Chip braid film coating temperature and pressure cooperative adjustment method based on fuzzy control
By combining fuzzy control and deep learning to coordinate temperature and pressure regulation, the chip tape-and-coating process can be monitored and optimized in real time, solving the problem of difficult detection of temperature and pressure fluctuations in existing technologies and improving product consistency and production stability.
Patent Information
- Application Number
- CN202511877695.6
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-12-12
- Publication Date
- 2026-02-13
AI Technical Summary
Existing chip tape-and-coating processes struggle to detect minute fluctuations in temperature and pressure in real time, making it difficult to promptly identify and address defects during the coating process. Furthermore, the lack of in-depth data mining and self-learning capabilities based on historical operating conditions negatively impacts product consistency and yield.
A temperature and pressure coordinated regulation method based on fuzzy control is adopted. By setting up multiple sensors and vision acquisition components at the coating processing end, the surface temperature and morphology of the carrier tape are monitored in real time. By combining the analysis of historical data with a deep learning model, temperature and pressure regulation parameters are generated, and the heating area and acquisition equipment are dynamically adjusted to achieve anomaly detection and adaptive optimization.
It achieves dynamic and precise control of the coating process, reduces defect rate, improves packaging yield, enhances defect detection rate and production stability, reduces energy consumption, and has self-learning and self-correction capabilities.
Smart Images
Figure CN121523480A_ABST
Abstract
Description
TECHNICAL FIELD
[0001] The present application relates to the technical field of data conditioning, in particular to a chip ribbon covering film temperature and pressure collaborative conditioning method based on fuzzy control. BACKGROUND
[0002] Chip product ribbon covering film is the last link of chip packaging, which realizes the sealing protection of chips by covering heat-sealing film on the surface of the carrier tape, and directly affects the storage life and transportation safety of the chips. The chip ribbon process is widely used in batch production and logistics management of high-precision microelectronic components such as semiconductor chips and integrated circuits. With the miniaturization of chips, the increasing of functional integration and the increasingly stringent requirements for high reliability, the chip ribbon covering film process puts forward higher requirements for fine collaborative conditioning of temperature and pressure parameters. In the covering film process, the dynamic balance of temperature and pressure directly affects the adhesion strength of the carrier tape and the chip, the integrity of the covering film and the probability of surface defect occurrence, and further affects the consistency and pass rate of the product.
[0003] The existing process mostly adopts fixed parameters or simple PID closed-loop adjustment, and cannot realize real-time sensing of the ribbon surface state and defect evolution. Only relying on the feedback of the processing sensor, it is difficult to respond to the slight fluctuations of temperature and pressure caused by material batch differences, environmental disturbances and other factors in the covering film process in a timely and accurate manner, resulting in deviations between the actual working conditions and the set values, and making it difficult to realize continuous and detailed abnormal detection of the whole covering film process, so that some small defects are difficult to be found and feedback adjusted in a timely manner. It lacks the ability of deep mining and self-learning of historical working condition data, and lacks the ability of analyzing and tracing the abnormal trend, reason and spatial distribution. SUMMARY
[0004] (I) Technical problems solved In view of the above-mentioned shortcomings of the prior art, the present application provides a chip ribbon covering film temperature and pressure collaborative conditioning method based on fuzzy control, which can effectively solve the problems of the prior art.
[0005] (II) Technical solutions In order to achieve the above-mentioned purposes, the present application is realized by the following technical solutions: The present application discloses a chip ribbon covering film temperature and pressure collaborative conditioning method based on fuzzy control, comprising the following steps: Step 1: multiple hot-pressing sensors are arranged at the covering film processing end, and the real-time temperature and pressure at the processing end are synchronously collected. At the same time, monitoring points are arranged at equal intervals along the displacement direction of the carrier tape, and infrared temperature measuring points and visual acquisition components are arranged to continuously collect the surface temperature of the carrier tape after covering film and the shape of the material outside; Step 2: collect the distribution characteristics of the ribbon hole to generate a dynamic heating area, and reduce the power of the heating unit corresponding to the non-covering film area of the covering film processing end to a preset safety threshold; Step 3: Plan an identification area on the surface of the carrier tape with a preset edge range as the collection range of temperature and morphology data, and establish a displacement coordinate system synchronized with the coding of the carrier tape; Step 4: During the displacement process of the coated carrier tape, the surface temperature and surface morphology profile of the planned identification area of the carrier tape are continuously obtained through a plurality of monitoring points, and the temperature decay gradient and morphology fluctuation variance are calculated; Step 5: Compare the temperature decay gradient and morphology fluctuation variance with the preset threshold value, and if any parameter exceeds the threshold value, trigger an abnormality flag; Step 6: Generate a temperature pressure compensation initial value through a fuzzy PID controller, input it into a pre-trained deep learning model, fuse historical working condition data analysis, and output the adjustment parameter of the temperature and pressure setting instruction of the current coating processing end; Step 7: After applying the adjustment parameter, monitor the verification data of the future several production cycles, if the abnormality still exists, expand the edge range of the identification area according to the abnormality amplitude, synchronously adjust the pitch angle and sampling frequency of the image acquisition device, and re-execute steps 3 to 6; Step 8: When the adjustment cycle reaches the preset upper limit and the abnormality still exists, generate an analysis report including the position coordinates and abnormality parameters and send it to the control center.
[0006] Further, the generation process of the dynamic heating area in step 2 is: Obtain the spatial coordinate data of the carrier tape hole distribution, identify the effective hole coordinate set corresponding to the coating processing area; Generate a dynamic heating area topology map according to the effective hole coordinate set, and divide the several heating units of the coating processing end into two control groups of coating area units and non-coating area units; Generate adjustment instructions for the non-coating area unit group, and reduce its heating power to a preset safe maintenance power.
[0007] Further, the division process of the identification area in step 3 is: Set a preset width of the edge buffer zone on the periphery of the material coating area on the surface of the carrier tape based on the dynamic heating area map; The material coating area and the edge buffer zone together define the identification area boundary; The length direction of the identification area is consistent with the displacement direction of the carrier tape, and the width direction covers the material width and the extended area on both sides; The range of the identification area needs to ensure that it completely contains the coating heat affected zone and the adjacent transition area; The established displacement coordinate system is synchronized with the carrier tape encoder in real time, so that the spatial position of the identification area dynamically matches the displacement of the carrier tape.
[0008] Further, the calculation process of the temperature attenuation gradient in step 4 is: for each monitoring point, the surface temperature difference between the current time and the previous sampling time is obtained, and the temperature change per unit time is obtained by dividing the sampling time interval; at the same sampling time, the surface temperature difference between adjacent monitoring points is calculated, and the temperature change per unit distance is obtained by dividing the actual physical distance between adjacent monitoring points; the temperature change per unit time and the temperature change per unit distance are superimposed according to the preset weight ratio, and multiplied by the material thermal diffusion characteristic compensation coefficient to generate the final temperature attenuation gradient value.
[0009] Further, the calculation process of the shape fluctuation variance in step 4 includes: During the tape displacement process, the visual acquisition component acquires the material surface profile image in the identification area at a preset sampling frequency; The edge enhancement processing is performed on each image, and the continuous boundary curve of the outer contour of the coated material is extracted; The boundary curve is equally divided into several segments along the width direction of the carrier tape, and the three-dimensional coordinate data of each segmented point is recorded; Taking the current monitoring point as the center, the coordinate data set of continuous multiple sampling times is taken, and the average value of the coordinates of each segmented point at the same time is taken as the reference plane. The vertical distance between the actual coordinates of each segmented point and the reference plane is calculated, and the variance value of the vertical distance sequence of continuous multiple times is calculated. The obtained variance value is taken as a quantitative index representing the shape fluctuation of the carrier tape surface.
[0010] Further, the continuous multiple sampling times cover at least 3 complete hot pressing periods, and when the distance deviation of a single segmented point exceeds the material thermal expansion limit value, it is considered as a noise point and filtered.
[0011] Further, in step 5, if any comparison result of the temperature attenuation gradient and the shape fluctuation variance with the preset threshold does not exceed the threshold, it is marked as normal, and jumps to step 4 for further operation, otherwise, if any comparison result exceeds the threshold, it is marked as abnormal, and jumps to step 5 for further operation, and generates an abnormal record backup to the cloud database.
[0012] Further, the working logic of the comprehensive identification model in step 6 is: Collect historical coating processing data to construct a training set, construct a deep neural network architecture, and use a spatial analysis branch to extract local distortion features of the coating surface shape using a dilated convolution layer. The time series analysis branch captures the dynamic joint relationship of temperature and pressure using a gated recurrent unit, encodes the fuzzy PID control rule library as network prior knowledge, and injects a fully connected decision layer through a residual connection; The temperature decay gradient and morphological fluctuation variance are collected in real time and input into the fuzzy PID controller, which generates primary instructions including temperature compensation base values and pressure compensation base values based on the rule base. The primary instructions and the current processing end condition data are input into the pre-trained comprehensive recognition model, and the spatiotemporal analysis branch analyzes the edge stripping index of the morphological image and the non-uniform distribution coefficient of the temperature field respectively. The decision-making layer integrates the output features of the time series analysis branch and the spatial analysis branch, and combines them with the comparison results of similar defect patterns in historical operating conditions to generate the optimized correction amount of the initial compensation value.
[0013] Furthermore, the process of expanding the edge range of the identified region in step 7 is as follows: ; In the formula, Represents the expanded edge width. Representing the The original preset edge width baseline value, Represents the comprehensive ratio coefficient of abnormal amplitude. Represents the temperature decay gradient calculated in real time. This represents a preset threshold for the temperature decay gradient. Represents the variance of morphological fluctuations calculated in real time. The variance of the representative morphological fluctuation is set to a preset threshold. and These represent the corresponding weight coefficients.
[0014] Furthermore, the analysis report generated in step 8 includes: the timestamp of the first trigger of the anomaly and the timestamp of the most recent trigger; the specific number or location identifier of the monitoring point that triggered the anomaly flag; the operating status parameters of the key equipment at the coating processing end during the period of anomaly; and the identification of the tape encoding segment corresponding to the location coordinates.
[0015] (III) Beneficial Effects Compared with the known prior art, the technical solution provided by this invention has the following beneficial effects: 1. By synchronously monitoring the temperature and pressure data of the coating end and the surface morphology of the carrier tape, a fuzzy PID controller is used to generate compensation initial values in real time. Then, the adjustment parameters are optimized by integrating historical working condition data through a deep learning model, realizing dynamic coordinated control of temperature and pressure. This effectively reduces the processing defect rate and significantly improves the chip packaging yield. The distribution characteristics of the taping hole are dynamically collected to generate a heating area map, accurately shutting down the heating unit in the non-coated area and reducing its power to a safe threshold, thereby reducing ineffective energy consumption.
[0016] 2. By adopting an adaptive adjustment mechanism for the identification area based on dynamic planning of the edge range, the problem of missed detection caused by uneven heat conduction in the edge area of the coating is solved. The monitoring boundary is expanded in real time through abnormal amplitude ratio analysis, and the pitch angle and sampling frequency of the image acquisition device are adjusted simultaneously to ensure that the identification area always covers the heat deformation sensitive area, thereby improving the edge defect detection rate and avoiding batch scrapping of materials due to minor undetected defects.
[0017] 3. Through a collaborative architecture of fuzzy control and deep learning models, fuzzy PID controllers rapidly respond to local abnormal fluctuations, while pre-trained deep learning models mine temperature and pressure correlation patterns from historical operating conditions to output high-precision adjustment parameters. This joint mechanism reduces the prediction error of adjustment parameters, and automatically triggers an alarm system when multiple adjustments fail to eliminate the abnormality, accurately locating the defect coordinates and causes, reducing manual troubleshooting time, and further improving the continuous and stable operation capability of the production line. Attached Figure Description
[0018] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the accompanying drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are merely some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without any creative effort.
[0019] Figure 1 This is a schematic diagram of the process of the present invention. Detailed Implementation
[0020] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of the present invention. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without creative effort are within the scope of protection of the present invention.
[0021] The present invention will be further described below with reference to embodiments.
[0022] This embodiment presents a method for coordinated adjustment of chip tape coating temperature and pressure based on fuzzy control, such as... Figure 1 As shown, it includes the following steps: Step 1: Install multiple thermo-pressure sensors at the coating processing end to simultaneously collect real-time temperature and pressure data. Simultaneously, set up monitoring points at equal intervals along the carrier belt displacement direction, and arrange infrared temperature measurement points and visual acquisition components to continuously collect data on the surface temperature of the carrier belt and the surface morphology of the material after coating. This achieves comprehensive real-time acquisition of temperature, pressure, and surface morphology at the processing end, quickly identifying localized overheating, stress concentration, and minor defects.
[0023] Step 2: Collect the distribution characteristics of the taping hole positions to generate a dynamic heating area, and reduce the power of the heating unit in the corresponding non-coated area at the coating processing end to a preset safety threshold; the process of generating the dynamic heating area is as follows: Obtain spatial coordinate data of the taping hole distribution and identify the effective hole coordinate set corresponding to the coating processing area; A dynamic heating area topology map is generated based on the effective hole position coordinate set, and several heating units at the coating processing end are divided into two control groups: coating area units and non-coating area units. An adjustment command is generated for the non-coated area unit group to reduce its heating power to a preset safe maintenance power. The system monitors the changes in hole distribution during the tape reel displacement process in real time. The encoder acquires the tape displacement in real time, and the hole distribution is rescanned every unit distance advanced. Heating is activated for newly added coated areas, and power is reduced for units moving out of the coated area. The heating area topology is dynamically updated, and the heating unit control group division is adjusted synchronously. The safety maintenance power does not exceed 20% of the base heating power. Combined with real-time encoder feedback, the heating topology is dynamically updated to ensure heating uniformity and consistency for different hole positions and different batches of products.
[0024] Step 3: Plan the recognition area on the carrier tape surface with a preset edge range as the acquisition range for temperature and morphological data, and establish a displacement coordinate system synchronized with the tape encoding; the process of dividing the recognition area is as follows: Based on the dynamic heating zone diagram, a preset width edge buffer zone is set around the material coating area on the carrier belt surface; The material-coated area and the edge buffer zone are jointly defined as the boundary of the identification area; The length of the identification area is consistent with the direction of the carrier belt displacement, and the width direction covers the width of the material and its extended areas on both sides. The scope of the identification area must be set to ensure that it completely includes the heat-affected zone of the film coating and the adjacent transition area; The established displacement coordinate system is synchronized with the tape encoder in real time, enabling the spatial position of the recognition area to dynamically match the displacement of the carrier tape. An edge buffer strip is used to cover the heat-affected zone and transition zone, avoiding blind spots in fixed small area monitoring; the displacement coordinate system synchronized with the encoder ensures that the recognition area is always aligned with the actual position of the carrier tape, improving detection and positioning accuracy.
[0025] Step 4: During the tape displacement process after lamination, the surface temperature and surface morphology profile of the planned identification area are continuously acquired through several monitoring points. The temperature decay gradient and morphological fluctuation variance are calculated. Continuous time-series data output helps to accurately calculate the temperature decay gradient and morphological fluctuation variance, providing high-confidence features for subsequent judgment.
[0026] Step 5: Compare the temperature decay gradient and morphological fluctuation variance with preset thresholds. If either parameter exceeds the limit, an anomaly flag is triggered. In Step 5, if neither the temperature decay gradient nor the morphological fluctuation variance exceeds the preset threshold, it is marked as normal, and the process proceeds to Step 4 for further execution. Conversely, if either comparison result exceeds the limit, it is marked as abnormal, and the process proceeds to Step 5 for further execution, generating an anomaly record and backing it up to the cloud database. The linkage between temperature and morphological indicators allows for the simultaneous capture of both temperature and morphological anomalies. Anomaly records are backed up in the cloud, enabling traceable management throughout the entire production process and providing a reliable basis for quality tracking and fault analysis.
[0027] Step 6: Generate initial values for temperature and pressure compensation using a fuzzy PID controller, input them into a pre-trained deep learning model, integrate historical operating data analysis, and output the adjustment parameters for the current temperature and pressure setting commands at the coating processing end; the overall recognition model's working logic is as follows: Historical coating processing data was collected to construct a training set, including time-series records of temperature and pressure at the processing end, temperature decay curves of the carrier surface captured by infrared thermometers, morphological change sequences collected by visual acquisition components, and final quality inspection results. These were integrated into the training set to construct a deep neural network architecture. The spatial analysis branch used a dilated convolutional layer to extract local distortion features of the coating surface morphology, while the temporal analysis branch used a gated recurrent unit to capture the dynamic joint relationship between temperature and pressure. The fuzzy PID control rule base was encoded into network prior knowledge and injected into the fully connected decision layer through residual connections. The spatial analysis branch set up three-level feature extraction channels: the first level identified pixel clustering regions of bubbles and wrinkles, the second level detected edge warping curvature, and the third level quantified the proportion of coating peeling area. The temperature decay gradient and morphological fluctuation variance are collected in real time and input into the fuzzy PID controller, which generates primary instructions including temperature compensation base values and pressure compensation base values based on the rule base. The primary instructions and the current processing end condition data are input into the pre-trained comprehensive recognition model, and the spatiotemporal analysis branch analyzes the edge stripping index of the morphological image and the non-uniform distribution coefficient of the temperature field respectively. The decision layer integrates the output features of the temporal and spatial analysis branches, and combines them with the comparison results of similar defect patterns in historical operating conditions to generate optimized correction values for the initial compensation. The decision layer sets up a dynamic weight allocator; when the detected peeling area ratio exceeds 5%, the weight of morphological features is automatically increased to 1.5 times the weight of temperature features. Fuzzy PID control is used to quickly respond to local deviations, and deep networks are used to mine the spatiotemporal features of historical data.
[0028] Step 7: After applying the adjusted parameters, monitor the verification data for several future production cycles. If the anomaly persists, analyze and expand the edge range of the recognition area according to the proportion of the anomaly magnitude. Simultaneously adjust the pitch angle and sampling frequency of the image acquisition device, and repeat steps 3 to 6. Dynamically expanding the monitoring range according to the anomaly magnitude and adjusting the device's viewing angle and sampling frequency further improves the sensitivity of secondary detection. Periodic verification and parameter feedback loops enable the system to have self-correction and self-learning capabilities, continuously and stably optimizing the production process.
[0029] Step 8: If an anomaly persists after the preset maximum number of cycles has been reached, the maximum number of cycles is dynamically set based on the tape length. An analysis report, including location coordinates and anomaly parameters, is generated and sent to the control center. The analysis report includes: the timestamps of the first and most recent anomaly triggers; the specific number or location identifier of the monitoring point that triggered the anomaly; the operating status parameters of key equipment at the lamination processing end during the anomaly's duration; and the tape encoding segment identifier corresponding to the location coordinates. A maximum number of cycles is set to prevent infinite loops. Once the threshold is exceeded, a detailed report containing time sequence, spatial information, and equipment status is automatically generated, triggering a real-time alert from the control center.
[0030] Compared with existing technologies, this technology saves energy and ensures processing stability by dynamically generating heating topology in real time and intelligently shutting down heating units in non-coated areas. It integrates fuzzy PID and deep neural networks to collaboratively optimize temperature and pressure regulation, improves anomaly detection sensitivity and defect detection rate, adaptively expands the recognition area and adjusts the acquisition strategy to ensure accurate coverage, and improves coating quality consistency and production reliability through multi-cycle closed-loop verification and cloud-based anomaly reporting mechanisms.
[0031] At other levels, this embodiment provides a calculation process for the temperature decay gradient, specifically: for each monitoring point, the surface temperature difference between the current time and the previous sampling time is obtained, and divided by the sampling time interval to obtain the temperature change per unit time; at the same sampling time, the surface temperature difference between adjacent monitoring points is calculated, and divided by the actual physical distance between adjacent monitoring points to obtain the temperature change per unit distance; the temperature change per unit time and the temperature change per unit distance are superimposed according to a preset weight ratio, and multiplied by a material thermal diffusivity compensation coefficient to generate the final temperature decay gradient value. By simultaneously considering the cooling rate after coating and the material's heat distribution state, it can effectively address the problems of material embrittlement caused by excessively rapid cooling after coating and coating warping caused by a sharp drop in temperature at the edge of the heated area. The compensation coefficient is set based on the thermal diffusivity of the tape substrate, solving the misjudgment caused by the temperature conduction difference of different materials under the same process.
[0032] The calculation process for morphological fluctuation variance includes: During the tape reel displacement process, the vision acquisition unit acquires the material surface contour image within the recognition area at a preset sampling frequency; Edge enhancement processing is performed on each frame of the image to extract the continuous boundary curve of the outer contour of the coated material; Divide the boundary curve into several equal segments along the width of the carrier tape, and record the three-dimensional coordinate data of each segment point; Centered on the current monitoring point, coordinate datasets from multiple consecutive sampling times are collected. The average coordinates of each segment point at the same time are used as the reference plane. The vertical distance between the actual coordinates of each segment point and the reference plane is calculated, and the variance of the vertical distance sequence at multiple consecutive times is obtained. The multiple consecutive sampling times cover at least three complete hot-pressing cycles. When the distance deviation of a single segment point exceeds the thermal expansion limit of the material, it is regarded as a noise point and filtered out. The obtained variance value is used as a quantitative index to characterize the surface morphology fluctuation of the carrier tape. By quantifying surface undulations through vertical distance, hidden defects missed by traditional methods can be identified, the anti-interference ability is enhanced, and the quantitative index drives parameter adjustment.
[0033] Compared with existing technologies, this method uses a weighted superposition of time and spatial gradients combined with material thermal diffusivity compensation to accurately define the temperature decay trend, effectively eliminate misjudgments caused by differences in heat transfer between different substrates, and avoid overcooling embrittlement and edge warping. The morphological fluctuation variance is calculated based on the variance of three-dimensional coordinate data at multiple times and dynamically filters out noise, quantifies surface undulations, captures hidden defects in one go, and enhances anti-interference capabilities. Compared with traditional single-dimensional or empirical parameter methods, this method improves the detection sensitivity, accuracy, and process stability of temperature and morphological anomalies.
[0034] This embodiment provides an extended calculation process for the edge range of the identification region, as shown below: ; In the formula, Represents the expanded edge width. Representing the The original preset edge width baseline value, This represents the comprehensive proportion coefficient of the abnormal amplitude. When the comprehensive proportion coefficient of the abnormal amplitude is >0.2, it is judged as a significant abnormality. Represents the temperature decay gradient calculated in real time. This represents a preset threshold for the temperature decay gradient. Represents the variance of morphological fluctuations calculated in real time. The variance of the representative morphological fluctuation is set to a preset threshold. and These represent the corresponding weight coefficients, satisfying... + =1 and ≥0.6.
[0035] In summary, this invention collects temperature and pressure data from the coating process end, as well as surface temperature and material morphology data of the carrier tape after coating; combines this with the current distribution characteristics of the taping holes to dynamically plan the switching of the heating area at the processing end, reducing the heating power in non-coated areas; and uses preset edge range data to plan the identification area of the taping surface, acquiring temperature and morphology data. In sequence, for designated identification areas on the braided surface, through several monitoring points, the surface temperature and surface morphology of the coated braided tape are continuously identified during the tape displacement process. The decay trend of the surface temperature and the fluctuation trend of the surface morphology of any braided tape during the displacement process are identified. The current temperature decay trend and morphology fluctuation trend are compared with a preset standard threshold to determine whether there is any abnormality. If an anomaly is detected, a comprehensive identification model is constructed by combining a fuzzy PID collaborative control strategy with a deep learning algorithm. The comprehensive identification model is trained using historical data, and the temperature decay trend and morphological fluctuation trend are input into the model for combined analysis. The output is the adjustment parameters for the temperature and pressure setting commands of the current processing end. If the adjusted parameters are successfully applied and pass the verification within a preset period, the adjustment is considered successful. Otherwise, if the verification data remains abnormal, the preset edge range data will be adjusted based on the extent of the abnormality, the recognition area of the tape surface will be replanned, and adjustment instructions for the angle and frequency settings of the acquisition device will be generated. The detection will be performed again. If the abnormality is still found, the adjustment parameters will be repeated. After the preset frequency of detection and adjustment cycle is reached, if the detection is still judged as abnormal, alarm data will be generated and submitted to the control center.
[0036] The above embodiments are only used to illustrate the technical solutions of the present invention, and are not intended to limit it. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions will not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.
Claims
1. A method for coordinated regulation of chip tape coating temperature and pressure based on fuzzy control, characterized in that, Includes the following steps: Step 1: Set up multiple hot-pressure sensors at the coating processing end to synchronously collect real-time temperature and pressure at the processing end. At the same time, set up monitoring points at equal intervals along the displacement direction of the carrier belt, arrange infrared temperature measurement points and vision acquisition components, and continuously collect the surface temperature of the carrier belt and the surface morphology of the material after coating. Step 2: Collect the distribution characteristics of the taping hole to generate a dynamic heating area, and reduce the power of the heating unit in the corresponding non-coated area at the coating processing end to a preset safety threshold; Step 3: Plan the recognition area on the carrier tape surface with a preset edge range as the collection range for temperature and morphology data, and establish a displacement coordinate system synchronized with the tape encoding; Step 4: During the tape displacement process after coating, the surface temperature and surface morphology profile of the planned identification area of the tape are continuously acquired through several monitoring points, and the temperature attenuation gradient and morphology fluctuation variance are calculated. Step 5: Compare the temperature decay gradient and morphological fluctuation variance with preset thresholds. If any parameter exceeds the limit, an anomaly flag is triggered. Step 6: Generate initial values for temperature and pressure compensation using a fuzzy PID controller, input them into a pre-trained deep learning model, integrate historical working condition data analysis, and output the adjustment parameters for the current temperature and pressure setting commands at the coating processing end; Step 7: After applying the adjustment parameters, monitor the verification data for several future production cycles. If the anomaly persists, analyze the edge range of the expanded recognition area according to the proportion of the anomaly amplitude, and simultaneously adjust the pitch angle and sampling frequency of the image acquisition device. Then, repeat steps 3 to 6. Step 8: If an anomaly still exists after the number of adjustment cycles reaches the preset upper limit, generate an analysis report including location coordinates and anomaly parameters and send it to the control center.
2. The method for coordinated adjustment of chip tape coating temperature and pressure based on fuzzy control according to claim 1, characterized in that, The process of generating the dynamic heating region in step 2 is as follows: Obtain spatial coordinate data of the taping hole distribution and identify the effective hole coordinate set corresponding to the coating processing area; A dynamic heating area topology map is generated based on the effective hole position coordinate set, and several heating units at the coating processing end are divided into two control groups: coating area units and non-coating area units. An adjustment command is generated for the non-coated area unit group to reduce its heating power to a preset safe maintenance power.
3. The method for coordinated regulation of chip tape coating temperature and pressure based on fuzzy control according to claim 1, characterized in that, The process of dividing the identification region in step 3 is as follows: Based on the dynamic heating zone diagram, a preset width edge buffer zone is set around the material coating area on the carrier belt surface; The material-coated area and the edge buffer zone are jointly defined as the boundary of the identification area; The length of the identification area is consistent with the direction of the carrier belt displacement, and the width direction covers the width of the material and its extended areas on both sides. The scope of the identification area must be set to ensure that it completely includes the heat-affected zone of the film coating and the adjacent transition area; The established displacement coordinate system is synchronized with the tape encoder in real time, so that the spatial position of the recognition area is dynamically matched with the displacement of the carrier tape.
4. The method for coordinated adjustment of chip tape coating temperature and pressure based on fuzzy control according to claim 1, characterized in that, The calculation process of the temperature decay gradient in step 4 is as follows: For each monitoring point, obtain the surface temperature difference between the current time and the previous sampling time, divide it by the sampling time interval to obtain the temperature change per unit time; at the same sampling time, calculate the surface temperature difference between adjacent monitoring points, divide it by the actual physical distance between adjacent monitoring points to obtain the temperature change per unit distance. The temperature change per unit time and the temperature change per unit distance are superimposed according to a preset weight ratio, and then multiplied by the material thermal diffusion characteristic compensation coefficient to generate the final temperature decay gradient value.
5. The method for coordinated adjustment of chip tape coating temperature and pressure based on fuzzy control according to claim 1, characterized in that, The calculation process for the morphological fluctuation variance in step 4 includes: During the tape reel displacement process, the vision acquisition unit acquires the material surface contour image within the recognition area at a preset sampling frequency; Edge enhancement processing is performed on each frame of the image to extract the continuous boundary curve of the outer contour of the coated material; Divide the boundary curve into several equal segments along the width of the carrier tape, and record the three-dimensional coordinate data of each segment point; Centered on the current monitoring point, take the coordinate dataset of multiple consecutive sampling times, use the average coordinate of each segment point at the same time as the reference plane, calculate the vertical distance between the actual coordinate of each segment point and the reference plane, and obtain the variance value of the vertical distance sequence of multiple consecutive times. The obtained variance value is used as a quantitative indicator to characterize the surface morphology fluctuation of the carrier tape.
6. The method for coordinated regulation of chip tape-and-reel coating temperature and pressure based on fuzzy control according to claim 5, characterized in that, The continuous sampling times cover at least 3 complete hot pressing cycles. When the distance deviation of a single segment point exceeds the thermal expansion limit of the material, it is regarded as a noise point and filtered out.
7. The method for coordinated adjustment of chip tape coating temperature and pressure based on fuzzy control according to claim 1, characterized in that, In step 5, if neither the temperature decay gradient nor the morphological fluctuation variance exceeds the preset threshold, it is marked as normal and the process proceeds to step 4 for further operation. Conversely, if either comparison result exceeds the threshold, it is marked as abnormal and the process proceeds to step 5 for further operation, and an abnormal record is generated and backed up to the cloud database.
8. The method for coordinated regulation of chip tape coating temperature and pressure based on fuzzy control according to claim 1, characterized in that, The working logic of the comprehensive recognition model in step 6 is as follows: Historical coating processing data is collected to build a training set, and a deep neural network architecture is constructed. The spatial analysis branch uses a dilated convolutional layer to extract local distortion features of the coating surface morphology, and the temporal analysis branch uses a gated recurrent unit to capture the dynamic joint relationship between temperature and pressure. The fuzzy PID control rule base is encoded into network prior knowledge and injected into the fully connected decision layer through residual connections. The temperature decay gradient and morphological fluctuation variance are collected in real time and input into the fuzzy PID controller, which generates primary instructions including temperature compensation base values and pressure compensation base values based on the rule base. The primary instructions and the current processing end condition data are input into the pre-trained comprehensive recognition model, and the spatiotemporal analysis branch analyzes the edge stripping index of the morphological image and the non-uniform distribution coefficient of the temperature field respectively. The decision-making layer integrates the output features of the time series analysis branch and the spatial analysis branch, and combines them with the comparison results of similar defect patterns in historical operating conditions to generate the optimized correction amount of the initial compensation value.
9. The method for coordinated regulation of chip tape coating temperature and pressure based on fuzzy control according to claim 1, characterized in that, The process of expanding the edge range of the identified region in step 7 is as follows: ； In the formula, Represents the expanded edge width. Representing the The original preset edge width baseline value, Represents the comprehensive ratio coefficient of abnormal amplitude. Represents the temperature decay gradient calculated in real time. This represents a preset threshold for the temperature decay gradient. Represents the variance of morphological fluctuations calculated in real time. The variance of the representative morphological fluctuation is set to a preset threshold. and These represent the corresponding weight coefficients.
10. The method for coordinated regulation of chip tape coating temperature and pressure based on fuzzy control according to claim 1, characterized in that, The analysis report generated in step 8 includes: the timestamp of the first trigger of the anomaly and the timestamp of the most recent trigger; the specific number or location identifier of the monitoring point that triggered the anomaly flag; the operating status parameters of the key equipment at the coating processing end during the period of anomaly; and the identification of the tape encoding segment corresponding to the location coordinates.