Hot pressing mold control method and device based on Internet of Things

Abstract

The invention relates to the technical field of data processing, in particular to a hot-pressing mold control method and device based on the Internet of Things, and the method comprises the steps: obtaining hot-pressing molding equipment data corresponding to a current unqualified carbon fiber product; obtaining a first probability and a second probability according to the hot press molding equipment data; if the first probability is smaller than or equal to a first probability threshold value and the second probability is smaller than or equal to a second probability threshold value, short-time step testing is conducted on each hot-press forming device, and a short-time step testing result of each hot-press forming device is obtained; determining the defect reason of the current unqualified carbon fiber product by using the short-time step test result of each hot press molding device; and according to the defect reason of the current unqualified carbon fiber product, the hot pressing mold is regulated and controlled. According to the invention, the control accuracy of the carbon fiber hot-press forming mold can be improved.

Description

Technical Field

[0001] This invention relates to the field of data processing technology, and more specifically to a hot pressing mold control method and device based on the Internet of Things. Background Technology

[0002] In the carbon fiber hot pressing process, controlling the hot pressing mold can effectively improve the yield of carbon fiber products.

[0003] Existing problems: Current methods for controlling carbon fiber hot pressing molds involve inspecting the quality of the carbon fiber products and adjusting the temperature and pressure of the mold accordingly. However, in actual production, improper process parameter settings and equipment malfunctions can both lead to quality problems in the carbon fiber products. Existing methods for controlling carbon fiber hot pressing molds only consider the impact of process parameters on the quality of the carbon fiber products, neglecting the influence of the hot pressing equipment. This reduces the accuracy of diagnosing the causes of quality degradation and consequently reduces the accuracy of controlling the carbon fiber hot pressing mold. Summary of the Invention

[0004] This invention provides a hot pressing mold control method and device based on the Internet of Things to solve existing problems.

[0005] The hot pressing mold control method and device based on the Internet of Things of the present invention adopts the following technical solution: One embodiment of the present invention provides a hot pressing mold control method based on the Internet of Things, the method comprising the following steps: Obtain the hot pressing equipment data corresponding to the currently substandard carbon fiber products; the hot pressing equipment data includes the hydraulic pump pressure data and the heating plate temperature data. The hydraulic pump pressure data includes the set pressure value and the actual pressure value of the hydraulic pump, and the heating plate temperature data includes the set temperature value and the actual temperature value of the heating plate. Based on the data from the hot pressing equipment, a first probability and a second probability are obtained; where the first probability represents the probability that the hot pressing equipment causes defects in the carbon fiber product, and the second probability represents the probability that the process parameters cause defects in the carbon fiber product. If the first probability is less than or equal to the first probability threshold, and the second probability is less than or equal to the second probability threshold, then a short-time step test is performed on each hot press molding equipment to obtain the short-time step test result for each hot press molding equipment; wherein, the short-time step test includes a short-time step heating test and a short-time step pressure test. By utilizing the short-time step test results of each hot-pressing molding equipment, the cause of defects in the current substandard carbon fiber products can be determined. Based on the causes of defects in the current substandard carbon fiber products, the hot pressing mold is adjusted.

[0006] Furthermore, the specific steps for obtaining the first probability and the second probability based on the data from the hot pressing molding equipment are as follows: The error rate of the hydraulic pump is determined by the ratio of the absolute value of the difference between the set pressure value and the actual pressure value to the set pressure value. The ratio of the absolute value of the difference between the set temperature value and the actual temperature value of the heating plate to the set temperature value is determined as the error rate of the heating plate. The larger error rate between the hydraulic pump and the heating plate is determined as the first probability, and (1 - the first probability) is determined as the second probability.

[0007] Furthermore, the specific steps for performing a short-time step test on each hot pressing molding device and obtaining the short-time step test results for each hot pressing molding device are as follows: A short-time step temperature rise test was performed on the heating plate to obtain the test results. A short-time step pressure test was performed on the hydraulic pump to obtain the test results.

[0008] Furthermore, the specific steps involved in performing a short-time step temperature rise test on the heating plate to obtain the test results are as follows: Obtain the maximum allowable temperature value of the heating plate; Multiple test stages are obtained based on the maximum allowable temperature value and the set temperature value of the heating plate; Based on the defective performance of the current substandard carbon fiber products in the actual production stage and in each test stage, we obtained the original data and the data for each test stage. Acquire historical data; including the first probability corresponding to historical substandard carbon fiber products; Based on historical data, obtain comparative data for each experimental phase; Based on the raw data, the data from each test stage, and the comparative data from each test stage, the short-time step temperature rise test results of the heating plate are obtained.

[0009] Furthermore, the specific steps for obtaining multiple test stages based on the maximum allowable temperature value and the set temperature value of the heating plate are as follows: Set the number of segments; The temperature gradient is determined by the ratio of the difference between the maximum allowable temperature value of the heating plate and the set temperature value to the number of segments. Based on a set temperature value, a temperature gradient is added sequentially to obtain multiple increased temperature values; the number of increased temperature values ​​is the same as the number of segments. Obtain the hot pressing cycle of the currently substandard carbon fiber products; Based on the hot pressing forming cycle, two hot pressing forming cycles are added sequentially to obtain multiple lifting cycles. Each lifting cycle is used as a test stage, thereby obtaining multiple test stages. The number of test stages is the same as the number of segments, and there is a one-to-one correspondence between the test stages and the lifting temperature values.

[0010] Furthermore, the specific steps for obtaining raw data and data for each testing stage based on the defective performance of the currently substandard carbon fiber products during the actual production stage and at each testing stage are as follows: The surface image of the carbon fiber product that is currently substandard after hot pressing is obtained, and the surface image is input into a trained neural network to obtain the defect type and severity of the carbon fiber product that is currently substandard in the actual production stage. The types and severity of defects in the current substandard carbon fiber products during the actual production stage are determined as the raw data; The surface images of the carbon fiber products that are currently substandard are obtained at each test stage, and each surface image is input into a trained neural network to obtain the defect type and defect severity of the carbon fiber products that are currently substandard at each test stage. The defect type and severity of the currently substandard carbon fiber products at each test stage are determined as data for each test stage.

[0011] Furthermore, the specific steps for obtaining the control data for each experimental phase based on historical data are as follows: Historical data with a probability less than the preset filtering threshold are selected as filtering data; Obtain surface images of substandard carbon fiber products at each test stage corresponding to the screening data, and input the surface images into a trained neural network to obtain the defect type and defect severity at each test stage corresponding to the screening data. The defect type and severity of each test stage corresponding to the screened data are used as the control data for each test stage.

[0012] Furthermore, the specific steps for obtaining the short-time step temperature rise test results of the heating plate based on the original data, the data from each test stage, and the comparison data from each test stage are as follows: The set of defect types in the original data and the set of defect types in the control data is defined as the defect type set. Determine whether the defect type in the data of each test phase belongs to the defect type set: if it does, then determine the defect severity enhancement rate for each test phase by the ratio of the difference between the defect severity in the data of each test phase and the defect severity in the original data to the defect severity in the data of each test phase. The average of the defect severity enhancement rates across all testing phases was determined as the first average. The defect severity enhancement rate was sorted according to the temperature increase value in the test phase from low to high, and a defect severity enhancement rate sequence was obtained. The slope of the fitted line was obtained by fitting the defect severity enhancement rate sequence using the least squares method. The slope of the fitted line is normalized to obtain the normalized value of the slope of the fitted line. The product of the first mean and the normalized value of the slope of the fitted line is determined as the defect probability value of the heating plate, and the defect probability value of the heating plate is used as the result of the short-time step temperature rise test of the heating plate.

[0013] Furthermore, the specific steps for determining the cause of defects in the currently substandard carbon fiber products using the short-time step test results of each hot-pressing molding equipment are as follows: The results of the short-time step pressure test of the hydraulic pump and the short-time step temperature test of the heating plate are judged: if the short-time step pressure test result of the hydraulic pump is less than the preset defect threshold, or the short-time step temperature test result of the heating plate is less than the preset defect threshold, then the defect of the current unqualified carbon fiber product is determined to be the hot pressing molding equipment. If the short-time step pressure test result of the hydraulic pump and the short-time step temperature test result of the heating plate are both greater than or equal to the preset defect threshold, then the defect of the current substandard carbon fiber product is determined to be due to process parameters.

[0014] One embodiment of the present invention provides a hot pressing mold control device based on the Internet of Things, the device comprising the following units: The data acquisition unit is used to acquire the hot pressing equipment data corresponding to the current substandard carbon fiber products; wherein, the hot pressing equipment data includes the pressure data of the hydraulic pump and the temperature data of the heating plate. The pressure data of the hydraulic pump includes the set pressure value and the actual pressure value of the hydraulic pump, and the temperature data of the heating plate includes the set temperature value and the actual temperature value of the heating plate. The cause analysis unit is used to obtain a first probability and a second probability based on data from the hot pressing equipment. The first probability represents the probability that the hot pressing equipment causes defects in the carbon fiber product, and the second probability represents the probability that process parameters cause defects. If the first probability is less than or equal to a first probability threshold, and the second probability is less than or equal to a second probability threshold, a short-time step test is performed on each hot pressing equipment to obtain the short-time step test result for each equipment. The short-time step test includes a short-time step temperature rise test and a short-time step pressure rise test. Using the short-time step test results of each hot pressing equipment, the cause of the defect in the currently substandard carbon fiber product is determined. The mold control unit is used to adjust the hot pressing mold according to the cause of defects in the current substandard carbon fiber products.

[0015] The beneficial effects of the technical solution of this invention are as follows: This invention proposes a hot-pressing mold control method and device based on the Internet of Things. It calculates the probability of defects caused by hot-pressing equipment and process parameters, and actively applies specific temperature / pressure excitation when the probability is low to amplify equipment characteristics and suppress parameter interference. Furthermore, it re-evaluates the cause of defects based on the defect evolution trend, achieving precise location of the defect cause. By constructing a closed-loop self-optimizing system, it ultimately forms a smart control method for hot-pressing molds that can significantly improve diagnostic accuracy and process optimization efficiency. This invention can improve the diagnostic accuracy of causes of quality degradation in carbon fiber products, thereby improving the control accuracy of carbon fiber hot-pressing molds. Attached Figure Description

[0016] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0017] Figure 1 This is a flowchart of the steps of the hot pressing mold control method based on the Internet of Things of the present invention; Figure 2 This is a block diagram of the hot pressing mold control device based on the Internet of Things of the present invention. Detailed Implementation

[0018] To further illustrate the technical means and effects adopted by the present invention to achieve its intended purpose, the following, in conjunction with the accompanying drawings and preferred embodiments, details the specific implementation, structure, features, and effects of the IoT-based hot pressing mold control method and apparatus proposed according to the present invention. In the following description, different "one embodiment" or "another embodiment" do not necessarily refer to the same embodiment. Furthermore, specific features, structures, or characteristics in one or more embodiments can be combined in any suitable form.

[0019] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention pertains.

[0020] The following description, in conjunction with the accompanying drawings, details the specific scheme of the IoT-based hot pressing mold control method and device provided by this invention.

[0021] In existing hot pressing processes for carbon fiber products, quality control primarily relies on inspecting the final product's quality and adjusting the temperature and pressure parameters of the hot pressing process accordingly. However, this feedback control method has inherent limitations: when quality defects originate from malfunctions in the equipment itself, simply adjusting the temperature / pressure setpoints is insufficient to fundamentally compensate for physical deviations at the equipment level. More seriously, in the early stages of equipment failure, minor anomalies often fail to reach system alarm thresholds and are therefore undetectable by conventional detection methods; however, these "sub-healthy" states are already sufficient to cause product defects. Therefore, the current method lacks sufficient control capability. To address these issues, this invention proposes a hot pressing mold control method and apparatus based on the Internet of Things (IoT).

[0022] Please see Figure 1 The diagram illustrates a flowchart of a hot-pressing mold control method based on the Internet of Things (IoT) according to an embodiment of the present invention. The method includes the following steps: Step S001: Obtain the hot pressing equipment data corresponding to the current substandard carbon fiber products; wherein, the hot pressing equipment data includes the pressure data of the hydraulic pump and the temperature data of the heating plate, the pressure data of the hydraulic pump includes the set pressure value and the actual pressure value of the hydraulic pump, and the temperature data of the heating plate includes the set temperature value and the actual temperature value of the heating plate.

[0023] It should be noted that the hot pressing process for carbon fiber products is as follows: Preparation stage: Workers lay carbon fiber prepreg on the mold, and then completely seal it with vacuum bags and sealing materials to form a closed space.

[0024] Mold assembly and closing: The packaged mold is sent into the machine frame, located between the upper and lower hot press plates. The hydraulic cylinder drives the upper hot press plate to move downward to close the mold.

[0025] Vacuuming stage: The vacuum pump is started, and the air and volatiles in the sealed space are removed through the tubing on the vacuum bag. This process is monitored by the control system, which uses pressure sensors and vacuum data.

[0026] Heating and pressurization stage: The heater starts working, heating the hot press plate and mold to the predetermined temperature, and the thermocouple provides real-time temperature feedback to the control system. When the temperature reaches the critical point, the hydraulic station provides power to the hydraulic cylinder, driving the hot press plate to apply enormous forming pressure to the mold.

[0027] Curing and heat preservation stage: Under the precise coordination of the control system, the system maintains the set temperature and time to allow the resin to fully cure, and the vacuum pump continues to work to maintain the vacuum environment.

[0028] Cooling and depressurization stage: After curing, the cooling system is activated to controllably cool the hot press plate; once the temperature drops to a safe range, the hydraulic system is depressurized and the vacuum pump stops working.

[0029] Mold opening and part removal: The hydraulic cylinder drives the upper hot platen to move upward to open the mold, and the worker takes out the fully formed carbon fiber product.

[0030] The carbon fiber products that are currently substandard are those whose defects require further analysis.

[0031] This embodiment uses sensors to monitor the actual data of the hot press mold: a pressure sensor is used to obtain the actual pressure value of the hydraulic pump; a temperature sensor is used to obtain the actual temperature value of the heating plate.

[0032] The set pressure value of the hydraulic pump and the set temperature value of the heating plate are set according to specific production requirements, and no specific limitation is made here.

[0033] Step S002: Based on the data from the hot pressing equipment, obtain the first probability and the second probability; wherein, the first probability represents the probability that the hot pressing equipment causes defects in the carbon fiber product, and the second probability represents the probability that the process parameters cause defects in the carbon fiber product.

[0034] Based on the data from the hot pressing equipment, obtain the first probability and the second probability, including steps S21-S23: Step S21: The ratio of the absolute value of the difference between the set pressure value and the actual pressure value of the hydraulic pump to the set pressure value is determined as the error rate of the hydraulic pump.

[0035] Step S22: The ratio of the absolute value of the difference between the set temperature value and the actual temperature value of the heating plate to the set temperature value is determined as the error rate of the heating plate.

[0036] Step S23: Determine the larger error rate between the hydraulic pump error rate and the heating plate error rate as the first probability, and determine (1 - first probability) as the second probability.

[0037] It should be noted that the first probability is denoted as p1, and the second probability is denoted as p2.

[0038] Step S003: If the first probability is less than or equal to the first probability threshold, and the second probability is less than or equal to the second probability threshold, then perform a short-time step test on each hot pressing molding equipment and obtain the short-time step test result for each hot pressing molding equipment; wherein, the short-time step test includes a short-time step heating test and a short-time step pressure test.

[0039] It should be noted that the first probability threshold is denoted as Y1, and the second probability threshold is denoted as Y2. The first and second probability thresholds are set according to specific circumstances, and are not limited here.

[0040] In this embodiment, both Y1 and Y2 are 0.7, which effectively avoids the risk of misjudgment caused by noise or random fluctuations. The system will only initiate targeted control adjustments when the probability of a certain factor is extremely overwhelming (>70%), thereby ensuring the stability of the production process and avoiding inappropriate or frequent interventions when the cause is unknown.

[0041] If either p1 or p2 is greater than the corresponding probability threshold, the corresponding event (the corresponding event for p1 refers to the equipment, and the corresponding event for p2 refers to the process parameters) is considered to have caused the defect in the carbon fiber product; if neither p1 nor p2 is greater than the corresponding probability threshold, a short-time step test is performed.

[0042] Short-time step testing, sometimes called step response testing, is a system identification method. Its basic idea is to apply a sudden, large-amplitude change in the input signal to the system (i.e., a "step"), and then observe and record the change in the system output over time (i.e., the "response") to understand the system's dynamic characteristics. In this embodiment, short-time step testing includes temperature and pressure step tests.

[0043] If the first probability is greater than the first probability threshold, it is considered that the defect in the carbon fiber product is caused by the hot pressing equipment, and the hot pressing equipment is located and repaired; if the second probability is greater than the second probability threshold, it is considered that the defect in the carbon fiber product is caused by the process parameters, and the process parameters are adjusted.

[0044] Perform a short-time step test on each hot press molding machine and obtain the short-time step test results for each hot press molding machine, including steps S31 and S32: Step S31: Perform a short-time step temperature rise test on the heating plate and obtain the short-time step temperature rise test results of the heating plate.

[0045] Perform a short-time step temperature rise test on the heating plate and obtain the test results, including steps S311-S316: Step S311: Obtain the maximum allowable temperature value of the heating plate.

[0046] It should be noted that the maximum allowable temperature value is denoted as T0.

[0047] Step S312: Obtain multiple test stages based on the maximum allowable temperature value and the set temperature value of the heating plate.

[0048] It should be noted that the set temperature value is denoted as T.

[0049] Based on the maximum allowable temperature value and the set temperature value of the heating plate, multiple test stages are obtained, including steps S3121-S3125: Step S3121: Set the number of segments.

[0050] It should be noted that users can set the number of segments according to their actual situation. In this embodiment, the number of segments is 8.

[0051] Step S3122: The temperature gradient is determined by the ratio of the difference between the maximum allowable temperature value of the heating plate and the set temperature value to the number of segments.

[0052] It should be noted that the temperature range between T and T0 is divided into 8 segments to obtain the temperature gradient. .

[0053] Step S3123: Based on the set temperature value, add a temperature gradient sequentially to obtain multiple temperature increase values; wherein the number of temperature increase values ​​is the same as the number of segments.

[0054] It should be noted that: obtaining multiple temperature increase values: obtaining the first temperature increase value T+ Obtain the second temperature increase value T+2× Obtain the third temperature increase value T+3× Obtain the 4th temperature increase value T+4× Obtain the 5th temperature increase value T+5× Obtain the 6th temperature increase value T+6× Obtain the 7th temperature increase value T+7× Obtain the 8th temperature increase value T+8× (That is, T0).

[0055] Step S3124: Obtain the hot pressing cycle of the carbon fiber product that is currently substandard.

[0056] It should be noted that the hot pressing cycle refers to the normal cycle A of a hot pressing process, for example: 300s.

[0057] Step S3125: Based on the hot pressing forming cycle, add two hot pressing forming cycles in sequence to obtain multiple lifting cycles. Each lifting cycle is used as a test stage to obtain multiple test stages. The number of test stages is the same as the number of segments, and there is a one-to-one correspondence between the test stages and the lifting temperature values.

[0058] It should be noted that, assuming the end time of the hot pressing cycle is t1, then the time t1+2×A is taken as the end time of the first lifting cycle, and the period from t1 to t1+2×A is considered the first lifting cycle, i.e., the first experimental stage. The first experimental stage corresponds to the first lifting temperature value T+ This indicates that the temperature of the first experimental phase needs to be increased from T to T+. .

[0059] The time t1+4×A is taken as the end time of the second lifting cycle, and the period from t1+2×A to the end of t1+4×A is defined as the second lifting cycle, i.e., the second experimental phase. The second experimental phase corresponds to the second lifting temperature value T+2×A. This indicates that the temperature for the second experimental phase needs to be changed from T+ Increase to T+2× Following this pattern, eight experimental stages can be obtained, with each stage corresponding to a temperature increase value.

[0060] Step S313: Based on the defective performance of the currently substandard carbon fiber products in the actual production stage and in each test stage, obtain the raw data and the data for each test stage.

[0061] It should be noted that the actual production stage refers to the stage used for the mass production of carbon fiber products. The testing stage is a small-scale testing phase conducted to analyze the causes of substandard products. The test objects in the testing stage are carbon fiber products that have already been found to be substandard.

[0062] Defect manifestations include defect type and defect severity.

[0063] Based on the defective performance of the currently substandard carbon fiber products during the actual production stage and at each testing stage, raw data and data for each testing stage are obtained, including steps S3131-S3134: Step S3131: Obtain a surface image of the carbon fiber product that is currently substandard after hot pressing, and input the surface image into the trained neural network to obtain the defect type and severity of the carbon fiber product that is currently substandard in the actual production stage.

[0064] It should be noted that when defects occur in carbon fiber products, it may be due to inappropriate process parameters or malfunctions in the equipment involved in the hot pressing process. Minor equipment malfunctions, influenced by factors such as machine noise, are difficult to identify. Therefore, this embodiment identifies the root cause of the defect and then adjusts the hot pressing process using different control methods. When minor equipment malfunctions cause defects, it is difficult to obtain high-quality carbon fiber products regardless of adjustments to the process parameters.

[0065] When a carbon fiber product produced in a certain batch is unqualified, i.e. has defects, the visible light image (surface image) of the carbon fiber product obtained in that batch is first obtained. The type and severity of the defect in the image are then obtained through a neural network recognition method.

[0066] The training process of a neural network is as follows: Objective: To construct an end-to-end network that takes a surface image of a carbon fiber product as input and outputs simultaneously: defect location (pixel level), defect type, and defect severity.

[0067] Network Architecture: A multi-task structure of "one encoder and three decoders" is adopted. The encoder uses ResNet (Residual Network) or HRNet (High-Resolution Network) as the backbone to extract general image features. Decoder 1 (Segmentation): Obtains defect locations through global pooling and fully connected layers. Decoder 2 (Classification): Identifies defect types through global pooling and fully connected layers. Decoder 3 (Regression): Calculates defect severity (e.g., area percentage) through global pooling and regression layers.

[0068] Data and Training: Dataset: It must contain accurate pixel-level annotations, defect type labels, and defect severity values ​​(the severity here is represented by the percentage of defect pixels).

[0069] Loss function: Total loss = weighted sum of Dice Loss (segmentation) + Focal Loss (classification) + Smooth L1 Loss (regression).

[0070] Training process: The encoder is initialized using transfer learning, followed by end-to-end joint training, which enables the network to learn to share features and complete multiple tasks.

[0071] Step S3132: Determine the defect type and severity of the currently substandard carbon fiber products in the actual production stage as raw data.

[0072] Step S3133: Obtain surface images of the currently substandard carbon fiber products at each test stage, and input each surface image into the trained neural network to obtain the defect type and defect severity of the currently substandard carbon fiber products at each test stage.

[0073] Step S3134: Determine the defect type and severity of the currently substandard carbon fiber products at each test stage, and obtain the data for each test stage.

[0074] Step S314: Obtain historical data; wherein, the historical data includes the first probability corresponding to historical substandard carbon fiber products.

[0075] It should be noted that if the defect is caused by the equipment, the temperature rise command will not be responded to in time or there will be no response. In this case, the formed carbon fiber product will not have any new defects or the defects will be similar to those before.

[0076] If the equipment is functioning normally and the defect is caused by process parameters, the system can quickly and accurately reach the new set value and stabilize rapidly after recovery. Defects in the formed carbon fiber products may temporarily change (e.g., flash due to overheating) or worsen on top of the original defects. By applying pressure / temperature stimulation to the equipment and observing the defect behavior of the carbon fiber products as the pressure / temperature gradually increases, the cause of the defect problem can be determined.

[0077] The first probability corresponding to historically substandard carbon fiber products is obtained in the same way as the first probability of currently substandard carbon fiber products.

[0078] Step S315: Based on historical data, obtain control data for each experimental phase.

[0079] Based on historical data, obtain control data for each experimental phase, including steps S3151-S3153: Step S3151: Historical data with a first probability less than the preset filtering threshold are identified as filtering data.

[0080] It should be noted that the preset filtering threshold is set according to specific circumstances, and is not specifically limited here. In this embodiment, historical data with a first probability of less than 0.1 are determined as the filtering data.

[0081] Step S3152: Obtain surface images of the substandard carbon fiber products under each test stage corresponding to the screening data, and input the surface images into the trained neural network to obtain the defect type and defect severity under each test stage corresponding to the screening data.

[0082] It should be noted that if the probability of the first screening data is very small, it means that the screening data is likely a defect caused by process parameters. When a short-term step test of pressure / temperature is performed on the hot pressing equipment corresponding to the screening data, the system can accurately and quickly reach the new set value. In this case, the type of defect in the corresponding carbon fiber product may change or the degree of defect may be aggravated.

[0083] Step S3153: Determine the defect type and defect severity for each test stage corresponding to the screened data as the control data for each test stage.

[0084] It should be noted that: the defect type and severity corresponding to the first test stage of the selected data are used as the control data for the first test stage; the defect type and severity corresponding to the second test stage of the selected data are used as the control data for the second test stage; and so on.

[0085] Step S316: Based on the original data, the data of each test stage, and the comparison data of each test stage, obtain the short-time step temperature rise test results of the heating plate.

[0086] Based on the raw data, data from each test stage, and comparative data from each test stage, the short-time step temperature rise test results of the heating plate are obtained, including steps S3161-S3167: Step S3161: Determine the set of defect types formed by the defect types in the original data and the defect types in the control data as the defect type set.

[0087] It should be noted that if the defect types in the experimental stage appear in both the control and original data, then the experimental stage data represents a subset of the defect types in these two data sets. Furthermore, if the severity of the corresponding defects in the experimental stage data is greater than that in the original data, then it is highly likely that the defect is caused by process parameters. Otherwise, the equipment should be inspected and repaired before proceeding with the production of carbon fiber products.

[0088] Obtain the defect types from the raw data and the control data to form a defect type set, such as: {porosity, delamination, warping}.

[0089] Step S3162: Determine whether the defect type in the data of each test stage belongs to the defect type set: If it does, then determine the defect severity enhancement rate for each test stage by the difference between the defect severity in the data of each test stage and the defect severity in the original data, and the ratio of the defect severity in the data of each test stage.

[0090] It should be noted that, taking the first experimental phase as an example: If only X belongs to the defect type set in the defect type data of the first experimental stage, then the ratio of the difference between the defect severity U1 of the defect type X in the experimental stage data and the defect severity u0 of the defect type X in the original data and U1 is recorded as the defect severity enhancement rate of the first experimental stage.

[0091] If both X and Y belong to the defect type set in the data of the first experimental phase, then the ratio of the difference between the defect severity U1 of the defect type X in the experimental phase data and the defect severity u0 of the original data to U1 is taken as the ratio of X; the ratio of the difference between the defect severity U1' of the defect type Y in the experimental phase data and the defect severity u0' of the original data to U1' is taken as the ratio of Y; and the mean of the ratios of X and Y is recorded as the defect severity enhancement rate of the first experimental phase.

[0092] If multiple defect types belong to the defect type set in the data of the first experimental phase, the average of the ratios of the multiple defect types is recorded as the defect severity enhancement rate of the first experimental phase.

[0093] The defect severity enhancement rate for each experimental stage was obtained using the method described above. It should be noted that if a certain experimental stage does not satisfy the subset relationship or if the U-value of the corresponding stage is greater than u0, then the defect severity enhancement rate for that experimental stage will not be calculated. The errors in this calculation are most likely due to other factors or random errors, and therefore are not included in the calculation to avoid noise interference with the results.

[0094] Step S3163: Determine the average of the defect severity enhancement rates of all test phases as the first average.

[0095] It should be noted that the more test stages there are, the more accurate the defect severity enhancement rate can be. The more obvious the increasing trend of the defect severity enhancement rate is as the test stages continue, the more likely it is that the defect is not caused by equipment, but by process parameters.

[0096] Let the first mean be denoted as h1. If all N test stages participate in the calculation of the defect severity enhancement rate, then the first mean is the mean of the N test stages.

[0097] Step S3164: Sort the defect severity enhancement rate of all test stages in order of increasing temperature value from low to high to obtain the defect severity enhancement rate sequence.

[0098] Step S3165: Fit the defect severity enhancement rate sequence using the least squares method to obtain the slope of the fitted line.

[0099] It should be noted that the slope of the fitted line is denoted as k. The larger the k is, the more obvious the increasing trend is.

[0100] Step S3166: Normalize the slope of the fitted line to obtain the normalized value of the slope of the fitted line.

[0101] It should be noted that a slope threshold k0 is preset. If k > k0, then k = 1; otherwise, the ratio of k to k0 is calculated to obtain the normalized k value.

[0102] Step S3167: The product of the first mean and the normalized value of the slope of the fitted line is determined as the defect probability value of the heating plate, and the defect probability value of the heating plate is used as the result of the short-time step temperature rise test of the heating plate.

[0103] It should be noted that the product of h1 and the normalized k value is used to obtain the defect probability value of the heating plate, denoted as w. The larger w is, the greater the probability that the defect is caused by process parameters.

[0104] Step S32: Perform a short-time step pressure test on the hydraulic pump and obtain the short-time step pressure test results of the hydraulic pump.

[0105] It should be noted that: following the process of conducting a short-time step temperature rise test on the heating plate, a short-time step pressure rise test is conducted on the hydraulic pump to obtain the defect probability value q of the hydraulic pump.

[0106] Step S004: Using the short-time step test results of each hot pressing molding equipment, determine the cause of the defect in the current substandard carbon fiber product.

[0107] Using the short-time step test results of each hot-pressing molding equipment, determine the cause of defects in the currently substandard carbon fiber products, including steps S41 and S42: Step S41: Judge the short-time step pressure test result of the hydraulic pump and the short-time step temperature test result of the heating plate: If the short-time step pressure test result of the hydraulic pump is less than the preset defect threshold, or the short-time step temperature test result of the heating plate is less than the preset defect threshold, then the defect of the current substandard carbon fiber product is determined to be the hot pressing molding equipment.

[0108] It should be noted that the preset defect threshold is set according to specific circumstances, and will not be described in detail here. In this embodiment, the preset defect threshold is 0.6.

[0109] Step S42: If the short-time step pressure test result of the hydraulic pump and the short-time step temperature test result of the heating plate are both greater than or equal to the preset defect threshold, then the defect of the current substandard carbon fiber product is determined to be due to process parameters.

[0110] Step S005: Adjust the hot pressing mold according to the cause of the defects in the current substandard carbon fiber products.

[0111] It should be noted that: once the cause of the equipment defect is determined, the next steps are: location and repair: immediately perform targeted maintenance on the malfunctioning equipment. Performance verification: after repair, run a no-load test to confirm that the equipment output accurately follows the set value. Production resumption: production can only be resumed after the equipment performance verification is passed.

[0112] Once the cause of the defect due to process parameters is identified, reverse parameter optimization is performed: based on the defect type, the core process parameters are adjusted in reverse. If pores / bubbles appear, reduce the heating rate, advance or increase the pressure; if delamination occurs, increase the pressure and extend the holding time; if insufficient curing occurs, increase the curing temperature and extend the curing time, etc.

[0113] Validation and consolidation: Conduct small-batch trial production using the new parameters. After passing the inspection, update the standard process documents.

[0114] Please see Figure 2 The diagram illustrates a block diagram of an IoT-based hot pressing mold control device according to an embodiment of the present invention, which includes the following units: The data acquisition unit 100 is used to acquire the hot pressing equipment data corresponding to the current substandard carbon fiber products; wherein, the hot pressing equipment data includes the pressure data of the hydraulic pump and the temperature data of the heating plate, the pressure data of the hydraulic pump includes the set pressure value and the actual pressure value of the hydraulic pump, and the temperature data of the heating plate includes the set temperature value and the actual temperature value of the heating plate.

[0115] The cause analysis unit 200 is used to obtain a first probability and a second probability based on data from the hot pressing equipment. The first probability represents the probability that the hot pressing equipment causes defects in the carbon fiber product, and the second probability represents the probability that process parameters cause defects in the carbon fiber product. If the first probability is less than or equal to a first probability threshold, and the second probability is less than or equal to a second probability threshold, a short-time step test is performed on each hot pressing equipment to obtain the short-time step test result for each equipment. The short-time step test includes a short-time step heating test and a short-time step pressurization test. Using the short-time step test results of each hot pressing equipment, the cause of the defect in the currently substandard carbon fiber product is determined.

[0116] The mold control unit 300 is used to control the hot pressing mold according to the cause of the defects in the current substandard carbon fiber products.

[0117] In summary, in this embodiment of the invention, by calculating the probability of defects caused by hot pressing equipment and process parameters, specific temperature / pressure excitation is actively applied when the probability is low to amplify equipment characteristics and suppress parameter interference; then, the cause of defects is re-evaluated according to the defect evolution trend to achieve accurate positioning of the cause of defects. By constructing a closed-loop self-optimization system, a smart control method for hot pressing molds that can significantly improve diagnostic accuracy and process optimization efficiency is finally formed.

[0118] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A hot press mold control method based on the Internet of Things, characterized by, The method comprises the following steps: Obtaining hot-press forming equipment data corresponding to the carbon fiber product of the current unqualified quality; wherein the hot-press forming equipment data comprises pressure data of a hydraulic pump and temperature data of a heating plate, the pressure data of the hydraulic pump comprises a set pressure value and an actual pressure value of the hydraulic pump, and the temperature data of the heating plate comprises a set temperature value and an actual temperature value of the heating plate; According to the hot-press forming equipment data, a first probability and a second probability are obtained; wherein the first probability represents the probability that the hot-press forming equipment causes defects of the carbon fiber product, and the second probability represents the probability that process parameters cause defects of the carbon fiber product; If the first probability is less than or equal to a first probability threshold value, and the second probability is less than or equal to a second probability threshold value, then a short-time step test is performed on each hot-press forming equipment to obtain a short-time step test result of each hot-press forming equipment; wherein the short-time step test comprises a short-time step temperature rising test and a short-time step pressure rising test; The short-time step test result of each hot-press forming equipment is used to determine the defect cause of the carbon fiber product of the current unqualified quality; According to the defect cause of the carbon fiber product of the current unqualified quality, the hot-press mold is regulated and controlled. 2.The hot press mold control method based on the Internet of Things according to claim 1, wherein, The method comprises the following steps: The absolute value of the difference between the set pressure value and the actual pressure value of the hydraulic pump is divided by the set pressure value to determine the error rate of the hydraulic pump; The absolute value of the difference between the set temperature value and the actual temperature value of the heating plate is divided by the set temperature value to determine the error rate of the heating plate; The larger error rate between the error rate of the hydraulic pump and the error rate of the heating plate is determined as the first probability, and (1-the first probability) is determined as the second probability. 3.The hot press mold control method based on the Internet of Things according to claim 1, wherein, The method comprises the following steps: The heating plate is subjected to a short-time step temperature rising test to obtain a short-time step temperature rising test result of the heating plate; The hydraulic pump is subjected to a short-time step pressure rising test to obtain a short-time step pressure rising test result of the hydraulic pump. 4.The hot press mold control method based on the Internet of Things according to claim 3, wherein, The method comprises the following steps: A maximum allowable temperature value of the heating plate is obtained; Based on the maximum allowable temperature value and the set temperature value of the heating plate, a plurality of test stages are obtained; Based on the defect performance of the carbon fiber product of the current unqualified quality in the actual production stage and in each test stage, original data and each test stage data are obtained; Historical data are obtained; wherein the historical data comprise a first probability corresponding to a historical carbon fiber product of unqualified quality; Based on the historical data, contrast data of each test stage data are obtained; Based on the original data, each test stage data, and the contrast data of each test stage data, a short-time step temperature rising test result of the heating plate is obtained. 5.The hot press mold control method based on the Internet of Things according to claim 4, wherein, The method comprises the following steps: A segmentation number is set; The difference between the maximum allowable temperature value and the set temperature value of the heating plate is divided by the segmentation number to determine a temperature gradient. On the basis of the temperature value, a temperature gradient is sequentially increased to obtain a plurality of elevated temperature values; wherein the number of the elevated temperature values is the same as the number of the segments; A hot-pressing cycle of the carbon fiber product with the current unqualified quality is obtained; On the basis of the hot-pressing cycle, two hot-pressing cycles are sequentially increased to obtain a plurality of elevated cycles, and each elevated cycle is taken as a test stage, thereby obtaining a plurality of test stages; wherein the number of the test stages is the same as the number of the segments, and the test stages have a one-to-one correspondence with the elevated temperature values. 6.The hot press mold control method based on the Internet of Things according to claim 4, wherein, Based on the defect performance of the carbon fiber product with the current unqualified quality in the actual production stage and in each test stage, original data and each test stage data are obtained, including the following specific steps: A surface image of the carbon fiber product with the current unqualified quality after hot-pressing is obtained, and the surface image is input into the trained neural network to obtain the defect type and defect severity of the carbon fiber product with the current unqualified quality in the actual production stage; The defect type and defect severity of the carbon fiber product with the current unqualified quality in the actual production stage are determined as the original data; A surface image of the carbon fiber product with the current unqualified quality in each test stage is obtained, and each surface image is input into the trained neural network to obtain the defect type and defect severity of the carbon fiber product with the current unqualified quality in each test stage; The defect type and defect severity of the carbon fiber product with the current unqualified quality in each test stage are determined as the test stage data. 7.The hot press mold control method based on the Internet of Things according to claim 4, wherein, Based on the historical data, the comparison data of each test stage data is obtained, including the following specific steps: The historical data with a first probability less than a preset screening threshold are determined as screening data; A surface image of the carbon fiber product with unqualified quality in each test stage corresponding to the screening data is obtained, and the surface image is input into the trained neural network to obtain the defect type and defect severity in each test stage corresponding to the screening data; The defect type and defect severity in each test stage corresponding to the screening data are determined as the comparison data of each test stage data. 8.The hot press mold control method based on the Internet of Things according to claim 4, wherein, According to the original data, each test stage data, and the comparison data of each test stage data, a short-time step-up temperature test result of the heating plate is obtained, including the following specific steps: A set composed of the defect types in the original data and the defect types in the comparison data is determined as a defect type set; It is judged whether the defect type in each test stage data belongs to the defect type set: if it belongs, the difference between the defect severity in each test stage data and the defect severity in the original data and the ratio of the defect severity in each test stage data are determined as the defect severity enhancement rate of each test stage; The average of the defect severity enhancement rates of all test stages is determined as a first average; According to the order of the elevated temperature values of the test stages from low to high, the defect severity enhancement rates of all test stages are sorted to obtain a defect severity enhancement rate sequence; The least square method is used to fit the defect severity enhancement rate sequence to obtain a fitting straight line slope; The fitting straight line slope is normalized to obtain a normalized value of the fitting straight line slope; The product of the first mean value and the normalized value of the fitting straight line slope is determined as the defect probability value of the heating plate, and the defect probability value of the heating plate is used as the short-time step-up temperature test result of the heating plate. 9.The hot press mold control method based on the Internet of Things according to claim 1, wherein, The specific steps of determining the defect cause of the current substandard carbon fiber product by using the short-time step-up test result of each hot pressing forming equipment are as follows: The short-time step-up pressure test result of the hydraulic pump and the short-time step-up temperature test result of the heating plate are judged: if the short-time step-up pressure test result of the hydraulic pump is less than the preset defect threshold, or the short-time step-up temperature test result of the heating plate is less than the preset defect threshold, it is determined that the defect cause of the current substandard carbon fiber product is the hot pressing forming equipment; If the short-time step-up pressure test result of the hydraulic pump and the short-time step-up temperature test result of the heating plate are both greater than or equal to the preset defect threshold, it is determined that the defect cause of the current substandard carbon fiber product is the process parameter.

10. A hot press mold control device based on the Internet of Things, characterized by, The device comprises the following units: A data acquisition unit is configured to acquire hot pressing forming equipment data corresponding to the current substandard carbon fiber product; wherein the hot pressing forming equipment data comprises pressure data of the hydraulic pump and temperature data of the heating plate, the pressure data of the hydraulic pump comprises a set pressure value and an actual pressure value of the hydraulic pump, and the temperature data of the heating plate comprises a set temperature value and an actual temperature value of the heating plate; A cause analysis unit is configured to acquire a first probability and a second probability according to the hot pressing forming equipment data; wherein the first probability represents the probability that the hot pressing forming equipment causes defects in the carbon fiber product, and the second probability represents the probability that the process parameter causes defects in the carbon fiber product; if the first probability is less than or equal to a first probability threshold, and the second probability is less than or equal to a second probability threshold, a short-time step-up test is performed on each hot pressing forming equipment to obtain a short-time step-up test result of each hot pressing forming equipment; wherein the short-time step-up test comprises a short-time step-up temperature test and a short-time step-up pressure test; and the defect cause of the current substandard carbon fiber product is determined by using the short-time step-up test result of each hot pressing forming equipment; A mold control unit is configured to control the hot pressing mold according to the defect cause of the current substandard carbon fiber product.

Images