module, used to acquire the image data, thermal distribution data and stress monitoring data; A data processing and fusion module, used for preprocessing and fusion processing the data to generate a multi-modal feature sequence of defects; A repair task generation module, used to generate a repair task sequence based on the defect feature sequence; The intelligent repair module is used to execute the repair task sequence, monitor the repair status in real time through sensors, and feed back the real-time monitoring data of the repair to the repair task generation module.

[0016] The key innovative features of the present invention include: (1) Combination of multimodal data fusion and deep learning: This invention, for the first time, aligns visual, thermal and stress data in time and space, extracts multidimensional features through a deep learning network, and analyzes the properties and distribution of complex defects through multimodal fusion, thus overcoming the limitations of traditional single-modal detection.

[0017] (2) Dynamic repair task sequence generation and optimization: The repair path planning is dynamically adjusted by utilizing the defect feature sequence and stress distribution characteristics to generate a repair task sequence including repair methods, repair parameters, and optimized paths, thus realizing an intelligent repair process.

[0018] (3) Real-time feedback and closed-loop optimization mechanism: By monitoring the repair status in real time and feeding back the repair monitoring data to the path planning model for dynamic adjustment, a closed-loop system of detection, repair, and verification is formed, which significantly improves the accuracy and adaptability of the repair.

[0019] (4) Efficient application of mathematical models: Introducing complex mathematical calculations into the detection and repair process, including convolutional neural network feature extraction, gradient analysis, dynamic path planning optimization, etc., ensures the efficiency of data processing and model optimization, and lays the foundation for large-scale industrial applications.

[0020] The following are its main beneficial effects: (1) Improve detection accuracy: Through the spatiotemporal alignment and fusion processing of the three sensor data, namely vision, thermal sensing and stress, the surface and internal defect characteristics of the forgings are comprehensively analyzed, solving the problem of low detection accuracy of a single sensing mode. The deep learning network further extracts and classifies multi-dimensional features, significantly improving the accuracy and comprehensiveness of defect detection.

[0021] (2) Optimizing repair efficiency: Based on the defect feature sequence and combined with the stress distribution characteristics, an optimized repair path planning model is generated. By dynamically adjusting the repair method and parameters, the accuracy and efficiency of the repair process are improved and redundant repair steps are reduced.

[0022] (3) Enhance system adaptability: During the repair process, sensors are used to monitor the repair status in real time and dynamically feed back to the path planning model to adjust the repair parameters and path, ensuring that the repair process can adapt to complex defect types and different working environments.

[0023] (4) Reduce production costs: Through intelligent defect detection and repair processes, human involvement is reduced, the defective rate is reduced, and the use of repair resources is optimized, significantly reducing production and repair costs.

[0024] (5) Improve product quality: The system verifies the repair effect in real time during the repair process to ensure that the structural integrity and performance indicators of the repaired forgings meet the requirements and meet the strict quality standards of high-end manufacturing. BRIEF DESCRIPTION OF THE DRAWINGS

[0025] Figure 1 A flow chart of an intelligent defect detection and repair method for vacuum isothermal forgings based on multimodal sensing data provided in an embodiment of the present application.

[0026] Figure 2 A structural block diagram of an intelligent defect detection and repair system for vacuum isothermal forgings based on multimodal sensing data provided in an embodiment of the present application. DETAILED DESCRIPTION

[0027] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as those commonly understood by technicians in the technical field of the present application; the terms used in the specification of the application herein are only for the purpose of describing specific embodiments and are not intended to limit the present application; the terms "including" and "having" and any variations thereof in the specification and claims of the present application and the above-mentioned drawings are intended to cover non-exclusive. The terms "first", "second", etc. in the specification and claims of the present application and the above-mentioned drawings are used to distinguish different objects, not to describe a specific order.

[0028] Reference to "embodiment" herein means that the specific features and structural characteristics described in conjunction with the embodiment may be in at least one embodiment of the present application. The appearance of the phrase in various places in the specification does not necessarily refer to the same embodiment, nor is it an independent alternative embodiment that is mutually exclusive with other embodiments. It is explicitly and implicitly understood by those skilled in the art that the embodiments described herein may be combined with other embodiments.

[0029] Example 1: Reference Figure 1 , is a flow chart of an intelligent defect detection and repair method for vacuum isothermal forgings based on multimodal sensing data provided by an embodiment of the present invention, and the flow chart may at least include steps S100-S400: S100, acquiring image data of vacuum isothermal forgings through a visual sensor, acquiring thermal distribution data through a thermal sensor, acquiring stress monitoring data through a stress sensor, preprocessing the image data, thermal distribution data and stress monitoring data, and generating a time-space aligned multimodal data set.

[0030] S200, fusing the multimodal data set, extracting multidimensional features of defects through a deep learning network, and generating a defect feature sequence of defect location, category, and severity.

[0031] S300, based on the defect feature sequence and in combination with stress distribution characteristics, the repair path planning model is optimized to generate a repair task sequence of repair methods, repair parameters and path optimization solutions.

[0032] S400, inputting the repair task sequence into the intelligent repair equipment, performing repair operations on the vacuum isothermal forgings, and monitoring the repair status in real time through sensors, generating a repair quality report and feeding it back to the repair path planning model for dynamic adjustment.

[0033] Step S100 includes at least steps S110-S130: S110, acquiring image data through a visual sensor and performing preprocessing.

[0034] First, the image data I(x,y,t) of the vacuum isothermal forging is obtained by the visual sensor, where x and y are the two-dimensional coordinates of the image and t is the time series. The image data contains defect information (cracks, scratches, deformation, etc.) on the surface of the forging, but there may be noise interference at the same time.

[0035] Furthermore, noise filtering is performed by applying a two-dimensional Gaussian filter G(x,y;σ) to I(x,y,t). The filtering formula is: Where σ is the standard deviation of the filter, which is used to control the degree of smoothing. The obtained I'(x,y,t) is the image data after removing high-frequency noise.

[0036] Furthermore, feature enhancement processing is performed. The Laplacian operator is used to enhance the features of high gradient areas in the image. The formula is: Among them, I''(x, y, t) is the image data after feature enhancement, which is used to extract crack edge and surface defect features.

[0037] The preprocessed image data I''(x, y, t) is obtained as the input for subsequent multimodal data fusion.

[0038] S120, collecting heat distribution data through a thermal sensor and performing preprocessing.

[0039] First, obtain thermal distribution data. Specifically, obtain the thermal distribution data T(x, y, t) of the vacuum isothermal forging through a thermal sensor, where T represents the temperature value of the pixel position, x and y are spatial coordinates, and t is the time series. The abnormal temperature distribution caused by the forging process in the thermal distribution data can reflect the potential defects inside the forging.

[0040] Furthermore, denoising is performed. The time series of T(x,y,t) is smoothed using a sliding average filter. The formula is: Where n is the time window size. The filtered T'(x,y,t) eliminates high-frequency noise in the temperature sampling.

[0041] Furthermore, normalization is performed. Normalization is performed on T'(x,y,t) to normalize the temperature value to the range of [0,1]: in, and are the maximum and minimum values ​​of the thermal distribution data respectively.

[0042] The normalized thermal distribution data T''(x, y, t) is obtained as the input for subsequent multimodal data fusion.

[0043] S130, collecting stress monitoring data through stress sensors and performing preprocessing.

[0044] First, stress monitoring data is obtained. The stress monitoring data S(x, y, t) of the vacuum isothermal forging is collected by the stress sensor, where S represents the stress value at the pixel position, x and y are spatial coordinates, and t is the time series. The stress monitoring data can reflect the stress distribution inside the forging and can be used to identify potential structural defects.

[0045] Furthermore, the time series of S(x,y,t) are aligned with I''(x,y,t) and T''(x,y,t) to ensure that the three modal data are at the same time point. There is consistency in: Furthermore, a spatial gradient analysis is performed on S(x, y, t) to calculate the local variation characteristics in the stress distribution. The formula is: Among them, ∇S(x,y,t) is the gradient vector of stress distribution, which is used to locate the stress concentration area.

[0046] Furthermore, a spatiotemporal aligned multimodal dataset is generated that is consistent with the image data I''(x, y, t) and the thermal distribution data T''(x, y, t) The features of thermal distribution and stress distribution are uniformly preprocessed and input into the multimodal dataset D(x, y, t), providing a unified basis for feature extraction of subsequent deep learning.

[0047] Step S200 includes at least steps S210-S230: S210: Perform deep feature extraction on the image data in the multimodal data set to generate an image defect feature sequence.

[0048] First, the pre-processed image data I''(x, y, t) from S100 is input. A convolutional neural network (CNN) is used to extract features from I''(x, y, t). The calculation formula of the convolution kernel is: in, Represents the eigenvalue after convolution; k is the convolution kernel size; is the convolution kernel weight; b is the bias value.

[0049] Furthermore, the convolution feature map is input into the maximum pooling layer to extract key feature points and obtain the image defect feature matrix: in, It is an image defect feature sequence used to represent crack morphology and edge characteristics.

[0050] Connection Instructions: It will be used as the input of the subsequent S230 module and fused with other modal features.

[0051] S220, extracting features from the thermal distribution data in the multimodal data set, and generating a thermal distribution feature sequence based on the temperature anomaly area.

[0052] First, input the normalized thermal distribution data T''(x,y,t) from S100. Apply temperature gradient calculation to T''(x,y,t), expressed as: in, Represents the local gradient of temperature change and is used to locate areas of temperature anomalies.

[0053] Furthermore, by setting the temperature threshold , extract abnormal areas in thermal distribution data: in, is the coordinate set of the abnormal area.

[0054] Furthermore, based on the distribution characteristics of the abnormal area, a heat distribution feature sequence is generated: in, Represents local abnormal characteristics and gradient information of thermal distribution.

[0055] Connection Instructions: With the image feature sequence and stress characteristic sequence Fusion in the S230.

[0056] S230, performing multimodal fusion processing on the image defect feature sequence, the thermal distribution feature sequence and the stress feature sequence to generate a defect feature sequence of defect location, category and severity.

[0057] First, input the stress gradient feature sequence ∇S(x,y,t) from S130. Time-align the stress data with the image and thermal distribution data using the formula: in, To ensure the temporal consistency of the three-modal data after alignment.

[0058] Furthermore, FI, FT, ∇S(x, y, t) are input into the deep learning fusion network, and the multi-layer attention mechanism is used to calculate the correlation weights between modalities: in, are the fusion weights of image, thermal distribution, and stress features, respectively; , , is the feature transformation matrix.

[0059] Furthermore, the feature vectors of defect location, category and severity are calculated based on the fusion weights: in, Represents the fused defect feature sequence, including the position ,category and severity .

[0060] Connection Instructions: It will be directly used as input to the S300 module for planning and optimization of the repair path.

[0061] Step S300 includes at least steps S310-S330: S310, preliminarily generating a repair method and repair parameters based on the defect category and position in the defect feature sequence.

[0062] First, input the defect feature sequence output by S230 ={ , , }, in, =( , , ) is the spatial position of the defect; is the defect type (e.g. cracks, holes, overheated areas); The severity of the defect.

[0063] Further, according to and , choose the appropriate repair method , the specific rules are as follows: in, and is the threshold of defect severity.

[0064] Furthermore, the repair parameters are calculated. Specifically, for each repair method , calculate its corresponding parameter set , such as laser power Heat treatment temperature , the calculation formula is: Where f is the parameter generating function.

[0065] Connection Notes: Preliminary Generation of Repair Methods and repair parameters It will be used as input to S320 for path optimization.

[0066] S320. Optimize the repair path planning model based on stress distribution characteristics and defect locations, and generate a repair path optimization plan.

[0067] First, input the stress gradient characteristic ∇S(x,y,t) of S130 and the defect location , calculate the stress distribution around the defect : in, is the local volume where the defect is located.

[0068] Furthermore, based on the repair method and stress distribution , establish a repair path optimization model: in, To repair the path length; is the stress value of the area through which the path passes; , is the path weight coefficient.

[0069] Furthermore, the repair path optimization plan is generated: the path model is solved using a dynamic programming algorithm to generate a path optimization plan , its repair order and waypoints.

[0070] Connection Instructions: Repair Path Optimization Plan And repair method , Repair parameters Will serve as input for S330.

[0071] S330, verifying the repair path planning model, integrating the repair method, repair parameters and path optimization scheme to generate a repair task sequence.

[0072] First, use simulation tests to optimize the path Verify and calculate path feasibility in, is the cost of repairing the area through which the path passes; Stress changes after repair; , To verify the weight coefficient.

[0073] Furthermore, the repair method , Repair parameters and path optimization solutions Integration, generation of repair task sequence : Connection Notes: Repair Task Sequence It will be used as input to the S400 module for repairing the equipment execution.

[0074] Step S400 includes at least steps S410-S430: S410: Input the repair task sequence into the intelligent repair device to execute the repair operation.

[0075] First, enter the repair task sequence exported by S330 ,in, for repair methods (e.g. laser cladding, mechanical compression, local heat treatment); For repair parameters (such as laser power , Heat treatment temperature wait); Optimize the repair path, including path points and repair sequence.

[0076] Furthermore, the parameters Parse it into device instructions and execute specific repair operations through intelligent repair equipment. For laser repair, control the power of laser equipment and scanning speed , and its relationship is expressed as: in, is the laser energy per unit area.

[0077] Furthermore, according to the repair path , drives the repair equipment to move along the set path points and repairs the defect at the defect location Complete the repair point by point.

[0078] Connection description: After the repair operation is completed, the repair status is collected through the sensor as the input of S420.

[0079] S420, monitoring the repair status in real time through sensors, performing feedback processing on stress and heat distribution data of the repair area, and generating real-time monitoring data of the repair.

[0080] First, the image data of the repair area is obtained through the visual sensor , collect heat distribution data through thermal sensors , collect stress distribution data through stress sensors .

[0081] Furthermore, the collected image data Perform differential analysis with the data before restoration I''(x,y,t) to calculate the visual changes of the restoration area: Heat distribution data Compare with the data T''(x,y,t) before repair to locate the temperature change after the thermal anomaly is repaired: Stress distribution data Analyze the gradient change with the gradient ∇S(x,y,t) before repair: ΔS(x,y,t)=∇ (x,y,t)−∇S(x,y,t) Furthermore, the visual, thermal distribution and stress data changes are integrated to generate real-time monitoring data for repairs. : Connection Notes: Real-time Monitoring Data As input to S430, it is used to adjust the repair parameters and paths.

[0082] S430: Input the real-time repair monitoring data into a repair path planning model, dynamically adjust the repair method and parameters, and generate a final repair quality report.

[0083] First, real-time monitoring data Input the repair path planning model and adjust the repair parameters and path Perform dynamic adjustment, the adjustment formula is: in, is the repair parameter after adjustment; is the adjusted repair path; , is the adjustment factor.

[0084] Furthermore, the residual severity of the defect after repair is calculated based on the adjusted repair result. And the repair completion: like ≤ϵ, the repair is considered successful, where ϵ is the allowable error.

[0085] Furthermore, the adjusted repair parameters ,path , Repair quality report Integration completes the final repair loop.

[0086] Connection Notes: Repair Quality Report It will serve as the verification basis for the system closed loop and will be fed back to the previous module to optimize the model.

[0087] The key innovative features of the present invention include: (1) Combination of multimodal data fusion and deep learning: This invention, for the first time, aligns visual, thermal and stress data in time and space, extracts multidimensional features through a deep learning network, and analyzes the properties and distribution of complex defects through multimodal fusion, thus overcoming the limitations of traditional single-modal detection.

[0088] (2) Dynamic repair task sequence generation and optimization: The repair path planning is dynamically adjusted by utilizing the defect feature sequence and stress distribution characteristics to generate a repair task sequence including repair methods, repair parameters, and optimized paths, thus realizing an intelligent repair process.

[0089] (3) Real-time feedback and closed-loop optimization mechanism: By monitoring the repair status in real time and feeding back the repair monitoring data to the path planning model for dynamic adjustment, a closed-loop system of detection, repair, and verification is formed, which significantly improves the accuracy and adaptability of the repair.

[0090] (4) Efficient application of mathematical models: Introducing complex mathematical calculations into the detection and repair process, including convolutional neural network feature extraction, gradient analysis, dynamic path planning optimization, etc., ensures the efficiency of data processing and model optimization, and lays the foundation for large-scale industrial applications.

[0091] The following are its main beneficial effects: (1) Improve detection accuracy: Through the spatiotemporal alignment and fusion processing of the three sensor data, namely vision, thermal sensing and stress, the surface and internal defect characteristics of the forgings are comprehensively analyzed, solving the problem of low detection accuracy of a single sensing mode. The deep learning network further extracts and classifies multi-dimensional features, significantly improving the accuracy and comprehensiveness of defect detection.

[0092] (2) Optimizing repair efficiency: Based on the defect feature sequence and combined with the stress distribution characteristics, an optimized repair path planning model is generated. By dynamically adjusting the repair method and parameters, the accuracy and efficiency of the repair process are improved and redundant repair steps are reduced.

[0093] (3) Enhance system adaptability: During the repair process, sensors are used to monitor the repair status in real time and dynamically feed back to the path planning model to adjust the repair parameters and path, ensuring that the repair process can adapt to complex defect types and different working environments.

[0094] (4) Reduce production costs: Through intelligent defect detection and repair processes, human involvement is reduced, the defective rate is reduced, and the use of repair resources is optimized, significantly reducing production and repair costs.

[0095] (5) Improve product quality: The system verifies the repair effect in real time during the repair process to ensure that the structural integrity and performance indicators of the repaired forgings meet the requirements and meet the strict quality standards of high-end manufacturing.

[0096] Embodiment 2: Figure 2 FIG. 2 shows a block diagram of an intelligent defect detection and repair system for vacuum isothermal forgings based on multimodal sensing data according to an embodiment of the present invention. Figure 2 As shown, the system may include: The data acquisition module 10 acquires the image data I(x, y, t) of the vacuum isothermal forging through a visual sensor, and collects the defect information (such as cracks, scratches and deformation) on the surface of the forging.

[0097] The thermal distribution data T(x,y,t) of the forging is collected by thermal sensors, and the abnormal temperature distribution of the forging is captured to reflect internal defects.

[0098] The stress distribution data S(x,y,t) is collected by stress sensors to monitor the stress concentration of forgings and locate potential structural defects.

[0099] The data processing and fusion module 20 performs preprocessing on the collected data, including noise filtering and feature enhancement of image data, denoising and normalization of thermal distribution data, time alignment and gradient calculation of stress distribution data.

[0100] The preprocessed multimodal data Unify the input, perform deep learning fusion, and extract multi-dimensional features of defects.

[0101] Output defect feature sequence ={ , , },in is the defect location, is the defect category, The severity of the defect.

[0102] Repair task generation module 30, based on defect feature sequence , combined with the stress distribution characteristics ∇S(x,y,t), select the appropriate repair method (Laser cladding, mechanical compression, local heat treatment).

[0103] Generate repair parameters Such as laser power , Heat treatment temperature and optimized repair paths .

[0104] Export Repair Task Sequence , as input for subsequent repair operations.

[0105] Intelligent repair and feedback module 40, repair task sequence Input the intelligent repair device and perform specific repair operations according to the path RdR_{d}.

[0106] Monitor the repair status in real time through sensors and collect image data after repair , thermal distribution data , stress data .

[0107] Analyze visual, thermal distribution and stress changes in the repair area to generate real-time monitoring data .

[0108] Dynamically adjust the repair parameters and paths based on the monitoring data, and finally generate a repair quality report .

[0109] Beneficial effects of the embodiments: (1) Efficient defect detection and repair closed loop: The system realizes a full-process closed loop from defect detection and classification to repair operations, significantly improving the detection and repair efficiency.

[0110] (2) Multimodal data fusion improves accuracy: Through the multimodal fusion of visual, thermal and stress data, the system can comprehensively capture the surface and internal defect characteristics of forgings, thereby improving the accuracy of defect detection.

[0111] (3) Intelligent repair task planning: Combining deep learning algorithms and dynamic path optimization models, the system can generate customized repair tasks for different defect categories and severities, ensuring the accuracy and adaptability of the repair process.

[0112] (4) Dynamic feedback mechanism enhances reliability: The real-time monitoring and feedback mechanism enables the system to dynamically adjust the repair parameters and paths according to the repair effect, thereby improving the success rate of repair and the quality of forgings.

[0113] Obviously, the embodiments described above are only some embodiments of the present application, rather than all embodiments. The preferred embodiments of the present application are given in the accompanying drawings, but they do not limit the patent scope of the present application. The present application can be implemented in many different forms. On the contrary, the purpose of providing the embodiments is to make the understanding of the disclosure of the present application more thorough and comprehensive. Although the present application is described in detail with reference to the aforementioned embodiments, for those skilled in the art, it is still possible to modify the technical solutions recorded in the aforementioned specific implementation methods and to replace some of the technical features therein with equivalents. Any equivalent structure made using the contents of the specification and drawings of this application, directly or indirectly used in other related technical fields, is similarly within the scope of patent protection of this application.

Claims

1. An intelligent defect detection and repair method for vacuum isothermal forgings based on multimodal sensing data, characterized in that: The following steps are involved: Acquire image data of vacuum isothermal forgings through a visual sensor, collect thermal distribution data through a thermal sensor, collect stress monitoring data through a stress sensor, pre-process the image data, thermal distribution data and stress monitoring data, and generate a multi-modal data set aligned in time and space; The multimodal data set is fused and processed, and multidimensional features of defects are extracted through a deep learning network to generate a defect feature sequence of defect location, category, and severity; Based on the defect feature sequence and combined with stress distribution characteristics, the repair path planning model is optimized to generate a repair task sequence of repair methods, repair parameters and path optimization solutions; The repair task sequence is input into the intelligent repair equipment, the vacuum isothermal forging is repaired, and the repair status is monitored in real time through sensors, and a repair quality report is generated and fed back to the repair path planning model for dynamic adjustment.

2. The method according to claim 1, characterized in that The step of acquiring image data through a visual sensor includes: Noise filtering and feature enhancement processing are performed on the image data to generate preprocessed image data.

3. The method according to claim 2, characterized in that The step of collecting heat distribution data by means of a thermal sensor comprises: The heat distribution data is subjected to denoising and normalization processing to generate pre-processed heat distribution data.

4. The method according to claim 1, characterized in that The step of collecting stress monitoring data through the stress sensor includes: The stress monitoring data is subjected to time alignment and feature extraction processing to generate stress gradient features consistent with the image data and the thermal distribution data.

5. The method according to claim 1, characterized in that: The fusion processing step of the multimodal data set includes: A convolution operation is performed on the image data to extract image defect features, a gradient analysis is performed on the thermal distribution data to extract temperature anomaly features, and feature extraction is performed on the stress monitoring data to generate a multi-modal feature sequence of defects.

6. The method according to claim 5, characterized in that The step of extracting multi-dimensional features through the deep learning network includes: The weight of each modal feature is calculated through the attention mechanism, and the image features, thermal distribution features and stress features are weightedly fused to generate the defect location, category and severity.

7. The method according to claim 1, characterized in that The repair path planning model optimization step includes: The preliminary repair method and repair parameters are generated according to the defect feature sequence, the repair path is optimized based on the stress distribution characteristics, and the repair task sequence is generated.

8. The method according to claim 7, characterized in that The repair task sequence includes: Repair method, repair parameters and path optimization scheme, wherein the path optimization scheme includes the spatial position of the repair points and the repair sequence.

9. The method according to claim 1, characterized in that: The steps of executing and providing feedback on the repair task include: Repair operations are performed through intelligent repair equipment according to the repair task sequence, real-time data of the repair status is collected, repair monitoring data is generated and fed back to the path planning model, and the repair parameters and paths are dynamically adjusted.

10. A vacuum isothermal forging intelligent defect detection and repair system based on multimodal sensing data, characterized in that: include: A data acquisition module, used to acquire the image data, thermal distribution data and stress monitoring data; A data processing and fusion module, used for preprocessing and fusion processing the data to generate a multi-modal feature sequence of defects; A repair task generation module, used to generate a repair task sequence based on the defect feature sequence; The intelligent repair module is used to execute the repair task sequence, monitor the repair status in real time through sensors, and feed back the real-time monitoring data of the repair to the repair task generation module.

Images