Intelligent production method of corrosion-resistant pressing rivet nut

Abstract

The invention discloses an intelligent production method of a corrosion-resistant pressing rivet nut in the technical field of information, which comprises the following steps: carrying out surface treatment on the pressing rivet nut subjected to heat treatment, selecting a zinc plating or nickel plating process according to the corrosion resistance requirement of a product, and obtaining a uniform and compact plating layer by controlling the parameters of the temperature and the concentration of a plating solution and the electroplating time, the plating layer quality is detected by adopting a plating layer thickness meter and an adhesive force tester, the pre-plating treatment and post-plating treatment processes are optimized, and the protection performance and the appearance quality of the plating layer are improved; integrating a surface defect detection result with other performance indexes to serve as a comprehensive evaluation basis of product quality, establishing a correlation model of pressure riveting nut performance detection data and material attributes, a heat treatment process and a surface treatment process, mining an influence rule of each element on the product performance by adopting a correlation rule mining and time sequence prediction method, and determining the quality of the product. And predicting product performance distribution under different production parameter combinations.

Description

Technical Field

[0001] The present invention relates to the field of information technology, and in particular to a production technology of self-clinching nuts, and in particular to an intelligent production method of corrosion-resistant self-clinching nuts. Background Art

[0002] In the process of producing corrosion-resistant self-clinching nuts, the temperature and time control of the heat treatment process is a key technical problem. At present, most manufacturers use traditional heat treatment methods with low temperature control accuracy and unstable holding time, resulting in uneven hardness and strength of the products and large differences between batches. This not only affects the corrosion resistance of the self-clinching nuts, but also brings difficulties to the subsequent assembly. If the heat treatment temperature is too high or the holding time is too long, the grain size of the material will be coarsened, reducing the plasticity and toughness of the material; if the temperature is too low or the holding time is insufficient, the ideal strength and hardness cannot be achieved.

[0003] Therefore, how to achieve precise control of the heat treatment process while ensuring product performance is a technical problem that needs to be solved urgently. This requires innovation and optimization in heat treatment equipment, temperature measurement, automatic control, etc. to improve production efficiency and product quality stability. Summary of the invention

[0004] The present invention provides an intelligent production method for corrosion-resistant self-clinching nuts, comprising the following steps:

[0005] Step S1, obtaining the property parameters of grain size, plasticity and toughness of different materials from the material performance database of the self-clinching nut, establishing a mathematical model of material properties and heat treatment process parameters through the support vector machine algorithm, determining the best material selection and heat treatment process matching scheme, and measuring the material utilization rate by calculating the number of qualified self-clinching nuts that can be produced per unit of raw material, and improving this index while ensuring the mechanical properties of the product;

[0006] Step S2, according to the selected material and heat treatment process, multiple high-precision temperature sensors are installed in the heat treatment furnace to collect temperature data at different positions in the furnace in real time and transmit them to the process control system. If the deviation between the actual temperature and the target temperature exceeds a preset threshold, the heating power is automatically adjusted to ensure that the temperature field in the furnace is evenly distributed. At the same time, a thermal imager is used to monitor the temperature distribution in the furnace, obtain three-dimensional temperature field data, and perform temperature field simulation through finite element analysis method to optimize the furnace temperature uniformity;

[0007] Step S3, the process control system presets the holding time and temperature curve according to the optimal heat treatment process parameters of the product, uses a programmable logic controller to control the switching of heating and holding states, ensures that the heat treatment process of each batch of products is consistent, and combines the product size tolerance requirements, establishes a correlation model between the holding time and the size accuracy through experimental data collection and multivariate regression analysis, and uses a neural network algorithm to dynamically optimize the holding time. The specific implementation includes: the input layer receives the current process parameters and product size data, the hidden layer performs feature extraction and nonlinear mapping, the output layer predicts the optimal holding time, and the network weights are continuously adjusted through back propagation to achieve real-time optimization of the holding time and control the size deviation of the product within the tolerance range;

[0008] Step S4, performing surface treatment on the heat-treated rivet nuts, selecting a zinc plating or nickel plating process according to the corrosion resistance requirements of the product, obtaining a uniform and dense coating by controlling the parameters of the plating solution temperature, concentration and electroplating time, using a coating thickness meter and an adhesion tester to detect the coating quality, optimizing the pre-plating and post-plating treatment processes, and improving the protective performance and appearance quality of the coating;

[0009] Step S5, conduct a comprehensive inspection of the finished product after the coating treatment, including using precision measuring tools to measure key dimensions, using a universal material testing machine to test tensile strength and yield strength, and using a salt spray test chamber to evaluate corrosion resistance. At the same time, a machine vision system is used to detect defects on the surface of the rivet nut, identify scratches, pits, and coating shedding defects, and correlate the inspection data with process parameters. The heat treatment and surface treatment processes are continuously optimized through the random forest algorithm. The random forest algorithm constructs multiple decision trees, each tree uses a different feature subset and sample subset, and combines the prediction results of all trees to obtain the optimal process parameter combination, thereby improving the batch stability of product performance;

[0010] Step S6: Integrate the surface defect detection results with other performance indicators as a basis for comprehensive evaluation of product quality, establish a correlation model between the performance detection data of the rivet nut and material properties, heat treatment process and surface treatment process, use association rule mining and time series prediction methods to mine the influence of various factors on product performance, and predict the product performance distribution under different production parameter combinations;

[0011] Step S7: If the predicted performance exceeds the allowable range of technical specifications, the system will automatically issue an early warning and give process optimization suggestions. These optimization suggestions are applied to specific production links through a feedback mechanism: adjusting material selection criteria, updating heat treatment process parameters, and optimizing surface treatment formulas. The process control system automatically adjusts the corresponding parameters based on these suggestions to achieve closed-loop optimization of the production process and continuously improve product quality and production efficiency.

[0012] The technical solution provided by the embodiment of the present invention may have the following beneficial effects:

[0013] The present invention discloses an intelligent production method for corrosion-resistant self-clinching nuts. The system determines the optimal material and heat treatment scheme by establishing a mathematical model of material properties and heat treatment process parameters, thereby improving material utilization. During the heat treatment process, multiple sensors are used to monitor the temperature distribution in the furnace in real time, and finite element analysis is combined to optimize the uniformity of the temperature field. At the same time, a neural network algorithm is used to dynamically optimize the insulation time and control the dimensional accuracy of the product. After a comprehensive inspection of the finished product, a random forest algorithm is used to analyze the correlation between the inspection data and the process parameters, and the heat treatment and surface treatment processes are continuously optimized. The present invention realizes the intelligent closed-loop optimization of the production process of self-clinching nuts, significantly improving the product quality stability and production efficiency. DETAILED DESCRIPTION

[0014] In order to make the purpose, technical solutions and advantages of the present invention more clear, the present invention is described in detail below in conjunction with specific embodiments.

[0015] The present embodiment provides an intelligent production method for corrosion-resistant self-clinching nuts, which specifically includes the following steps:

[0016] Step S1, obtaining the property parameters of grain size, plasticity and toughness of different materials from the material performance database of self-clinching nuts. A mathematical model of material properties and heat treatment process parameters is established through the support vector machine algorithm to determine the best material selection and heat treatment process matching scheme. The material utilization rate is measured by calculating the number of qualified self-clinching nuts that can be produced per unit of raw material, and this indicator is improved while ensuring the mechanical properties of the product.

[0017] Step S1 includes: obtaining key property parameters of grain size, plasticity, and toughness of different materials in the material performance database of the self-clinching nut, preprocessing the material property parameters, eliminating outliers, and performing standardization processing; using a support vector machine algorithm, taking the material property parameters as input and the heat treatment process parameters as output, training and establishing a mathematical model of the material properties and the heat treatment process parameters; according to the mathematical model, inputting the property parameters of the target material, and predicting a matching combination of heat treatment process parameters; according to the predicted combination of heat treatment process parameters, performing a heat treatment test on the target material to test the mechanical properties of the heat-treated material; if the mechanical properties of the heat-treated material meet the requirements, calculating the number of qualified self-clinching nuts that can be produced per unit of raw material, and obtaining a material utilization index; comprehensively evaluating the material utilization index and the mechanical property test results, determining whether the target material and the heat treatment process parameter combination are the best matching scheme; if so, applying the best matching scheme to the production of self-clinching nut; if not, adjusting the property parameters of the target material, returning to the prediction matching scheme step, and re-predicting the matching scheme.

[0018] Specifically, the material performance database of rivet nuts is the key foundation, which contains the property parameters of various materials such as grain size, plasticity, and toughness. For example, the grain size of a certain carbon steel is 10μm, the plastic elongation is 20%, and the impact toughness is 80J / cm2. After obtaining these data, they need to be preprocessed, outliers removed and standardized. For example, the plastic elongation is standardized to the range of 0-1 to facilitate subsequent modeling. The support vector machine algorithm is used to establish a mathematical model of material properties and heat treatment process parameters. The input may include standardized property values ​​such as grain size, plasticity, and toughness, and the output is heat treatment process parameters such as quenching temperature and holding time. This model is able to learn complex nonlinear relationships, which is conducive to accurately predicting the optimal process parameters. According to the established model, the property parameters of the target material are input to predict the matching optimal heat treatment process parameter combination. For example, for a new type of alloy steel, the standardized grain size of 0.8, plasticity of 0.6, and toughness of 0.7 are input, and the model may output parameters such as the optimal quenching temperature of 850℃ and the holding time of 30 minutes. According to the predicted parameters, heat treatment tests are carried out to test whether the mechanical properties of the material meet the requirements. Possible test items include tensile strength, yield strength, elongation, etc. If the tensile strength of the material after heat treatment reaches 800mPa and the elongation is 15%, which meets the design requirements, the next step of evaluation is entered. Calculating material utilization is an important indicator for evaluating the pros and cons of the scheme. For example, 1 ton of raw materials can produce 100,000 qualified rivet nuts, and the material utilization rate is 95%. This indicator reflects the economy and environmental protection of the process, and has an important impact on production costs and resource utilization efficiency. Comprehensively evaluate the material utilization rate and mechanical property test results to determine the best matching scheme. If a scheme can not only ensure the strength and toughness of the rivet nut, but also achieve a material utilization rate of more than 95%, it can be regarded as the best solution and applied to production. If the expected goal is not achieved, it is necessary to return to adjust the material property parameters and re-predict the matching scheme. This iterative optimization process reflects the close integration of material science and process technology. Through data-driven methods, the optimal material-process matching scheme can be quickly found to improve product quality, reduce production costs, and achieve continuous improvement of the manufacturing process of rivet nuts. At the same time, this method also provides a reference for optimizing the manufacturing process of other metal parts.

[0019] Step S2: According to the selected material and heat treatment process, multiple high-precision temperature sensors are installed in the heat treatment furnace to collect temperature data at different positions in the furnace in real time and transmit them to the process control system. If the actual temperature deviates from the target temperature by more than the preset threshold, the heating power is automatically adjusted to ensure that the temperature field in the furnace is evenly distributed. At the same time, a thermal imager is used to monitor the temperature distribution in the furnace, obtain three-dimensional temperature field data, and perform temperature field simulation through finite element analysis to optimize the furnace temperature uniformity.

[0020] Step S2 includes: acquiring real-time temperature data of each position in the heat treatment furnace, the real-time temperature data being acquired by a plurality of high-precision temperature sensors arranged at different positions in the heat treatment furnace; transmitting the real-time temperature data to a process control system; after receiving the real-time temperature data, the process control system compares the real-time temperature data with a preset target temperature, and calculates a deviation value between the real-time temperature data and the target temperature; if the deviation value exceeds a preset threshold value, the process control system automatically adjusts the heating power of the heat treatment furnace according to the size and positive and negative of the deviation value; acquiring three-dimensional temperature field data in the heat treatment furnace, the three-dimensional temperature field data being acquired by a thermal imager that performs real-time monitoring of the heat treatment furnace; fusing the three-dimensional temperature field data with the real-time temperature data, and generating a high-precision three-dimensional temperature field model in the heat treatment furnace through a data fusion algorithm; simulating the high-precision three-dimensional temperature field model using a finite element analysis method to obtain a simulation result of furnace temperature uniformity; and optimizing the control parameters of the heat treatment furnace according to the simulation result.

[0021] Specifically, the temperature field monitoring of the heat treatment furnace is crucial to ensure the quality of the rivet nuts. To achieve accurate monitoring, multiple high-precision thermocouples, such as K-type thermocouples, can be arranged in the furnace, with a measurement range of -200℃ to 1300℃ and an accuracy of ±0.75%. These sensors can be installed at different heights and positions in the furnace, such as the furnace top, furnace bottom, furnace wall, etc., to form a three-dimensional monitoring network. The temperature data collected by the sensor is transmitted to the central control system in real time via the RS-485 bus or the wireless ZigBee network. After receiving the actual temperature data, the control system compares it with the preset target temperature. For example, if the target temperature is 850℃, the actual measured temperature at a certain point is 830℃, with a deviation of 20℃. The system presets the allowable deviation threshold of ±10℃. At this time, the deviation exceeds the limit and needs to be adjusted. The control system will automatically increase the heating power according to the size and direction of the deviation. Specifically, the PID control algorithm can be used to calculate the output power according to the deviation value, integral value and differential value to achieve fast and stable temperature adjustment. The introduction of thermal imagers further improves the accuracy and comprehensiveness of temperature field monitoring. High-end thermal imagers can measure temperatures from -40°C to 2000°C, with a thermal sensitivity of up to 0.03°C. By scanning the entire furnace cavity, the thermal imager can generate a high-resolution temperature distribution map with a resolution of up to 640×480 pixels. These data are fused with sensor data and algorithms such as Kalman filtering can generate a more accurate three-dimensional temperature field model. The generated three-dimensional temperature field model is simulated and analyzed using finite element analysis software such as ANSYS. By setting appropriate meshing, boundary conditions and material properties, the heat transfer process in the furnace can be simulated. The simulation results can intuitively display the temperature distribution cloud map, revealing temperature gradients and hot spots. For example, the simulation may find that there is a 20°C temperature drop in the corner of the furnace, based on which the layout of the heating elements can be optimized or the thickness of the insulation layer can be adjusted. Based on the simulation results, the heat treatment process parameters can be further optimized. If it is found that the heating rate is too fast and causes uneven temperature, the heating rate can be reduced from 10°C / min to 5°C / min. If the insulation time is not enough to eliminate the temperature gradient, the insulation time can be extended from 30min to 45min. Through these optimizations, the temperature field in the furnace can be made more uniform, the temperature deviation can be controlled within ±5°C, and the consistency and reliability of the heat treatment quality can be significantly improved. This closed-loop feedback control and optimization mechanism can not only ensure the quality of a single batch of heat treatment, but also continuously improve the heat treatment process parameter library through long-term data accumulation. For example, the optimal heat treatment curve for rivet nuts of different materials and sizes can be established, and intelligent recommendation and automatic adjustment of process parameters can be achieved, further improving production efficiency and product quality.

[0022] Step S3, the process control system presets the holding time and temperature curve according to the optimal heat treatment process parameters of the product, and uses a programmable logic controller to control the switching of heating and holding states to ensure that the heat treatment process of each batch of products is consistent. Combined with the product dimensional tolerance requirements, a correlation model between holding time and dimensional accuracy is established through experimental data collection and multivariate regression analysis. The neural network algorithm is used to dynamically optimize the holding time. The specific implementation includes: the input layer receives the current process parameters and product dimensional data, the hidden layer performs feature extraction and nonlinear mapping, the output layer predicts the optimal holding time, and the network weights are continuously adjusted through back propagation to achieve real-time optimization of the holding time and control the product dimensional deviation within the tolerance range.

[0023] Step S3 includes: obtaining preset temperature curve and holding time data as the control basis of PLC; establishing a mathematical correlation model between holding time and product size to obtain regression model coefficients; using the coefficients as the initial weights of the neural network, inputting the process parameters and product size data collected in real time, and outputting the optimized holding time through the neural network; transmitting the holding time back to the PLC, dynamically adjusting the holding time of the current batch of products, and controlling the product size deviation within the preset tolerance range through iterative optimization.

[0024] Specifically, the acquisition and application of heat treatment process parameters are the key to ensuring product quality. Taking the heat treatment of automobile engine crankshafts as an example, according to the material properties and usage requirements, the preset temperature curve may be: heating to 850℃ at a rate of 10℃ / min, keeping warm for 2 hours, and then dropping to 200℃ at a rate of 5℃ / min. These parameters will serve as the basis for PLC control. The process of PLC controlling heating equipment can be compared to a car cruise control system. Assuming the target temperature is 850℃ and the current temperature is 830℃, the PLC will determine the temperature deviation of 20℃ and increase the heating power accordingly. As the temperature gradually approaches the target value, the PLC will dynamically reduce the heating power to avoid temperature overshoot and finally stabilize at the target temperature. The control of the insulation stage is similar to the insulation function of an electric rice cooker. When the temperature reaches 850℃, the PLC switches to the insulation state and starts the timer to start the 2-hour timing. During this period, the PLC will maintain a constant temperature by fine-tuning the heating power to ensure that the crankshaft obtains a uniform heat treatment effect. The acquisition of product dimension data can use a laser measurement system. For example, for the crankshaft main journal diameter, multiple points can be measured before and after heat treatment, and the average value can be calculated and recorded. These data are stored in the database together with process parameters (such as temperature curve and holding time) to provide a basis for subsequent analysis. The multivariate regression analysis method is used to establish the correlation model between holding time and product size. Assume that there is a linear relationship between the holding time (x) and the change in the crankshaft main journal diameter (y): y = 0.02x + 0.05. This means that for every increase of 1 hour of holding time, the main journal diameter increases by an average of 0.02mm, and the basic expansion is 0.05mm. The use of the coefficients of the regression model as the initial weights of the neural network is similar to the application of transfer learning in image recognition. For example, in the optimization of crankshaft heat treatment, the input layer may include parameters such as material composition, initial size, and target size; the hidden layer performs feature extraction and nonlinear transformation, which may contain implicit information such as material properties and thermal expansion coefficient; and the output layer gives the optimized holding time. The holding time output by the neural network is transmitted back to the PLC for dynamic adjustment, which is similar to an adaptive cruise control system. For example, if the neural network predicts that the current batch of crankshafts will need 2.2 hours of insulation to reach the target size, the PLC will extend the insulation time accordingly. In this way, the system can adapt to the slight differences in raw materials from batch to batch to ensure the consistency of the final product size. The iterative optimization process can be likened to a golfer constantly adjusting his batting stance. After each heat treatment, the system analyzes the product dimensional deviation and fine-tunes the process parameters. As data accumulates and the model is optimized, the system's prediction accuracy will continue to improve, and ultimately the product size can be controlled within a smaller tolerance range, such as

[0025] ±0.01mm. The advantage of this intelligent heat treatment control system lies in its adaptive ability. It can cope with the influence of factors such as raw material batch differences and ambient temperature changes, continuously optimize process parameters, and improve product quality stability. At the same time, through data accumulation and analysis, the system can also provide valuable insights for process improvement, such as discovering the potential correlation between certain material components and heat treatment effects, providing a basis for material optimization.

[0026] Step S4, surface treatment is performed on the heat-treated rivet nut. According to the corrosion resistance requirements of the product, a galvanizing or nickel plating process is selected. A uniform and dense coating is obtained by controlling the parameters of the plating solution temperature, concentration and electroplating time. A coating thickness meter and an adhesion tester are used to detect the coating quality, optimize the pre-plating and post-plating treatment processes, and improve the protective performance and appearance quality of the coating.

[0027] Step S4 includes: obtaining the corrosion resistance requirements of the rivet nut, and determining whether to use a galvanizing process or a nickel plating process for surface treatment according to the corrosion resistance requirements; establishing a correlation model between the process parameters of the plating solution temperature, plating solution concentration and electroplating time and the coating quality for the galvanizing process or the nickel plating process; training the correlation model using a support vector machine or a decision tree algorithm to obtain an optimized correlation model; during the electroplating process, obtaining the plating solution temperature and concentration parameters, inputting the parameters into the optimized correlation model to obtain predicted coating uniformity and density; dynamically adjusting the process parameters according to the predicted coating uniformity and density until the coating quality meets the preset requirements; using a coating thickness meter to measure the coating thickness distribution, comparing the measured data with the target thickness, and judging whether the coating thickness is uniform. uniform; if the coating thickness is uneven, input the measurement data into the association model, and retrain and optimize the association model; use an adhesion tester to detect the bonding force between the coating and the substrate to obtain adhesion data; determine whether the adhesion is qualified according to a preset threshold; if the adhesion is unqualified, input the adhesion data into the pre-plating process optimization model, use a neural network algorithm to establish the association between the pre-plating and post-plating process parameters and defects, and optimize the pre-plating process and post-plating process; comprehensively evaluate the uniformity of the coating thickness, the adhesion and the appearance quality test results, and output a comprehensive quality score for the surface treatment process; based on the comprehensive quality score, continuously optimize the process parameters to improve the corrosion resistance and appearance quality of the rivet nut products to meet the use requirements of downstream customers.

[0028] Specifically, the corrosion resistance of the rivet nuts is one of the key indicators of product quality. According to different application environment requirements, galvanizing or nickel plating processes can be selected for surface treatment. For example, for rivet nuts used outdoors, hot-dip galvanizing is usually used, which can provide 15-20 years of corrosion protection; while for precision electronic equipment used indoors, electroplating nickel can be selected, which can provide good corrosion resistance and ensure dimensional accuracy. Establishing a correlation model between process parameters and coating quality is the key to optimizing the surface treatment process. Taking the galvanizing process as an example, a large amount of historical production data can be collected, including bath temperature (usually in the range of 45-55°C), bath concentration (zinc ion concentration is about 10-15g / L), current density (2-3A / dm 2) and other parameters, as well as the corresponding quality indicators such as coating thickness and uniformity. Using the support vector machine algorithm, nonlinear classification boundaries in multidimensional parameter space can be established to predict the coating quality grade under different parameter combinations. In the electroplating process, real-time monitoring and adjustment of process parameters are crucial. For example, the concentration of the plating solution is monitored by an online conductivity meter. When the concentration is detected to be lower than 10g / L, the refilling system is automatically started; the temperature of the plating solution is monitored in real time by an infrared thermometer, and the heating power is accurately controlled in combination with the PID control algorithm to control the temperature fluctuation within the range of ±1°C. These real-time adjustment measures can effectively improve the stability of the coating quality. The uniformity of the coating thickness distribution directly affects the corrosion resistance of the product. Using an X-ray fluorescence coating thickness meter, the coating thickness can be quickly measured at multiple measuring points. For example, for an M6-specification rivet nut, 3 measuring points can be selected at the top surface, side surface and internal thread of the nut to measure the coating thickness. If the thickness standard deviation of the 9 points measured is less than 0.5μm, and the average thickness is within the range of 8-12μm, it can be judged as qualified. The bonding strength between the coating and the substrate is another important quality indicator. Adhesion testing by cross-cutting or pulling methods can quantitatively evaluate the bonding strength. For example, use an automatic cross-cutting instrument to draw a 10×10 grid on the surface of the coating, then quickly peel it off after taping it, and observe the number of peeled squares. If the peeling area is less than 5%, it is judged to be qualified. This test method is simple and fast, and is suitable for quality control in mass production. In the case of coating quality problems, the production process can be optimized by analyzing the pre-plating and post-plating treatment processes. For example, if the coating adhesion is found to be insufficient, it may be caused by insufficient pickling before plating. By establishing a neural network model, inputting parameters such as pickling time and acid concentration, and outputting the adhesion prediction value, the optimal pickling process parameter combination can be found. Similarly, for post-plating passivation treatment, the corrosion resistance of the coating can be improved by adjusting the passivation solution concentration and immersion time. When comprehensively evaluating the coating quality, a weighted scoring method can be used. For example, thickness uniformity, adhesion and appearance quality are weighted 40%, 40% and 20% respectively, and the comprehensive score is calculated based on the measured values ​​of each indicator. By continuously collecting scoring data and feeding it back to the process optimization model, the surface treatment quality of the rivet nuts can be continuously improved, ultimately meeting the strict requirements of downstream customers, such as the 1000-hour neutral salt spray test standard required by the automotive industry.

[0029] Step S5: Conduct a comprehensive inspection of the finished product after coating treatment, including using precision measuring tools to measure key dimensions, using a universal material testing machine to test tensile strength and yield strength, and using a salt spray test chamber to evaluate corrosion resistance. At the same time, a machine vision system is used to detect defects on the surface of the rivet nuts to identify scratches, pits, and coating shedding defects. The inspection data is correlated with the process parameters, and the heat treatment and surface treatment processes are continuously optimized through the random forest algorithm. The random forest algorithm constructs multiple decision trees, each tree uses a different feature subset and sample subset, and combines the prediction results of all trees to derive the optimal process parameter combination and improve the batch stability of product performance.

[0030] Step S5 includes: obtaining technical requirements of the self-clinching nuts, determining target values ​​and tolerance ranges of key dimensions, tensile strength, yield strength and corrosion resistance, and establishing product quality evaluation standards; obtaining inspection data of the self-clinching nuts and heat treatment and surface treatment process parameters of corresponding batches, and constructing a data set; using a random forest algorithm to establish a prediction model between process parameters and product performance based on the data set; determining key process parameters through feature importance evaluation; optimizing heat treatment and surface treatment process parameters according to the prediction model, adjusting process setting values, and improving batch stability of product performance.

[0031] Specifically, the quality evaluation standard of self-clinching nuts is the basis of product performance control. Taking key dimensions as an example, the inner diameter of the thread is an important parameter that determines the accuracy of the fit between nuts and bolts. Assume that the target value of the inner diameter of the thread of a certain type of self-clinching nut is 6.0mm, and the tolerance range is ±0.05mm. Use a precision thread micrometer to measure the inner diameter of the thread. If the measured value is 6.03mm, it is judged to be qualified; if the measured value is 6.07mm, it exceeds the upper limit of the tolerance and needs to be marked as a dimensional defect. The tensile strength and yield strength reflect the bearing capacity of the self-clinching nut. Taking a certain type of self-clinching nut as an example, its tensile strength target value is 800MPa and its yield strength target value is 640MPa. Through the tensile test of the universal material testing machine, if the tensile strength is measured to be 785MPa and the yield strength is 630MPa, although it is slightly lower than the target value but within the allowable range, it can be judged to be qualified. If the tensile strength is only 750MPa, it needs to be marked as a mechanical property defect and analyze whether there is a problem with the heat treatment process. The corrosion resistance test simulates the performance of the self-clinching nut in the actual use environment. Taking galvanized rivet nuts as an example, a salt spray test can be carried out in a 5% sodium chloride solution for 96 hours. If there is no obvious red rust on the coating after 96 hours and only a small amount of white rust appears, the corrosion resistance is determined to be qualified. If a large area of ​​red rust appears, it needs to be marked as a corrosion resistance defect and the electroplating process parameters should be checked for abnormalities. The machine vision system can quickly detect surface defects of rivet nuts. Taking the scratch detection on the surface of the nut as an example, a line scan camera can be used to obtain a high-resolution image, and the edge features can be extracted through the image processing algorithm to identify the scratches. If a scratch longer than 0.5 mm is detected, it is determined to be a surface defect. This automated detection method can improve detection efficiency and accuracy and reduce human misjudgment. The random forest algorithm is used to establish a predictive model of process parameters and product performance. Taking the heat treatment process as an example, the quenching temperature, holding time, cooling rate, etc. can be used as input features, and the tensile strength and yield strength can be used as output targets. By training the random forest model, the degree of influence of each process parameter on the strength can be obtained. Assuming that the model shows that the feature importance of the quenching temperature is the highest, the quenching temperature can be adjusted first to improve the strength. Process parameter optimization is a dynamic process. Taking the surface treatment process as an example, if it is found that a batch of products has poor corrosion resistance, the electroplating time or current density can be adjusted according to the prediction of the random forest model. The optimized parameters are applied in subsequent production, and corrosion resistance test data is continuously collected. By continuously updating the model and optimizing the parameters, the stability of product performance can be gradually improved and the fluctuation between batches can be reduced. This data-driven quality control and process optimization method can effectively improve the overall quality level of self-clinching nuts. By establishing a complete detection system and prediction model, problems in the production process can be discovered and solved in a timely manner to ensure that the products continue to meet customer needs.

[0032] Step S6: Integrate the surface defect detection results with other performance indicators as a basis for comprehensive evaluation of product quality. Establish an association model between the performance test data of the rivet nut and material properties, heat treatment process and surface treatment process, use association rule mining and time series prediction methods to explore the influence of various factors on product performance, and predict the product performance distribution under different production parameter combinations.

[0033] Step S6 includes: obtaining surface defect detection result data and other performance index data of the rivet nut, performing data fusion processing on the surface defect detection result data and other performance index data, and obtaining comprehensive evaluation data of the quality of the rivet nut product. Obtaining the performance detection data, material attribute data, heat treatment process parameters and surface treatment process parameters of the rivet nut, and establishing an association model between the performance of the rivet nut and the production factors based on the performance detection data, material attribute data, heat treatment process parameters and surface treatment process parameters of the rivet nut. Based on the association model, an association rule mining algorithm is used to mine the association rules between the production factors of material attributes, heat treatment process parameters and surface treatment process parameters and the performance of the rivet nut, and the influence rules of each production factor on the performance of the rivet nut are obtained. A set of production parameter combination data to be predicted is obtained, and the production parameter combination data is input into the association model, and combined with the influence rules, a time series prediction algorithm is used to predict the performance distribution of the rivet nut product under the production parameter combination. According to the predicted performance distribution results, it is judged whether the production parameter combination meets the preset performance requirements. If it does, the production parameter combination is determined to be a feasible production parameter and applied to the production and manufacturing process of the rivet nut. In the production process of the self-clinching nut, the parameter data of each production link is collected in real time, and the real-time collected parameter data is input into the association model to dynamically predict the performance distribution of the self-clinching nut product, so as to realize real-time monitoring of the product performance. If the predicted performance distribution result does not meet the preset performance requirements, the production parameters are adjusted in time, and the above steps are re-executed until the predicted performance distribution result meets the preset performance requirements, thereby ensuring the quality stability of the self-clinching nut product.

[0034] Specifically, the quality assessment of rivet nuts involves data fusion in multiple dimensions. For example, surface defect detection may include indicators such as scratch depth and area percentage, while performance indicators include tensile strength, yield strength, corrosion resistance, etc. By standardizing and weighting these heterogeneous data, a comprehensive score can be obtained. Assuming that the surface defect area of ​​a batch of rivet nuts is 2%, the tensile strength is 500MPa, and the corrosion resistance is 96 hours, the comprehensive score obtained after data fusion processing is 92 points, which is higher than the qualified line of 90 points. Establishing an association model between the performance of rivet nuts and production factors is the key to optimizing the production process. This model needs to consider multiple factors such as material composition, heat treatment temperature, time, cooling rate, surface treatment process, etc. For example, by analyzing historical data, it may be found that the tensile strength of the product is most stable when the carbon content is in the range of 0.45%-0.50%, the heat treatment temperature is between 850℃-870℃, and the insulation time is 30-35 minutes. This association model can be constructed using machine learning algorithms such as random forests or neural networks. Association rule mining algorithms can discover hidden rules from massive production data. For example, it may be found that when the heat treatment temperature increases by 10°C, the hardness of the product will increase by 2HRC, but at the same time the toughness will decrease slightly. For another example, for every 5μm increase in the thickness of the galvanized layer, the corrosion resistance will increase by about 20 hours. These rules provide a quantitative basis for process optimization. Based on the established association model and the mined influence rules, the performance of new production parameter combinations can be predicted. Suppose there is a new set of parameters: carbon content 0.48%, heat treatment temperature 860°C, holding time 32 minutes, and galvanized layer thickness 15μm. Inputting these parameters into the model and combining them with the time series prediction algorithm may result in the following prediction results: mean tensile strength 520MPa, standard deviation 10MPa; mean hardness 52HRC, standard deviation 1HRC; mean corrosion resistance 110 hours, standard deviation 5 hours. If these prediction results meet the preset performance requirements, the parameter combination can be applied to actual production. In the production process, real-time monitoring and dynamic adjustment are essential. By installing sensors in key processes, parameters such as heat treatment furnace temperature curve, cooling rate, and surface treatment tank concentration can be collected in real time. By inputting these real-time data into the correlation model, product performance can be dynamically predicted. For example, if the temperature fluctuation of the heat treatment furnace is detected to exceed ±5°C, the model may predict that the hardness distribution of the product will be abnormal. At this time, the system will promptly alarm, and the operator can adjust the heating power or extend the insulation time to compensate for this deviation. This data-driven intelligent manufacturing method has significant advantages. First, it can make full use of historical production data and transform the empirical wisdom implicit in the data into a quantifiable basis for decision-making. Secondly, by establishing complex associations between multiple factors, the impact of process parameters on product performance can be more fully understood, avoiding the negative effects that may be caused by the optimization of a single factor.Furthermore, real-time monitoring and prediction capabilities make the production process more agile and allow timely intervention before problems expand. Finally, this method can be continuously self-improving. As data accumulates, the accuracy of the model will continue to improve, thereby achieving continuous optimization of the production process. However, there are also some challenges in implementing this method. Data quality is the primary issue, and it is necessary to ensure that the collected data is accurate and reliable. Secondly, the interpretability of the model is also important, especially when applying machine learning algorithms, it is necessary to be able to explain to engineers the reasons for the predicted results. In addition, how to balance the complexity and practicality of the model also needs to be carefully considered. Overly complex models may increase the computational burden and affect real-time performance. Finally, attention should also be paid to data security and privacy protection, especially when it comes to core process parameters.

[0035] Step S7: If the predicted performance exceeds the allowable range of technical specifications, the system automatically issues an early warning and gives process optimization suggestions. These optimization suggestions are applied to specific production links through a feedback mechanism: adjusting material selection criteria, updating heat treatment process parameters, and optimizing surface treatment formulas. The process control system automatically adjusts the corresponding parameters based on these suggestions to achieve closed-loop optimization of the production process and continuously improve product quality and production efficiency.

[0036] Step S7 includes: acquiring equipment operation data, the equipment operation data including temperature, pressure and vibration frequency; preprocessing the equipment operation data to obtain a standardized equipment operation data set; constructing a prediction model based on the standardized equipment operation data set using a support vector machine regression algorithm to obtain an equipment performance prediction value; if the equipment performance prediction value exceeds a preset threshold range, triggering an early warning signal; based on the early warning signal, obtaining preliminary process optimization rules through a decision tree algorithm; combining an expert knowledge base and Bayesian network reasoning to generate process optimization suggestions based on the preliminary process optimization rules, the process optimization suggestions including material selection, heat treatment temperature and time, and surface treatment solution concentration.

[0037] Specifically, the collection of equipment operation data is the basis of predictive maintenance. Taking the riveting nut production line as an example, key parameters such as heat treatment furnace temperature, press pressure, and vibration sensor frequency can be collected. Data preprocessing cleans and standardizes the original data, such as removing outliers, filling missing values, and normalization to ensure data quality. The support vector machine regression algorithm is suitable for small sample and nonlinear problems and can be used to predict equipment performance indicators. For example, based on the historical data training model, the maximum pressure value of the press in the next 24 hours is predicted. Comparing the performance prediction value with the threshold can timely discover potential problems. If the predicted pressure value exceeds the set threshold of 800kN, the system will generate an early warning signal and record the time and specific value of the excess. The decision tree algorithm can quickly locate the key parameter combination that causes performance abnormalities by analyzing historical data. For example, it is found that when the heat treatment temperature exceeds 550℃ and the holding time is less than 30 minutes, it is easy to cause the pressure to exceed the standard. The generation of process optimization suggestions combines data analysis results and expert experience. Bayesian networks can simulate the complex relationship between parameters and infer the optimal parameter combination. For example, according to the analysis results, it is recommended to adjust the heat treatment temperature to 530℃, extend the holding time to 40 minutes, and increase the concentration of the surface treatment solution by 5%. These optimization suggestions are sent to the control system through the communication interface to realize the automatic adjustment of production parameters. After receiving the optimization suggestions, the control system converts them into specific execution instructions. For example, adjust the temperature curve of the heat treatment furnace, modify the pressure setting value of the press, update the solution ratio of the surface treatment tank, etc. This closed-loop optimization process can continuously improve product quality and production efficiency. By continuously collecting new operating data, the system can self-learn and optimize to adapt to the dynamic changes of the production environment. The core advantage of this method is that it combines data-driven predictive analysis with expert knowledge to realize the intelligent and adaptive optimization of the production process. It can not only detect potential problems in a timely manner, but also provide targeted solutions, greatly improving the stability of production efficiency and product quality. At the same time, through continuous data accumulation and model optimization, the prediction accuracy and optimization effect of the system will continue to improve, providing strong support for the long-term development of the enterprise.

[0038] Based on the above embodiments of the present invention, relevant personnel can make various changes and modifications without departing from the technical concept of the present invention through the above description. The technical scope of the present invention is not limited to the content in the specification, and its technical scope must be determined according to the scope of the claims.

Claims

1. An intelligent production method for corrosion-resistant self-clinching nuts, characterized in that: The method comprises the following steps: Step S1, obtaining the property parameters of grain size, plasticity and toughness of different materials from the material performance database of the self-clinching nut, establishing a mathematical model of material properties and heat treatment process parameters through the support vector machine algorithm, determining the best material selection and heat treatment process matching scheme, and measuring the material utilization rate by calculating the number of qualified self-clinching nuts that can be produced per unit of raw material, and improving this index while ensuring the mechanical properties of the product; Step S2, according to the selected material and heat treatment process, multiple high-precision temperature sensors are installed in the heat treatment furnace to collect temperature data at different positions in the furnace in real time and transmit them to the process control system. If the deviation between the actual temperature and the target temperature exceeds a preset threshold, the heating power is automatically adjusted to ensure that the temperature field in the furnace is evenly distributed. At the same time, a thermal imager is used to monitor the temperature distribution in the furnace, obtain three-dimensional temperature field data, and perform temperature field simulation through finite element analysis method to optimize the furnace temperature uniformity; Step S3, the process control system presets the holding time and temperature curve according to the optimal heat treatment process parameters of the product, uses a programmable logic controller to control the switching of heating and holding states, ensures that the heat treatment process of each batch of products is consistent, and combines the product size tolerance requirements, establishes a correlation model between the holding time and the size accuracy through experimental data collection and multivariate regression analysis, and uses a neural network algorithm to dynamically optimize the holding time. The specific implementation includes: the input layer receives the current process parameters and product size data, the hidden layer performs feature extraction and nonlinear mapping, the output layer predicts the optimal holding time, and the network weights are continuously adjusted through back propagation to achieve real-time optimization of the holding time and control the size deviation of the product within the tolerance range; Step S4, performing surface treatment on the heat-treated rivet nuts, selecting a zinc plating or nickel plating process according to the corrosion resistance requirements of the product, obtaining a uniform and dense coating by controlling the parameters of the plating solution temperature, concentration and electroplating time, using a coating thickness meter and an adhesion tester to detect the coating quality, optimizing the pre-plating and post-plating treatment processes, and improving the protective performance and appearance quality of the coating; Step S5, conduct a comprehensive inspection of the finished product after the coating treatment, including using precision measuring tools to measure key dimensions, using a universal material testing machine to test tensile strength and yield strength, and using a salt spray test chamber to evaluate corrosion resistance. At the same time, a machine vision system is used to detect defects on the surface of the rivet nut, identify scratches, pits, and coating shedding defects, and correlate the inspection data with process parameters. The heat treatment and surface treatment processes are continuously optimized through the random forest algorithm. The random forest algorithm constructs multiple decision trees, each tree uses a different feature subset and sample subset, and combines the prediction results of all trees to obtain the optimal process parameter combination, thereby improving the batch stability of product performance; Step S6: Integrate the surface defect detection results with other performance indicators as a basis for comprehensive evaluation of product quality, establish a correlation model between the performance detection data of the rivet nut and material properties, heat treatment process and surface treatment process, use association rule mining and time series prediction methods to mine the influence of various factors on product performance, and predict the product performance distribution under different production parameter combinations; Step S7: If the predicted performance exceeds the allowable range of technical specifications, the system will automatically issue an early warning and give process optimization suggestions. These optimization suggestions are applied to specific production links through a feedback mechanism: adjusting material selection criteria, updating heat treatment process parameters, and optimizing surface treatment formulas. The process control system automatically adjusts the corresponding parameters based on these suggestions to achieve closed-loop optimization of the production process and continuously improve product quality and production efficiency.

2. According to claim 1, the intelligent production method of corrosion-resistant self-clinching nuts is characterized in that: The step S1 comprises: Obtain key property parameters of grain size, plasticity and toughness of different materials in the material performance database of self-clinching nuts, pre-process the material property parameters, remove abnormal values, and perform standardization; A support vector machine algorithm is used to train and establish a mathematical model of the material properties and the heat treatment process parameters, with the material property parameters as input and the heat treatment process parameters as output; According to the mathematical model, the property parameters of the target material are input to predict a matching combination of heat treatment process parameters; According to the predicted heat treatment process parameter combination, a heat treatment test is performed on the target material to test the mechanical properties of the material after the heat treatment; If the mechanical properties of the heat-treated material meet the requirements, the number of qualified self-clinching nuts that can be produced per unit of raw material is calculated to obtain the material utilization index; Comprehensively evaluate the material utilization index and the mechanical property test results to determine whether the target material and the heat treatment process parameter combination is the best matching solution; If yes, then the optimal matching solution is applied to the production of rivet nuts; If not, the property parameters of the target material are adjusted, and the process returns to the step of predicting a matching solution to re-predict a matching solution.

3. According to claim 1, the intelligent production method of corrosion-resistant self-clinching nuts is characterized in that: The step S2 comprises: Acquiring real-time temperature data at various positions in the heat treatment furnace, wherein the real-time temperature data is collected by a plurality of high-precision temperature sensors arranged at different positions in the heat treatment furnace; transmitting the real-time temperature data to a process control system; After receiving the real-time temperature data, the process control system compares the real-time temperature data with a preset target temperature, and calculates a deviation value between the real-time temperature data and the target temperature; If the deviation value exceeds a preset threshold, the process control system automatically adjusts the heating power of the heat treatment furnace according to the size and positive and negative of the deviation value; Acquiring three-dimensional temperature field data in the heat treatment furnace, wherein the three-dimensional temperature field data is collected by a thermal imager that performs real-time monitoring of the heat treatment furnace; The three-dimensional temperature field data is fused with the real-time temperature data, and a high-precision three-dimensional temperature field model in the heat treatment furnace is generated by a data fusion algorithm; The high-precision three-dimensional temperature field model is simulated by using a finite element analysis method to obtain a simulation result of furnace temperature uniformity; According to the simulation results, the control parameters of the heat treatment furnace are optimized.

4. According to claim 1, the intelligent production method of corrosion-resistant self-clinching nuts is characterized in that: Described step S4 comprises: Obtaining the corrosion resistance requirements of the rivet nut, and determining whether to use a galvanizing process or a nickel plating process for surface treatment according to the corrosion resistance requirements; For the zinc plating process or nickel plating process, a correlation model between the plating solution temperature, plating solution concentration and electroplating time process parameters and the coating quality is established; Using a support vector machine or a decision tree algorithm to train the association model to obtain an optimized association model; During the electroplating process, obtaining the temperature and concentration parameters of the plating solution, inputting the parameters into the optimized correlation model, and obtaining the predicted uniformity and density of the coating; According to the predicted uniformity and density of the coating, dynamically adjust the process parameters until the quality of the coating meets the preset requirements; Using a coating thickness meter to measure the coating thickness distribution, comparing the measured data with the target thickness, and determining whether the coating thickness is uniform; If the coating thickness is uneven, input the measurement data into the correlation model, and retrain and optimize the correlation model; Adhesion tester is used to detect the bonding force between the coating and the substrate to obtain adhesion data; Determining whether the adhesion is qualified according to a preset threshold; If the adhesion is unqualified, the adhesion data is input into a pre-plating process optimization model, a neural network algorithm is used to establish a correlation between pre-plating and post-plating process parameters and defects, and the pre-plating process and post-plating process are optimized; Comprehensively evaluate the uniformity of the coating thickness and the adhesion and appearance quality test results, and output a comprehensive quality score for the surface treatment process; Based on the comprehensive quality score, the process parameters are continuously optimized to improve the corrosion resistance and appearance quality of the rivet nut products to meet the usage requirements of downstream customers.

5. An intelligent production method for corrosion-resistant self-clinching nuts according to any one of claims 1 to 4, characterized in that: The step S6 comprises: Acquire surface defect detection result data and other performance index data of the rivet nut, perform data fusion processing on the surface defect detection result data and other performance index data, and obtain comprehensive evaluation data of the product quality of the rivet nut; Acquire performance test data, material property data, heat treatment process parameters and surface treatment process parameters of the self-clinching nut, and establish a correlation model between the performance of the self-clinching nut and production factors based on the performance test data, material property data, heat treatment process parameters and surface treatment process parameters of the self-clinching nut; Based on the association model, an association rule mining algorithm is used to mine the association rules between the production factors of material properties, heat treatment process parameters and surface treatment process parameters and the performance of the rivet nut, and the influence of each production factor on the performance of the rivet nut is obtained; A set of production parameter combination data to be predicted is obtained, the production parameter combination data is input into the association model, and the performance distribution of the rivet nut product under the production parameter combination is predicted by using a time series prediction algorithm in combination with the influence rule; According to the predicted performance distribution results, it is judged whether the production parameter combination meets the preset performance requirements. If so, the production parameter combination is determined to be a feasible production parameter and applied to the production process of the self-clinching nut; In the production process of the self-clinching nut, parameter data of each production link is collected in real time, and the parameter data collected in real time is input into the association model to dynamically predict the performance distribution of the self-clinching nut product and realize real-time monitoring of the product performance; If the predicted performance distribution results do not meet the preset performance requirements, the production parameters are adjusted in time and the above steps are re-executed until the predicted performance distribution results meet the preset performance requirements, thereby ensuring the quality stability of the self-clinching nut products.

6. An intelligent production method for corrosion-resistant self-clinching nuts according to any one of claims 1 to 4, characterized in that: The step S7 comprises: Acquiring equipment operation data, wherein the equipment operation data includes temperature, pressure and vibration frequency; Preprocessing the equipment operation data to obtain a standardized equipment operation data set; Based on the standardized equipment operation data set, a prediction model is constructed using a support vector machine regression algorithm to obtain a predicted value of equipment performance; If the predicted value of the equipment performance exceeds a preset threshold range, a warning signal is triggered; According to the early warning signal, a preliminary process optimization rule is obtained through a decision tree algorithm; Combining the expert knowledge base and Bayesian network reasoning, a process optimization suggestion is generated based on the preliminary process optimization rules, and the process optimization suggestion includes material selection, heat treatment temperature and time, and surface treatment solution concentration.