A method and system for evaluating casting surface treatment quality

Abstract

The present invention discloses a method and system for assessing the quality of surface treatment of castings, which relates to the field of intelligent manufacturing and quality control technology, including: collecting multimodal data of castings and performing preprocessing; analyzing the preprocessed data to construct a comprehensive quality assessment data set; constructing a quality assessment model to identify abnormalities in surface treatment and generate quality assessment results; classifying the quality of castings according to the quality assessment results, and generating an assessment report to provide quality status determination and related optimization suggestions; adjusting and optimizing the casting surface treatment process to improve the overall manufacturing quality and achieve continuous improvement in quality assessment. The present invention improves the detection accuracy and process optimization capabilities of the casting surface treatment quality. Using data-driven closed-loop control, it can accurately identify quality anomalies, provide optimization suggestions, realize the intelligence of the manufacturing process, improve stability and long-term quality improvement, effectively reduce the defect rate, and improve manufacturing consistency.

Description

Technical Field

[0001] The present invention relates to the technical field of intelligent manufacturing and quality control, and in particular to a casting surface treatment quality assessment method and system. Background Art

[0002] In recent years, casting surface quality assessment technology has gradually evolved toward intelligent and multimodal integration. Traditional methods primarily rely on manual visual inspection, contact measurement (such as roughness testing), and qualitative grading based on industry standards (such as ISO 8062 and GB / T6060.1). Quality assessment is based on visual comparison of surface defects (such as pores, slag inclusions, and cold shuts) or measurement of local physical parameters. Advances in nondestructive testing (NDT) have led to the introduction of technologies such as 3D topography scanning, infrared thermal imaging, and ultrasonic testing, enabling quantitative analysis of surface topography, temperature distribution, and internal microdefects. For example, machine vision-based surface defect recognition technology has achieved partial automation, while acoustic emission testing is used to monitor crack propagation during the casting process. Furthermore, some research has attempted to combine machine learning algorithms (such as convolutional neural networks) with multi-sensor data to improve defect classification accuracy. However, existing technologies still primarily rely on single-modal data and lack the ability to deeply integrate and dynamically model multi-dimensional data, making them difficult to meet the needs of comprehensive surface quality assessment for complex castings.

[0003] Currently, traditional inspection methods rely on manual visual inspection or single sensors (such as roughness meters and infrared thermal imagers), resulting in isolated data on key parameters such as surface topography, microstructure, and thermal stress. For example, while 3D visual inspection can capture surface geometric features, it cannot simultaneously analyze changes in the material's microstructure. Infrared thermal imaging can monitor temperature distribution, but lacks the ability to model the relationship between surface residual stress and process parameters (such as heat treatment temperature and cooling rate). This fragmented data leads to incomplete assessment results and makes it difficult to reveal the root causes of quality issues. Existing methods are often based on fixed thresholds or empirical rules (e.g., defect area ratio ≤5% is acceptable), and are unable to dynamically respond to process fluctuations or differences in material properties. For example, differential cooling rates caused by uneven wall thickness in a casting can lead to localized stress concentrations, but traditional models cannot predict these hidden defects because they do not incorporate process parameters. Furthermore, static models are difficult to optimize through incremental learning, resulting in a gradual degradation of assessment accuracy as production conditions change.

[0004] Furthermore, existing technologies focus on defect identification rather than process optimization, and are unable to generate dynamic adjustment recommendations based on evaluation results (such as adjusting shot peening intensity or optimizing heat treatment curves). For example, insufficient surface hardness of a casting may be related to the flow rate of the quenching medium, but existing methods lack the ability to jointly analyze multimodal data (hardness, temperature, flow rate), making it impossible to establish a closed-loop "detection-feedback-optimization" mechanism. Furthermore, the heterogeneity of multi-source data (such as topographic point clouds and thermal time series data) makes feature fusion difficult, making it difficult for existing algorithms to achieve unified representation and collaborative optimization of cross-modal data. Summary of the Invention

[0005] In view of the above-mentioned problems, the present invention is proposed.

[0006] Therefore, the technical problems solved by the present invention are: the existing casting surface quality assessment methods have the problems of multimodal data fragmentation resulting in one-sided assessment, static models cannot adapt to dynamic process changes, and lack a closed-loop feedback mechanism to achieve process optimization; and how to achieve accurate, adaptive and continuous improvement of quality assessment through intelligent data fusion and dynamic modeling.

[0007] In order to solve the above technical problems, the present invention provides the following technical solutions:

[0008] In a first aspect, an embodiment of the present invention provides a method for evaluating the surface treatment quality of a casting, comprising collecting multimodal data of the casting and performing preprocessing;

[0009] Analyze the pre-processed data to extract characteristic information that affects the quality of casting surface treatment, and fuse different types of data to construct a comprehensive quality assessment data set;

[0010] Build a quality assessment model to identify abnormalities in surface treatment and generate quality assessment results;

[0011] Based on the quality assessment results, the castings are classified and an assessment report is generated, providing quality status determination and related optimization suggestions;

[0012] Adjust and optimize casting surface treatment processes to improve overall manufacturing quality and achieve continuous improvement in quality assessment.

[0013] As a preferred embodiment of the method for evaluating the surface treatment quality of castings of the present invention, the multimodal data includes morphological data, thermal data and process characteristic data;

[0014] The topography data includes the casting's geometry, surface roughness, machining texture, and coating thickness information;

[0015] The thermal data includes temperature distribution information on the casting surface, characteristics of the heat treatment area and thermal stability of the coating;

[0016] The process characteristic data includes material information, processing parameters, stress distribution, environmental factors and quality inspection history data of the casting.

[0017] As a preferred embodiment of the casting surface treatment quality assessment method of the present invention, the data preprocessing includes format conversion, denoising, coordinate alignment and normalization of the multimodal data to meet the requirements of subsequent analysis, wherein:

[0018] The format conversion includes standardizing the data formats of the collected morphological data, thermal data, and process characteristic data so that different data sources have a unified data structure to adapt to subsequent calculations and analysis;

[0019] The denoising process includes filtering out measurement errors in topographic data, environmental interference in thermal data, and outliers in process characteristic data to improve data accuracy and stability.

[0020] The coordinate alignment includes matching the spatial coordinates of the topographic data and the thermal data so that different data types can be fused and calculated in a unified coordinate system to ensure data consistency and accuracy of comparative analysis.

[0021] The normalization process includes numerically standardizing data of different units and scales to eliminate calculation deviations caused by differences in data dimensions and ensure the stability of subsequent feature extraction and model calculations.

[0022] As a preferred embodiment of the method for evaluating the surface treatment quality of castings according to the present invention, the extraction of characteristic information affecting the surface treatment quality of castings includes extracting geometric deviation, surface roughness change rate, processing texture pattern and coating thickness distribution from the topography data to characterize the structural integrity and processing uniformity of the casting surface;

[0023] Extract temperature gradient changes, regional average temperature distribution, local hot spot anomalies, and cooling uniformity from thermal data to characterize heat treatment consistency and coating stability during casting surface treatment;

[0024] Processing parameter deviations, stress distribution trends, production environment influencing factors, and quality inspection history statistics are extracted from process feature data to correlate the relationship between process parameters and surface quality.

[0025] As a preferred embodiment of the casting surface treatment quality assessment method of the present invention, the construction of a comprehensive quality assessment data set includes correlating and matching feature information from different sources to establish a mapping relationship between morphological features, thermal features, and process features, wherein:

[0026] By correlating topographic data with thermal data, the effect of surface roughness changes on heat treatment uniformity is analyzed, and the coupling between coating thickness distribution and temperature distribution is evaluated.

[0027] By correlating topography data with process characteristic data, the influence of processing parameters on the geometric accuracy of castings can be evaluated, and the degree of optimization of shot peening intensity and coating deposition process on surface topography can be determined;

[0028] By correlating thermal data with process characteristic data, the relationship between heat treatment temperature control and stress distribution is analyzed, and the heat treatment stability under different production environment conditions is determined;

[0029] A multimodal data fusion method is used to uniformly express the extracted feature information and construct a comprehensive quality assessment dataset;

[0030] The fused quality assessment dataset is stored in a structured manner and converted into a standard data format suitable for quality assessment model calculation.

[0031] As a preferred embodiment of the casting surface treatment quality assessment method of the present invention, the construction of the quality assessment model includes calculating the surface feature quality score based on the surface feature data, process parameter data and thermal data. , Process parameters affect the score and thermal data impact ratings , and normalize each score.

[0032] As a preferred embodiment of the method for evaluating the quality of casting surface treatment according to the present invention, the surface feature quality score is The calculation formula is:

[0033] ;

[0034] in, is the total number of surface measurement points; is the surface feature index, indicating the Data of measurement points; For the The weight factor of each measurement point, For the The surface characteristic function of the measurement points, is the actual surface characteristic value, is the ideal target value of the surface characteristics, is the exponential decay factor, is the gamma function, which calculates the rationality of the distribution of feature data; is the time influence function, is the upper limit of the time integral, is the time variable; is a small time increment, which represents an infinitesimal time step in the integration process;

[0035] The process parameters affect the score The calculation formula is:

[0036] ;

[0037] in, is the process parameter index, indicating the process parameters; is the total number of process parameters, For the The adjustment factor of each process parameter, is the process parameter variable, is the process parameter influence function, is the process disturbance function, is the process parameter stability correction function, is the upper limit of the process parameter integral calculation, is the lower limit of the process parameter integral calculation, Process disturbance variables, is the process disturbance variable Small increments of

[0038] The thermal data affects the score The calculation formula is:

[0039] ;

[0040] in, is the thermal measurement point index, is the total number of temperature measurement points, is the temperature weighting factor, For the Temperature data at is the exponential inhibition factor, is the temperature uniformity parameter, is the temperature disturbance influence function, is the temperature stability correction function, is the upper limit of points, is the lower limit of the integral, is the temperature disturbance variable, is the temperature disturbance variable Small increments of

[0041] based on 、 and Calculate the comprehensive deviation of feature data , measure the matching degree between surface characteristics, process parameters and thermal data, comprehensive deviation The calculation formula is:

[0042] ;

[0043] in, is the total number of comprehensive evaluation data points, is the index value calculated for the comprehensive deviation, is the mean, is the standard deviation;

[0044] Based on calculated 、 、 and comprehensive deviation , calculate the final casting surface treatment quality assessment value , used to characterize the surface quality grade of castings, the calculation formula is:

[0045] ;

[0046] in, It is an index correction item to control the impact of deviation on the score; is the regulating factor.

[0047] As a preferred embodiment of the method for evaluating the quality of casting surface treatment according to the present invention, the quality classification of castings includes: When the quality level is judged to be excellent, the system automatically archives the data and forms a quality feature standard library for reference in subsequent process optimization;

[0048] When quality rating When the quality level is determined to be good, the influencing factors are analyzed in combination with historical data, and the process parameters are adjusted for fine-tuning and optimization;

[0049] When quality rating When the quality level is determined to be acceptable, the system performs multi-dimensional correlation analysis, identifies the main quality-influencing parameters, and recommends optimization paths;

[0050] When quality rating When the quality level is judged as unqualified, the system starts fault tracing analysis, determines the cause of the abnormality, and triggers corresponding quality optimization measures.

[0051] As a preferred embodiment of the casting surface treatment quality assessment method of the present invention, the adjustment and optimization of the casting surface treatment process includes dynamically optimizing process parameters based on the quality assessment results, and verifying and iteratively updating the optimization effect;

[0052] Through long-term quality data analysis, a trend prediction model is established to monitor the stability of production batches, identify the long-term trend of the impact of process parameters on quality, and adaptively adjust the manufacturing process;

[0053] Implement fault diagnosis and precise optimization for quality anomalies, identify surface defects, process deviations, and equipment anomalies, take appropriate corrective measures, and optimize process flows;

[0054] Establish a closed-loop quality control mechanism, combine historical optimization data, automatically generate process improvement plans, and provide decision support through data visualization to ensure the sustainability of quality optimization.

[0055] In a second aspect, an embodiment of the present invention provides a casting surface treatment quality assessment system, comprising:

[0056] Data acquisition and preprocessing module: collect multimodal data of castings and perform preprocessing;

[0057] Quality feature extraction and data fusion module: Analyzes pre-processed data, extracts feature information that affects the quality of casting surface treatment, and fuses different types of data to construct a comprehensive quality assessment data set;

[0058] Quality assessment module: builds a quality assessment model, identifies abnormalities in surface treatment, and generates quality assessment results;

[0059] Quality classification module: classifies castings according to quality assessment results, generates assessment reports, and provides quality status determination and related optimization suggestions;

[0060] Quality Optimization and Process Adjustment Module: Adjust and optimize the casting surface treatment process to improve the overall manufacturing quality and achieve continuous improvement in quality assessment.

[0061] Beneficial effects of the present invention: The present invention breaks through the limitations of a single detection method and realizes a comprehensive analysis of the surface characteristics of castings, heat treatment quality and process parameters. An intelligent quality assessment model is constructed to improve assessment accuracy, calculate surface defects, uneven coatings, heat treatment anomalies, etc., and realize automated quality classification and anomaly detection. According to the quality assessment results, the shot peening intensity, heat treatment temperature, and cooling rate are automatically adjusted, and the production consistency is improved and quality fluctuations are reduced through intelligent optimization algorithms. A report containing quality scores, defect analysis, and optimization suggestions is generated, and quality trends are visually displayed to provide decision support and reduce manual intervention. Through intelligent assessment + process optimization closed-loop control, the rework rate is reduced, the level of production automation is improved, production efficiency is improved, and quality management costs are reduced. BRIEF DESCRIPTION OF THE DRAWINGS

[0062] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following briefly introduces the drawings required for describing the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. Those skilled in the art can also derive other drawings based on these drawings without inventive work, among which:

[0063] Figure 1 This is an overall flow chart of a casting surface treatment quality assessment method provided by the first embodiment of the present invention. DETAILED DESCRIPTION

[0064] To make the above-mentioned objects, features, and advantages of the present invention more clearly understood, the following detailed description of the specific embodiments of the present invention is given in conjunction with the accompanying drawings. It is obvious that the described embodiments are only part of the embodiments of the present invention, not all of them. Based on the embodiments of the present invention, all other embodiments obtained by ordinary persons in this field without creative work should fall within the scope of protection of the present invention.

[0065] Example 1, reference Figure 1 , as one embodiment of the present invention, provides a method for evaluating the surface treatment quality of a casting, comprising:

[0066] S1: Multimodal data acquisition and preprocessing of castings.

[0067] Multimodal data includes morphological data, thermal data, and process characteristic data;

[0068] Morphological data includes the casting's geometry, surface roughness, machining texture, and coating thickness information. "Geometry" refers to the casting's external parameters, such as length, width, height, surface flatness, and angularity, which affect its assembly accuracy and contact stress distribution. "Surface roughness" directly determines the casting's friction characteristics and wear resistance. Traditional manual measurement errors are large, and this invention uses 3D point cloud scanning to provide high-precision measurement. "Machining texture" reflects the tool trajectory or surface shot peening effect during the machining process. Its uniformity and directionality affect the casting's fatigue life. "Coating thickness information" directly affects the casting's corrosion resistance. Too thin a coating may lead to failure, while too thick a coating may increase mass and affect fit accuracy.

[0069] Thermal data includes temperature distribution information on the casting surface, characteristics of the heat-treated area, and thermal stability of the coating. "Temperature distribution information" reflects the heat conduction and heat dissipation of the casting during the heat treatment process. Excessive temperature gradients can increase residual stress within the material. "Heat-treated area characteristics" primarily focus on the localized heating areas during the heat treatment process, ensuring uniform hardness distribution in the casting and avoiding localized embrittlement or softening. "Thermal stability of the coating" affects the casting's resistance to high temperatures and thermal fatigue. Poor thermal stability can lead to coating flaking or performance degradation.

[0070] Process characteristic data includes the casting's material information, processing parameters, stress distribution, environmental factors, and quality inspection history data. "Material information" requires clarifying the material composition (such as carbon steel, alloy steel) and its corresponding mechanical properties to ensure that the subsequent evaluation model can accurately match the quality standards of different materials. "Processing parameters" involve key variables such as shot peening intensity, heat treatment time, and cooling rate. These parameters directly affect the final mechanical properties of the casting. "Stress distribution" is related to the reliability of the casting in long-term operation. Excessive residual stress may cause crack initiation. "Environmental factors" include variables such as temperature, humidity, and equipment vibration that affect production stability. These variables need to be included in the data set to improve the adaptability of the evaluation model. "Quality inspection history data" provides quality change trends of castings under different process conditions, providing an important reference for the training of machine learning models.

[0071] It should be noted that the combination of morphological data and thermal data can evaluate the uniformity of surface treatment, such as the relationship between stress distribution and temperature field after shot peening, which helps to optimize the heat treatment process; by collecting historical quality inspection data, a data-driven evaluation model can be established to improve the applicability of the system under different production conditions.

[0072] Data preprocessing includes format conversion, denoising, coordinate alignment, and normalization of the multimodal data to meet the requirements of subsequent analysis, including:

[0073] Format conversion involves standardizing the collected topographical, thermal, and process characteristic data, ensuring a unified data structure across diverse data sources for subsequent calculations and analysis. Topographical data is typically in point cloud format, thermal data is a temperature matrix, and process data is often discrete variables. This standardization facilitates subsequent calculations by converting all data. This unified data format reduces errors caused by varying data sources and improves the accuracy of evaluation results.

[0074] De-noising includes filtering out measurement errors in topographic data, environmental interference in thermal data, and outliers in process characteristic data to improve data accuracy and stability.

[0075] Coordinate alignment involves matching the spatial coordinates of morphological data and thermal data, so that different data types can be fused and calculated in a unified coordinate system to ensure data consistency and accuracy of comparative analysis.

[0076] Normalization processing includes numerical standardization of data of different units and scales to eliminate calculation bias caused by differences in data dimensions and ensure the stability of subsequent feature extraction and model calculation.

[0077] S2: Analyze the preprocessed data, extract the characteristic information that affects the surface treatment quality of castings, and fuse different types of data to construct a comprehensive quality assessment data set.

[0078] Extract geometric deviation, surface roughness variation rate, machining texture pattern and coating thickness distribution from topography data to characterize the structural integrity and machining uniformity of the casting surface;

[0079] Extract temperature gradient changes, regional average temperature distribution, local hot spot anomalies, and cooling uniformity from thermal data to characterize heat treatment consistency and coating stability during casting surface treatment;

[0080] Processing parameter deviations, stress distribution trends, production environment influencing factors, and quality inspection history statistics are extracted from process feature data to correlate the relationship between process parameters and surface quality.

[0081] Correlate and match the feature information from different sources to establish a mapping relationship between morphological features, thermal features, and process features, where:

[0082] By correlating topographic data with thermal data, the effect of surface roughness changes on heat treatment uniformity is analyzed, and the coupling between coating thickness distribution and temperature distribution is evaluated.

[0083] By correlating topography data with process characteristic data, the influence of processing parameters on the geometric accuracy of castings can be evaluated, and the degree of optimization of shot peening intensity and coating deposition process on surface topography can be determined;

[0084] By correlating thermal data with process characteristic data, the relationship between heat treatment temperature control and stress distribution is analyzed, and the heat treatment stability under different production environment conditions is determined;

[0085] A multimodal data fusion method is used to uniformly express the extracted feature information and construct a comprehensive quality assessment dataset;

[0086] The fused quality assessment dataset is stored in a structured manner and converted into a standard data format suitable for quality assessment model calculation.

[0087] It should be noted that existing detection methods focus solely on a single data source, making it difficult to comprehensively assess casting surface quality. For example, relying solely on 3D scanning cannot detect heat treatment uniformity, while using only infrared imaging makes it difficult to assess surface roughness. This invention overcomes these limitations and enables comprehensive assessment. By calculating key indicators such as geometric deviation, temperature gradient, and machining parameter deviation, it quantifies the casting quality status and reduces human error. Utilizing real-time data stream analysis, this invention can provide early warning of anomalies in the production process, preventing the accumulation of quality defects.

[0088] Furthermore, existing methods fail to quantitatively analyze the matching of coating thickness and heat treatment temperature. The present invention establishes a mathematical model to optimize the temperature control strategy and improve the stability of the coating; it improves the adaptability of shot peening to heat treatment and reduces the poor heat treatment phenomenon caused by uneven distribution of shot peening stress.

[0089] S3: Build a quality assessment model to identify anomalies in surface treatment and generate quality assessment results.

[0090] Constructing a quality assessment model includes calculating surface feature quality scores based on surface feature data, process parameter data, and thermal data. , Process parameters affect the score and thermal data impact ratings , and normalize each score.

[0091] Surface feature quality score The calculation formula is:

[0092] ;

[0093] in, is the total number of surface measurement points; is the surface feature index, indicating the Data of measurement points; For the The weight factor of each measurement point, For the The surface characteristic function of the measurement points, is the actual surface characteristic value, is the ideal target value of the surface characteristics, is the exponential decay factor, is the gamma function, which calculates the rationality of the distribution of feature data; is the time influence function, is the upper limit of the time integral, is the time variable; is a small time increment, representing an infinitesimal time step in the integration process.

[0094] Further, is a polynomial fitting function used to characterize the The surface characteristic state of each measurement point, such as roughness, coating thickness, geometric error, etc. The polynomial form of this function can better adapt to different surface characteristic changes and improve the flexibility of the model. It is specifically expressed as:

[0095] ;

[0096] In this function, is the actual surface feature value, corresponding to the input value of a certain surface feature (such as height, texture depth, etc.); is a constant term, indicating that when hour, The initial value of , which usually represents the baseline value of the surface feature; is the linear coefficient, indicating that the surface eigenvalue varies with Linear trends of change, such as the uniformity of surface roughness along a certain direction; is the coefficient of the high-order term, which describes the nonlinear change trend of the surface characteristics. The higher the order of the term, the more accurately it can fit the complex surface morphology. is the order of the polynomial, which determines the fitting complexity of the function. It can better fit complex surface features, but may introduce the risk of overfitting.

[0097] Time impact function , the expanded form is:

[0098] ;

[0099] In this function, is a time variable, which represents the current time point and is a time parameter that affects the measurement data; is an integral variable representing the small time step during the time integration process and is used to calculate the cumulative impact. This function calculates the impact of time on the measured data, preventing systematic deviations in the quality score over time. This ensures that data from different time periods are appropriately weighted during data fusion, reducing noise interference caused by time factors.

[0100] Gamma function It is used to extend the concept of factorial to non-integer range and is defined as follows:

[0101] ;

[0102] In this function, To control the order of the gamma function and determine the convergence characteristics of the calculation results, it is used to adjust the data distribution shape in this formula; The integral variable represents the numerical range of a continuous variable, which changes continuously during the integration process. This function adjusts the distribution of surface quality score data to conform to statistical laws and improve the stability of the model. By smoothing the data distribution, calculation deviations caused by individual outliers are reduced, ensuring a more robust evaluation result.

[0103] Process parameter impact score The calculation formula is:

[0104] ;

[0105] in, is the process parameter index, indicating the process parameters; is the total number of process parameters, For the The adjustment factor of each process parameter, is the process parameter variable, is the process parameter influence function, is the process disturbance function, is the process parameter stability correction function, is the upper limit of the process parameter integral calculation, is the lower limit of the process parameter integral calculation, Process disturbance variables, is the process disturbance variable small increments.

[0106] Furthermore, the expanded form of the specific function is:

[0107] ;

[0108] ;

[0109] ;

[0110] In the above function, is the process parameter influence function, which characterizes the current process parameters Relative to process parameter reference values the degree of deviation; It is the maximum value allowed for the process parameter, usually determined by the process specification; This is the minimum value allowed for the process parameter, ensuring that the process parameter is not lower than the process requirements.

[0111] is the process disturbance function, which measures the changing trend of process parameters under disturbance conditions; is a disturbance variable, which indicates the fluctuation of process parameters due to equipment or environmental factors during the production process; is the reference value of process disturbance, that is, the stable disturbance level that should be maintained in theory; It is the attenuation factor that controls the sensitivity of the disturbance. A larger value will severely suppress the disturbance far from the reference value.

[0112] is the process parameter stability correction function, used to evaluate the current process parameters Stability to ensure that the process meets the optimal process state; is the ideal process parameter value, indicating the optimal setting value.

[0113] The thermal data affects the score The calculation formula is:

[0114] ;

[0115] in, is the thermal measurement point index, is the total number of temperature measurement points, is the temperature weighting factor, For the Temperature data at is the exponential inhibition factor, is the temperature uniformity parameter, is the temperature disturbance influence function, is the temperature stability correction function, is the upper limit of points, is the lower limit of the integral, is the temperature disturbance variable, is the temperature disturbance variable Small increments of

[0116] Furthermore, the expanded form of its specific function is:

[0117] ;

[0118] ;

[0119] In the above function, is the temperature disturbance influence function, which measures the influence of temperature disturbance on the surface quality of castings; It is an exponential decay term, which suppresses the influence of smaller temperature disturbances and makes the influence of extreme deviations greater. is an attenuation factor that controls the sensitivity to temperature disturbances. This function is used to quantify the impact of temperature disturbances, ensuring that temperature fluctuations during heat treatment are not ignored, thereby improving the accuracy of quality assessment.

[0120] It is the temperature stability correction function, which measures the current temperature of the casting. Is it close to the optimal heat treatment temperature? This function is used to compensate for the deviation of temperature data so that the final quality evaluation value can more accurately reflect the stability of heat treatment quality.

[0121] based on 、 and Calculate the comprehensive deviation of feature data , measure the matching degree between surface characteristics, process parameters and thermal data, comprehensive deviation The calculation formula is:

[0122] ;

[0123] in, is the total number of comprehensive evaluation data points, is the index value calculated for the comprehensive deviation, is the mean, is the standard deviation.

[0124] Based on calculated 、 、 and comprehensive deviation , calculate the final casting surface treatment quality assessment value , used to characterize the surface quality grade of castings, the calculation formula is:

[0125] ;

[0126] in, It is an index correction item to control the impact of deviation on the score; is the regulating factor.

[0127] It should be noted that the final casting quality grade is calculated by combining all scores and comprehensive deviations, which is used to intelligently determine product quality status. This improves the accuracy of quality grading, avoids human misjudgment, and enhances the scientific nature of quality management. It also optimizes the control strategy of the manufacturing process, automatically adjusts process parameters based on the quality score, and achieves closed-loop production optimization.

[0128] S4: Based on the quality assessment results, the castings are classified and an assessment report is generated to provide quality status judgment and related optimization suggestions.

[0129] When quality rating When the quality grade is judged to be excellent, an automatic confirmation mechanism is adopted to ensure that the surface treatment quality of the casting reaches the ideal target. The system automatically archives the data and forms a quality feature standard library for reference in subsequent process optimization.

[0130] Enter the batch tracking mode and confirm the consistency of the production of this batch through historical data analysis. If multiple batches maintain this level, the system will automatically optimize the production parameters and recommend expanding the production process to other batches.

[0131] If the test data conforms to the long-term stable trend, the process will be automatically set as the recommended process for the subsequent manufacturing process of similar products.

[0132] The evaluation report is automatically archived and quality maintenance suggestions are pushed to process engineers, such as maintaining current processing parameters and optimizing production rhythm.

[0133] When quality rating When the quality grade is judged to be good, the local optimization mode is entered to analyze the key influencing factors in the casting quality score and determine the main influencing parameters (such as shot peening intensity and cooling rate).

[0134] The system compares the process parameters of excellent quality batches, calculates the optimization space based on the parameter deviation correction model, and outputs recommended adjustment ranges, such as shot peening pressure adjustment of ±5%, heat treatment temperature adjustment of ±10℃, etc.

[0135] If the optimization range is within an acceptable range (i.e., it does not affect production cycle and cost), the production process parameters are automatically adjusted and a verification test is performed. After the verification is passed, the system automatically corrects the process settings.

[0136] Generate an optimization suggestion report for process engineers to confirm and synchronize to the quality management system.

[0137] When quality rating When the quality level is determined to be acceptable, the system enters active intervention mode, starts root cause analysis, and uses a multi-dimensional quality correlation analysis model to evaluate the deviation relationship between surface characteristics, process parameters, and thermal data to determine the main quality risk points.

[0138] If a parameter (such as shot peening time, heat treatment cooling rate) is used to score the quality If the contribution of a parameter exceeds 50%, the system will automatically mark it as a key parameter and adjust the production process for experimental verification.

[0139] The system recommends two optimization paths:

[0140] Local adjustment strategy: fine-tune within the existing process parameters (such as adjusting the shot peening pressure by 3%).

[0141] Global optimization strategy: It is recommended to change process parameters (such as using different coating materials).

[0142] Generate a quality improvement report and send it to the process engineer, who will decide the final optimization path.

[0143] When quality rating When the quality level is judged as unqualified, the system enters the fault tracing mode, starts the quality anomaly detection process, analyzes the anomaly points, and determines the type of fault (such as surface roughness exceeding the standard, coating peeling, uneven heat treatment).

[0144] A deep fault learning model is used to determine whether the anomaly is a single point failure (i.e., an anomaly occurs in a single casting) or a batch failure (anomalies occur in most products in a production batch) based on historical data comparison.

[0145] If it is a single point of failure:

[0146] Adopt intelligent recycling mechanism to send unqualified castings to rework process, such as secondary shot peening, secondary heat treatment, and recalculate the quality score .

[0147] like , it is allowed to be re-stocked, otherwise it will be scrapped.

[0148] If it is a batch failure:

[0149] The system automatically triggers the emergency process adjustment mechanism, suspends the current process, and analyzes whether there is any abnormality in the production equipment (such as insufficient spray force of the shot peening equipment).

[0150] If the equipment is abnormal, the system generates equipment maintenance instructions and notifies maintenance personnel to perform calibration.

[0151] If the process parameters are abnormal, the system will look back at historical data to find the parameters of the last successful batch and recommend restoring to the most recent stable process settings.

[0152] Generate detailed failure analysis reports for review by the quality management team and recommend short-term and long-term optimization measures.

[0153] Furthermore, the quality assessment system automatically generates an assessment report based on data calculations, including: casting surface feature quality analysis, process parameter matching analysis, quality score trend chart and key optimization suggestions.

[0154] The assessment report can be automatically sent to the quality management system and pushed to the relevant responsible persons.

[0155] Using parameter regression analysis models, we calculate the optimal process parameter range based on historical data and automatically recommend optimization solutions:

[0156] If the adjustment range of the optimization plan is within ±5%, the system will automatically adjust the production process.

[0157] If the adjustment exceeds 5%, manual confirmation is required before execution.

[0158] The quality assessment system adopts real-time monitoring mode. If the quality score of a production batch is If the value is continuously below 0.7, a quality warning will be automatically triggered.

[0159] Warning level:

[0160] Level 1 warning: single batch , it is recommended to optimize the process parameters;

[0161] Level 2 warning: three consecutive batches , recommend equipment maintenance or process adjustment;

[0162] Level 3 warning: five consecutive batches , forced to suspend production and start troubleshooting.

[0163] It should be noted that the different quality rating ranges Trigger different quality management strategies, avoiding simple pass / fail decisions and improving refined quality management. Parameter deviation correction models and deep fault learning models automatically analyze and optimize production processes, improving manufacturing stability. Quality anomaly warning mechanisms proactively identify quality trends, prevent large-scale quality issues, and enhance the intelligence of production lines.

[0164] S5: Adjust and optimize the casting surface treatment process to improve the overall manufacturing quality and achieve continuous improvement in quality assessment.

[0165] S51: Based on the quality assessment results, dynamically optimize process parameters, and verify and iteratively update the optimization effects.

[0166] During the production process, the quality of casting surface treatment is affected by a variety of process parameters (such as shot peening intensity, heat treatment temperature, and cooling rate). In order to improve the overall manufacturing quality, this paper adopts a data-driven dynamic optimization strategy:

[0167] Based on real-time quality assessment results, adjust key process parameters, such as:

[0168] When the surface roughness exceeds the allowable range, the shot peening pressure and medium particle size are automatically optimized;

[0169] When the coating thickness is uneven, adjust the coating speed and curing time;

[0170] When the heat treatment temperature distribution is abnormal, optimize the heating time or cooling method.

[0171] Through adaptive parameter adjustment, the surface quality of castings is stabilized within the optimal range, improving manufacturing consistency.

[0172] Furthermore, the lag of traditional manual experience adjustment is eliminated, and data-driven optimization methods are used to ensure timely and accurate process adjustments. The adaptability of process parameters is improved, avoiding quality fluctuations caused by fixed parameter settings and improving manufacturing consistency.

[0173] Whether the optimized process parameters are effective must be fed back through a quality verification mechanism:

[0174] Collect quality data after optimization, compare it with the quality assessment results before optimization, and calculate the degree of quality improvement;

[0175] Use error correction algorithms to modify optimization parameters, such as:

[0176] If the surface uniformity still does not meet the standard after optimization, further adjust the processing path or coating process;

[0177] If microcracks still exist in the casting after heat treatment, optimize the cooling rate or control the temperature gradient in the furnace.

[0178] Through the iterative update mechanism, the optimized parameters undergo multiple rounds of corrections and ultimately reach a stable optimal process state.

[0179] S52: Through long-term quality data analysis, establish trend prediction models, monitor production batch stability, identify long-term trends in the impact of process parameters on quality, and adaptively adjust manufacturing processes.

[0180] Quality optimization is not only about improving a single batch, but also requires predicting future quality trends based on long-term data analysis.

[0181] Use machine learning models for trend prediction, such as:

[0182] Monitor the quality stability of long-term production batches and evaluate the changing trends of process parameters;

[0183] Identify potential process drift (such as aging of shot peening media, equipment wear, etc.) and provide early warning of quality degradation risks;

[0184] Based on historical data, predict the quality deviation of different batches in the future and optimize process parameters in advance.

[0185] For example, if the system predicts that the shot peening pressure of a certain equipment will be low next month, equipment maintenance can be performed in advance; if historical data shows that cooling rate fluctuations will lead to unstable coating thickness, the cooling strategy can be adjusted.

[0186] S53: Implement fault diagnosis and precise optimization for quality anomalies, identify surface defects, process deviations and equipment anomalies, take corresponding corrective measures, and optimize the process flow.

[0187] When abnormal surface quality of castings is detected, such as excessive roughness, coating peeling, cracks, etc., it is necessary to accurately identify the cause of the abnormality and implement optimization:

[0188] Through data retrospective analysis, find the root cause of the anomaly, such as:

[0189] If the surface roughness is abnormal, it may be caused by changes in the particle size of the shot peening medium or pressure fluctuations; if the coating adhesion is insufficient, it may be caused by unstable heat treatment temperature or coating material problems.

[0190] Intelligent diagnostic algorithms are used to attribute abnormal problems to specific process parameters, such as:

[0191] Calculate the correlation between process parameters and quality anomalies to screen out key influencing factors; analyze which parameters need to be optimized through experimental data and physical models; implement precise optimization, adjust corresponding processing steps, and improve defect repair capabilities.

[0192] S54: Establish a closed-loop quality control mechanism, combine historical optimization data, automatically generate process improvement plans, and provide decision support through data visualization to ensure the sustainability of quality optimization.

[0193] In order to ensure the sustainability of quality optimization, the present invention establishes a closed-loop control system:

[0194] Automatically optimize process parameters: Combine historical optimization data to form an optimal process database for future optimization reference.

[0195] After a quality problem occurs, the system analyzes past similar cases and automatically recommends feasible optimization solutions. Combined with real-time quality assessment data, the optimization solution is dynamically adjusted to adapt it to current production conditions.

[0196] Heat maps, trend curves, and abnormal warning maps are used to intuitively display quality assessment results; production managers can quickly make optimization decisions based on data visualization reports.

[0197] Example 2 is the second embodiment of the present invention, which is different from the previous embodiment in that:

[0198] If the functions are implemented as software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, or the portion that contributes to the prior art or the portion of the current technical solution, can be embodied in the form of a software product. The current computer software product is stored in a storage medium and includes several instructions for causing a computer device (which can be a personal computer, server, or network device, etc.) to execute all or part of the steps of the method described in each embodiment of the present invention. The aforementioned storage medium includes various media that can store program code, such as USB flash drives, mobile hard drives, read-only memories (ROMs), random access memories (RAMs), magnetic disks, or optical disks.

[0199] The logic and / or steps represented in the flowcharts or otherwise described herein, for example, can be considered as a sequenced list of executable instructions for implementing the logical functions, and can be embodied in any computer-readable medium for use by, or in conjunction with, an instruction execution system, apparatus, or device (e.g., a computer-based system, a system including a processor, or other system that can fetch and execute instructions from an instruction execution system, apparatus, or device). For purposes of this specification, a "computer-readable medium" can be any device that can contain, store, communicate, propagate, or transport a program for use by, or in conjunction with, an instruction execution system, apparatus, or device.

[0200] More specific examples (a non-exhaustive list) of computer-readable media include the following: an electrical connection with one or more wires (electronic devices), a portable computer disk cartridge (magnetic devices), a random access memory (RAM), a read-only memory (ROM), an erasable and programmable read-only memory (EPROM or flash memory), an optical fiber device, and a portable compact disc read-only memory (CDROM). In addition, the computer-readable medium may even be paper or other suitable medium on which the program is printed, since the program may be obtained electronically, for example, by optically scanning the paper or other medium, and then editing, interpreting, or processing in another suitable manner as necessary, and then storing it in a computer memory.

[0201] Example 3 is an embodiment of the present invention, which provides a casting surface treatment quality assessment system, including a data acquisition and preprocessing module, a quality feature extraction and data fusion module, a quality assessment module, a quality classification module and a quality optimization and process adjustment module.

[0202] Data acquisition and preprocessing module: collect multimodal data of castings and perform preprocessing;

[0203] Quality feature extraction and data fusion module: Analyzes pre-processed data, extracts feature information that affects the quality of casting surface treatment, and fuses different types of data to construct a comprehensive quality assessment data set;

[0204] Quality assessment module: builds a quality assessment model, identifies abnormalities in surface treatment, and generates quality assessment results;

[0205] Quality classification module: classifies castings according to quality assessment results, generates assessment reports, and provides quality status determination and related optimization suggestions;

[0206] Quality Optimization and Process Adjustment Module: Adjust and optimize the casting surface treatment process to improve the overall manufacturing quality and achieve continuous improvement in quality assessment.

[0207] Example 4 is an embodiment of the present invention, which provides a method for evaluating the surface treatment quality of castings. In order to verify the beneficial effects of the present invention, scientific demonstration is carried out through economic benefit calculation and simulation / comparative experiments.

[0208] This experiment aimed to verify the effectiveness of a casting surface treatment quality assessment method in terms of process optimization, quality assessment accuracy, and manufacturing stability. The experiment quantified the optimization effect of the method by comprehensively analyzing surface roughness, coating thickness, heat treatment temperature, cooling rate, stress distribution uniformity, and final quality score.

[0209] The experimental subjects were 8 groups of casting samples, each group of samples was produced under different process conditions, focusing on analyzing the effects of different coating thicknesses, heat treatment temperatures, and cooling rates on the final quality. The following equipment was used for data collection:

[0210] 3D laser scanner: measure surface roughness, geometric deviation and coating thickness;

[0211] Infrared thermal imager: monitors heat treatment temperature distribution and evaluates temperature uniformity;

[0212] Stress testing system: analyzes stress distribution uniformity to ensure processing quality;

[0213] High-precision data acquisition system: real-time recording of parameters such as cooling rate and quality score.

[0214] 3D laser scanning was used to record the surface roughness and coating thickness of the castings;

[0215] Record the heat treatment temperature distribution of castings by infrared thermography;

[0216] Use stress sensors to record stress distribution data and calculate uniformity indicators;

[0217] Perform data denoising and normalization processing to eliminate environmental interference and improve data accuracy.

[0218] Extract key features such as surface roughness, coating thickness, temperature gradient, and stress uniformity;

[0219] Calculate the coupling relationship between coating thickness distribution and heat treatment temperature, and analyze the effect of heat treatment on coating stability;

[0220] A multimodal data fusion algorithm is used to unify data from different sources and make them useful for quality assessment.

[0221] A machine learning model is used to train the quality assessment system, which inputs surface characteristics, process parameters, and temperature data to calculate the quality score;

[0222] Identify the main factors affecting quality through regression analysis and optimize the evaluation algorithm.

[0223] Castings are classified according to their quality score:

[0224] Rating ≥90: high quality;

[0225] Score 80-89: Pass;

[0226] Score <80: Needs optimization.

[0227] Generate quality assessment reports and provide optimization suggestions.

[0228] Through quality analysis, it was found that some samples had room for optimization in parameters such as cooling rate and shot peening pressure;

[0229] Adjust heat treatment parameters to improve coating thickness uniformity and reduce defect rate;

[0230] Retest the optimized samples and compare the changes in quality scores.

[0231] The experimental reference data is shown in Table 1.

[0232] Table 1 Experimental data records

[0233]

[0234] The advantages of the quality assessment method of the present invention can be observed through experimental data:

[0235] Through multimodal data fusion, the evaluation model combines surface roughness, coating thickness, heat treatment temperature and cooling rate to achieve comprehensive quality assessment and improve accuracy.

[0236] For example, the T003 test sample scored 92, indicating that its surface quality was good, while the T004 score was only 67. Data analysis showed that its fast cooling rate led to uneven stress distribution, which in turn affected the quality.

[0237] Test data show that some castings have uneven coating thickness due to fluctuations in shot peening pressure and heat treatment temperature, which affects the surface roughness.

[0238] For example, the T002 test sample received a score of 75, which is lower than the standard value of 80, and analysis showed that its coating thickness had local deviations (±5μm).

[0239] After adopting the machine learning model, a trend prediction model is built based on historical data, which can detect potential quality problems in advance.

[0240] For example, in the T005-T008 trial group, the score range was stable at 85-94, indicating that the method can effectively identify and optimize quality parameters.

[0241] This experiment verified the effectiveness of the quality assessment method: through multimodal data fusion, the accuracy of quality assessment was improved, so that the scoring was no longer limited to a single indicator, but a comprehensive evaluation was conducted in combination with multiple process parameters; the method could accurately identify quality issues. For example, when the T004 score was too low, it could be traced back to an excessively fast cooling rate, and quality could be improved through process optimization; using machine learning modeling, historical data could be automatically analyzed, trend prediction models could be constructed, quality risks could be discovered in advance, and manufacturing consistency could be improved.

[0242] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit the present invention. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solutions of the present invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical solutions of the present invention, which should all be included in the scope of the claims of the present invention.

Claims

1. A method for evaluating the quality of casting surface treatment, characterized in that: include: Collect multimodal data of castings and perform preprocessing; Analyze the pre-processed data to extract characteristic information that affects the quality of casting surface treatment, and fuse different types of data to construct a comprehensive quality assessment data set; Build a quality assessment model to identify abnormalities in surface treatment and generate quality assessment results; Based on the quality assessment results, the castings are classified and an assessment report is generated, providing quality status determination and related optimization suggestions; Adjust and optimize casting surface treatment processes to improve overall manufacturing quality and achieve continuous improvement in quality assessment; The construction of the quality assessment model includes calculating the surface feature quality score based on the surface feature data, process parameter data and thermal data. , Process parameters affect the score and thermal data impact ratings , and normalize each score; The surface feature quality score The calculation formula is: ； in, is the total number of surface measurement points; is the surface feature index, indicating the Data of measurement points; For the The weight factor of each measurement point, For the The surface characteristic function of the measurement points, is the actual surface characteristic value, is the ideal target value of the surface characteristics, is the exponential decay factor, is the gamma function, which calculates the rationality of the distribution of feature data; is the time influence function, is the upper limit of the time integral, is the time variable; is a small time increment, which represents an infinitesimal time step in the integration process; Time impact function , the expanded form is: ； in, is a time variable, which represents the current time point and is a time parameter that affects the measurement data; is the integration variable, which represents the small time step in the time integration process and is used to calculate the cumulative effect.

2. The casting surface treatment quality assessment method according to claim 1, wherein: The multimodal data includes morphological data, thermal data and process characteristic data; The topography data includes the casting's geometry, surface roughness, machining texture, and coating thickness information; The thermal data includes temperature distribution information on the casting surface, characteristics of the heat treatment area and thermal stability of the coating; The process characteristic data includes material information, processing parameters, stress distribution, environmental factors and quality inspection history data of the casting.

3. The casting surface treatment quality assessment method according to claim 2, wherein: The preprocessing includes format conversion, denoising, coordinate alignment, and normalization of the multimodal data to meet the requirements of subsequent analysis, wherein: The format conversion includes standardizing the data formats of the collected morphological data, thermal data, and process characteristic data so that different data sources have a unified data structure to adapt to subsequent calculations and analysis; The denoising process includes filtering out measurement errors in topographic data, environmental interference in thermal data, and outliers in process characteristic data to improve data accuracy and stability. The coordinate alignment includes matching the spatial coordinates of the topographic data and the thermal data so that different data types can be fused and calculated in a unified coordinate system to ensure data consistency and accuracy of comparative analysis. The normalization process includes numerically standardizing data of different units and scales to eliminate calculation deviations caused by differences in data dimensions and ensure the stability of subsequent feature extraction and model calculations.

4. The casting surface treatment quality assessment method according to claim 3, wherein: The extraction of characteristic information affecting the surface treatment quality of the casting includes extracting geometric deviation, surface roughness change rate, processing texture pattern and coating thickness distribution from the topography data to characterize the structural integrity and processing uniformity of the casting surface; Extract temperature gradient changes, regional average temperature distribution, local hot spot anomalies, and cooling uniformity from thermal data to characterize heat treatment consistency and coating stability during casting surface treatment; Processing parameter deviations, stress distribution trends, production environment influencing factors, and quality inspection history statistics are extracted from process feature data to correlate the relationship between process parameters and surface quality.

5. The method for evaluating casting surface treatment quality according to claim 4, wherein: The construction of a comprehensive quality assessment data set includes correlating and matching feature information from different sources and establishing a mapping relationship between morphological features, thermal features, and process features, wherein: By correlating topographic data with thermal data, the effect of surface roughness changes on heat treatment uniformity is analyzed, and the coupling between coating thickness distribution and temperature distribution is evaluated. By correlating topography data with process characteristic data, the influence of processing parameters on the geometric accuracy of castings can be evaluated, and the degree of optimization of shot peening intensity and coating deposition process on surface topography can be determined; By correlating thermal data with process characteristic data, the relationship between heat treatment temperature control and stress distribution is analyzed, and the heat treatment stability under different production environment conditions is determined; A multimodal data fusion method is used to uniformly express the extracted feature information and construct a comprehensive quality assessment dataset; The fused quality assessment dataset is stored in a structured manner and converted into a standard data format suitable for quality assessment model calculation.

6. The method for evaluating casting surface treatment quality according to claim 5, wherein: The process parameters affect the score The calculation formula is: ； in, is the process parameter index, indicating the process parameters; is the total number of process parameters, For the The adjustment factor of each process parameter, is the process parameter variable, is the process parameter influence function, is the process disturbance function, is the process parameter stability correction function, is the upper limit of the process parameter integral calculation, is the lower limit of the process parameter integral calculation, Process disturbance variables, is the process disturbance variable Small increments of The thermal data affects the score The calculation formula is: ； in, is the thermal measurement point index, is the total number of temperature measurement points, is the temperature weighting factor, For the Temperature data at is the exponential inhibition factor, is the temperature uniformity parameter, is the temperature disturbance influence function, is the temperature stability correction function, is the upper limit of points, is the lower limit of the integral, is the temperature disturbance variable, is the temperature disturbance variable Small increments of based on 、 and Calculate the comprehensive deviation of feature data , measure the matching degree between surface characteristics, process parameters and thermal data, comprehensive deviation The calculation formula is: ； in, is the total number of comprehensive evaluation data points, is the index value calculated for the comprehensive deviation, is the mean, is the standard deviation; Based on calculated 、 、 and comprehensive deviation , calculate the final casting surface treatment quality assessment value , used to characterize the surface quality grade of castings, the calculation formula is: ； in, It is an index correction item to control the impact of deviation on the score; is the regulating factor.

7. The method for evaluating casting surface treatment quality according to claim 6, wherein: The quality classification of castings includes: When the quality level is judged to be excellent, the system automatically archives the data and forms a quality feature standard library for reference in subsequent process optimization; When quality rating When the quality level is determined to be good, the influencing factors are analyzed in combination with historical data, and the process parameters are adjusted for fine-tuning and optimization; When quality rating When the quality level is determined to be acceptable, the system performs multi-dimensional correlation analysis, identifies the main quality-influencing parameters, and recommends optimization paths; When quality rating When the quality level is judged as unqualified, the system starts fault tracing analysis, determines the cause of the abnormality, and triggers corresponding quality optimization measures.

8. The method for evaluating casting surface treatment quality according to claim 7, wherein: The adjustment and optimization of the casting surface treatment process includes dynamically optimizing process parameters based on quality assessment results, and verifying and iteratively updating the optimization effect; Through long-term quality data analysis, a trend prediction model is established to monitor the stability of production batches, identify the long-term trend of the impact of process parameters on quality, and adaptively adjust the manufacturing process; Implement fault diagnosis and precise optimization for quality anomalies, identify surface defects, process deviations, and equipment anomalies, take appropriate corrective measures, and optimize process flows; Establish a closed-loop quality control mechanism, combine historical optimization data, automatically generate process improvement plans, and provide decision support through data visualization to ensure the sustainability of quality optimization.

9. A casting surface treatment quality assessment system for implementing the casting surface treatment quality assessment method according to any one of claims 1 to 8, characterized in that: include: Data acquisition and preprocessing module: collect multimodal data of castings and perform preprocessing; Quality feature extraction and data fusion module: Analyzes pre-processed data, extracts feature information that affects the quality of casting surface treatment, and fuses different types of data to construct a comprehensive quality assessment data set; Quality assessment module: builds a quality assessment model, identifies abnormalities in surface treatment, and generates quality assessment results; Quality classification module: classifies castings according to quality assessment results, generates assessment reports, and provides quality status determination and related optimization suggestions; Quality Optimization and Process Adjustment Module: Adjust and optimize the casting surface treatment process to improve the overall manufacturing quality and achieve continuous improvement in quality assessment.

Images