A method and system for intelligent adjustable medical cotton ball forming

Abstract

This invention relates to the field of intelligent molding technology, specifically to an intelligent adjustable medical cotton ball molding method and system, comprising the following steps: collecting pressure data through a distributed pressure sensor array and generating a pressure feature vector through Gaussian filtering for noise reduction; analyzing the pressure change trend based on the feature vector and outputting a gradient adjustment command; controlling a valve to adjust the air pressure to obtain compensation parameters; driving a servo motor to adjust the mold gap and collecting offset data to generate a feedback signal; and combining the feedback and adjustment command inputs to a neural network to evaluate the molding result. In this invention, data is collected through a distributed sensor array, multi-dimensional noise reduction enhances feature extraction, dynamic functions analyze pressure trends, offset adjustment parameters are identified, collaborative control achieves closed-loop compensation, displacement sensors provide feedback on gap changes, parameter modeling optimizes accuracy and consistency, convolutional networks extract temporal features, and fusion analysis constructs an adaptive system to achieve intelligent prediction and optimization of molding quality.

Description

Technical Field

[0001] This invention relates to the field of intelligent molding technology, and in particular to an intelligent adjustable medical cotton ball molding method and system. Background Technology

[0002] The field of intelligent molding technology integrates medical material processing equipment with automated control systems. Its core lies in achieving precise control of material morphology through real-time feedback of sensor data and dynamic adjustment of actuators. In the medical dressing production process, it mainly involves three key technologies: cotton fiber orientation control, lamination structure molding precision adjustment, and porous structure density gradient control. The technological development in this field shows an evolutionary path from mechanical mold shaping to intelligent parameter adaptive adjustment. At present, the technology system covers three basic components: pressure gradient control module, temperature and humidity collaborative control module, and visual inspection feedback module.

[0003] The intelligent adjustable medical cotton ball forming method refers to optimizing and controlling the forming process of medical cotton balls through intelligent adjustment. Addressing the shortcomings of existing cotton ball forming methods, this method proposes a way to precisely control the size, density, and shape of cotton balls using intelligent adjustment equipment. Specifically, this method utilizes intelligent sensors and a control system to monitor the forming status of the cotton balls in real time, automatically adjusting key parameters such as forming pressure and temperature to ensure that the quality of each cotton ball meets requirements. This method also incorporates advanced automated forming equipment, achieving standardized production of cotton balls through precisely controlled robotic arms and forming molds.

[0004] Existing technologies rely on manual experience to set fixed molding parameters, lacking a real-time monitoring and feedback mechanism for dynamic changes in pressure distribution. This results in the inability to promptly correct quality deviations caused by material property fluctuations or environmental interference during the molding process. Traditional pressure sensors, using a single-point detection mode, struggle to comprehensively reflect the spatial distribution characteristics of the pressure field within the mold cavity, easily leading to localized pressure overload or uneven molding density. Mechanical adjustment systems often employ open-loop control strategies, with a lack of coordination between valve opening and mold gap adjustment. The coupling effect of air pressure fluctuations and mechanical deformation can easily cause dimensional deviations in the molded product. Quality assessment relies on offline sampling and manual visual inspection, failing to capture transient abnormal signals during the molding process, increasing the risk of defective products entering subsequent stages. For example, if a micron-level shift in the mold gap is not detected in time during the molding of medical cotton balls, the cotton balls will not achieve the required fluffiness, affecting the hemostatic effect. Existing technologies, lacking a high-precision closed-loop feedback system, struggle to avoid such problems. Summary of the Invention

[0005] To address the shortcomings of existing technologies that rely on manual experience to set fixed molding parameters and lack real-time monitoring and feedback mechanisms for dynamic changes in pressure distribution, resulting in the inability to promptly correct quality deviations caused by material property fluctuations or environmental interference during the molding process, traditional pressure sensors, employing single-point detection, struggle to comprehensively reflect the spatial distribution characteristics of the pressure field within the mold cavity, easily leading to localized pressure overload or uneven molding density. Mechanical adjustment systems often employ open-loop control strategies, lacking coordination between valve opening and mold gap adjustment; the coupling effect of air pressure fluctuations and mechanical deformation can easily cause dimensional deviations in the molded product. Quality assessment relies on offline sampling and manual visual inspection, failing to capture transient abnormal signals during the molding process, increasing the risk of defective products entering subsequent stages. For example, if micron-level deviations in the mold gap are not detected in time during the molding of medical cotton balls, the cotton balls will not achieve the required fluffiness, affecting the hemostatic effect. Due to the lack of a high-precision closed-loop feedback system in existing technologies, this invention provides an intelligent adjustable medical cotton ball molding method and system. The technical solution is as follows:

[0006] On the one hand, a method for intelligently adjustable medical cotton ball forming is provided, the method comprising:

[0007] S1: Collect regional pressure data through a distributed pressure sensor array, input the collected raw pressure data into a Gaussian filtering algorithm that dynamically adjusts kernel parameters based on noise level, and perform multi-dimensional spatial denoising to generate regional pressure feature vectors.

[0008] S2: Based on the regional pressure feature vector, call the dynamic gradient determination function to analyze the directionality of the pressure change trend in the region. When the trend change satisfies the continuous direction offset condition, output the dynamic gradient adjustment command.

[0009] S3: According to the dynamic gradient adjustment command, control the valve opening of the compressed gas channel in the target area, adjust the gas pressure in the mold cavity, and modify the current valve opening parameter by adjusting the amplitude to obtain the molding pressure compensation parameter;

[0010] S4: Based on the molding pressure compensation parameters, drive the servo motor to adjust the mold gap, collect the adjusted gap data in real time through the displacement sensor, calculate the offset with the reference gap value, and obtain the molding state feedback signal by combining the offset direction and the rate of change.

[0011] S5: Call the forming state feedback signal and dynamic gradient adjustment command, input them into the trained convolutional neural network model, perform temporal feature extraction and analysis, and output the forming evaluation result.

[0012] As a further aspect of the present invention, the dynamic gradient determination function selects a time window based on the noise level, and the continuous directional offset condition is that the gradient direction continuously changes within no less than three consecutive time steps, and the offset angle exceeds a set threshold. The set threshold is optimized on historical data samples using the gradient descent method.

[0013] The offset direction and rate of change are respectively standardized using Z-score to unify the dimensions.

[0014] The convolutional neural network model includes two convolutional extraction layers, one LSTM temporal processing layer, and two fully connected layers. The training is based on a batch of original, pre-formed data samples from the experimental setup.

[0015] The regional pressure feature vector includes a pressure distribution matrix, contact area coordinates, and effective pressure value. The dynamic gradient adjustment command specifically includes pressure gradient direction angle, cumulative displacement, and critical rate threshold. The molding pressure compensation parameters include valve opening linear displacement, cavity pressure gradient coefficient, and pulse width modulation parameters. The molding state feedback signal specifically refers to gap displacement deviation, real-time displacement rate, and deformation convergence rate. The molding result evaluation includes porosity distribution map, density standard deviation, and stress concentration factor distribution.

[0016] As a further aspect of the present invention, the specific steps of S1 include:

[0017] S101: Collect regional pressure data through a distributed pressure sensor array, collect pressure data at multiple points within the region according to a preset fixed period, record the pressure value and timestamp of the sensor node, and summarize the sampled values ​​to form a data set with a unified structure, thus obtaining the original pressure data set;

[0018] S102: Based on the original pressure data set, call the Gaussian filtering algorithm to perform spatial domain denoising, identify the Gaussian weight distribution of each data point and its neighboring points, calculate the pressure response value and adjust the node pressure value to obtain denoised pressure distribution data;

[0019] The kernel size and standard deviation of the Gaussian filtering algorithm are configured based on the sensor noise level and the target resolution;

[0020] S103: Based on the denoised pressure distribution data, extract the spatial distribution features of the nodes, analyze the high-dimensional feature space under the positional relationship between the nodes, calculate the coordinate values ​​of the nodes in the feature space and unify them into a standard vector form, and integrate the pressure feature vector of the node output area.

[0021] As a further aspect of the present invention, the specific steps of S2 include:

[0022] S201: Based on the regional pressure feature vector, extract the pressure values ​​at the time nodes, construct the pressure change matrix, call the dynamic gradient determination function to calculate the directional offset of adjacent units, and generate a sequence of directional change indexes.

[0023] S202: Based on the directional change index sequence, extract continuous directional change segments, filter segments that meet the angle change and length thresholds, calculate the segment directional switching frequency and angle offset, and obtain the directional offset trend parameter set.

[0024] S203: Call the paragraph change parameters in the directional offset trend parameter set, compare them with the corresponding time period directional codes in the regional pressure feature vector, record the start and end times of the paragraphs that meet the continuous directional offset conditions and the regional numbers, and obtain the dynamic gradient adjustment command.

[0025] As a further aspect of the present invention, the specific steps of S3 include:

[0026] S301: Based on the dynamic gradient adjustment command, call the current opening parameter of the compressed gas channel in the target area, and adjust the opening parameter within the set range according to the adjustment amplitude to generate the valve adjustment opening value;

[0027] S302: Adjust the opening value of the valve according to the valve, control the real-time opening and closing state of the compressed gas channel valve in the mold cavity, collect the gas channel flow rate data after opening and closing and calculate the interval difference between it and the initial value, and generate the flow rate change value.

[0028] S303: Based on the change value of the flow rate, and combined with the mold cavity volume data, normalization processing is performed to calculate the amplification value corresponding to the change rate of the flow rate per unit volume and the set pressure adjustment coefficient, so as to obtain the molding pressure compensation parameter.

[0029] As a further aspect of the present invention, the change in flow rate is calculated using the following formula:

[0030]

[0031] Where ΔQ represents the change in flow rate, in units of L / min, Q t1 This represents the throughput data value at time t1, in L / min, Q. t0 The flow rate at initial time t0 is represented by V, in L / min; V represents the volume of the mold cavity, in L; λ j γ represents the real-time flow rate deviation of the compressed gas channel during the j-th cycle, in L / min. j η represents the instantaneous opening and closing amplitude of the valve within the j-th cycle, expressed as a percentage; k represents the total number of sampling cycles; and η represents the gas compression response sensitivity coefficient, a dimensionless parameter.

[0032] As a further aspect of the present invention, the specific steps of S4 include:

[0033] S401: Based on the molding pressure compensation parameters, drive the servo motor to adjust the mold gap, extract the difference between the control displacement and the target command, perform synchronous correction operation and record the response rate and feedback delay, and generate the mold gap adjustment value;

[0034] S402: Call the mold gap adjustment value, collect mold gap change data through displacement sensor, calculate the offset of mold gap from reference gap, and combine the data acquisition time points to perform time sequence organization to obtain the mold gap offset change.

[0035] S403: Based on the change in mold gap offset, extract the direction and rate of change of gap, construct a trend offset map by combining time series data, perform key trend identification and state determination operations of the map, and generate a molding state feedback signal.

[0036] As a further aspect of the present invention, the specific steps of S5 include:

[0037] S501: Based on the forming state feedback signal and the dynamic gradient adjustment command, they are synchronously input into the trained convolutional neural network model. The feedback signal is spatially expanded through the convolutional layer to extract the grayscale difference between adjacent pixels. Mean filtering is performed by region to calculate the trend of pixel difference value change in region and generate signal gradient change rate.

[0038] S502: Call the signal gradient change rate, reconstruct the gradient change sequence in the time axis direction according to the time sequence recognition module in the convolutional neural network model, superimpose the node weight ratio in the dynamic gradient adjustment instruction, calculate the horizontal offset of the node, and generate the node gradient offset.

[0039] S503: Call the node gradient offset, compare the node offset result with the gradient tolerance limit value according to the evaluation function parameters at the model output, determine whether the offset magnitude deviates, extract the segment deviation ratio to fit the overall trend, and obtain the forming evaluation result.

[0040] As a further aspect of the present invention, the gradient change rate of the molding signal is calculated using the following formula:

[0041]

[0042] Where ΔG represents the rate of change of the signal gradient, P i X represents the grayscale value of the i-th pixel. i Y represents the x-coordinate of the i-th pixel. i w represents the ordinate value of the i-th pixel. iThe weight factor representing the i-th pixel, the adaptive weight factor w i The feature parameters learned by the model are dynamically set and adjusted according to the importance of different pixels. n represents the number of pixels in the image region and i is the pixel index.

[0043] On the other hand, an intelligent adjustable medical cotton ball forming system is provided, the intelligent adjustable medical cotton ball forming system being used to perform the above-described intelligent adjustable medical cotton ball forming method, the system comprising:

[0044] The pressure acquisition module is used to acquire real-time pressure data of the target area of ​​the mold cavity through a distributed pressure sensor array, and input the raw pressure data into a Gaussian filtering algorithm for spatial dimension noise reduction to generate a regional pressure feature vector, which is then transmitted to the pressure trend determination module.

[0045] The pressure trend determination module is used to receive the pressure feature vector of the region and, by calling the dynamic gradient determination function, determine the direction of the pressure change trend in the region, determine whether the feature vector satisfies the continuous directional offset condition, and when the condition is met, output the dynamic gradient adjustment command and transmit it to the pressure adjustment module.

[0046] The pressure regulation module is used to receive the dynamic gradient regulation command, perform amplitude regulation on the compressed gas control valve, calculate the molding pressure compensation parameters, and transmit them to the gap adjustment module.

[0047] The gap adjustment module is used to receive the molding pressure compensation parameters, drive the servo motor to adjust the mold gap based on the parameters, collect the adjusted gap data through the displacement sensor and perform offset calculation with the reference gap value to generate a molding status feedback signal, which is then transmitted to the molding evaluation module.

[0048] The molding evaluation module is used to receive the molding state feedback signal and dynamic gradient adjustment command, input them into the convolutional neural network model for temporal feature extraction and analysis, and output the molding evaluation result.

[0049] The beneficial effects of the technical solutions provided by the embodiments of the present invention include at least the following:

[0050] Data is collected at fixed intervals using a distributed pressure sensor array. Combined with multi-dimensional spatial denoising processing, the accuracy of pressure feature extraction is improved, eliminating environmental noise interference in the molding process. A dynamic gradient judgment function provides directional analysis of pressure change trends, accurately identifying continuous offset conditions and enabling real-time dynamic adjustment of molding parameters. This avoids the hysteresis errors caused by traditional static threshold control. A closed-loop pressure compensation mechanism is formed through the coordinated control of the compressed gas channel valve opening and the servo motor, ensuring the synchronization of air pressure and gap adjustment within the mold cavity and enhancing the stability of the molding process. Displacement sensors collect gap data in real time and perform offset calculations with reference values. Combined with the rate of change, feedback signals are generated to establish a multi-dimensional parameter correlation model, optimizing molding accuracy and consistency. Convolutional neural networks perform deep extraction of temporal features, fusing and analyzing dynamic adjustment commands and feedback signals to construct an adaptive evaluation system. This overcomes the limitations of traditional experience-dependent process optimization, achieving intelligent prediction and closed-loop optimization of molding quality. Attached Figure Description

[0051] Figure 1 This is a schematic diagram of the workflow of the present invention;

[0052] Figure 2 This is a system flowchart of the present invention. Detailed Implementation

[0053] The technical solution of the present invention will now be described with reference to the accompanying drawings.

[0054] In embodiments of the present invention, words such as "exemplarily," "for example," etc., are used to indicate that something is an example, illustration, or description. Any embodiment or design described as "exemplary" in the present invention should not be construed as being more preferred or advantageous than other embodiments or designs. Specifically, the use of the word "exemplary" is intended to present the concept in a concrete manner. Furthermore, in embodiments of the present invention, the meaning expressed by "and / or" can be both, or either one.

[0055] In the embodiments of this invention, the terms "image" and "picture" may sometimes be used interchangeably. It should be noted that, without emphasizing the distinction between them, their intended meanings are consistent. Similarly, the terms "of," "corresponding (relevant)," and "corresponding" may sometimes be used interchangeably. It should be noted that, without emphasizing the distinction between them, their intended meanings are consistent.

[0056] In this embodiment of the invention, sometimes a subscript such as W1 may be written in a non-subscript form such as W1. When the difference is not emphasized, the meaning they express is the same.

[0057] To make the technical problems, technical solutions and advantages of the present invention clearer, a detailed description will be given below in conjunction with the accompanying drawings and specific embodiments.

[0058] Please see Figure 1 This invention provides an intelligent adjustable medical cotton ball forming method, the processing flow of which may include the following steps:

[0059] S1: Collect regional pressure data through a distributed pressure sensor array, input the collected raw pressure data into a Gaussian filtering algorithm that dynamically adjusts kernel parameters based on noise level, and perform multi-dimensional spatial denoising to generate regional pressure feature vectors.

[0060] S2: Based on the regional pressure feature vector, the dynamic gradient determination function is called to analyze the directionality of the pressure change trend in the region. When the trend change meets the continuous direction offset condition, the dynamic gradient adjustment command is output.

[0061] S3: According to the dynamic gradient adjustment command, control the valve opening of the compressed gas channel in the target area, adjust the gas pressure in the mold cavity, and obtain the molding pressure compensation parameters by adjusting the amplitude to modify the current valve opening parameter.

[0062] S4: Based on the molding pressure compensation parameters, drive the servo motor to adjust the mold gap, collect the adjusted gap data in real time through the displacement sensor, calculate the offset with the reference gap value, and obtain the molding state feedback signal by combining the offset direction and the rate of change.

[0063] S5: Call the forming state feedback signal and dynamic gradient adjustment command, input them into the trained convolutional neural network model, perform temporal feature extraction and analysis, and output the forming evaluation result;

[0064] The regional pressure feature vector includes the pressure distribution matrix, contact area coordinates, and effective pressure value. The dynamic gradient adjustment command specifically includes the pressure gradient direction angle, cumulative displacement, and critical rate threshold. The molding pressure compensation parameters include the valve opening linear displacement, cavity pressure gradient coefficient, and pulse width modulation parameters. The molding state feedback signal specifically refers to the gap displacement deviation, real-time displacement rate, and deformation convergence rate. The molding result evaluation includes the porosity distribution map, density standard deviation, and stress concentration factor distribution.

[0065] Specifically, the steps of S1 are as follows:

[0066] S101: Collect regional pressure data through a distributed pressure sensor array, collect pressure data at multiple points within the region according to a preset fixed period, record the pressure value and timestamp of the sensor node, and summarize the sampled values ​​to form a data set with a unified structure, thus obtaining the original pressure data set;

[0067] First, the sensor array layout must be reasonable. In medical cotton ball forming equipment, the sensor array can be evenly distributed on the inner surface of the mold to ensure comprehensive monitoring of pressure changes throughout the pressing area. For example, assuming the spacing between sensor nodes is set to 1cm, the distribution of each sensor will be uniform, ensuring that pressure changes in each area are collected in a timely manner. The acquisition cycle needs to be optimized based on the operating speed of the equipment and the dynamic characteristics of the forming process. Setting the acquisition cycle to 100ms in the equipment means acquiring data 10 times per second. Assuming that in a practical application, sensor node A records a pressure value of 15.2kPa and a timestamp of 1050ms, each sensor will record its pressure value and the corresponding timestamp. For example, if node A1 has a pressure value of 15.2kPa at 1050ms, the corresponding data item is (A1, 1050, 15.2). The data collected by all sensor nodes is transmitted to the central processing unit via serial communication and aggregated into a triplet data set containing the node number, timestamp, and pressure value. This data set is updated once per acquisition cycle.

[0068] S102: Based on the original pressure data set, the Gaussian filtering algorithm is called to perform spatial domain denoising, identify the Gaussian weight distribution of each data point and its neighboring points, calculate the pressure response value and adjust the node pressure value to obtain the denoised pressure distribution data;

[0069] The kernel size and standard deviation of the Gaussian filtering algorithm are configured based on the sensor noise level and the target resolution;

[0070] In practice, the kernel size and standard deviation σ of the Gaussian filter must first be determined. A kernel size of 3×3 is chosen, and σ is set to 1.0 to accommodate the spatial resolution of the sensor array. For each sensor node, its surrounding 3×3 neighborhood data needs to be extracted, and the corresponding Gaussian weight matrix calculated. Assuming the weight of the central node is 0.1336, the weights of neighboring nodes decrease with increasing distance from the central node. For example, the pressure values ​​of the neighborhood surrounding the central node B2 are shown in Table 1.

[0071] Table 1: Gaussian filter neighborhood pressure values

[0072]

[0073]

[0074] As shown in Table 1, although the pressure values ​​do not differ significantly, fine weighting can buffer abnormal noise or local abrupt changes, thereby improving the stability and representativeness of the overall pressure data.

[0075] According to the calculation rules of Gaussian filtering, the pressure value of the central node B2 is obtained by multiplying the neighborhood data by the corresponding Gaussian weights and then summing the results, as expressed by the formula:

[0076]

[0077] Among them, P B2 w represents the pressure value at the central node. ij P represents the Gaussian weight of the node in the i-th row and j-th column. ij Let be the pressure value of the node in the i-th row and j-th column.

[0078] For example, the weight values ​​around node B2 are as follows:

[0079] w 11 =0.1336, w 12 =0.1065, w 13 =0.0755, w 21 =0.1065, w 22 =

[0080] 0.1592, w 23 =0.1065, w 31 =0.0755, w 32 =0.1065, w 33 =0.1336;

[0081] After substituting the pressure value, the denoising pressure value of node B2 was calculated to be 15.2 kPa. This value replaced the original pressure value, thus completing the denoising process.

[0082] S103: Based on the denoised pressure distribution data, extract the spatial distribution features of the nodes, analyze the high-dimensional feature space under the positional relationship between the nodes, calculate the coordinate values ​​of the nodes in the feature space and unify them into a standard vector form, and integrate the pressure feature vector of the node output area.

[0083] The spatial locations of the sensor nodes are mapped to a two-dimensional coordinate system. For example, node A1 is set to (0, 0), node A2 to (1, 0), and so on, forming the spatial coordinates of the nodes. For each node, a vector containing its coordinates and the denoised pressure value is constructed. For example, the vector for node A1 can be represented as (0, 0, 14.8). Next, the vectors of all nodes are combined into a matrix, and then principal component analysis (PCA) is used for dimensionality reduction to extract the main eigenvalues. Assuming that the PCA analysis finds that the first two principal components can explain 95% of the data variance, these two principal components are selected as the basis for the eigenvectors. The original three-dimensional vectors are projected onto these two principal components to obtain the standardized eigenvectors. For example, the eigenvector projection result for node A1 is (0.85, -0.12). Finally, the eigenvectors of all nodes are combined into a regional pressure eigenvector for subsequent analysis and processing.

[0084] Specifically, the steps of S2 are as follows:

[0085] S201: Based on the regional pressure feature vector, extract the pressure values ​​at the time node, construct the pressure change matrix, call the dynamic gradient determination function to calculate the directional offset of adjacent units, and generate a sequence of directional change indexes.

[0086] First, by acquiring pressure feature vector data within the region, and using time nodes as a reference, the pressure values ​​at the corresponding time nodes are extracted. The pressure values ​​are obtained through a real-time monitoring and data acquisition system, where pressure data is acquired periodically by sensors installed at pressure points in different areas. Assuming that at a certain moment, the pressure values ​​within the region are P1 = 10 Pa, P2 = 15 Pa, P3 = 18 Pa, a pressure matrix is ​​formed. For example, for data at consecutive time nodes t1, t2, t3, ..., the pressure value corresponding to each node is recorded and formed into a vector. The pressure matrix is ​​then {[P1(t1), P2(t1), P3(t1)], [P1(t2), P2(t2), P3(t2)], ...}, which represents the pressure distribution at multiple points within the region at different time nodes.

[0087] Next, the dynamic gradient within the region is calculated, and the pressure change is calculated using a gradient determination function. Assuming the pressure change in adjacent regions is from P1(t1) to P2(t2), its directional shift can be determined by comparing the pressure changes of adjacent cells. For example, if the pressure change is greater than a certain preset threshold, it is considered that a large shift has occurred in that direction. The dynamic gradient determination function calculates the gradient value based on the change between specific pressure values ​​and generates a sequence of directional change indicators. The key to this process is dynamically adjusting the pressure gradient within the region to ensure accurate pressure value changes. Assuming the gradient change rate ΔP = P2 - P1 = 15Pa - 10Pa = 5Pa, the directional change of the shift will be calculated using this difference.

[0088] S202: Based on the sequence of directional change indicators, extract paragraphs with continuous directional changes, filter out segments that meet the thresholds for angle change and length, calculate the frequency of paragraph directional switching and the amount of angle offset, and obtain the set of directional offset trend parameters.

[0089] By extracting the sequence of directional change indicators, we can identify paragraphs with continuous directional changes. The core of this process lies in defining what constitutes a "continuous directional change." Generally, a paragraph is considered a valid directional change paragraph if the angle change exceeds a preset threshold (e.g., an angle change greater than 15°) and the duration exceeds a certain length (e.g., more than 10 seconds).

[0090] Next, paragraphs that meet the requirements for angle change and length threshold are selected. Suppose that the direction of a paragraph changes from θ1 = 10° to θ2 = 30° within the time interval [t1, t2], and its angle change is 20°. At this time, the angle change of the paragraph meets the preset threshold and can be selected.

[0091] Then, calculate the direction switching frequency and angle offset for each paragraph. Assuming a paragraph contains multiple smaller sections with direction switching, and the angle offsets for each switch are Δθ1 = 5° and Δθ2 = 3° respectively, the formula for calculating the direction switching frequency is:

[0092]

[0093] Where Δt is the time interval between each switch, and T is the total duration of the paragraph. Assuming the total paragraph duration is 30 seconds and the number of switches is 3, then the switching frequency is:

[0094]

[0095] This result reflects the frequency of directional changes within the paragraph. The directional offset is calculated by accumulating the angular offset of each switch to obtain the total angular change.

[0096] S203: Call the paragraph change parameters in the directional offset trend parameter set, compare them with the corresponding time period directional codes in the regional pressure feature vector, record the start and end times of the paragraphs that meet the continuous directional offset conditions and the regional numbers, and obtain the dynamic gradient adjustment instructions.

[0097] By calling the paragraph change parameters from the directional offset trend parameter set and combining them with the directional code in the regional pressure feature vector, a alignment comparison is performed. The key to this process is to accurately align the directional offset parameters with the actual pressure changes over the time period, ensuring that pressure changes within different time periods can be precisely matched with directional changes. Assuming the directional code corresponds to a time period of t1 to t3 for a certain region, the directional code for this period is [θ1, θ2, θ3], and the corresponding pressure data is [P1, P2, P3]. By comparing the directional changes with the pressure data, it can be determined whether the paragraph meets the continuous directional offset condition.

[0098] If the conditions are met, the start and end times of the segment and its region number are recorded to generate dynamic gradient adjustment instructions. This method allows for precise control of pressure changes within a region and the generation of corresponding adjustment instructions. For example, if a region experiences a continuous directional shift between time periods t2 and t4, and the pressure changes meet the requirements, adjustment instructions will be generated based on this information to optimize the region's pressure state.

[0099] Specifically, the steps of S3 are as follows:

[0100] S301: Based on the dynamic gradient adjustment command, call the current opening parameter of the compressed gas channel in the target area, adjust the opening parameter within the set range according to the adjustment amplitude, and generate the valve adjustment opening value;

[0101] Based on the dynamic gradient adjustment command, the specific location of the compressed gas channel in the target area referenced in the adjustment command is first identified, its control valve is located, and the current opening parameter is read. Assuming the mold number is M01 and the opening is read as 30%, with the adjustment amplitude set to 10%, the current opening is checked within the set adjustment range of 20%-70% to determine if it is within the allowable range. If 30% is found to be slightly below the midpoint of the set range, the opening parameter is adjusted upwards. During the adjustment process, the current opening value A is... d The values ​​are superimposed with the adjustment amplitude ΔA, and the calculation formula is as follows:

[0102] A x =A d +ΔA;

[0103] Where A d =30%, ΔA=10%, then the calculation is:

[0104] A d=30% + 10% = 40%;

[0105] If this value falls within the set range of 20%-70%, it is determined that the adjustment condition is met. This adjusted value is recorded and applied as the current new valve opening. For mold number M02, its current opening is 45%, the adjustment range is -5%, and the set range is 30%-60%. Therefore:

[0106] A x =45% - 5% = 40%,

[0107] It also satisfies the set interval constraints; if there exists an instance, such as M03, where the current opening is 60% and the adjustment amplitude is 15%, then the values ​​are directly added together to obtain:

[0108] A x =60% + 15% = 75%,

[0109] Although this value is greater than the initial value, it has not exceeded the maximum threshold of 80%, and is still a valid adjustment value. If it exceeds the maximum value, it needs to be limited to not exceed the upper limit of the range. During execution, the current opening data is obtained by reading the valve's current status register, and arithmetic operations are performed through the adjustment amplitude field in the instruction to limit the result to the set range. If it exceeds the range, a trimming process is performed, that is: if the new opening is greater than the maximum value, the maximum value is taken; if it is less than the minimum value, the minimum value is taken. The final output of this adjustment operation is the valve's adjusted opening value, which is used to drive the actuator to perform the next mechanical response operation.

[0110] Table 2: Valve Adjustment Data Table

[0111]

[0112]

[0113] As shown in Table 2, by comparing the given adjustment amplitude with the set range, the target opening value after each execution of the dynamic gradient adjustment command can be clearly obtained, ensuring that it is within the control range and providing basic parameters for subsequent valve opening and closing control and flow rate analysis.

[0114] S302: Based on the valve adjustment opening value, control the real-time opening and closing status of the compressed gas channel valve in the mold cavity, collect the gas channel flow rate data after opening and closing and calculate the interval difference between it and the initial value to generate the flow rate change value.

[0115] Upon receiving and executing the new valve opening adjustment value, the real-time control mechanism for the compressed gas channel within the mold cavity is immediately activated. The control mechanism first reads the control ID corresponding to this channel in the control bus and synchronously transmits the valve drive signal to the physical actuator via the control logic unit. The actuator sets its rotation angle according to the input opening value, thereby physically changing the effective cross-sectional area of ​​the compressed gas channel. Assuming the current mold is M01 and the adjusted opening value is 40%, the actuator controls the valve core to the corresponding opening position. Feedback elements such as Hall effect sensors confirm in real time whether the current physical position of the valve core has reached the target opening. If the target is not reached, the drive is repeated and monitored until it is achieved. After completing the valve's physical position adjustment, real-time acquisition of the compressed gas flow rate data begins. Sensors are placed in the middle of the channel, and the actual gas flow volume per minute is obtained through differential pressure measurement and thermal flow detection. Assuming the initial flow rate Q... t0 =100L / min, the current flow rate Q is measured. t1 =110L / min, mold cavity volume V = 2.0L, total sampling period k = 3, compressed gas flow rate deviation λ within each period j The valve opening / closing amplitude γ is 0.3, 0.2, and 0.1 respectively. j The percentages are 10%, 8%, and 12% respectively, with a compression response sensitivity coefficient η = 1.5. The change in flow rate is calculated using the formula:

[0116]

[0117] Where ΔQ represents the change in flow rate, in units of L / min, Q t1 This represents the throughput data value at time t1, in L / min, Q. t0 The flow rate at initial time t0 is represented by V, in L / min; V represents the volume of the mold cavity, in L; λ j γ represents the real-time flow rate deviation of the compressed gas channel during the j-th cycle, in L / min. j η represents the instantaneous opening and closing amplitude of the valve within the j-th cycle, expressed as a percentage; k represents the total number of sampling cycles; and η represents the gas compression response sensitivity coefficient, a dimensionless parameter.

[0118] First, calculate the average of the deviation terms:

[0119]

[0120] Then calculate the principal difference term:

[0121]

[0122] ΔQ=|7.07+2.90|=9.97L / min;

[0123] This value represents the change in flow rate per unit condition in the compressed gas passage caused by valve opening adjustment. This change value will be used as key input data in subsequent pressure regulation and compensation. Therefore, by controlling the valve opening and sensing the resulting flow rate change in real time, and combining this with valve dynamic behavior data over multiple cycles, the instantaneous and average flow rate changes caused by the opening adjustment can be accurately reconstructed, thus providing quantitative input for the mold gas pressure system.

[0124] S303: Based on the change value of flow rate, combined with the mold cavity volume data, normalization processing is performed to calculate the amplification value corresponding to the change rate of flow rate per unit volume and the set pressure adjustment coefficient, and the molding pressure compensation parameter is obtained.

[0125] The calculated flow rate change value ΔQ = 9.97 L / min is normalized to the volume parameters of the mold cavity to quantify the degree of flow rate change per unit volume. Here, the mold volume is V = 2.0 L, so the flow rate change per unit volume is: ΔQ / V = 9.97 / 2.0 = 4.985 L / min / L. This result indicates that the gas flow rate per liter of mold space changes by approximately 4.985 liters per minute. Subsequently, a pressure adjustment coefficient K = 1.2 is introduced, representing the pressure amplification response ratio of the gas in the mold to the flow rate change. The molding pressure compensation value P... b The calculation is as follows: P b =4.985×1.2=5.982L / min / L. This value is the reference value of the additional pressure applied during the mold forming process according to the current valve adjustment. It is used to adjust the output setting value of the pressure control system to compensate for the pressure fluctuation of the mold cavity caused by the change in flow rate. By comparing this value with the predetermined target pressure change range (such as 4.5~6.5L / min / L), it can be seen that the result is within the target range. This indicates that the pressure change caused by the current opening adjustment can be effectively accepted and can maintain a stable pressure state in the mold cavity within the expected control range, thus providing a reliable guarantee for subsequent molding operations.

[0126] Specifically, the steps of S4 are as follows:

[0127] S401: Based on the molding pressure compensation parameters, drive the servo motor to adjust the mold gap, extract the difference between the control displacement and the target command, perform synchronous correction operation and record the response rate and feedback delay, and generate the mold gap adjustment value;

[0128] Based on molding pressure compensation parameter P b =5.982L / min / L, the mold gap adjustment mechanism is driven by the servo motor control module. First, the target displacement command D of the servo motor is read. m=0.5mm, and simultaneously the actual displacement value D of the current mold clearance is collected in real time by the encoder. s =0.45mm, calculate the displacement difference ΔD = |0.5-0.45| = 0.05mm. If the difference exceeds the preset threshold of 0.03mm, a correction operation is triggered. The correction amount is ΔD×H = 0.05×1.2 = 0.06mm (where the correction coefficient H = 1.2 is set according to the servo motor response characteristics). Record the correction command issuance time t1 = 10:00:00.000 and the actuator feedback completion time t2 = 10:00:00.150. Calculate the response delay Δt = 150ms. The response rate is v = ΔD / Δt = 0.06mm / 0.15s = 0.4mm / s. Finally, the mold clearance adjustment value D is generated. t =0.45+0.06=0.51mm, this value is used as the control input for the next stage.

[0129] Table 3: Die Clearance Adjustment Parameter Table

[0130] Parameter name numerical values unit Target displacement 0.50 mm actual displacement 0.45 mm Correction coefficient 1.2 none Response delay 150 ms

[0131] As shown in Table 3, by calculating the displacement difference and applying the correction coefficient, the dynamic adjustment process of the servo motor can be accurately quantified, and a data basis can be provided for subsequent clearance offset analysis.

[0132] S402: Call the mold clearance adjustment value, collect mold clearance change data through the displacement sensor, calculate the offset between the mold clearance and the reference clearance, and combine the data acquisition time points to perform time sequence organization to obtain the mold clearance offset change.

[0133] Call the mold clearance adjustment value D t =0.51mm, gap data at 5 time points were continuously collected using a laser displacement sensor at a sampling frequency of 1000Hz:

[0134] [0.50, 0.52, 0.49, 0.53, 0.51]mm;

[0135] Reference gap D c =0.50mm, calculate the time point offset Δd i =|d i -0.50|, resulting in [0.00, 0.02, 0.01, 0.03, 0.01] mm;

[0136] Data is organized according to the time series t = [0, 0.1, 0.2, 0.3, 0.4] s, generating a time series offset sequence {(0, 0.00), (0.1, 0.02), (0.2, 0.01), (0.3, 0.03), (0.4, 0.01)}. The average offset rate ∑Δd is calculated through linear interpolation. i / ∑ti =0.07 / 1.0=0.07mm / s, which indicates that the mold gap exhibits a fluctuating convergence trend after adjustment.

[0137] S403: Based on the change in mold clearance offset, extract the direction and rate of change of clearance, construct a trend offset map by combining time series data, perform key trend identification and state determination operations of the map, and generate molding state feedback signal.

[0138] Based on the time-series offset sequence data, the direction of the interval change is extracted as alternating positive and negative (sign sequence is 0, +1, -1, +1, -1), and the rate of change between adjacent time points is calculated:

[0139] r = [NaN, +0.2, -0.15, +0.4, -0.2] mm / s, taking the average absolute value of the rate of change after removing invalid values. When constructing the trend shift map, the stability threshold was set to 0.25 mm / s (based on the 90th percentile of the original data). Because... The current gap state is determined to be a "stable transition stage," and a molding state feedback signal S=1 is generated (coded 1 to represent that production can continue). This triggers a warning signal. The advantage of the formula lies in its ability to accurately characterize the dynamic behavior of the mold by quantifying the coupling relationship between the direction and rate of gap change, thus avoiding the limitations of a single parameter criterion.

[0140] Specifically, the steps of S5 are as follows:

[0141] S501: Based on the forming state feedback signal and dynamic gradient adjustment command, it synchronously inputs them into the trained convolutional neural network model, performs spatial expansion of the feedback signal through the convolutional layer, extracts the gray level difference of adjacent pixels, performs mean filtering by region, calculates the changing trend of regional pixel difference values, and generates the signal gradient change rate.

[0142] Based on the feedback signal state S=1 and the node weight ratio W=[0.3, 0.5, 0.2] set in the dynamic gradient adjustment command, the current feedback flux data is spatially expanded. Three monitoring flux data points Q=[1.2, 1.5, 1.3]L / s are selected to represent the measured values ​​of consecutive nodes, and the corresponding coordinates are set as (X1, Y1)=(1, 1), (X2, Y2)=(2, 1), and (X3, Y3)=(3, 1). In the convolutional layer of the convolutional neural network, the local gradient change rate between adjacent data points is calculated using the formula:

[0143]

[0144] Where ΔG represents the rate of change of the signal gradient, P iX represents the grayscale value of the i-th pixel. i Y represents the x-coordinate of the i-th pixel. i w represents the ordinate value of the i-th pixel. i The weight factor representing the i-th pixel, the adaptive weight factor w i The feature parameters learned by the model are dynamically set and adjusted according to the importance of different pixels. n represents the number of pixels in the image region and i is the pixel index.

[0145] In practice, the first step is to calculate the flux change between point 2 and point 1: P2 = 1.5, P1 = 1.2, and the lateral distance between the two points is... The local rate of change is |(1.5-1.2) / 1.0|=0.3. Combined with the weight w2=0.5, the weighted value for this segment is 0.3×0.5=0.15. The second step calculates the rate of change between point 3 and point 2: P3=1.3, P2=1.5, and the lateral distance remains 1.0. The flux rate of change is |(1.3-1.5) / 1.0|=0.2, corresponding to a weight w3=0.2. The weighted value is 0.2×0.2=0.04. The final signal gradient rate of change is the sum of the two:

[0146] ΔG = 0.15 + 0.04 = 0.19 L / s 2 ;

[0147] Meanwhile, to ensure numerical rationality, a standard filtering window was used for moving average processing, with a window length of 2. The mean filtering results for the difference sequence [0.3, 0.2] were calculated as follows:

[0148]

[0149] Since the second term lacks subsequent data to pad with zeros, and considering trend analysis, the slope is calculated to be (0.1-0.25) / 1 = -0.15. Taking the absolute value as the overall trend assessment index of 0.15, this value is close to the calculated result of 0.19, verifying the accuracy of its spatial gradient reflection. The table below provides the core parameters and results from the signal processing process:

[0150] Table 4: Calculation Table of Signal Gradient Change Rate

[0151] Parameter name numerical values unit Flux monitoring point data 1.2，1.5，1.3 L / s Pixel coordinates (1，1)，(2，1)，(3，1) grid Difference value 0.3，0.2 L / s Distance calculation 1.0，1.0 grid weight value 0.5，0.2 none Weighted gradient term 0.15，0.04 <![CDATA[L / s 2 ]]> Filtering results 0.25，0.1 L / s Gradient rate of change 0.19 <![CDATA[L / s 2 ]]>

[0152] As shown in Table 4, the gradient change rate ΔG = 0.19 L / s 2 The result is obtained by dividing the flux difference by the distance and then summing the results with weights. This result is used in subsequent temporal gradient reconstruction operations.

[0153] S502: Call the signal gradient change rate, reconstruct the gradient change sequence in the time axis direction according to the time sequence recognition module in the convolutional neural network model, superimpose the node weight ratio in the dynamic gradient adjustment instruction, calculate the horizontal offset of the node, and generate the node gradient offset.

[0154] The signal gradient change rate ΔG obtained from the front-end calculation is 0.19L / s. 2 In the time series recognition module, time series information is introduced. Following the timestamp order t = [0, 0.5, 1.0]s, the corresponding gradient monitoring value sequence S = [0.10, 0.15, 0.12] is organized. The node weight ratio W = [0.3, 0.5, 0.2] set in the dynamic gradient adjustment command is then superimposed, and a weighted calculation is performed.

[0155] G=0.3×0.10+0.5×0.15+0.2×0.12=0.03+0.075+0.024=0.129;

[0156] The weighted results are used for interpolation reconstruction of the time series curve. A linear interpolation method is employed to fit the gradient changes between discrete time nodes, thereby analyzing local offset characteristics. Taking the node between t = 0.5s and t = 1.0s as an example, the corresponding gradient values ​​are 0.15 and 0.12, and the lateral offset magnitude is the gradient difference multiplied by a scaling factor, i.e.

[0157] Δx=|0.15-0.12|×10=0.03×10=0.3mm,

[0158] This yields a node gradient offset of 0.3 mm, which reflects the degree of local disturbance in the forming trajectory within a specified time interval.

[0159] S503: Call the node gradient offset, compare the node offset result with the gradient tolerance limit value according to the evaluation function parameters at the model output, determine whether the offset magnitude deviates, extract the segment deviation ratio to fit the overall trend, and obtain the forming evaluation result.

[0160] Read the gradient tolerance limit value T set in the evaluation module. r=0.5mm, compared with the currently calculated offset of 0.3mm. Since 0.3 < 0.5, it is determined that the current node has not exceeded the boundary and is marked as "no deviation". To further evaluate the stability of the overall forming process, the deviation ratio data of three consecutive time periods [0.1, 0.2, 0.25] are extracted and their changing trend is calculated. Through linear regression, the trend slope k = (0.2-0.1) + (0.25-0.2) / 2 = 0.075 is estimated. This value is lower than the set threshold of 0.1 (based on the 85th percentile value obtained from historical data statistics), indicating that the current forming state has trend stability. Therefore, the forming evaluation result is E = qualified. This process effectively realizes the joint judgment of forming dynamic characteristics and accuracy through time-series analysis of one-dimensional sequence, improving the accuracy of evaluation while ensuring response speed.

[0161] like Figure 2 As shown, an intelligent adjustable medical cotton ball forming system includes:

[0162] The pressure acquisition module is used to acquire real-time pressure data of the target area of ​​the mold cavity through a distributed pressure sensor array, and input the raw pressure data into a Gaussian filtering algorithm for spatial dimension noise reduction to generate a regional pressure feature vector, which is then transmitted to the pressure trend determination module.

[0163] The pressure trend determination module is used to receive the regional pressure feature vector and, by calling the dynamic gradient determination function, to determine the direction of the pressure change trend in the region, and to determine whether the feature vector meets the continuous directional offset condition. When the condition is met, the dynamic gradient adjustment command is output and transmitted to the pressure adjustment module.

[0164] The pressure regulation module is used to receive dynamic gradient regulation commands, perform amplitude regulation on the compressed gas control valve, calculate the molding pressure compensation parameters, and transmit them to the gap adjustment module.

[0165] The gap adjustment module is used to receive molding pressure compensation parameters and drive the servo motor to adjust the mold gap based on the parameters. It collects the adjusted gap data through the displacement sensor and performs offset calculation with the reference gap value to generate a molding status feedback signal, which is then transmitted to the molding evaluation module.

[0166] The molding evaluation module receives molding status feedback signals and dynamic gradient adjustment instructions, inputs them into a convolutional neural network model for temporal feature extraction and analysis, and outputs molding evaluation results.

[0167] The above description is merely a specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the technical scope disclosed in the present invention should be included within the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be determined by the scope of the claims.

Claims

1. A method of forming a smart regulated medical cotton ball, comprising: The method comprises the following steps: S1: collecting regional pressure data through a distributed pressure sensor array, inputting the collected original pressure data into a Gaussian filtering algorithm based on dynamic adjustment of noise level kernel parameters, and performing multi-dimensional space denoising processing to generate a regional pressure feature vector; S2: based on the regional pressure feature vector, calling a dynamic gradient determination function to analyze the direction of the trend of the pressure change in the region, and outputting a dynamic gradient adjustment instruction when the trend change meets the continuous directional offset condition; S3: controlling the valve opening degree of the target region compression gas channel according to the dynamic gradient adjustment instruction, adjusting the gas pressure in the mold cavity, modifying the current valve opening degree parameter by adjusting the amplitude, and obtaining a molding pressure compensation parameter; S4: based on the molding pressure compensation parameter, driving a servo motor to adjust the mold gap, collecting the adjusted gap data in real time through a displacement sensor, calculating the offset with a reference gap value, combining the offset direction and the change rate, and obtaining a molding state feedback signal; S5: calling the molding state feedback signal and the dynamic gradient adjustment instruction, inputting them into a trained convolutional neural network model, performing time series feature extraction and analysis, and outputting a molding evaluation result.

2. The intelligent adjustable medical cotton ball forming method according to claim 1, characterized in that, The dynamic gradient determination function selects a time window according to the noise level, and the continuous directional offset condition is that the gradient direction continuously changes in not less than three consecutive time steps, and the offset angle exceeds a set threshold value, which is optimized on historical data samples using the gradient descent method; The offset direction and the change rate are respectively subjected to Z-score standardization processing to unify the dimensions; The convolutional neural network model comprises two convolution extraction layers, one LSTM time series processing layer, and two fully connected layers, and the training is based on batch original molding data samples of an experimental device; The regional pressure feature vector includes a pressure distribution matrix, contact region coordinates, and effective pressure values, the dynamic gradient adjustment instruction specifically includes a pressure gradient direction angle, a displacement cumulative amount, and a critical rate threshold value, the molding pressure compensation parameter includes a valve opening degree linear displacement, a cavity pressure gradient coefficient, and a pulse width modulation parameter, the molding state feedback signal specifically refers to a gap displacement deviation, a real-time displacement rate, and a deformation convergence rate, and the molding evaluation result includes a porosity distribution graph, a density standard deviation, and a stress concentration coefficient distribution.

3. The intelligent adjustable medical cotton ball forming method according to claim 1, characterized in that, The specific steps of S1 include: S101: collecting regional pressure data through a distributed pressure sensor array, collecting the pressure of multiple points in the region according to a preset fixed period, recording the pressure value and time stamp of the sensor node, and forming a unified structure data set by collecting the sample value to obtain an original pressure data set; S102: based on the original pressure data set, calling a Gaussian filtering algorithm for spatial domain denoising, identifying the Gaussian weight distribution of each data point and its adjacent points, calculating the pressure response value and adjusting the node pressure value, and obtaining denoised pressure distribution data; The kernel size and standard deviation of the Gaussian filtering algorithm are configured according to the sensor noise level and the target resolution; S103: Based on the de-noised pressure distribution data, the spatial distribution characteristics of the nodes are extracted, the high-dimensional feature space under the position relationship between the nodes is analyzed, the coordinate values of the nodes in the feature space are calculated and unified into a standard vector form, and the node output area pressure feature vector is integrated.

4. The method of claim 3, wherein the step of forming the medical ball is performed by a molding machine. The specific steps of S2 include: S201: Based on the area pressure feature vector, the pressure values at the time nodes are extracted, a pressure change matrix is constructed, a dynamic gradient determination function is called to calculate the direction offset of adjacent units, and a directional change index sequence is generated; S202: According to the directional change index sequence, the continuous directional change paragraphs are extracted, the sections meeting the angle change and length threshold are screened, the paragraph direction switching frequency and angle offset are calculated, and the direction offset trend parameter set is obtained; S203: The paragraph change parameters in the direction offset trend parameter set are called, and the corresponding time period direction encoding in the area pressure feature vector is compared, the start and end time and area number of the paragraph meeting the continuous direction offset condition are recorded, and the dynamic gradient adjustment instruction is obtained.

5. The method of claim 4, wherein the step of forming the medical ball is performed by a molding machine. The specific steps of S3 include: S301: Based on the dynamic gradient adjustment instruction, the current opening parameter of the compressed gas channel of the target area is called, the opening parameter is adjusted within the set interval according to the adjustment amplitude, and the valve adjustment opening value is generated; S302: According to the valve adjustment opening value, the real-time opening and closing state of the compressed gas channel valve in the mold cavity is controlled, the flow data of the gas channel after opening and closing is collected and the interval difference between the initial value is calculated, and the flow change value is generated; S303: Based on the flow change value, the unit volume flow rate change rate and the amplification amplitude corresponding to the set pressure adjustment coefficient are calculated by normalizing the mold cavity volume data, and the molding pressure compensation parameter is obtained.

6. The method of claim 5, wherein the step of forming the medical ball is performed by a molding machine. The formula for calculating the flow change value is: ； wherein, represents the change value of the flow rate, unit: L / min, Q t1 represents the current flow rate data value at time t1, unit: L / min, Q t0 represents the initial flow rate data value at time t0, unit: L / min, V represents the mold cavity volume data, unit: L, λ j represents the real-time flow rate deviation of the compressed gas channel in the jth period, unit: L / min, Y j represents the instantaneous opening and closing amplitude of the valve in the jth period, unit: percentage, k represents the total number of sampling periods, η represents the gas compression response sensitivity coefficient, which is a dimensionless parameter.

7. The method of claim 5, wherein the step of forming the medical ball comprises the step of: The specific steps of S4 include: ​ S401: Based on the molding pressure compensation parameter, the servo motor is driven to adjust the mold gap, the difference between the control displacement and the target instruction is extracted, the synchronous correction operation is performed and the response rate and feedback delay are recorded, and the mold gap adjustment value is generated; S402: The mold gap adjustment value is called, the mold gap change data is collected through the displacement sensor, the offset of the mold gap and the reference gap is calculated, and the time sequence is organized in combination with the data collection time point, and the mold gap offset change amount is obtained; S403: According to the mold gap offset change amount, the gap change direction and change rate are extracted, the trend offset atlas is constructed in combination with the time sequence data, the key trend recognition and state determination operation is performed, and the molding state feedback signal is generated.

8. The method of claim 7, wherein the step of forming the smart regulated medical ball is characterized by, The specific steps of S5 include: S501: Based on the molding state feedback signal and dynamic gradient adjustment instruction, they are input into the trained convolutional neural network model, the feedback signal is expanded in space through the convolutional layer, the gray difference of adjacent pixels is extracted, the mean filter is performed by region, the regional pixel difference value change trend is calculated, and the signal gradient change rate is generated; S502: Call the signal gradient change rate, reconstruct the gradient change sequence in the time axis direction according to the time sequence recognition module in the convolutional neural network model, superimpose the node weight ratio in the dynamic gradient adjustment instruction, calculate the node lateral offset amplitude, and generate the node gradient offset amount; S503: Call the node gradient offset amount, compare the node offset result with the gradient tolerance limit value according to the evaluation function parameter of the model output end, judge whether the offset amplitude deviates, extract the section deviation ratio for overall trend fitting, and obtain the forming evaluation result.

9. The method of claim 8, wherein the step of forming the medical ball is performed by a molding machine. The signal gradient change rate is calculated by using the formula: ； wherein, representing the signal gradient change rate, P i representing the gray value of the i-th pixel, X i representing the horizontal coordinate value of the i-th pixel, Y i representing the vertical coordinate value of the i-th pixel, W i representing the weight factor of the i-th pixel, adaptive weight factor W i The feature parameters learned according to the model are dynamically set, and the key of the differentiated pixel points is controlled, n represents the number of pixels in the image region, and i is the pixel index.

10. An intelligent medical swab forming system, comprising: The system is used to realize the intelligent adjustment type medical cotton ball forming method of any one of claims 1-9, and the system comprises: A pressure acquisition module is configured to acquire real-time pressure data of a target region of a mold cavity through a distributed pressure sensor array, input original pressure data to a Gaussian filtering algorithm for spatial dimension denoising processing, generate a regional pressure feature vector, and deliver the regional pressure feature vector to a pressure trend judgment module; The pressure trend judgment module is configured to receive the regional pressure feature vector, call a dynamic gradient judgment function to judge the direction of the pressure change trend in the region, judge whether the feature vector meets the continuous directional offset condition, output a dynamic gradient adjustment instruction when the condition is met, and deliver the dynamic gradient adjustment instruction to a pressure adjustment module; The pressure adjustment module is configured to receive the dynamic gradient adjustment instruction, perform amplitude adjustment on the compressed gas control valve, calculate the forming pressure compensation parameter, and deliver the forming pressure compensation parameter to a gap adjustment module; The gap adjustment module is configured to receive the forming pressure compensation parameter, drive a servo motor to adjust the mold gap based on the parameter, collect the adjusted gap data through a displacement sensor, perform offset operation on the adjusted gap data and a reference gap value, generate a forming state feedback signal, and deliver the forming state feedback signal to a forming evaluation module; The forming evaluation module is configured to receive the forming state feedback signal and the dynamic gradient adjustment instruction, input the forming state feedback signal and the dynamic gradient adjustment instruction to a convolutional neural network model for time sequence feature extraction and analysis, and output a forming evaluation result.

Images