Operation and maintenance management method and system for vehicle frame production

Abstract

The invention discloses an operation and maintenance management method and system for vehicle frame production, and relates to the technical field of production operation and maintenance. An operation and maintenance management system for vehicle frame production comprises a vehicle frame production information acquisition module and a vehicle frame production operation and maintenance management module. According to the method, internal part sensing information and external image information are utilized, the equipment operation state and historical process parameters are fused, the appearance quality of a product is concerned, dynamic changes in the machining process are also concerned, unified analysis of the product quality, the process stability and the equipment health state is achieved, and the product quality is improved. The middleware state identification precision and real-time performance are effectively improved; a refined grading mechanism is introduced, so that the problem of excessive maintenance or misjudgment in traditional operation and maintenance management is avoided; and particularly, the critical frame middleware is re-divided and analyzed, so that repairable intermediate-state products at the edge of a fault can be identified more accurately, on-demand processing and dynamic scheduling are realized, and resource waste is avoided.

Description

Technical Field

[0001] The present invention relates to the field of production operation and maintenance technology, and in particular to an operation and maintenance management method and system for frame production. Background Art

[0002] In traditional vehicle frame production, quality inspection often relies on manual inspection or single-sensor monitoring. This leads to issues such as delayed response, inaccurate anomaly identification, and crude defect localization, making it difficult to achieve refined quality control and predictive equipment maintenance. Existing technologies, while employing image recognition and equipment status prediction methods, often rely on single-point data and lack comprehensive assessment capabilities of frame structural components, equipment health, and process complexity. This makes it impossible to accurately identify critical conditions or provide actionable control recommendations.

[0003] Therefore, a maintenance management method for vehicle frame production is needed that can integrate internal sensor information and external image information, redistribute abnormal middleware through predictive models, and make multi-dimensional judgments based on vehicle frame structural parts and risk trends. It can not only identify critical states, but also perform closed-loop repairs through maintenance strategies, significantly improving the intelligent, flexible, and traceable maintenance capabilities of the vehicle frame production line. Summary of the Invention

[0004] The present invention aims to provide an operation and maintenance management method and system for vehicle frame production, to perform operation and maintenance management on defective states that are easily overlooked in vehicle frame production, and to build an intelligent and traceable vehicle frame production line.

[0005] An operation and maintenance management method for vehicle frame production includes the following steps: Obtaining production process information and corresponding production equipment information from the vehicle frame production middleware; the production process information includes internal component sensor information and external component image information; performing detection based on the external component image information to obtain vehicle frame production and maintenance information; and identifying the vehicle frame part labels from the vehicle frame production middleware. Add a frame operation and maintenance tag to the frame production middleware based on the frame production and operation and frame part tags; the frame operation and maintenance tag includes frame middleware to be operated and maintained, abnormal frame middleware, and normal frame middleware; For abnormal frame middleware, production prediction is performed based on internal component sensor information and production equipment information to obtain frame production prediction information; based on the frame production prediction information, abnormal frame middleware is subdivided into critical frame middleware or frame middleware to be maintained; the frame production prediction information marked as critical frame middleware is analyzed to obtain production equipment operation and maintenance information; For the frame middleware to be maintained, perform abnormal operation and maintenance operations on the frame middleware to be maintained; for the normal frame middleware, continue to perform the next production operation on the normal frame middleware.

[0006] As a preferred technical solution of the present invention, the specific steps of adding a frame operation and maintenance tag to the frame production middleware based on the frame production and operation and maintenance information and the frame part tag include: Compare the frame production and maintenance information with the preset frame part operation and maintenance standards to obtain the frame production comparison results; Frame production deviation is identified based on the frame production comparison results and frame part labels to obtain the frame production deviation factor; Based on the frame production deviation factor, the frame production middleware is marked as frame middleware to be maintained, abnormal frame middleware and normal frame middleware.

[0007] As a preferred technical solution of the present invention, the specific steps of performing production prediction based on internal component sensor information and production equipment information include: Align the internal component sensor information and production equipment information in time series to obtain a fusion sequence of middleware processing information; process the fusion sequence of middleware processing information using a sliding window of indefinite length to obtain several continuous sliding middleware processing information slices; Extract the change trend, coefficient of variation and perturbation ratio of the sliding middleware processing information piece; Obtain historical data on vehicle frame production; determine equipment health factors and process complexity index based on historical data; The change trend, coefficient of variation, perturbation ratio, equipment health factor, and process complexity index are integrated to obtain a frame processing fusion vector. Production prediction is performed based on the frame processing fusion vector to obtain frame production prediction information. The frame production forecast information includes production reliability forecast value, production abnormality risk trend and frame production abnormality forecast location.

[0008] As a preferred technical solution of the present invention, the specific steps of reclassifying abnormal frame middleware based on frame production prediction information include: Perform interval mapping based on the production reliability prediction value in the frame production prediction information to obtain production reliability level area information; The production anomaly risk trend in the frame production forecast information is judged. If the production anomaly risk trend is in the trend enhancement range but does not exceed the severe anomaly threshold, the critical frame weighted information is obtained; otherwise, the weighted information of the frame to be operated and maintained is obtained. Critical judgment is made based on the predicted parts of frame production abnormality in the frame production prediction information: Step S1: Identify the importance level of the predicted location of the frame production anomaly; perform a comprehensive judgment based on the importance level, production reliability level area information, and critical frame weighted information or weighted information of the frame to be maintained, and obtain first frame middleware judgment information; Step S2: performing a judgment based on the vehicle frame production prediction information and the pre-trained enhanced fuzzy judgment model to obtain second vehicle frame middleware judgment information; According to the first frame middleware determination information and the second frame middleware determination information, it is determined that the abnormal frame middleware belongs to the critical frame middleware or the frame middleware to be maintained.

[0009] As a preferred technical solution of the present invention, the specific steps of analyzing the frame production forecast information marked as critical frame middleware to obtain production equipment operation and maintenance information include: Based on the production anomaly risk trend and the predicted location of the frame production anomaly, reverse judgment is made to match similar known internal equipment faults. If the match is successful, the production equipment operation and maintenance information is generated based on the known internal equipment faults. If the match fails, the production equipment is readjusted based on the corresponding production equipment information to obtain the production equipment operation and maintenance information. After adjusting the corresponding production equipment using the production equipment operation and maintenance information, the critical frame middleware is reprocessed until the normal frame middleware is obtained.

[0010] As a preferred technical solution of the present invention, the specific steps of performing detection based on external component image information to obtain frame production and operation information include: using a trained frame defect detection and recognition model to perform feature recognition on external component image information to obtain frame production and operation information; and training the frame defect detection and recognition model based on a CNN model.

[0011] An operation and maintenance management system for vehicle frame production, comprising: The frame production information acquisition module includes a data acquisition unit; the data acquisition unit is used to obtain production process information and corresponding production equipment information of the frame production middleware; the production process information includes internal component sensor information and external component image information; based on the external component image information, detection is performed to obtain frame production and maintenance information; and the frame part labels of the frame production middleware are identified; The frame production and operation and maintenance management module includes an operation and maintenance identification unit and an operation and maintenance judgment unit; the operation and maintenance identification unit is used to add frame operation and maintenance labels to the frame production middleware based on the frame production and operation and maintenance information and the frame part label; the frame operation and maintenance label includes the frame middleware to be operated and maintained, abnormal frame middleware and normal frame middleware; the operation and maintenance judgment unit is used to make production predictions for the abnormal frame middleware based on the internal component sensor information and the production equipment information to obtain the frame production prediction information; based on the frame production prediction information, the abnormal frame middleware is subdivided to obtain the critical frame middleware or the frame middleware to be operated and maintained; the frame production prediction information marked as the critical frame middleware is analyzed to obtain the production equipment operation and maintenance information; for the frame middleware to be operated and maintained, the operation and maintenance abnormality operation is performed on the frame middleware to be operated and maintained; for the normal frame middleware, the next production operation is continued to be performed on the normal frame middleware.

[0012] The present invention has the following advantages: 1. The present invention utilizes internal component sensor information and external image information, integrates equipment operating status and historical process parameters, and focuses not only on the product's appearance quality but also on dynamic changes during the processing process, thereby achieving a unified analysis of product quality, process stability, and equipment health status, and effectively improving the accuracy and real-time performance of middleware status identification. By introducing a refined grading mechanism, the problem of excessive maintenance or misjudgment in traditional operation and maintenance management is avoided. In particular, the redivision and analysis of critical frame middleware can more accurately identify intermediate products that are on the verge of failure but can be repaired, realize on-demand processing and dynamic scheduling, and avoid waste of resources.

[0013] 2. By constructing a fusion feature vector and a trained prediction model, the present invention can make advance judgments on production trends before defects occur, thereby realizing early identification of abnormal trends; in particular, the introduction of equipment health factors and process complexity indexes makes the model more context-sensitive, improving the interpretability and adaptability of predictions; after identifying the critical state, it can automatically match the equipment failure mode, or reversely analyze and adjust the equipment parameters, complete automatic repair or fine-tuning, and reprocess and verify the middleware, forming a complete processing control closed loop, effectively reducing the scrap rate, and improving the stable operation capability of the production line. BRIEF DESCRIPTION OF THE DRAWINGS

[0014] Figure 1 This is a schematic structural diagram of an operation and maintenance management system for vehicle frame production adopted in an embodiment of the present invention. DETAILED DESCRIPTION

[0015] In order to enable those skilled in the art to better understand the technical solutions in the present invention, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the accompanying drawings in the embodiments of the present invention.

[0016] Example 1, an operation and maintenance management method for frame production, comprising the following steps: Obtaining production process information and corresponding production equipment information from the vehicle frame production middleware; the production process information includes internal component sensor information and external component image information; performing detection based on the external component image information to obtain vehicle frame production and maintenance information; and identifying the vehicle frame part labels from the vehicle frame production middleware. The primary purpose of identifying frame part labels in vehicle frame production middleware is to achieve precise defect location and resolution, structured quality analysis, personalized maintenance strategies, and automated production decision-making. In the production process of complex vehicle frames and diverse processes, the processing requirements, tolerance standards, and quality control points of different parts vary significantly. Simply identifying a problem in the middleware is not enough; the specific structural part must be identified. Production equipment information refers to a collection of information such as various static properties, dynamic states, operating histories, and process configurations related to the equipment involved in the frame production process. It reflects the technical capabilities of the equipment itself, its current operating status, and its impact on the production process. It specifically refers to the specific process parameters used by the equipment when performing different processing tasks, such as the current and voltage of the welding machine, the speed and angle of the cutting equipment, the clamping pressure of the assembly equipment, etc. These parameters are direct variables that affect product quality.

[0017] The operation and maintenance management methods of vehicle frame production involve multiple data collection and analysis links. Internal component sensor information refers to the data collected in real time by various sensors embedded in the vehicle frame fixtures, processing equipment or middleware during the frame production process. This information usually includes parameters such as temperature, pressure, torque, vibration, current, voltage, etc. generated during processing, which are used to reflect the state changes of the internal structure during the processing; the method of obtaining this sensor information relies on IoT sensors installed in key positions, which realize continuous monitoring of the internal operating status during the processing through real-time collection and data upload operations.

[0018] External component image information refers to the appearance images of the frame middleware obtained through the industrial vision system. These images can cover key parts such as welds, connection points, and shell surfaces, and are used to detect whether there are external visible defects such as cracks, deformation, burns, incomplete welding, etc.; image acquisition usually uses multi-angle industrial cameras, which automatically shoot after the frame completes a certain key process, and cooperates with the light source system to improve image clarity; the obtained image information is then sent as input to the frame defect detection and recognition model for identification and analysis.

[0019] The specific steps for detecting and obtaining frame production and maintenance information based on external component image information include: The trained frame defect detection and recognition model is used to perform feature recognition on external component image information to obtain frame production and operation information. The frame defect detection and recognition model is trained based on the CNN model. Frame production and maintenance information refers to a collection of data obtained during the frame production process based on real-time monitoring and analysis of the status of intermediate components (semi-finished frames) to guide subsequent maintenance judgments and decisions. This data includes information on the presence of visual defects such as cracks, deformation, and poor solder joints (derived from external image recognition), as well as internal parameters such as stress anomalies, temperature anomalies, and current offsets (derived from sensor data). Training a vehicle frame defect detection and recognition model primarily relies on building a high-quality dataset. To ensure the model can accurately identify various external defects that may occur in the vehicle frame, a large number of frame image samples from actual production lines must be collected. These images should cover different process stages, different part areas, and contain multiple types of defects, such as weld cracks, cold welds, burns, deformation, and foreign matter adhesion. After image acquisition, experienced quality inspection engineers must annotate the images, clearly indicating the location, range, and type of defects, forming structured image-label pairs that serve as the basic data for model training. The model is built based on a convolutional neural network architecture, typically using a deep network structure with strong feature extraction capabilities, such as a residual connection-based network structure or a lightweight feature pyramid structure, to better extract detailed features such as welds and edges. During the model training phase, preprocessed image data is input along with the annotated defect labels, and parameter optimization is performed using a backpropagation algorithm. During training, a cross-entropy loss function or a multi-task loss function is used to jointly optimize target detection accuracy and localization accuracy. A positive-negative sample balancing strategy is also introduced to prevent the model from being overly biased against large background areas. During the training process, a validation set will be set up to monitor the model's performance on non-training samples to prevent overfitting. When the model performs stably on the validation set and reaches the expected accuracy standard, training can be stopped and the model weights can be saved to obtain a final frame defect detection and recognition model that can be used in a production environment.

[0020] Add a frame operation and maintenance tag to the frame production middleware based on the frame production and operation and frame part tags; the frame operation and maintenance tag includes frame middleware to be operated and maintained, abnormal frame middleware, and normal frame middleware; The specific steps for adding frame operation and maintenance tags to the frame production middleware based on the frame production and operation information and frame part tags include: Compare the frame production and maintenance information with the preset frame part operation and maintenance standards to obtain the frame production comparison results; The frame operation and maintenance standards are a quality threshold system for different structural parts established based on historical production data, quality specifications, process requirements and empirical knowledge, such as the allowable range of weld width, maximum stress value at the connection, vibration signal fluctuation limit, etc.; when performing comparisons, the system will compare the external image detection results and internal sensor data with these standards item by item to determine whether each indicator exceeds the normal tolerance range, and output the comparison results in a structured form to provide a basis for subsequent deviation analysis.

[0021] Frame production deviation is identified based on the frame production comparison results and frame part labels to obtain the frame production deviation factor; After obtaining the frame production comparison results, combined with the identified frame part labels, the system will conduct refined deviation identification on the current middleware production quality status. This process not only determines whether a certain indicator exceeds the standard, but also comprehensively considers the location of the deviation, deviation amplitude, process steps and historical abnormality distribution. For example, the same temperature anomaly may appear in the bottom tube and the transfer tube, and its impact may be different. By building a set of rule bases and deviation calculation models, the system combines the comparison results and part labels to calculate the frame production deviation factor that reflects the degree of deviation and the intensity of the impact, and clearly points out the severity of the problem and the scope of impact.

[0022] Based on the frame production deviation factor, the frame production middleware is marked as frame middleware to be maintained, abnormal frame middleware and normal frame middleware; After obtaining the frame production deviation factor, the system will classify the middleware into operation and maintenance labels based on its numerical range, trend judgment and contextual process conditions; if the deviation factor is low, the fluctuation is stable and there is no obvious development trend, the middleware will be marked as normal frame middleware; if the deviation factor is in the medium range, it may be controllable in the short term but there is an uncertainty risk, and it will be marked as frame middleware to be maintained for further observation or reprocessing; if the deviation factor exceeds the serious threshold, or there is a significant deviation in the key parts, it will be directly marked as abnormal frame middleware, which requires rework or intervention.

[0023] For abnormal frame middleware, production prediction is performed based on internal component sensor information and production equipment information to obtain frame production prediction information; based on the frame production prediction information, abnormal frame middleware is subdivided into critical frame middleware or frame middleware to be maintained; the frame production prediction information marked as critical frame middleware is analyzed to obtain production equipment operation and maintenance information; The specific steps for making production predictions based on internal component sensor information and production equipment information include: Align the internal component sensor information and production equipment information in time series to obtain a fusion sequence of middleware processing information; process the fusion sequence of middleware processing information using a sliding window of indefinite length to obtain several continuous sliding middleware processing information slices; These two types of information are aligned and processed along a unified timeline to form a fusion sequence of middleware processing information covering the entire processing cycle. This sequence integrates sensor values ​​such as temperature, vibration, and current during the processing, as well as equipment operating status, processing rhythm, load changes, and other content to construct a dynamic process dataset that reflects the actual processing status of the frame middleware within a specific period of time, laying the foundation for subsequent feature extraction and trend analysis.

[0024] After generating a fusion sequence of middleware processing information, the system slices it using a sliding window of variable length. The length of the sliding window can be flexibly set based on different process stages or sensor data fluctuations, ensuring that each information slice covers key change periods. Through window sliding, the system generates a series of continuous middleware processing information slices, each representing a local operating condition variation interval. This process helps capture subtle but significant trends in the processing process, while avoiding misjudgments caused by data dilution or noise interference, and enhancing the sensitivity and stability of local predictions.

[0025] Extract the change trend, coefficient of variation and perturbation ratio of the sliding middleware processing information piece; After obtaining the sliding middleware processing information piece, feature extraction is performed on each segment, focusing on calculating its change trend, coefficient of variation and perturbation ratio; the change trend is used to determine whether the processing parameters in the segment tend to be stable, increasing or decreasing, reflecting the evolution direction of the process; the coefficient of variation is used to measure the relative amplitude of data fluctuations in the segment and evaluate the stability of the processing process; the perturbation ratio focuses on the degree of mutation between the current segment and the previous segment, and is an important indicator for identifying sudden anomalies; through the extraction of these features, the system can construct a multi-dimensional dynamic characterization of the current middleware processing status.

[0026] Obtain historical data on vehicle frame production; determine equipment health factors and process complexity index based on historical data; The equipment health factor is a comprehensive indicator reflecting the equipment's current operating status, performance stability, and failure risk level. It measures whether a particular piece of equipment is in a healthy and safe processing state. Its value is typically a standardized score ranging from 0 to 100, with higher values ​​indicating healthier equipment and lower values ​​indicating deteriorating operating conditions and higher risks. The process complexity index measures the degree of dependence and sensitivity of a particular frame processing step on equipment, operating accuracy, and parameter control requirements during execution. It reflects the difficulty and risk level of quality control for the middleware's current process and is typically a standardized score. Higher values ​​indicate a more complex process and higher system requirements.

[0027] By calling the frame production history database, historical production data related to the current middleware process node, equipment type and material batch is obtained; based on this historical data, the system conducts a comprehensive assessment of the equipment currently involved in the processing and generates an equipment health factor that reflects factors such as its operating status, failure frequency, maintenance record and usage time; at the same time, a process complexity index is generated based on indicators such as process route complexity, processing accuracy requirements, and control parameter fluctuation tolerance; these two types of factors jointly reflect the background conditions that affect the current production environment on the frame middleware, which helps to improve the model's ability to judge boundary states and critical risks.

[0028] The change trend, coefficient of variation, perturbation ratio, equipment health factor, and process complexity index are integrated to obtain a frame processing fusion vector. Production prediction is performed based on the frame processing fusion vector to obtain frame production prediction information. The change trend, coefficient of variation, and perturbation ratio extracted from the processing information slice are integrated with the equipment health factor and process complexity index in a multi-dimensional feature analysis to form a complete frame processing fusion vector. This fusion vector not only incorporates real-time processing dynamics but also incorporates equipment and process contextual factors, providing greater expressiveness and predictive value. This fusion vector is then input into a trained production prediction model, which, by learning from historically annotated data, outputs the current frame middleware processing quality status and risk trends.

[0029] The training set for the production prediction model is derived from historical vehicle frame production data. This data includes sensor information collected during middleware processing (such as temperature, current, and vibration), equipment operating status (such as operating time and fault records), process parameters, and their corresponding final quality judgment results (qualified or specific defect type). This data is extracted and standardized by the Manufacturing Execution System and Quality Management System. The model training utilizes a deep neural network-based architecture, taking the fused middleware processing features as input and outputting the quality status as a label. The system iteratively adjusts the model parameters to ensure that the predicted results are as consistent as possible with the actual labels. Cross-validation is used during training to evaluate the model's performance on unseen samples and ensure good generalization. Training is terminated when validation set performance stops improving over multiple training rounds, the model reaches the preset accuracy requirement, or the number of training rounds reaches a set limit. The model parameters with the best validation results are retained and used for middleware status prediction and maintenance judgment in actual vehicle frame production. Either a CNN or a BP neural network model can be used as the basis for the production prediction model.

[0030] The frame production prediction information includes the production reliability prediction value, the production abnormality risk trend and the predicted location of the frame production abnormality; intelligent prediction is performed based on the frame processing fusion vector to generate the frame production prediction information; this information mainly includes three aspects: the first is the production reliability prediction value, which indicates whether the middleware has acceptable processing stability under the current process environment; the second is the production abnormality risk trend, which is used to assess whether the middleware is in a potential deterioration trend, especially the critical situation close to the edge of failure; the third is the predicted location of the frame production abnormality, that is, combining the location label to determine the structural location where the abnormality is most likely to occur. These prediction results will directly serve as an important basis for the subsequent division of middleware status, optimization of equipment parameters and scheduling maintenance strategies.

[0031] The specific steps for reclassifying abnormal frame middleware based on frame production prediction information include: Perform interval mapping based on the production reliability prediction value in the frame production prediction information to obtain production reliability level area information; The production reliability prediction values ​​in the frame production forecast information are mapped to intervals. The system categorizes the production reliability prediction values ​​according to preset scoring intervals, such as high reliability, medium reliability, and low reliability. These intervals are determined based on historical annotation data and quality stability analysis results, reflecting the overall stability of the current middleware under production conditions. After interval mapping, the system can obtain information on the production reliability level range of the middleware, providing a basic reference for subsequent risk assessment.

[0032] The production anomaly risk trend in the frame production forecast information is judged. If the production anomaly risk trend is in the trend enhancement range but does not exceed the severe anomaly threshold, the critical frame weighted information is obtained; otherwise, the weighted information of the frame to be operated and maintained is obtained. Risk trends are curves generated based on sensor data and device status changes, used to determine whether the middleware's status is deteriorating. If the trend is clearly rising, meaning the middleware's status is developing towards instability but has not yet exceeded the critical anomaly warning threshold, the system will deem the middleware to be potentially risky but still recoverable, and will generate weighted critical frame information. This information indicates that the middleware is in a controllable but dangerous boundary. If the trend is clearly abnormal or approaching the critical upper limit, the system will generate weighted maintenance frame information, indicating that the middleware is likely to require rework or repair.

[0033] Critical judgment is made based on the predicted parts of frame production abnormality in the frame production prediction information: Step S1: Identify the importance level of the predicted location of the frame production anomaly; perform a comprehensive judgment based on the importance level, production reliability level area information, and critical frame weighted information or weighted information of the frame to be maintained, and obtain first frame middleware judgment information; Criticality is determined based on structural safety, load-bearing capacity, and impact on product performance. For example, main welds and support connections are classified as high-criticality, while auxiliary decorative structures are classified as low-criticality. Once identified, the system combines the criticality of the component, the aforementioned production reliability zone information, and the generated weighted information for critical or maintenance-required frames to perform a comprehensive logical analysis and derive the first frame middleware determination information. This information represents a preliminary classification recommendation for the middleware based on engineering rules and risk level criteria.

[0034] Step S2: Based on the frame production prediction information and the pre-trained enhanced fuzzy judgment model, a judgment is made to obtain the second frame middleware judgment information; in order to enhance the flexibility and accuracy of the judgment, the system also inputs the production prediction information of the frame middleware into a pre-trained enhanced fuzzy judgment model for auxiliary judgment; the model is based on the principle of fuzzy logic reasoning, combined with the fuzzy membership function and historical sample data, and through adaptive weight allocation, it performs fuzzy attribution scoring on the input features and outputs a critical state probability as the second frame middleware judgment information; the model is particularly sensitive to boundary states and is suitable for processing fuzzy critical situations that cannot be judged by hard rules. The second frame middleware judgment information it outputs provides the system with a supplementary judgment of human-like decision-making.

[0035] Determining, based on the first frame middleware determination information and the second frame middleware determination information, that the abnormal frame middleware is a critical frame middleware or a frame middleware to be maintained; The judgment information of the first frame middleware is cross-compared and logically integrated with the judgment information of the second frame middleware to form a final decision; if the two are consistent, the middleware is directly classified into the corresponding category; if there is a disagreement, the system will adopt a set weight mechanism or introduce confidence analysis to make a fusion judgment on the two types of results; thus, the abnormal frame middleware will be finally classified as critical frame middleware or frame middleware to be maintained, and will serve as the basis for subsequent processing.

[0036] The specific steps for analyzing the production forecast information of the frame marked as critical frame middleware to obtain the production equipment operation and maintenance information include: Based on the production anomaly risk trend and the predicted location of the frame production anomaly, reverse judgment is made to match similar known internal equipment faults. If the match is successful, the production equipment operation and maintenance information is generated based on the known internal equipment faults. If the match fails, the production equipment is readjusted based on the corresponding production equipment information to obtain the production equipment operation and maintenance information. When processing a frame middleware marked as critical, the system first conducts an in-depth analysis of the frame production forecast information corresponding to the middleware, specifically conducting reverse reasoning on the production anomaly risk trends and the predicted locations of frame production anomalies. The core purpose of this reverse reasoning is to infer the possible equipment causes that caused the middleware to enter a critical state. The system compares the current risk trend characteristics and predicted anomaly locations with equipment failure patterns in the historical fault knowledge base to identify known internal equipment failure cases with similar characteristics, such as insufficient heat input due to welding current drift, or weld offset due to feed mechanism jamming. After adjusting the corresponding production equipment using the production equipment operation and maintenance information, reprocess the critical frame middleware until the normal frame middleware is obtained; If the system matches an equipment failure case in the historical fault database that is highly consistent with the current risk characteristics, it means that the critical state is likely caused by an internal problem in the equipment; at this time, the system will extract the handling measures, maintenance suggestions, faulty component information, etc. associated with the matching fault, and generate complete production equipment operation and maintenance information; this operation and maintenance information includes the equipment number that needs to be repaired, the recommended maintenance location, and recommended maintenance operations (such as replacing the welding gun, recalibrating the electronic control unit, calibrating the fixture pressure, etc.), and may be accompanied by a priority level to guide maintenance personnel to accurately locate and quickly handle it.

[0037] If no clearly corresponding historical failure cases are matched in the fault knowledge base, it means that the current critical state may not belong to a known equipment failure type. At this time, the system will analyze the information of the production equipment used in the middleware processing, extract its process parameters, operating time, loading method, current load and motion curve and other data, compare it with the historical optimal equipment configuration for difference analysis, and then infer potential setting problems that may cause critical states. Based on this, the system will generate a set of optimization adjustment suggestions for the equipment as new production equipment operation and maintenance information output. Such adjustment suggestions may include fine-tuning of welding current, slightly slowing down the feeding speed, optimizing the fixture synchronization time, etc. The goal is to restore the equipment processing stability through soft control.

[0038] After the operation and maintenance information is generated, the system will apply it to the adjustment operation of the corresponding equipment; the equipment adjustment can be completed automatically by the system, or it can be manually fine-tuned on-site according to suggestions; after the adjustment is completed, the system will return the middleware originally marked as critical to the corresponding workstation for reprocessing; by applying new equipment parameters for reprocessing and performing quality monitoring simultaneously, the system can determine whether the middleware has been transferred from a critical state to a normal state; if the test results meet the standards, it is determined that the middleware has been restored to a normal frame middleware, and the process ends; if there are still signs of criticality, the equipment fine-tuning and reprocessing process can be entered again until the middleware meets the quality requirements, realizing a closed-loop processing mode of precise repair and equipment self-correction.

[0039] For the frame middleware to be maintained, perform abnormal operation and maintenance operations on the frame middleware to be maintained; for the normal frame middleware, continue to perform the next production operation on the normal frame middleware; When the vehicle frame middleware is determined by the system to be in the state of waiting for operation and maintenance, it means that a relatively clear abnormality has occurred in the middleware during the production process, which has exceeded the process tolerance range and cannot flow directly into the next process, but has not yet reached the severity of being scrapped or returned to the factory; therefore, the system will immediately perform abnormal operation and maintenance operations on it; first, the middleware will be automatically separated from the main production rhythm and transferred to a dedicated abnormality processing station or cache area; according to the defect type, defect location and operating status of the middleware, the system calls the operation and maintenance processing solution library to generate exclusive reprocessing strategies or repair instructions; for example, in the case of loose welding, the system will assign repositioning, local repair welding and secondary quality inspection; if it is a slight deformation, it will enter the correction process; the entire abnormal operation process will be recorded by the system, including the repair time, responsible equipment, processing method and result feedback, as an important data source for equipment traceability and quality statistics; after the operation and maintenance processing is completed and verified, the middleware can be reintegrated into the production process.

[0040] For parts that are judged to be normal frame middleware, the system will directly allow them to enter the next standardized production process without any intervention or correction operations; the system will mark the processing quality status of the middleware as qualified, and at the same time update its status code in the production traceability system. According to the current process route and planned scheduling, the middleware will be automatically transferred to the processing station corresponding to the next process, such as from welding to polishing, from assembly to painting, etc.; during the transfer process, the system will simultaneously carry the complete production data package of the middleware, including process parameters, inspection records and equipment numbers, to ensure that subsequent processes can continue to perform seamless processing operations based on the current status, and ultimately achieve a stable and efficient production closed loop.

[0041] In this embodiment, a feasible method for managing vehicle frame production and maintenance is presented. For example, during the production process of the main weld joint area of ​​a vehicle frame, the system first obtains internal component sensor information and equipment operation data generated by the middleware during the welding phase. The sensors collect parameters such as real-time temperature, current, voltage, and vibration, as well as equipment operating time, process settings, fixture pressure, and welding tact time. Combined with external image information captured by an industrial vision system, the system detects a slight discontinuity in the main weld area of ​​the middleware using a defect recognition model trained using a convolutional neural network. The system then identifies the defect as located in the main weld joint area based on a standardized vehicle frame structure model and establishes a corresponding label with historical quality data.

[0042] The system further aligns the sensor data with the equipment operating status in time, constructs a fusion sequence of middleware processing information, and applies a sliding window of indefinite length to extract change trends, volatility and disturbance indicators. At the same time, it generates equipment health factors and process complexity indexes based on the process accuracy requirements and equipment maintenance conditions corresponding to the process. The fusion features are input into the trained frame production prediction model, and the system outputs prediction results of a medium-to-low reliability score, an abnormal trend of an upward trend, and abnormal risk positioning in the weld connection area. Based on these results, the system preliminarily marks the middleware as abnormal and further executes the re-division process.

[0043] The system mapped the production reliability score to a medium-to-low range, and considering that the abnormal trend had not yet exceeded the critical threshold, it generated weighted information for a critical frame. It also identified the main weld joint area as a high-structurally important area. Combining its criticality and rising risk trend, the system used rule-based reasoning and fuzzy judgment models to dually verify and determine that the middleware was critical. The system then reverse-searched the historical fault knowledge base, matching similar welding current drop fault records and automatically generating recommended maintenance instructions, prompting adjustments to the current control module and recalibration of the welding gun trajectory. After the equipment adjustments were completed, the middleware was placed in the repair process for local repair welding and re-inspection. It ultimately passed the inspection and returned to normal status, continuing to the subsequent polishing process.

[0044] Example 2, an operation and maintenance management system for frame production, see Figure 1 As shown, including: The frame production information acquisition module includes a data acquisition unit; the data acquisition unit is used to obtain production process information and corresponding production equipment information of the frame production middleware; the production process information includes internal component sensor information and external component image information; based on the external component image information, detection is performed to obtain frame production and maintenance information; and the frame part labels of the frame production middleware are identified; The frame production and operation and maintenance management module includes an operation and maintenance identification unit and an operation and maintenance judgment unit; the operation and maintenance identification unit is used to add frame operation and maintenance labels to the frame production middleware based on the frame production and operation and maintenance information and the frame part label; the frame operation and maintenance label includes the frame middleware to be operated and maintained, abnormal frame middleware and normal frame middleware; the operation and maintenance judgment unit is used to make production predictions for the abnormal frame middleware based on the internal component sensor information and the production equipment information to obtain the frame production prediction information; based on the frame production prediction information, the abnormal frame middleware is subdivided to obtain the critical frame middleware or the frame middleware to be operated and maintained; the frame production prediction information marked as the critical frame middleware is analyzed to obtain the production equipment operation and maintenance information; for the frame middleware to be operated and maintained, the operation and maintenance abnormality operation is performed on the frame middleware to be operated and maintained; for the normal frame middleware, the next production operation is continued to be performed on the normal frame middleware.

[0045] It should be understood that those skilled in the art may make improvements or modifications based on the above description, and all such improvements and modifications shall fall within the scope of protection of the appended claims. Any portion of this specification not described in detail is prior art known to those skilled in the art.

Claims

1. An operation and maintenance management method for frame production, characterized in that: The following steps are involved: Obtain the production process information and corresponding production equipment information of the frame production middleware; The production process information includes internal component sensor information and external component image information; Detection based on external component image information to obtain frame production and maintenance information; identify frame part labels in frame production middleware; Add a frame operation and maintenance tag to the frame production middleware based on the frame production and operation and frame part tags; the frame operation and maintenance tag includes frame middleware to be operated and maintained, abnormal frame middleware, and normal frame middleware; For abnormal frame middleware, production prediction is performed based on internal component sensor information and production equipment information to obtain frame production prediction information; based on the frame production prediction information, abnormal frame middleware is subdivided into critical frame middleware or frame middleware to be maintained; the frame production prediction information marked as critical frame middleware is analyzed to obtain production equipment operation and maintenance information; For the frame middleware to be maintained, perform abnormal operation and maintenance operations on the frame middleware to be maintained; for the normal frame middleware, continue to perform the next production operation on the normal frame middleware.

2. The operation and maintenance management method for vehicle frame production according to claim 1, characterized in that: The specific steps for adding frame operation and maintenance tags to the frame production middleware based on the frame production and operation information and frame part tags include: Compare the frame production and maintenance information with the preset frame part operation and maintenance standards to obtain the frame production comparison results; Frame production deviation is identified based on the frame production comparison results and frame part labels to obtain the frame production deviation factor; Based on the frame production deviation factor, the frame production middleware is marked as frame middleware to be maintained, abnormal frame middleware and normal frame middleware.

3. The operation and maintenance management method for vehicle frame production according to claim 2, characterized in that: The specific steps for making production predictions based on internal component sensor information and production equipment information include: Align the internal component sensor information and production equipment information in time series to obtain a fusion sequence of middleware processing information; process the fusion sequence of middleware processing information using a sliding window of indefinite length to obtain several continuous sliding middleware processing information slices; Extract the change trend, coefficient of variation and perturbation ratio of the sliding middleware processing information piece; Obtain historical data on vehicle frame production; determine equipment health factors and process complexity index based on historical data; The change trend, coefficient of variation, perturbation ratio, equipment health factor, and process complexity index are integrated to obtain a frame processing fusion vector. Production prediction is performed based on the frame processing fusion vector to obtain frame production prediction information. The frame production forecast information includes production reliability forecast value, production abnormality risk trend and frame production abnormality forecast location.

4. The operation and maintenance management method for vehicle frame production according to claim 3, characterized in that: The specific steps for reclassifying abnormal frame middleware based on frame production prediction information include: Perform interval mapping based on the production reliability prediction value in the frame production prediction information to obtain production reliability level area information; The production anomaly risk trend in the frame production forecast information is judged. If the production anomaly risk trend is in the trend enhancement range but does not exceed the severe anomaly threshold, the critical frame weighted information is obtained; otherwise, the weighted information of the frame to be operated and maintained is obtained. Critical judgment is made based on the predicted parts of frame production abnormality in the frame production prediction information: Step S1: Identify the importance level of the predicted location of the frame production anomaly; perform a comprehensive judgment based on the importance level, production reliability level area information, and critical frame weighted information or weighted information of the frame to be maintained, and obtain first frame middleware judgment information; Step S2: performing a judgment based on the vehicle frame production prediction information and the pre-trained enhanced fuzzy judgment model to obtain second vehicle frame middleware judgment information; According to the first frame middleware determination information and the second frame middleware determination information, it is determined that the abnormal frame middleware belongs to the critical frame middleware or the frame middleware to be maintained.

5. The operation and maintenance management method for vehicle frame production according to claim 4, characterized in that: The specific steps for analyzing the production forecast information of the frame marked as critical frame middleware to obtain the production equipment operation and maintenance information include: Based on the production anomaly risk trend and the predicted location of the frame production anomaly, reverse judgment is made to match similar known internal equipment faults. If the match is successful, the production equipment operation and maintenance information is generated based on the known internal equipment faults. If the match fails, the production equipment is readjusted based on the corresponding production equipment information to obtain the production equipment operation and maintenance information. After adjusting the corresponding production equipment using the production equipment operation and maintenance information, the critical frame middleware is reprocessed until the normal frame middleware is obtained.

6. The operation and maintenance management method for vehicle frame production according to claim 5, characterized in that: The specific steps of performing detection based on external component image information to obtain frame production and operation information include: using a trained frame defect detection and recognition model to perform feature recognition on the external component image information to obtain frame production and operation information; and training the frame defect detection and recognition model based on a CNN model.

7. An operation and maintenance management system for vehicle frame production, characterized in that: The system applies the operation and maintenance management method for vehicle frame production according to any one of claims 1 to 6, including: The frame production information acquisition module includes a data acquisition unit; the data acquisition unit is used to obtain production process information and corresponding production equipment information of the frame production middleware; the production process information includes internal component sensor information and external component image information; based on the external component image information, detection is performed to obtain frame production and maintenance information; and the frame part labels of the frame production middleware are identified; The frame production and operation and maintenance management module includes an operation and maintenance identification unit and an operation and maintenance judgment unit; the operation and maintenance identification unit is used to add frame operation and maintenance labels to the frame production middleware based on the frame production and operation and maintenance information and the frame part label; the frame operation and maintenance label includes the frame middleware to be operated and maintained, abnormal frame middleware and normal frame middleware; the operation and maintenance judgment unit is used to make production predictions for the abnormal frame middleware based on the internal component sensor information and the production equipment information to obtain the frame production prediction information; based on the frame production prediction information, the abnormal frame middleware is subdivided to obtain the critical frame middleware or the frame middleware to be operated and maintained; the frame production prediction information marked as the critical frame middleware is analyzed to obtain the production equipment operation and maintenance information; for the frame middleware to be operated and maintained, the operation and maintenance abnormality operation is performed on the frame middleware to be operated and maintained; for the normal frame middleware, the next production operation is continued to be performed on the normal frame middleware.

Images