A method and system for adaptive control of a high viscosity film blowing machine air ring

Abstract

The application belongs to the technical field of data processing, and proposes a high-adhesion film blowing machine air ring adaptive control method and system, specifically: first, a thermal imaging monitoring module is arranged on the film blowing machine to collect real-time images, the real-time images are preprocessed and thermal stripe features are extracted, then the thermal stripe layering risk is dynamically calculated through the thermal stripe features, and finally the film blowing machine air ring is adaptively controlled according to the thermal stripe layering risk. Through accurate quantification of the local thermal energy retention leading to the potential insufficient cooling of the cold air duct, the imbalance of internal stress and the interface slip trend caused by the asynchronous cooling between the film layers, the central input of the closed-loop adaptive adjustment feedback is formed, so that the air ring cold air duct control becomes active intervention based on the thermal field evolution trend of the film bubble; the continuous identification system of the risk trend implements feedforward cooling intervention before the actual occurrence of the layering, thereby significantly improving the stability and accuracy of the film blowing process, and ensuring the consistency and yield of the multi-layer structure of the high-adhesion film.

Description

Technical Field

[0001] This invention belongs to the field of data processing technology, specifically relating to an adaptive control method and system for the air ring of a high-viscosity blown film machine. Background Technology

[0002] Blown film technology is a key process in the plastic film manufacturing industry, widely used in packaging, agriculture, building materials and other industries. Among them, multilayer co-extrusion blown film technology can combine different polymer materials in a layered manner to achieve multiple functions such as barrier, mechanical and heat sealing. High-viscosity polymers have the advantages of barrier and rigidity and are often placed in the middle layer of co-extruded films. However, these materials have the characteristics of high heat capacity and high viscosity, which leads to thermal inertia and flow lag. This is because, under the same cooling conditions, the heat release rate of the high-viscosity layer with high heat capacity is much slower than that of the low-viscosity layer. The resulting cooling delay will maintain a soft flow stage for a longer period of time. The velocity difference and stress difference between the high-viscosity layer and the surrounding shaped or semi-shaped low-viscosity layer will cause micro-interlayer misalignment, making the function or key parameters such as thickness of the co-extruded film unstable or uneven. When high-viscosity polymers are used as intermediate layers in high-speed blown film production, the effects are more pronounced, especially during the initial stabilization stage of the film bubble. Increased interlayer slippage and deformation accumulation lead to some intermediate layers being dragged by the earlier-shaped outer layers and the unshaped high-viscosity layers, resulting in tensile stretching. To address the issue of poor cooling of high-viscosity materials, existing technologies primarily employ high-power integrated air rings to enhance overall cooling capacity. This allows each layer to meet cooling requirements within a short time by providing cooling capabilities far exceeding actual needs. However, this results in significant energy consumption, typically exceeding that of blown film production using air rings that do not involve high-viscosity polymers by more than 50%. This extremely high energy consumption is often difficult to reduce because the thermal inertia response of high-viscosity layers is not linearly controllable. They must be allowed to solidify together with other film layers within a sufficiently short time. Even when the energy consumption of blown film production meets preset conditions, layer misalignment can still occur. This is because the single air outlet of the air ring causes a lag in the crucial air temperature regulation during the blown film process, a common drawback in dynamic temperature control. However, in high-power air rings, this deficiency in temperature control precision is amplified, resulting in some layer misalignment issues still present in the finished high-viscosity film. The dual-channel automatic air ring, developed to address this issue, enables dual-source film blowing. This method eliminates the need for high-power blowing techniques to complete high-viscosity film blowing. Traditional single-source blowing requires coordinated air pressure and temperature to simultaneously provide stretching and cooling effects, thus relying on high power to improve the later-stage setting speed of high-viscosity films. In contrast, the dual-channel automatic air ring delivers stretching and cooling to different air sources, allowing high-viscosity films to quickly complete bubble setting based on the physically matched temperature difference provided by the cooling air.However, even though dual-source blown film extrusion can balance the cooling efficiency of the bubble tube and the power consumption of the air ring, the layering phenomenon still exists in its application. This is because in the early stage of film extrusion, high-pressure shaping using hot air ducts initially forms the bubble tube. During this process, the air pressure is high but the shape is not cooled and solidified. Therefore, in the early stage of the film shape formation, such as the first fifth of the bubble tube, leopard-print-like striped hot areas can be identified from infrared thermography. These striped hot areas are caused by high-pressure hot air. In the early stage, the bubble tube only exists in the form of temperature difference. The principle behind its appearance is that the airflow disturbance of the hot air flow brings about significant nonlinear and uneven heat accumulation during the stretching process of the initial bubble tube. Moreover, the higher the density of the striped hot areas, the more difficult it is to eliminate the layering effect in the later cooling and forming process. Therefore, there is an urgent need for an adaptive control method for the air ring of a high-viscosity blown film extrusion machine. Summary of the Invention

[0003] The purpose of this invention is to propose an adaptive control method and system for the air ring of a high-viscosity blown film machine, so as to solve one or more technical problems existing in the prior art, and at least provide a beneficial option or create conditions.

[0004] To achieve the above objectives, according to one aspect of the present invention, an adaptive control method for the air ring of a high-viscosity blown film machine is provided, the method comprising the following steps:

[0005] S100, a thermal imaging monitoring module is installed on the blown film machine to collect real-time images;

[0006] S200 preprocesses real-time images and extracts thermal ripple features;

[0007] S300 dynamically calculates the risk of thermal pattern misalignment based on thermal pattern features;

[0008] S400, adaptively controls the blown film machine air ring based on the risk of thermal creasing and misalignment;

[0009] Further, in step S100, the method of arranging a thermal imaging monitoring module in the blown film machine and acquiring real-time images is as follows: an infrared thermal imaging monitoring module is arranged in the film bubble forming area of ​​the blown film machine. The monitoring module acquires thermal imaging images of the film bubble from the die head to the front rising area in real time and records them as real-time images. The thermal imaging images cover the initial blowing area of ​​the film bubble.

[0010] Furthermore, in step S200, the method for preprocessing the real-time image and extracting thermal features is as follows: the preprocessing of the real-time image includes grayscale processing and selection of the bubble region.

[0011] The process of selecting the membrane vesicle region is achieved through either a foreground extraction algorithm or an edge detection algorithm. The default is the Otsu threshold segmentation method in the foreground extraction algorithm, which is implemented by the OpenCV function cv2.findContours().

[0012] Further, in step S200, the method for preprocessing the real-time image and extracting thermal ripple features is as follows: the step of extracting thermal ripple features is to identify thermal ripple regions from the real-time image using the MSER algorithm or the Peak Detection algorithm;

[0013] The MSER algorithm for identifying hotspot regions from real-time images refers to the Maximum Stable Extreme Region method. It obtains pre-selected regions using the cv2.MSER_create() function of OpenCV, calculates the gray level of each pre-selected region, and identifies the pre-selected regions with gray levels higher than the overall average performance as hotspot regions. Alternatively, the Peak Detection method can be used. The processing flow is to improve the contrast between hotspots and the background by using the CLAHE local contrast enhancement operator cv2.createCLAHE(), and to extract the hotspot regions by using a sliding window with a side length of 5 and applying a threshold to the local peak regions.

[0014] The MSER algorithm has a higher accuracy in thermal pattern screening, but it suffers from insufficient coverage of thermal pattern areas, which affects the calculation of thermal pattern extensibility. The Peak Detection method, on the other hand, has a strong ability to identify thermal patterns that are often accompanied by drastic local temperature changes, but its accuracy in thermal pattern screening is relatively low because its algorithm relies on the significance of local changes, which can easily reduce its ability to identify low-level thermal pattern areas.

[0015] The center lines of the stripes in the heat ripple region are obtained by morphological skeletonization. The distance between any stripe center line and the center line of the stripe below it is the center line spacing. The weighted average of the center line spacings is calculated as the density trend value, with the distance between each stripe center line and the top of the image as the weight.

[0016] The stripe center line is calculated using the v2.ximgproc.thinning() function, which gives each thermal ripple region the corresponding Y or ordinate value, which is the stripe center line. The center line spacing is the distance between the stripe center lines of two thermal ripple regions, in pixels. The area below any stripe center line refers to the direction of reverse blown film. The top of the image refers to the side away from the mold head in the real-time image.

[0017] The maximum thickness among the various hot-strip regions is used as the stretch; the thickness of a hot-strip region is defined as the difference between the maximum and minimum values ​​of the vertical axis within a hot-strip region; the vertical axis is the coordinate value along the blown film direction.

[0018] The average gray value of pixels above the upper quartile in the thermal ripple region is used as the thermal overcoming value. The weighted average of the thermal overcoming levels of each thermal ripple region is calculated using the number of pixels in the thermal ripple region as the weight. The thermal ripple feature is composed of the density trend value, the stretching degree, and the thermal overcoming value.

[0019] Furthermore, in step S300, the method for dynamically calculating the risk of thermal pattern misalignment through thermal pattern features is as follows:

[0020] Set a time window, and use the time window that is reversed from time at any given moment as the sliding window interval;

[0021] To accurately reflect the changing characteristics of the operating status of the high-viscosity blown film machine, the dense trend value D at each moment within the sliding window interval is used. i and stretch S i The characteristic oscillation factor is calculated using the following method: OS t =|D t -mean{D i1}|+|S t -mean{S i2}|; where i1 and i2 are the index variables of the trend value and stretch of the sliding window interval, and mean{} is the average function, D t S t These represent the density trend value and extension at the current moment, respectively, used to dynamically capture the instantaneous fluctuations of system characteristics. This ensures that statistical analysis can be performed from the current state back to a certain period in the past, balancing the timeliness and stability of the data. Based on this characteristic oscillation factor, a thermal ripple misalignment risk model is constructed, and the thermal ripple misalignment risk R is calculated by weighting the difference between the oscillation factor and the historical peak value of the thermal overcoming value. t The specific expression is: R t =OS t ·(H max -H t ) / H max ;where H t H represents the thermal overload value at the current moment. max This represents the historical thermal overcoming value peak, which is the maximum value among all thermal overcoming values. The thermal overcoming value peak can be dynamically updated according to changes in the system's operating status.

[0022] The thermal ripple misalignment risk model sets a reverse-time sliding window to retrospectively analyze the density trend and extension of the membrane bubble within that time period, balancing data timeliness and historical stability, and effectively avoiding the disturbance interference of single-point anomalies on system judgment. A highly nonlinear characteristic oscillation factor is constructed by superimposing the absolute difference between the current density trend and extension and their historical averages, quantifying the transient fluctuations and dynamic instability of the membrane bubble's microscopic thermal structure and molecular chain stress field. The oscillation factor is weighted and superimposed using the normalized historical peak-to-valley ratio of thermal overcoming values ​​as a weighting adjustment factor, refining the dynamic deviation of the current thermal energy state from the historical extreme peak value to assess the magnitude of the thermal ripple misalignment risk. This risk model can reflect the dynamic changes of the internal thermal structure of the membrane bubble in real time, effectively capturing the uneven distribution of thermal energy and its corresponding misalignment defect risk, further providing a basis for adaptive control to offset the phenomenon that the cooling rate of the cold air duct cannot match the current operating conditions, thereby improving the product quality stability of high-viscosity blown film.

[0023] Because the process of obtaining thermal ripple misalignment risk based on characteristic oscillation factors involves a balance between dense trend values ​​and extension weights, it leads to a lack of a systematic global basic reference. This results in the inability to effectively identify and quantify some mildly occurring, slowly growing thermal ripple phenomena. However, existing technologies cannot solve the problem of missing global basic references. To better address this issue and eliminate the failure to identify slowly growing thermal ripple phenomena, this invention proposes a more preferred solution as follows:

[0024] Furthermore, in step S300, the method for dynamically calculating the risk of thermal ripple misalignment through thermal ripple features is as follows: Let the current time be t.

[0025] Define the dense trend value at the current moment as D. t Locate the dense trend peak DP corresponding to the first maximum value along the reverse time direction. t The thermal density perturbation factor D is obtained by calculating the difference rate between the dense trend and the dense trend peak. t Thermal density perturbation factor ΔD t Based on the complex and variable energy distribution of the internal thermal field, the thermal density perturbation factor, as the relative difference between the current density trend and the density trend peak DPt, expresses the non-uniform oscillation of energy distribution and local thermal field structural anomalies within the thermal ripple zone, thereby identifying potential outbreaks of misalignment risks. This indicator keenly captures abnormal fluctuations in the density of the thermal ripple structure, reflects the influence mechanism of microscopic thermal field changes on misalignment risks, and reveals the instantaneous imbalance of energy distribution within the thermal ripple.

[0026] In the degree of stretch S t Gradient calculation is introduced in the processing to quantify the spatial state evolution of thermal ripples. The gradient of the stretch at the current time and the previous time is calculated to obtain the structural strain response factor ΔS. tThe structural strain response factor, quantified by a gradient operator, reflects the difference between the current extension and the previous step in the evolution of the thermal morphology. It is used to express the spatiotemporal stretching and contraction dynamics of microscopic molecular chain orientation and stress field, achieving a combination of spatial and temporal perspectives. This gradient characterizes the extension or contraction trend of the thermal ridge structure, reflecting the dynamic adjustment of microscopic molecular chain orientation and stress transmission process, and providing a direct spatiotemporal characterization variable for the risk of interlayer misalignment in thermal ridges.

[0027] Thermal strength H t The processing employs an exponentially weighted moving average method for dynamic smoothing, calculating the heat feedback adjustment factor ΔH based on the current and previous heat recovery values, with the rate of change of the dense trend as the weighting value. t The thermal feedback adjustment factor simulates the intrinsic temporal evolution of molecular thermal motion and heat transfer through weighted filtering, achieving a stable and robust response to thermal intensity fluctuations and greatly suppressing the interference of instantaneous disturbances on the control algorithm. This processing conforms to the laws of energy reduction and state feedback in nature. By weighted fusion of recent situation and historical memory, it achieves a robust response to changes in thermal intensity and simulates the temporal evolution of molecular thermal motion and heat transfer characteristics in the thermal ripple region.

[0028] By incorporating the thermal density disturbance factor, structural strain response factor, and thermal feedback adjustment factor into the sub-thermal ripple misalignment risk model, the sub-thermal ripple misalignment risk is obtained.

[0029] A time interval is set as the backtesting segment RTRange. Within the backtesting segment, the median value of thermal overcoming is defined as the thermal overcoming level. If the thermal overcoming value at a certain moment is greater than or equal to the thermal overcoming level, then that moment is defined as the overcoming point. The predicted value of the sub-thermal ripple misalignment risk at the current moment is obtained through an autoregressive model. If the predicted value of the sub-thermal ripple misalignment risk at an overcoming point is less than the sub-thermal ripple misalignment risk, then the overcoming point is removed, i.e., it is no longer considered an overcoming point. The number of moments between the overcoming point and the current moment is the scale distance. The negative exponential function of the scale distance is used as the weight term to calculate the average value of the sub-thermal ripple misalignment risk corresponding to each overcoming point, which is recorded as the thermal ripple misalignment risk.

[0030] The predicted value of sub-thermal ripple fault risk at the current moment is obtained by using an autoregressive model. This process is achieved by directly obtaining the autoregressive model through the statsmodels.tsa.ar_model.AutoReg() function, and then using the obtained autoregressive model to obtain the corresponding predicted value at the current moment. The accuracy is better if the autoregressive model is finely adjusted.

[0031] If the autoregressive model is not tuned, the average value of the sub-thermal ripple misalignment risk is taken as the predicted value of the sub-thermal ripple misalignment risk.

[0032] Using a negative exponential function of the scale distance as a weight term means that the weight term is exp(-ntick), where exp() is an exponential function with base e and ntick represents the scale distance.

[0033] The sub-thermal ripple misalignment risk model deeply integrates thermal density disturbance, structural deformation, and thermal feedback adjustment factors through complex nonlinear operators, achieving high-precision coupling and quantitative analysis of microscopic thermal field dynamics and macroscopic processing state. It uses the median of dynamically calculated thermal overcoming values ​​as a dynamic benchmark to screen key overcoming points and eliminate abnormal error signals. Furthermore, it utilizes autoregressive time series modeling combined with a negative exponential weighting function to reduce the confidence level of historical sample timeliness, thereby improving the real-time performance of thermal ripple misalignment risk prediction. This model demonstrates the deep integration of thermodynamics and molecular dynamics theories in the field of intelligent adaptive control of high-viscosity blown film.

[0034] Beneficial Effects: By calculating the risk of thermal ripple misalignment, a quantitative mapping between the thermal anomaly characteristics of the membrane bubble surface and the potential misalignment evolution trend is achieved. This solves the technical problem of the inability to detect misalignment risks in advance and the delayed response during high-viscosity membrane blown film production. It accurately quantifies the precursor state of misalignment, where localized heat retention leads to insufficient cooling of the cold air duct. This risk value represents the internal stress imbalance and interface slippage trend caused by asynchronous cooling between membrane layers, and is an abstract expression of the physical mechanism of structural instability caused by the thermal inertia of high-viscosity materials. This constitutes the central input of the closed-loop adaptive adjustment feedback, transforming the control of the air ring cold air duct into an active intervention based on the evolution trend of the membrane bubble's thermal field. Through the continuous identification system of risk trends, feedforward cooling intervention can be implemented before misalignment actually occurs, thereby significantly improving the stability and accuracy of the blown film process and ensuring the consistency and yield of multilayer structures of high-viscosity membranes.

[0035] Further, in step S400, the method for adaptively controlling the blown film machine air ring based on the thermal crease misalignment risk is as follows: a time period is set as the reference window STT, and its value range is STT∈[5,60] minutes; the set of all thermal crease misalignment risks within the reference window is the reference set; if the thermal crease misalignment risk at the current moment is greater than the upper quartile of the reference set, it is defined as the incremental state; if the thermal crease misalignment risk at the current moment is less than the median of the reference set, it is defined as the slow-release state.

[0036] If the current time is in an incremental state, the power of the cooling duct in the air ring will be increased by 1%-10%, with a default value of 5%.

[0037] If the current state is a slow release state, reduce the power of the cooling duct in the air ring by 1%-10%, with a default value of 5%;

[0038] Preferably, all undefined variables in this invention, if not explicitly defined, can be manually set thresholds.

[0039] This invention also provides an adaptive control system for the airflow of a high-viscosity blown film machine. The adaptive control system includes a processor, a memory, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, it implements the steps in the adaptive control method for the airflow of the high-viscosity blown film machine. The adaptive control system can run on computing devices such as desktop computers, laptops, handheld computers, and cloud data centers. The runnable system may include, but is not limited to, processors, memory, and server clusters. The processor executes the computer program within the following system units:

[0040] A thermal imaging monitoring unit is used to deploy a thermal imaging monitoring module in a blown film machine and acquire real-time images.

[0041] The graphic feature extraction unit is used for real-time image preprocessing and extracting thermal pattern features;

[0042] The layered model building unit is used to dynamically calculate the layered risk of thermal ripples based on thermal ripple features.

[0043] A dynamic feedback unit is used to adaptively control the blown film machine's air ring based on the risk of thermal creasing and misalignment.

[0044] The beneficial effects of this invention are as follows: This invention provides an adaptive control method and system for the air ring of a high-viscosity blown film machine. By adaptively adjusting the power of the cold air duct of the dual-source air ring, the thermal ripples in the initial formation zone of the film bubble can quickly dissipate before the interface is set, effectively suppressing the interlayer misalignment behavior caused by thermal inertia hysteresis in the high-viscosity layer. Its main objective is to accurately quantify the potential misalignment precursor state caused by local heat retention leading to insufficient cooling of the cold air duct, and the internal stress imbalance and interface slippage trend caused by asynchronous cooling between film layers. This constitutes the central input of the closed-loop adaptive adjustment feedback, making the control of the air ring cold air duct an active intervention based on the evolution trend of the film bubble's thermal field. This significantly improves the stability and accuracy of the blown film process, ensuring the consistency and yield of the multilayer structure of the high-viscosity film. In macroscopic terms, it establishes a closed-loop adjustment mechanism between cooling intervention and thermal ripple evolution, making heat release and melt setting more synchronized, reducing the probability of the intermediate high-viscosity layer being dragged and deformed by the outer layer, and significantly improving the smoothness and thickness uniformity of the film interface. Especially in the high-viscosity film preparation environment of high-speed blown film, it can significantly reduce the probability of the formation of misaligned stress accumulation points, thereby improving the problems of local warping and functional degradation caused by the lag in the traditional blown film and thus ensuring the quality of film formation. Attached Figure Description

[0045] The above and other features of the present invention will become more apparent from the detailed description of the embodiments shown in conjunction with the accompanying drawings. In the accompanying drawings, the same reference numerals denote the same or similar elements. Obviously, the drawings described below are merely some embodiments of the present invention. For those skilled in the art, other drawings can be obtained from these drawings without any creative effort. In the drawings:

[0046] Figure 1 The diagram shows a flowchart of an adaptive control method for the air ring of a high-viscosity blown film machine.

[0047] Figure 2 The diagram shows the structure of an adaptive control system for the air ring of a high-viscosity blown film machine. Detailed Implementation

[0048] The following will provide a clear and complete description of the concept, specific structure, and technical effects of the present invention in conjunction with the embodiments and accompanying drawings, so as to fully understand the purpose, solution, and effects of the present invention. It should be noted that, unless otherwise specified, the embodiments and features described in this application can be combined with each other.

[0049] like Figure 1 The diagram shows a flowchart of an adaptive control method for the air ring of a high-viscosity blown film machine. The following section will discuss this method in conjunction with... Figure 1 This invention describes an adaptive control method for the air ring of a high-viscosity blown film machine according to an embodiment of the present invention. The method includes the following steps:

[0050] S100, a thermal imaging monitoring module is installed on the blown film machine to collect real-time images;

[0051] S200 preprocesses real-time images and extracts thermal ripple features;

[0052] S300 dynamically calculates the risk of thermal pattern misalignment based on thermal pattern features;

[0053] S400, adaptively controls the blown film machine air ring based on the risk of thermal creasing and misalignment;

[0054] Further, in step S100, the method of arranging a thermal imaging monitoring module in the blown film machine and acquiring real-time images is as follows: an infrared thermal imaging monitoring module is arranged in the film bubble forming area of ​​the blown film machine. The monitoring module acquires thermal imaging images of the film bubble from the die head to the front rising area in real time and records them as real-time images. The thermal imaging images cover the initial blowing area of ​​the film bubble.

[0055] The initial inflation region refers to the area from the initial inflation to a height of 1 / 5 of the total height of the membrane bubble; the frame rate of the infrared thermal imaging module is not less than 1 frame / second, and the image resolution is not less than 2000*2000, to ensure the continuity and traceability of the membrane bubble temperature field changes.

[0056] The infrared thermal imaging module is preferably installed on the side of the air ring. By arranging an infrared thermal imager on the side of the bubble, thermal images of the bubble surface are acquired laterally. The image processing system unfolds the thermal image of the arc-shaped bubble surface into a two-dimensional thermal distribution map, or covers the area from when the bubble is extruded from the die to about 1 / 5 of its rising height. Since the initial stage of the bubble is a high-temperature shaping stage and the melt layers have not yet solidified, the heat flow exhibits significantly high density and strong disturbance characteristics. Therefore, the thermal field structure in this region directly reflects the interlayer cooling coordination of the bubble.

[0057] The infrared thermal imaging module can be either an infrared imager or a thermal imager, with optional models being FLIR A35 or Optris PI640.

[0058] Furthermore, in step S200, the method for preprocessing the real-time image and extracting thermal features is as follows: the preprocessing of the real-time image includes grayscale processing and selection of the bubble region.

[0059] If the real-time image acquired by the infrared thermal imaging module is a pseudo-color image, grayscale processing is performed through the brightness channel. The preferred algorithm is the cv2.cvtColor() function in OpenCV. If the real-time image acquired by the infrared thermal imaging module is a temperature matrix image, the temperature values ​​are directly linearly mapped to the 0–255 range to achieve grayscale processing of the temperature data. This is a common method and will not be elaborated further.

[0060] The process of selecting the membrane vesicle region is achieved through either a foreground extraction algorithm or an edge detection algorithm. The default is the Otsu threshold segmentation method in the foreground extraction algorithm, which is implemented by the OpenCV function cv2.findContours().

[0061] Further, in step S200, the method for preprocessing the real-time image and extracting thermal ripple features is as follows: the step of extracting thermal ripple features is to identify thermal ripple regions from the real-time image using the MSER algorithm or the Peak Detection algorithm;

[0062] The MSER algorithm for identifying hotspot regions from real-time images refers to the Maximum Stable Extreme Region method. It obtains pre-selected regions using the cv2.MSER_create() function of OpenCV, calculates the gray level of each pre-selected region, and identifies the pre-selected regions with gray levels higher than the overall average performance as hotspot regions. Alternatively, the Peak Detection method can be used. The processing flow is to improve the contrast between hotspots and the background by using the CLAHE local contrast enhancement operator cv2.createCLAHE(), and to extract the hotspot regions by using a sliding window with a side length of 5 and applying a threshold to the local peak regions.

[0063] The center lines of the stripes in the heat ripple region are obtained by morphological skeletonization. The distance between any stripe center line and the center line of the stripe below it is the center line spacing. The weighted average of the center line spacings is calculated as the density trend value, with the distance between each stripe center line and the top of the image as the weight.

[0064] The stripe center line is calculated using the v2.ximgproc.thinning() function, which gives each thermal ripple region the corresponding Y or ordinate value, which is the stripe center line. The center line spacing is the distance between the stripe center lines of two thermal ripple regions, in pixels. The area below any stripe center line refers to the direction of reverse blown film. The top of the image refers to the side away from the mold head in the real-time image.

[0065] The maximum thickness among the various hot-strip regions is used as the stretch; the thickness of a hot-strip region is defined as the difference between the maximum and minimum values ​​of the vertical axis within a hot-strip region; the vertical axis is the coordinate value along the blown film direction.

[0066] The average gray value of pixels above the upper quartile in the thermal ripple region is used as the thermal overcoming value. The weighted average of the thermal overcoming levels of each thermal ripple region is calculated using the number of pixels in the thermal ripple region as the weight. The thermal ripple feature is composed of the density trend value, the stretching degree, and the thermal overcoming value.

[0067] Furthermore, in step S300, the method for dynamically calculating the risk of thermal pattern misalignment through thermal pattern features is as follows:

[0068] The time window is set as SRTrange, and its value range is the time length of SRTrange∈[5,10] minutes. The time window in the opposite direction of time at any moment is used as the sliding window interval.

[0069] To accurately reflect the changing characteristics of the operating status of the high-viscosity blown film machine, the dense trend value D at each moment within the sliding window interval is used. i and stretch S i The characteristic oscillation factor is calculated using the following method: OS t =|D t -mean{D i1}|+|S t -mean{S i2}|; where i1 and i2 are the index variables of the trend value and stretch of the sliding window interval, and mean{} is the average function, D t S tThese represent the density trend value and extension at the current moment, respectively, used to dynamically capture the instantaneous fluctuations of system characteristics. This ensures that statistical analysis can be performed from the current state back to a certain period in the past, balancing the timeliness and stability of the data. Based on this characteristic oscillation factor, a thermal ripple misalignment risk model is constructed, and the thermal ripple misalignment risk R is calculated by weighting the difference between the oscillation factor and the historical peak value of the thermal overcoming value. t The specific expression is: R t =OS t ·(H max -H t ) / H max ;where H t H represents the thermal overload value at the current moment. max This represents the historical thermal overcoming value peak, which is the maximum value among all thermal overcoming values. The thermal overcoming value peak can be dynamically updated according to changes in the system's operating status.

[0070] Furthermore, in step S300, the method for dynamically calculating the risk of thermal pattern misalignment through thermal pattern features is as follows: define the current time as t, and collect thermal pattern features at each time, including density trend value, extension degree and thermal overcoming value;

[0071] Define the dense trend value at the current moment as D. t Locate the dense trend peak DP corresponding to the first maximum value along the reverse time direction. t The thermal density perturbation factor D is obtained by calculating the difference rate between the dense trend and the dense trend peak. t Its mathematical expression is: ΔD t =|D t -DP t | / DP t ;

[0072] In the degree of stretch S t Gradient calculation is introduced in the processing to quantify the spatial state evolution of thermal ripples. The gradient of the stretch at the current time and the previous time is calculated to obtain the structural strain response factor ΔS. t Its mathematical expression is: ΔS t = S t - S t-1 S t Represents the degree of stretch.

[0073] Thermal strength H t The processing employs an exponentially weighted moving average method for dynamic smoothing, calculating the heat feedback adjustment factor ΔH based on the current and previous heat recovery values, with the rate of change of the dense trend as the weighting value. t The update rules are as follows: ΔH t =ΔD t · H t +(1-ΔD t)·ΔH t-1 H t-1 This represents the thermal feedback adjustment factor at time t-1.

[0074] By incorporating the thermal density perturbation factor, structural strain response factor, and thermal feedback adjustment factor into the sub-thermal ripple misalignment risk model, the sub-thermal ripple misalignment risk Sub.R is obtained. t Rt = (1 / sqrt(ΔSt)) * (ln(ΔHt + 1) / (ΔDt)^2); where ln() is the logarithmic function of the natural constant e as its base.

[0075] Set a time interval as the backtesting segment RTRange, whose value range is RTRange∈[10,30] minutes;

[0076] Within the backtesting segment, the median value of thermal overcoming is defined as the thermal overcoming level. If the thermal overcoming value at a certain moment is greater than or equal to the thermal overcoming level, then that moment is defined as the overcoming point. The predicted value of the sub-thermal ripple misalignment risk at the current moment is obtained through an autoregressive model. If the predicted value of the sub-thermal ripple misalignment risk at an overcoming point is less than the sub-thermal ripple misalignment risk, then the overcoming point is removed, i.e., it is no longer considered an overcoming point. The number of moments between the overcoming point and the current moment is the scale distance. The negative exponential function of the scale distance is used as the weight term to calculate the average value of the sub-thermal ripple misalignment risk corresponding to each overcoming point, which is recorded as the thermal ripple misalignment risk.

[0077] The predicted value of sub-thermal ripple fault risk at the current moment is obtained by using an autoregressive model. This process is achieved by directly obtaining the autoregressive model through the statsmodels.tsa.ar_model.AutoReg() function, and then using the obtained autoregressive model to obtain the corresponding predicted value at the current moment. The accuracy is better if the autoregressive model is finely adjusted.

[0078] If the autoregressive model is not tuned, the average value of the sub-thermal ripple misalignment risk is taken as the predicted value of the sub-thermal ripple misalignment risk.

[0079] Using a negative exponential function of the scale distance as a weight term means that the weight term is exp(-ntick), where exp() is an exponential function with base e and ntick represents the scale distance.

[0080] Further, in step S400, the method for adaptively controlling the blown film machine air ring based on the thermal crease misalignment risk is as follows: a time period is set as the reference window STT, and its value range is STT∈[5,60] minutes; the set of all thermal crease misalignment risks within the reference window is the reference set; if the thermal crease misalignment risk at the current moment is greater than the upper quartile of the reference set, it is defined as the incremental state; if the thermal crease misalignment risk at the current moment is less than the median of the reference set, it is defined as the slow-release state.

[0081] The blown film machine features a dual-source air ring, comprising a hot air duct and a cold air duct. The hot air duct provides high-temperature, high-pressure airflow during the initial stage of bubble formation to support the bubble structure and assist in its expansion, maintaining stability. The cold air duct rapidly cools the bubble after initial formation, ensuring quick film set and preventing flow misalignment at high temperatures. The hot air duct is located in the inner ring, while the cold air duct is located in the outer ring or arranged in separate zones. The cold air duct has the ability to independently adjust air pressure and temperature. The cold air duct primarily operates during the stage when the bubble rises to a stable height and the film transitions from a hot to a solid state.

[0082] If the current time is in an incremental state, the power of the cooling duct in the air ring will be increased by 1%-10%, with a default value of 5%.

[0083] If the current state is a slow release state, reduce the power of the cooling duct in the air ring by 1%-10%, with a default value of 5%;

[0084] The set of all thermal crevice misalignment risks within the assessment window is a reference set, excluding the thermal crevice misalignment risks acquired at the current moment. Increasing or decreasing the power of the cold air duct in the air ring refers to using the default power of the cold air duct in the air ring as the base value for increasing or decreasing the percentage; the maximum increase in the power of the cold air duct can be 200%, and the minimum decrease can be 90%; the upper limit of 200% is considered to be the structural withstand limit, exceeding this value will generate excessive wind pressure and disturbance on the membrane bubble surface, which can easily cause membrane bubble deformation or rupture. The lower limit of 90% is set as a safe cooling threshold: below this value, it may lead to insufficient cooling, and the high viscosity layer may remain in a hot state for a long time, thus failing to achieve synchronous structural stabilization and forming a risk of drag-and-pull-out misalignment.

[0085] In the dynamic adjustment process of this invention, the power of the cooling duct in the air ring is adjusted based on its default operating power, and is increased or decreased by a percentage. The default power is generally a stable power that can meet ordinary cooling requirements according to the process settings. The adjustment percentage range is set to a fine adjustment step of 1% to 10% to ensure sensitive response and smooth control, meet the flexibility of adaptive control, and avoid causing sudden changes in airflow that could lead to changes in the membrane bubble structure.

[0086] An embodiment of the present invention provides an adaptive control system for the airflow of a high-viscosity blown film machine, such as... Figure 2 The diagram shown is a structural diagram of an adaptive airflow control system for a high-viscosity blown film machine according to the present invention. This embodiment of the adaptive airflow control system for a high-viscosity blown film machine includes: a processor, a memory, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, it implements the steps in the above embodiment of the adaptive airflow control method for a high-viscosity blown film machine.

[0087] The system includes: a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program in a unit of the following system:

[0088] A thermal imaging monitoring unit is used to deploy a thermal imaging monitoring module in a blown film machine and acquire real-time images.

[0089] The graphic feature extraction unit is used for real-time image preprocessing and extracting thermal pattern features;

[0090] The layered model building unit is used to dynamically calculate the layered risk of thermal ripples based on thermal ripple features.

[0091] A dynamic feedback unit is used to adaptively control the blown film machine's air ring based on the risk of thermal creasing and misalignment.

[0092] The aforementioned adaptive airflow control system for a high-viscosity blown film extrusion machine can run on computing devices such as desktop computers, laptops, handheld computers, and cloud servers. The system that can run on this high-viscosity blown film extrusion machine adaptive airflow control system may include, but is not limited to, processors and memory. Those skilled in the art will understand that the above example is merely an illustration of an adaptive airflow control system for a high-viscosity blown film extrusion machine and does not constitute a limitation on such a system. It may include more or fewer components, or a combination of certain components, or different components. For example, the aforementioned adaptive airflow control system for a high-viscosity blown film extrusion machine may also include input / output devices, network access devices, buses, etc.

[0093] The processor referred to can be a Central Processing Unit (CPU), or other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. The general-purpose processor can be a microprocessor or any conventional processor. The processor is the control center of the high-viscosity blown film machine air-ring adaptive control system, connecting various parts of the system through various interfaces and lines.

[0094] The memory can be used to store the computer program and / or modules. The processor implements various functions of the high-viscosity blown film machine air ring adaptive control system by running or executing the computer program and / or modules stored in the memory and calling the data stored in the memory. The memory may mainly include a program storage area and a data storage area. The program storage area may store the operating system, at least one application program required for a function (such as sound playback function, image playback function, etc.), etc.; the data storage area may store data created according to the use of the mobile phone (such as audio data, phone book, etc.). In addition, the memory may include high-speed random access memory, and may also include non-volatile memory, such as hard disk, memory, plug-in hard disk, smart media card (SMC), secure digital (SD) card, flash card, at least one disk storage device, flash memory device, or other volatile solid-state storage device.

[0095] Although the invention has been described in considerable detail and particularly with regard to several of the described embodiments, it is not intended to limit itself to any of these details or embodiments or any particular embodiment, thereby effectively covering the intended scope of the invention. Furthermore, the invention has been described above with respect to embodiments foreseeable by the inventors in order to provide a useful description, and non-substantial modifications to the invention that have not yet been foreseen may still represent equivalent modifications.

Claims

1. A method for adaptive control of the air ring in a high-viscosity blown film machine, characterized in that, The method includes the following steps: S100, a thermal imaging monitoring module is installed on the blown film machine to collect real-time images; S200 preprocesses real-time images and extracts thermal ripple features; S300 dynamically calculates the risk of thermal pattern misalignment based on thermal pattern features; S400, adaptively controls the blown film machine air ring based on the risk of thermal creasing and misalignment; In step S200, the method for preprocessing the real-time image and extracting thermal ripple features is as follows: thermal ripple regions are identified from the real-time image using the MSER algorithm or Peak Detection algorithm; the center lines of the stripes in the thermal ripple regions are obtained through morphological skeletonization processing, and the distance between any stripe center line and the center line of the stripe below it is the center line spacing. The weighted average of the center line spacings is calculated as the density trend value, using the distance between each stripe center line and the top of the image as the weight; the maximum value of the thickness of each thermal ripple region is used as the extensibility; the average gray value of each pixel in the thermal ripple region that is greater than the upper quartile is used as the thermal overcoming value, and the weighted average of the thermal overcoming level corresponding to each thermal ripple region is calculated as the thermal overcoming value, using the number of pixels in the thermal ripple region as the weight; the thermal ripple features are composed of the density trend value, extensibility, and thermal overcoming value. In step S300, the method for dynamically calculating the risk of thermal pattern misalignment based on thermal pattern features is as follows: a time window is set, with the time window in the reverse direction at any given moment serving as the sliding window interval; the current characteristic oscillation factor OS is calculated based on the density trend value and extension at each moment within the sliding window interval. t The risk of thermal ripple misalignment (R) is calculated by weighting the historical peak differences between the characteristic oscillation factor and the thermal overcoming value. t R t =OS t ·(H max -H t ) / H max ;where H t H represents the thermal overload value at the current moment. max Indicates the historical thermal overload peak value; In step S400, the method for adaptively controlling the blown film machine air ring based on the thermal crease misalignment risk is as follows: a time period is set as the reference window STT, and its value range is STT∈[5,60] minutes; the set of all thermal crease misalignment risks within the reference window is the reference set; if the thermal crease misalignment risk at the current moment is greater than the upper quartile of the reference set, it is defined as the incremental state; if the thermal crease misalignment risk at the current moment is less than the median of the reference set, it is defined as the slow-release state. If the current moment is in an incremental state, increase the power of the cooling duct in the air ring by 1%-10%; If the current state is a slow release state, reduce the power of the cooling duct in the air ring by 1%-10%.

2. The adaptive control method for the air ring of a high-viscosity blown film machine according to claim 1, characterized in that, In step S100, the method of arranging a thermal imaging monitoring module and acquiring real-time images in the blown film machine is as follows: an infrared thermal imaging monitoring module is arranged in the film bubble forming area of ​​the blown film machine. The monitoring module acquires thermal imaging images of the film bubble from the die head to the front rising area in real time and records them as real-time images. The real-time images cover the initial blowing area of ​​the film bubble.

3. The adaptive control method for the air ring of a high-viscosity blown film machine according to claim 1, characterized in that, In step S200, the method for preprocessing the real-time image and extracting thermal ripple features is as follows: the preprocessing of the real-time image includes grayscale processing and selection of the bubble region.

4. The adaptive control method for the air ring of a high-viscosity blown film machine according to claim 1, characterized in that, Furthermore, in step S300, the method for dynamically calculating the risk of thermal pattern misalignment through thermal pattern features is as follows: define the current time as t, and define the dense trend value at the current time as D. t Locate the dense trend peak DP corresponding to the first maximum value along the reverse time direction. t The thermal density perturbation factor D is obtained by calculating the difference rate between the dense trend and the dense trend peak. t Its mathematical expression is: ΔD t =|D t -DP t | / DP t The structural strain response factor ΔS is obtained by calculating the gradient of the stretch between the current and previous time steps. t Its mathematical expression is: ΔS t = S t - S t-1 S t Represents stretchability; Based on the current and previous thermal overcoming values, and using the dense trend change rate as the weighting value, the thermal feedback adjustment factor ΔH is calculated. t The update rules are as follows: ΔH t =ΔD t · H t +(1-ΔD t )·ΔH t-1 H t-1 The heat feedback adjustment factor represents time t-1. By incorporating the thermal density perturbation factor, structural strain response factor, and thermal feedback adjustment factor into the sub-thermal ripple misalignment risk model, the sub-thermal ripple misalignment risk Sub.R is obtained. t Rt = (1 / sqrt(ΔSt)) (ln(ΔHt + 1) / (ΔDt)^2); A time interval is set as the backtesting range RTRange. Within the backtesting range, the median value of thermal overcoming is defined as the thermal overcoming level. If the thermal overcoming value at a certain moment is greater than or equal to the thermal overcoming level, then that moment is defined as the overcoming point. The predicted value of the sub-thermal ripple misalignment risk at the current moment is obtained through an autoregressive model. If the predicted value of the sub-thermal ripple misalignment risk at an overcoming point is less than the sub-thermal ripple misalignment risk, then the overcoming point is removed, i.e., it is no longer considered an overcoming point. The number of moments between the overcoming point and the current moment is the scale distance. The negative exponential function of the scale distance is used as the weight term to calculate the average value of the sub-thermal ripple misalignment risk corresponding to each overcoming point, which is recorded as the thermal ripple misalignment risk.

5. An adaptive control system for the air ring of a high-viscosity blown film machine, characterized in that, The high-viscosity blown film machine air ring adaptive control system includes: a processor, a memory, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, it implements the steps of the high-viscosity blown film machine air ring adaptive control method according to any one of claims 1-4. The high-viscosity blown film machine air ring adaptive control system operates on computing devices such as desktop computers, laptops, handheld computers, and cloud data centers.

Images