A method and system for optimizing a vacuum compounding process

Abstract

The application provides a vacuum mixing process optimization method and system, which comprises collecting mixing container cavity pressure, jacket temperature, material torque, stirring speed, and air extraction flow signal, calculating pressure change rate, torque fluctuation amplitude, air extraction amount per unit time, and pressure fluctuation characteristic quantity, dividing the vacuum mixing process into four stages of initial exhaust, bubble release, viscosity climb, and densification stability; in the initial exhaust section, the vacuum valve opening degree and time length are controlled according to the air extraction amount and the pressure change rate, and the pressure response delay is recorded; in the bubble release section, an air extraction and stop alternating sequence is generated, and the air extraction and stop time length is corrected in combination with the delay; in the viscosity climb section, the target pressure interval is determined according to the torque, temperature, and pressure change rate, and the speed and valve opening degree are cooperatively adjusted to stabilize the pressure; in the densification stability section, a stability criterion is constructed, and the valve opening degree is controlled as needed; when adjacent stages are switched, the target pressure starting point, valve opening degree initial value, speed initial value, and control period of the next stage are corrected according to the end parameters of the previous stage.

Description

Technical Field

[0001] This application belongs to the field of mixing process optimization, and in particular relates to a method and system for optimizing a vacuum mixing process. Background Technology

[0002] Vacuum mixing of high-viscosity or solid-liquid mixtures is a thermodynamic and fluid dynamic process. As powders wet, solvents evaporate, and polymer chains crosslink or extend, the apparent viscosity, internal flowability, thermal conductivity, and gas-liquid separation characteristics of the material all change. Using a single vacuum level or a fixed pumping rate can easily lead to serious consequences: in the initial stages of mixing, excessively rapid pumping and excessively high vacuum levels can cause shallow gas expansion, resulting in the loss of core formulation components; in the middle and later stages of mixing, as the material viscosity increases, the microbubbles trapped inside cannot overcome the surface tension of the high-viscosity fluid and overflow, leading to micropores in the product, weakening the mechanical strength, electrical conductivity, and batch consistency of the finished product. Vacuum mixing equipment combines sensing and control technology to achieve multi-stage vacuum adjustment by monitoring the pressure inside the container and setting different rotation speeds.

[0003] However, phased control, based on fixed schedules set by operators or single absolute pressure thresholds, completely severs the intrinsic connection between pressure changes and the actual rheological state of the material, leading to a disconnect between phase switching and the actual degassing requirements of the material. During the bubble release and viscosity ramp-up phases, the crude method of continuously opening valves to extract air overlooks the inherent pressure response delay characteristics of the gas path system, fails to intervene in pressure fluctuations, and is prone to pressure overshoot and continuous liquid level oscillations, making it impossible to maintain the required optimal pressure range when viscosity increases. Existing control methods use hard switching logic during the transition between adjacent process stages, assigning a fixed setpoint to the next stage without utilizing historical parameters such as pressure state, valve opening, and rotational speed from the previous stage for transitional correction. This type of stage jump leads to a mismatch between the initial control value and the actual system state, causing sudden pressure changes and mechanical disturbances, disrupting the newly established stable flow field and degassing environment, prolonging the system's recovery time, resulting in high energy consumption in the mixing process and an inability to guarantee product yield. Summary of the Invention

[0004] To address the problems of mismatched degassing requirements, pressure overshoot, and system disturbances caused by the disconnect between stage division and the actual rheological state of materials, neglect of gas pressure response delay, and use of hard switching logic between stages in existing technologies.

[0005] In a first aspect, the present invention provides a method for optimizing a vacuum mixing process, comprising the following steps: The pressure inside the mixing container, jacket temperature, material torque, stirring speed and air flow rate are collected. The pressure change rate, torque fluctuation amplitude, air pumping volume per unit time and pressure fluctuation characteristics are calculated. The vacuum mixing process is divided into the initial exhaust section, bubble release section, viscosity rise section and dense stabilization section. In the initial exhaust section, the vacuum intensity is determined based on the pumping volume per unit time and the pressure change rate, the opening degree of the vacuum valve and the duration of its action are adjusted, and the pressure response delay time after the opening degree change is recorded; in the bubble release section, an alternating control sequence of pumping and stopping is generated based on the pressure fluctuation characteristics, and the actual pumping duration and the actual stopping duration are corrected in combination with the pressure response delay time. During the viscosity ramp-up phase, the target pressure holding range is determined based on the torque fluctuation amplitude, jacket temperature, and pressure change rate. The stirring speed and vacuum valve opening are coordinated and adjusted to maintain the pressure within the target pressure holding range. During the dense and stable phase, a stability criterion is constructed based on the average pressure, pressure fluctuation degree, and continuous range. When the stability conditions are met, the target vacuum level is maintained. When the stability conditions are not met, the vacuum valve opening is adjusted based on the pressure difference direction. When switching between adjacent stages, the target pressure starting point, vacuum valve opening initial value, stirring speed initial value and control cycle of the next stage are corrected based on the end state parameters of the previous stage.

[0006] In a second aspect, the present invention also provides a vacuum mixing process optimization system, comprising the following modules: The module is used to collect signals of pressure inside the mixing container, jacket temperature, material torque, stirring speed and air flow rate, calculate pressure change rate, torque fluctuation amplitude, air volume per unit time and pressure fluctuation characteristics, and divide the vacuum mixing process into initial exhaust section, bubble release section, viscosity rise section and dense stabilization section. The first correction module is used to determine the vacuum intensity based on the pumping volume per unit time and the pressure change rate in the initial exhaust section, adjust the opening degree and duration of the vacuum valve, and record the pressure response delay time after the opening degree changes; in the bubble release section, it generates an alternating control sequence of pumping and stopping based on the pressure fluctuation characteristics, and corrects the actual pumping duration and actual stopping duration in combination with the pressure response delay time. The determination module is used to determine the target pressure holding range based on the torque fluctuation amplitude, jacket temperature and pressure change rate during the viscosity ramp-up section, and coordinate the stirring speed and vacuum valve opening to maintain the pressure within the target pressure holding range; during the dense and stable section, it constructs a stability criterion based on the average pressure, pressure fluctuation degree and continuous range, and maintains the target vacuum level when the stability conditions are met, and adjusts the vacuum valve opening based on the pressure difference direction when the stability conditions are not met. The second correction module is used to correct the target pressure start point, vacuum valve opening initial value, stirring speed initial value, and control cycle of the next stage based on the end state parameters of the previous stage when switching between adjacent stages.

[0007] This invention comprehensively collects multi-dimensional state signals, dividing the vacuum mixing process into four specific stages and implementing differentiated control strategies. In the initial venting and bubble release stage, response delay time correction and alternating pumping and stopping control ensure complete bubble breakage and removal. In the viscosity ramp-up stage, multi-parameter coordinated adjustment of stirring speed and vacuum valve opening maintains stable target pressure. In the compaction and stabilization stage, a stability criterion is established to adjust the valve opening, ensuring consistent product quality. During stage transitions, the initial parameters of the subsequent stage are corrected using the end-stage state parameters, achieving smooth transitions between different process stages and improving the overall control accuracy and system stability of the vacuum mixing process. Attached Figure Description

[0008] Figure 1 A flowchart of an optimization method for the vacuum mixing process; Figure 2 This is a time-series curve of chamber pressure versus pressure change rate during the vacuum mixing process. Figure 3 A comparison chart of pumping and stopping times before and after pressure response delay compensation; Figure 4 This is a comparison chart of the total process time and residual bubble amount between the experimental group and the control group. Detailed Implementation

[0009] The exemplary embodiments of this disclosure are described below with reference to the accompanying drawings, including various details of the embodiments to aid understanding, and should be considered merely exemplary. Therefore, those skilled in the art will recognize that various changes and modifications can be made to the embodiments described herein without departing from the scope of this disclosure. Similarly, for clarity and brevity, descriptions of well-known functions and structures are omitted in the following description.

[0010] In this application, an optimization method for a vacuum mixing process is provided, such as... Figure 1 As shown, it includes the following steps: S1 collects signals of pressure inside the mixing container, jacket temperature, material torque, stirring speed and air flow rate, calculates pressure change rate, torque fluctuation amplitude, air pumping volume per unit time and pressure fluctuation characteristics, and divides the vacuum mixing process into initial exhaust section, bubble release section, viscosity ramp-up section and dense stabilization section.

[0011] The internal pressure signal is acquired by a high-frequency pressure sensor installed in the mixing container, the jacket temperature signal is acquired by a thermocouple sensor, the material torque signal is acquired by a torque sensor, the stirring speed signal is acquired by a servo motor encoder, and the suction flow rate signal is acquired by a mass flow meter. All analog signals are input to a programmable logic controller via an Ethernet bus for analog-to-digital conversion and low-pass filtering for noise reduction.

[0012] The pressure change rate is calculated by differentiating the discrete pressure time series using the `gradient` function from Python's NumPy library. The torque signal envelope is extracted using the `signal` module from the SciPy library, and the difference between peaks and troughs is calculated as the torque fluctuation amplitude. The pumping flow rate is numerically integrated within a unit time window using NumPy's `trapz` function to calculate the pumping volume per unit time. A Fast Fourier Transform is performed on the discrete sampled values ​​of the internal pressure within the current control cycle, extracting the maximum amplitude in the 0-5Hz frequency band as the pressure fluctuation feature. Based on this, the Support Vector Machine (SVM) classifier module from the scikit-learn library is used to construct a four-dimensional feature vector from the pressure change rate, torque fluctuation amplitude, pumping volume per unit time, and pressure fluctuation feature. This vector is then input into a pre-trained SVM model using historical batch mixing data. The SVM model calculates the distance between the feature vector and each decision boundary using the radial basis function kernel, outputting the current stage label, thus dividing the vacuum mixing process into the initial exhaust stage, bubble release stage, viscosity ramp-up stage, and compaction and stabilization stage in real time.

[0013] The Support Vector Machine (SVM) classifier uses a radial basis function (RBF) kernel to nonlinearly map the input four-dimensional feature vectors—pressure change rate, torque fluctuation amplitude, pumping volume per unit time, and pressure fluctuation features—to a high-dimensional feature space. In this space, a quadratic programming problem is solved to find the optimal hyperplane that maximizes the classification margin, thereby distinguishing samples from different mixing stages: the initial venting stage, the bubble release stage, the viscosity ramp-up stage, and the dense and stable stage. The training process uses historical batches of mixing data. The feature vectors at each sampling time are combined with manually labeled stage tags to form the training sample set. Cross-validation is used to determine kernel function parameters such as bandwidth γ and penalty coefficient C. Then, a sequential minimum optimization algorithm is used to solve for the support vectors and decision boundaries, ultimately resulting in an offline classification model that can output stage labels based on real-time feature vectors.

[0014] In some implementations, the acquisition of signals such as the pressure inside the mixing container, jacket temperature, material torque, stirring speed, and air extraction flow rate, and the calculation of pressure change rate, torque fluctuation amplitude, air extraction volume per unit time, and pressure fluctuation characteristic quantities, includes: Discrete sampled values ​​of the current internal cavity pressure and material torque are extracted synchronously at fixed time intervals; A first-order backward difference equation is established to calculate the rate of change of internal cavity pressure; The current sampling time is shifted forward by 3 seconds to form the sliding sampling interval. The maximum and minimum values ​​of the material torque within the sliding sampling interval are extracted, and the difference between the two values ​​is used to obtain the torque fluctuation amplitude. The pumping flow rate signal is numerically integrated within a unit time window to calculate the pumping volume per unit time. Perform a Fast Fourier Transform on the discrete sampled values ​​of the internal pressure within the current control cycle, and extract the maximum amplitude in the 0 to 5 Hz frequency band as the pressure fluctuation characteristic.

[0015] In actual operation, the industrial-grade PLC data acquisition card triggers a hardware interrupt at a pre-set 0.1s high frequency, converting the real-time captured internal pressure and motor torque into a high-resolution discrete time series. By executing a first-order backward differential algorithm, it can detect short-term pressure transients, such as when the internal pressure drops instantaneously from 0.50MPa to 0.45MPa, calculating the rate of change of -0.5MPa / s, thus avoiding the timing delay problem caused by the traditional average sliding filter algorithm. The high-frequency sampled value of the mixing container's internal pressure within 0~3s, and the time series curve of the pressure change rate calculated by the corresponding first-order differential algorithm are shown below. Figure 2 As shown, the continuous decay trend and instantaneous fluctuation characteristics of the internal cavity pressure are illustrated. Simultaneously, a sliding window with a width of 3 seconds that continuously moves forward is maintained in memory to locate the extreme points of material torque within a dataset containing 30 consecutive discrete sampling points. For example, within a certain period, the highest torque is extracted as 120 N·m, and the lowest torque as 110 N·m. The difference is calculated to obtain a fluctuation amplitude of 10 N·m, which represents the real-time fluctuation of viscous resistance and dispersion uniformity during the material mixing process. The suction flow rate signal is integrated using a trapezoidal numerical method within a 1-second unit time window to calculate the suction rate per unit time. For example, if the current suction flow rate is measured to be 200 L / min, then the suction rate per unit time is 3.33 L / s.

[0016] Perform a Fast Fourier Transform (FFT) on the discrete sampling sequence of internal cavity pressure within the current control cycle to convert the pressure data in the time domain into frequency domain information, and extract the energy peak value in the frequency range of 0 to 5 Hz.

[0017] S2, in the initial exhaust section, the vacuum intensity is determined based on the pumping volume per unit time and the pressure change rate, the opening degree of the vacuum valve and the duration of its action are adjusted, and the pressure response delay time after the opening degree change is recorded.

[0018] A fuzzy logic controller is constructed. The membership function of the pumping volume and pressure change rate per unit time is defined by calling the skfuzzy library. Fuzzy inference is performed using a preset fuzzy rule library, and the centroid method is used to defuzzify and output the vacuum intensity command value.

[0019] The target vacuum valve opening is linearly calculated based on the vacuum intensity command value. The valve action duration is obtained by dividing the difference between the current opening and the target opening by the valve action rate. The programmable logic controller (PLC) converts the target vacuum valve opening into a 4-20mA analog control signal and sends it to the proportional servo vacuum valve. A hardware timer is started the instant the control signal is sent to monitor the pressure signal inside the mixing container in real time. The timer stops when the absolute decrease in internal pressure exceeds a set threshold for five consecutive sampling cycles. The timer count is read as the pressure response delay time after the opening change.

[0020] In some implementations, the process of determining the vacuum intensity based on the pumping volume per unit time and the pressure change rate in the initial exhaust section, adjusting the vacuum valve opening and duration of action, and recording the pressure response delay time after the opening change includes: Based on the pumping volume Q per unit time and the pressure change rate Calculate the vacuum strength The formula is: in, and The set weighting coefficients, This represents the rated maximum pumping capacity per unit time. This is the rated maximum value of the pressure change rate; Target vacuum valve opening calculated based on vacuum pumping intensity. ,in This is the proportionality coefficient. Basic constants; Based on the target opening Current opening degree and valve operating rate Calculate valve action duration : Record the time difference from the moment the valve action command is issued until the absolute value of the rate of change of internal pressure exceeds a set threshold, and use this time difference as the pressure response delay time. .

[0021] The controller acquires the pumping volume Q per unit time from the pipeline flow meter in real time, for example, if the measured value is 200 L / min. The value is 500 L / min and the absolute value of the real-time pressure change rate calculated. For example, 0.1 MPa / s, this equipment The pressure is 0.5 MPa / s, and both are normalized and multiplied by weighting coefficients pre-calibrated through bench tests. Under the preferred implementation parameters, Set to 0.6. The value is set to 0.4, and the current comprehensive vacuum intensity index is calculated using a weighted fusion algorithm. This index will be used as the independent variable, combined with a preset proportional coefficient. For example, set to 0.8 and the basic opening constant. For example, 10%, the hardware-executable target vacuum valve opening is calculated through linear mapping. For example, the calculated actual target opening is 35.6%.

[0022] After entering the hardware execution level, retrieve the current actual opening degree of the vacuum proportional valve. Assuming the current rate is 40%, and considering the inherent and constant operating rate of this type of valve... For example, 5% movement per second. By dividing the absolute value of the difference between the target and the current opening by the movement rate, the command duration for maintaining the valve movement can be calculated. In the example above, the action needs to last 0.88 seconds. Due to the delay inherent in industrial exhaust piping and gas hydrodynamics, the controller triggers a microsecond-level timer simultaneously with the action command. The timer stops when the high-frequency pressure sensor detects that the absolute value of the internal pressure change rate exceeds the preset response threshold of 0.05 MPa / s. The time difference obtained at this point, if measured to be 1.2 seconds, is the pressure response delay time. .

[0023] S3, in the bubble release section, an alternating control sequence of pumping and stopping is generated based on the pressure fluctuation characteristics, and the actual pumping duration and actual stopping duration are corrected by combining the pressure response delay time.

[0024] A linear mapping function between pressure fluctuation characteristics and duty cycle is established. When the pressure fluctuation characteristics increase, the pumping duty cycle is reduced proportionally to prevent material boiling and overflow. Based on the calculated duty cycle, a pulse width modulation generation algorithm is used to construct a basic square wave control sequence that alternates between high-level pumping states and low-level stop pumping states.

[0025] By using a time compensation mechanism, the pumping time in the basic sequence is added to the pressure response delay time recorded in the previous stage to obtain the corrected actual pumping time. At the same time, the pumping stop time in the basic sequence is added to the pressure response delay time to obtain the corrected actual pumping stop time. This ensures that the actual pumping energy acting on the vacuum chamber is consistent with the theoretical set value. The corrected alternating control sequence is then used to periodically open and close the solenoid coil of the vacuum isolation valve through a digital output module.

[0026] In some implementations, during the bubble release phase, generating an alternating control sequence for pumping and stopping pumping based on pressure fluctuation characteristics, and correcting the actual pumping duration and actual stopping duration using the pressure response delay time, includes: Based on pressure fluctuation characteristics Match the preset control range to obtain the basic pumping time. With basic stop pumping duration ; Utilizing pressure response delay time The base duration is adjusted to calculate the actual pumping time. Actual duration of pumping stop : in, and The compensation coefficient is greater than 0, which is used to generate an alternating control sequence and execute it.

[0027] In the bubble release section of material degassing, the tiny gases encapsulated in the high-viscosity liquid are prone to rapid expansion and accumulation due to external pressure drop. Continuous vacuuming can easily lead to severe material overflow. A square wave pulse control sequence alternating between vacuuming and stopping is used. The pressure fluctuation characteristics of the previous sampling period are evaluated. This data is then compared and looked up in a multi-segment mapping table pre-stored within the controller. For example, when the characteristic value indicates severe fluctuations, such as... When the pressure reaches 0.03 MPa, a set of conservative time parameters is extracted, including the baseline pumping time. The basic pause duration is a relatively short 5.0 seconds. The longer duration of 8.0s allows sufficient relaxation buffer time for the bubbles inside the polymer solution to rise and burst.

[0028] The stress response delay time obtained during the initial phase is invoked. Assuming a time interval of 1.2 seconds, time-domain compensation is performed. The algorithm utilizes a dedicated compensation coefficient greater than 0. and In practical engineering, the best option is selected. Set to 1.0 to fully compensate for the air extraction delay. The timeout for stopping the airtightness is set to 0.5 to compensate for half of the delay, and the calculated physical lag delay is weighted and added to the base duration. After correction, the actual duration of the generated air extraction command is... Updated to 6.2s, actual duration of the stop pumping command. Updated to 8.6s. The controller issues valve on / off commands according to the corrected alternating timing sequence. Comparison of the basic and compensated command durations for evacuation and evacuation stop actions before and after pressure response delay compensation: Figure 3 As shown, the control command after delay compensation can match the characteristics of pipeline pressure response and eliminate control errors caused by timing deviations.

[0029] S4, in the viscosity ramp-up section, the target pressure holding range is determined based on the torque fluctuation amplitude, jacket temperature and pressure change rate, and the stirring speed and vacuum valve opening are coordinated and adjusted to maintain the pressure within the target pressure holding range.

[0030] The PyTorch deep learning framework is used to load a pre-trained multilayer perceptron neural network model. The torque fluctuation amplitude, jacket temperature, and pressure change rate at the current moment are normalized to their maximum and minimum values ​​before being input into the input layer of the multilayer perceptron neural network. After forward propagation through multiple hidden layers with linear rectified activation functions, the output layer outputs the upper and lower limits of the target pressure maintenance range suitable for the current material viscosity.

[0031] The structure of the multilayer perceptron neural network model is as follows: the input layer contains 3 nodes, corresponding to the normalized torque fluctuation amplitude, jacket temperature, and pressure change rate, respectively; the hidden layers typically consist of 2-3 layers, each containing 32-64 neurons, using the ReLU activation function to introduce nonlinearity and alleviate gradient vanishing; the output layer contains 2 nodes, outputting the lower limit of the target pressure maintenance interval. and upper limit value The output layer does not use an activation function to allow continuous values ​​to be output. The training method involves collecting input features and corresponding manually labeled optimal pressure ranges at various points in the historical mixing process. These ranges are calibrated by process experts based on degassing effects, product quality, etc., forming a training set. Mean squared error is used as the loss function, and the Adam optimizer is used for mini-batch gradient descent training. Weights and biases are updated through backpropagation. Early stopping and validation set monitoring can be used during training to prevent overfitting, ultimately resulting in a regression model that can output the target pressure range in real time.

[0032] A model predictive controller is constructed by calling the CasADi optimization library. The objective function is to minimize the mean square error between the predicted internal pressure and the target pressure within the range of the median. The maximum motor power and the maximum valve action rate are used as hard constraints. The constrained quadratic programming problem is solved using the interior point method. The optimal stirring speed increment and the optimal vacuum valve opening increment in the future control sequence are calculated in real time. The first set of optimal increments is sent to the frequency converter and servo valve actuator to achieve multivariable decoupling and coordinated control.

[0033] In some implementations, the step of determining a target pressure holding range based on the torque fluctuation amplitude, jacket temperature, and pressure change rate during the viscosity ramp-up phase, and coordinating the adjustment of the stirring speed and vacuum valve opening to maintain the pressure within the target pressure holding range, includes: Utilizing jacket temperature Calculate the temperature compensation coefficient Combined with torque fluctuation amplitude and pressure change rate Determine the lower limit of the target pressure holding range. and upper limit : in, This is the ultimate pressure threshold. This is the torque influence coefficient. For bias pressure, For intervals with a magnification ratio greater than 1, The coefficient representing the influence of the rate of change of pressure; When the real-time internal pressure At that time, the vacuum valve opening is increased in fixed steps; When the real-time internal pressure At that time, the vacuum valve opening is reduced by a fixed step size, and the stirring speed is reduced proportionally until the internal pressure rises back to the target pressure holding range.

[0034] When the mixing process progresses to the viscosity ramp-up stage, the material properties undergo phase change or polymerization, and the heat generated by molecular friction within the fluid intensifies. A vacuum maintenance window needs to be established to prevent the erroneous extraction of low-boiling-point solvents or surface crusting caused by high temperatures. During this stage, the real-time jacket temperature transmitted by the thermocouples is read. For example, if the temperature is detected to have risen to 55°C, a temperature compensation coefficient can be calculated based on the equipment's thermal conductivity characteristics. Let's assume the value here is 1.1. Combined with real-time detection of torque fluctuation amplitude... If the fluctuation reaches 15 N·m, and the current rate of change of pressure... For example, -0.02 MPa / s, take its absolute value of 0.02 MPa / s, within the absolute pressure threshold of the equipment. For example, deducting safety redundancy from a base of 0.01 MPa. This is based on the set torque influence coefficient. For example, 0.0002 and bias pressure For example, 0.002 MPa, and the influence coefficient of pressure change rate. For example, under a pressure of 0.05, through multiplication chain calculations, a lower pressure limit is obtained that neither boils nor fails to remove residual gas. For example, 0.0055 MPa. Then use a range magnification ratio greater than 1. The preferred value is between 1.2 and 1.5, for example, set to 1.3, to calculate the upper pressure limit. It is 0.00715 MPa. In the above formula, when the torque fluctuation amplitude... An increase in pressure indicates increased material viscosity or increased mixing resistance. In this case, a lower pressure limit is required, i.e., a stronger vacuum, to help the bubbles inside the highly viscous material overcome surface tension and escape; hence, this term is negative. When the pressure drops too quickly, i.e. When the value is negative and has a large absolute value, to prevent excessive suction from causing a large loss of low-boiling-point components, by... The lower limit of pressure is appropriately raised to provide safety compensation.

[0035] After defining a pressure maintenance range of 0.0055~0.00715MPa, if the real-time pressure of the container decreases due to the release of residual microbubbles... If the pressure rises above the 0.00715MPa upper limit, the controller will not change the status of the main stirring motor, but will continuously open the vacuum valve in a fixed step size per cycle, such as 2%, to smooth out the sudden increase in gas pressure. Conversely, if the strong suction causes the gas pressure to drop sharply and fall below the 0.0055MPa lower limit, the warning mechanism will determine that there is a large loss of low-boiling-point components and a risk of dehydration. It will instruct the vacuum valve to reduce its opening by 3% each time to prevent damage, and will simultaneously switch to the dual protection mode. This will force the inverter of the main stirring motor to reduce the speed by a predetermined ratio, such as 10% each time, reducing the original stirring speed of 50rpm to 45rpm to reduce the phase breakage induced by high shear, until the high-frequency sensor reports that the internal pressure has safely climbed up and stabilized within the calculated target range.

[0036] S5. In the dense and stable section, a stability criterion is constructed based on the average pressure, the degree of pressure fluctuation, and the continuous range. When the stability condition is met, the target vacuum level is maintained. When the stability condition is not met, the opening of the vacuum valve is adjusted based on the direction of the pressure difference.

[0037] A first-in-first-out (FIFO) sliding data queue of length 100 is used to store the latest acquired intracavity pressure time series. A statistical algorithm library is called to calculate the arithmetic mean of the data in this queue as the pressure mean, the standard deviation of the data as the pressure fluctuation level, and the difference between the maximum and minimum pressure values ​​within a window as the continuous range. A multi-condition joint logic criterion is constructed. When the absolute value of the deviation between the pressure mean and the target vacuum level is less than a set threshold, the pressure fluctuation level is less than a set stability threshold, and the continuous range is less than a set stability threshold, the stability condition is considered met. At this time, a position locking command is triggered to fix the vacuum valve opening at the current position to maintain the target vacuum level. If any of the above conditions are not met, the stability condition is considered not met. In this case, the algebraic difference between the current pressure and the target pressure is calculated as the pressure difference direction, and a linear proportional adjustment algorithm is used to calculate the vacuum valve opening adjustment amount based on the pressure difference direction.

[0038] In some implementations, in the dense and stable section, a stability criterion is constructed based on the average pressure, the degree of pressure fluctuation, and the continuous range. When the stability condition is met, the target vacuum level is maintained; when the stability condition is not met, the vacuum valve opening is adjusted based on the pressure difference direction. This includes: Within the set historical sampling window, extract N pressure sampling points. Calculate the average pressure With pressure standard deviation : Extract the maximum and minimum pressure values ​​within the window and calculate the continuous range. ; when , and When all values ​​are less than the set stable threshold, the stability condition is determined to be met, and the current vacuum valve opening is locked. When the stability condition is not met, based on the current pressure With target pressure Calculate the deviation direction of the vacuum valve opening adjustment amount : in, The current initial opening degree, This is the directional feedback gain coefficient.

[0039] When the material enters the compaction and stabilization stage of the mixing process, its internal structure has become highly homogeneous and dense. External intervention needs to be minimized. Therefore, the focus shifts from extraction to small disturbance mitigation control based on statistical analysis of large sample data. In the actual calculation stage, the controller opens a sliding historical sampling window of constant length in its internal memory, such as capturing N=100 consecutive pressure sampling points generated within the last 10 seconds. Offline rapid statistical analysis was performed on this set of high-frequency sampled data to calculate the overall average pressure within the current window in real time. For example, 0.0035 MPa and the pressure standard deviation used to represent the tendency to deviate from the center. For example, 0.00015 MPa. To detect sudden changes in the polarity of instantaneous noise, the absolute maximum and minimum values ​​are simultaneously selected from the same window group and subtracted to obtain the continuous range representing the amplitude boundary. For example, the measured range is 0.0004 MPa.

[0040] when , and If all values ​​are below the set extreme stability threshold, the valve opening is locked to maintain a tight and stable state; if internal leakage in the vacuum system or trace gas production causes statistical indicators to exceed the standard and disrupt the stability criterion, the current pressure will be adjusted. With the established target limit pressure If the target pressure is to remain stable at 0.0030 MPa, and the current deviation is 0.0040 MPa, input the above formula, along with the current initial opening. Assuming it is 12%, the gain coefficient is obtained through a very high resolution feedback coefficient with positive and negative directional characteristics. For example, setting the amplification factor to 2000% / MPa, the correction increment is calculated. Based on this, it is determined that the system needs to issue an adjustment command of 14%. It uses a gentle and targeted micro-motion correction to push the system back to a dense and stable state.

[0041] S6, when switching between adjacent stages, corrects the target pressure starting point, vacuum valve opening initial value, stirring speed initial value and control cycle of the next stage based on the end state parameters of the previous stage.

[0042] The state transition interrupt service routine is triggered the instant the stage label output by the support vector machine model changes. The average pressure change rate, average vacuum valve opening, and average stirring speed of the last five control cycles of the previous stage are extracted as the end-stage state parameters. A smooth transition algorithm based on exponentially weighted moving average is invoked. The preset target pressure starting point of the next stage is added to the actual pressure value at the end of the previous stage according to weights to obtain the corrected target pressure starting point. The average vacuum valve opening of the previous stage is multiplied by a gain coefficient and added to the empirical opening value of the next stage to calculate the corrected initial value of the vacuum valve opening. Similarly, the average stirring speed is used to perform weighted interpolation with the process reference value to obtain the corrected initial value of the stirring speed. Simultaneously, the Euclidean distance between the state space vectors before and after the stage switch is calculated. The basic control cycle parameters of the next stage are scaled inversely according to the magnitude of this Euclidean distance, resulting in a smooth actuator motion curve during the cross-stage transition to avoid material defects caused by sudden pressure changes.

[0043] In some implementations, when switching between adjacent stages, the target pressure starting point, initial value of vacuum valve opening, initial value of stirring speed, and control cycle of the next stage are corrected based on the end-state parameters of the previous stage, including: Obtain the end cavity pressure at the end of the previous stage Terminal vacuum valve opening End stirring speed and end-point control cycle ; The initial parameters of the next stage are calculated and assigned values ​​using a preset stage state transition matrix. The target pressure starting point for the next stage is revised to ,in This is the preset differential pressure compensation amount for the corresponding stage; The initial value of the vacuum valve opening in the next stage is corrected to: ,in This is the openness inheritance coefficient; The initial value of the stirring speed in the next stage is corrected to ,in This is the speed inheritance coefficient; The control cycle for the next stage is revised to ,in This is the periodic adjustment coefficient.

[0044] Within a millisecond interval after the current step's end condition is met, triggering a segment switch, all key equipment physical parameters from the previous stage's end are extracted. For example, at the split point from the bubble release stage to the viscosity ramp-up stage, the controller latches the end-cavity pressure. The pressure is 0.06 MPa, and the opening degree of the terminal vacuum valve is... Remain at 30%, end stirring speed The speed is fixed at 60 rpm, and the end control cycle of the previous application is also fixed. It lasts for 5 seconds.

[0045] After obtaining the initial image parameters, a seamless assignment mapping is performed using the state transition operation matrix specifically designed for the target switching phase. The target pressure starting point of the new segment is changed from the original... Superimposed micro-negative pressure difference compensation amount in the matrix For example, -0.005MPa is flexibly corrected to 0.055MPa; similarly, the initial vacuum valve command of the new process step applies the opening inheritance coefficient. The initial opening was set to a more moderate 0.8, adjusting the original 30% opening to a new initial value of 24%. Regarding the internal dynamic transmission of materials, a rotational speed inheritance coefficient of 0.9 was used. The inverter is controlled to smoothly continue the previous kinetic energy input at a new speed of 54 rpm. Furthermore, given the differences in data refresh rates required at different stages, a cycle adjustment coefficient of 0.4 is applied. The original 5-second long control cycle is compressed and mapped to a 2-second control frequency band. This process avoids the oscillation dead zone caused by zeroing and restarting, and achieves smooth integration and relay across process stages.

[0046] Two identical industrial-grade 500-liter dual planetary mixing units were configured as the experimental and control groups, respectively, to process the same batch and formulation of high-viscosity thermally conductive structural adhesive for power batteries. The initial viscosity of both materials was 150,000 mPa·s, the jacket heating temperature was set to 55°C, the basic stirring speed was set to 50 rpm, and the air extraction flow rate was 300 liters per minute. The experimental group adopted the multi-stage vacuum decoupling control scheme of this invention, which has pressure response delay compensation and state boundary inheritance mechanisms; the control group adopted a traditional fixed-sequence single-loop proportional-integral-derivative control scheme, which lacks time-difference blind zone feedforward and pressure interval decoupling mechanisms. The entire test was conducted synchronously under standard atmospheric pressure at 25°C.

[0047] During the degassing and degassing stages, the experimental group calculated a pressure response delay of 1.2s using high-frequency acquisition features. This was corrected to generate an alternating sequence of 6.2s evacuation and 8.6s evacuation stop. During the viscosity ramp-up phase, the target pressure was stably locked within the range of 0.0055 to 0.00715 MPa. The opening degree remained at 24% during process switching, with no liquid overflow throughout the process. The total degassing time for a single batch in the experimental group was only 42 minutes, and the residual microbubble count in the finished product was 3 per 100 grams. The control group, due to mechanical lag and inability to correct the timing, performed a fixed 5s evacuation stop during the bubble release phase, resulting in a brief pressure overload and slight material sticking. The opening degree was reset to zero and restarted during the transition, causing the pressure to drop to 0.002 MPa. The residual microbubble count in the finished product reached 21 per 100 grams, and the total time for a single batch was extended to 65 minutes due to the handling of overflowing material. The comparison results between the experimental group and the control group using the traditional fixed-timing PID scheme in terms of the two core indicators—total process time per batch and residual microbubble count in the finished product—are as follows: Figure 4 As shown.

[0048] The calculation of pressure fluctuation characteristics and response delay cancels out the physical time difference blind zone. The cross-segment inherited state transition matrix eliminates the step-change dead zone, ensuring that low-boiling-point solvents are not accidentally pumped and that the internal hydrodynamic state remains consistent. Compared with traditional methods, this application reduces the degassing operation time by 35%, reduces the residual defect rate of microbubbles in the finished product by 85%, and achieves stable operation of the rheological mixing equipment under complex working conditions.

[0049] In this application, a vacuum mixing process optimization system includes the following modules: The module is used to collect signals of pressure inside the mixing container, jacket temperature, material torque, stirring speed and air flow rate, calculate pressure change rate, torque fluctuation amplitude, air volume per unit time and pressure fluctuation characteristics, and divide the vacuum mixing process into initial exhaust section, bubble release section, viscosity rise section and dense stabilization section. The first correction module is used to determine the vacuum intensity based on the pumping volume per unit time and the pressure change rate in the initial exhaust section, adjust the opening degree and duration of the vacuum valve, and record the pressure response delay time after the opening degree changes; in the bubble release section, it generates an alternating control sequence of pumping and stopping based on the pressure fluctuation characteristics, and corrects the actual pumping duration and actual stopping duration in combination with the pressure response delay time. The determination module is used to determine the target pressure holding range based on the torque fluctuation amplitude, jacket temperature and pressure change rate during the viscosity ramp-up section, and coordinate the stirring speed and vacuum valve opening to maintain the pressure within the target pressure holding range; during the dense and stable section, it constructs a stability criterion based on the average pressure, pressure fluctuation degree and continuous range, and maintains the target vacuum level when the stability conditions are met, and adjusts the vacuum valve opening based on the pressure difference direction when the stability conditions are not met. The second correction module is used to correct the target pressure start point, vacuum valve opening initial value, stirring speed initial value, and control cycle of the next stage based on the end state parameters of the previous stage when switching between adjacent stages.

[0050] In some implementations, the acquisition of signals such as the pressure inside the mixing container, jacket temperature, material torque, stirring speed, and air extraction flow rate, and the calculation of pressure change rate, torque fluctuation amplitude, air extraction volume per unit time, and pressure fluctuation characteristic quantities, includes: Discrete sampled values ​​of the current internal cavity pressure and material torque are extracted synchronously at fixed time intervals; A first-order backward difference equation is established to calculate the rate of change of internal cavity pressure; The current sampling time is shifted forward by 3 seconds to form the sliding sampling interval. The maximum and minimum values ​​of the material torque within the sliding sampling interval are extracted, and the difference between the two values ​​is used to obtain the torque fluctuation amplitude. The pumping flow rate signal is numerically integrated within a unit time window to calculate the pumping volume per unit time. Perform a Fast Fourier Transform on the discrete sampled values ​​of the internal pressure within the current control cycle, and extract the maximum amplitude in the 0 to 5 Hz frequency band as the pressure fluctuation characteristic.

[0051] In some implementations, the process of determining the vacuum intensity based on the pumping volume per unit time and the pressure change rate in the initial exhaust section, adjusting the vacuum valve opening and duration of action, and recording the pressure response delay time after the opening change includes: Based on the pumping volume Q per unit time and the pressure change rate Calculate the vacuum strength The formula is: in, and The set weighting coefficients, This represents the rated maximum pumping capacity per unit time. This is the rated maximum value of the pressure change rate; Target vacuum valve opening calculated based on vacuum pumping intensity. ,in This is the proportionality coefficient. Basic constants; Based on the target opening Current opening degree and valve operating rate Calculate valve action duration : Record the time difference from the moment the valve action command is issued until the absolute value of the rate of change of internal pressure exceeds a set threshold, and use this time difference as the pressure response delay time. .

[0052] In some implementations, during the bubble release phase, generating an alternating control sequence for pumping and stopping pumping based on pressure fluctuation characteristics, and correcting the actual pumping duration and actual stopping duration using the pressure response delay time, includes: Based on pressure fluctuation characteristics Match the preset control range to obtain the basic pumping time. With basic stop pumping duration ; Utilizing pressure response delay time The base duration is adjusted to calculate the actual pumping time. Actual duration of pumping stop : in, and The compensation coefficient is greater than 0, which is used to generate an alternating control sequence and execute it.

[0053] In some implementations, the step of determining a target pressure holding range based on the torque fluctuation amplitude, jacket temperature, and pressure change rate during the viscosity ramp-up phase, and coordinating the adjustment of the stirring speed and vacuum valve opening to maintain the pressure within the target pressure holding range, includes: Utilizing jacket temperature Calculate the temperature compensation coefficient Combined with torque fluctuation amplitude and pressure change rate Determine the lower limit of the target pressure holding range. and upper limit : in, This is the ultimate pressure threshold. This is the torque influence coefficient. For bias pressure, For intervals with a magnification ratio greater than 1, The coefficient representing the influence of the rate of change of pressure; When the real-time internal pressure At that time, the vacuum valve opening is increased in fixed steps; When the real-time internal pressure At that time, the vacuum valve opening is reduced by a fixed step size, and the stirring speed is reduced proportionally until the internal pressure rises back to the target pressure holding range.

[0054] In some implementations, in the dense and stable section, a stability criterion is constructed based on the average pressure, the degree of pressure fluctuation, and the continuous range. When the stability condition is met, the target vacuum level is maintained; when the stability condition is not met, the vacuum valve opening is adjusted based on the pressure difference direction. This includes: Within the set historical sampling window, extract N pressure sampling points. Calculate the average pressure With pressure standard deviation : Extract the maximum and minimum pressure values ​​within the window and calculate the continuous range. ; when , and When all values ​​are less than the set stable threshold, the stability condition is determined to be met, and the current vacuum valve opening is locked. When the stability condition is not met, based on the current pressure With target pressure Calculate the deviation direction of the vacuum valve opening adjustment amount : in, The current initial opening degree, This is the directional feedback gain coefficient.

[0055] In some implementations, when switching between adjacent stages, the target pressure starting point, initial value of vacuum valve opening, initial value of stirring speed, and control cycle of the next stage are corrected based on the end-state parameters of the previous stage, including: Obtain the end cavity pressure at the end of the previous stage Terminal vacuum valve opening End stirring speed and end-point control cycle ; The initial parameters of the next stage are calculated and assigned values ​​using a preset stage state transition matrix. The target pressure starting point for the next stage is revised to ,in This is the preset differential pressure compensation amount for the corresponding stage; The initial value of the vacuum valve opening in the next stage is corrected to: ,in This is the openness inheritance coefficient; The initial value of the stirring speed in the next stage is corrected to ,in This is the speed inheritance coefficient; The control cycle for the next stage is revised to ,in This is the periodic adjustment coefficient.

[0056] It should be understood that the various forms of processes shown above can be used to rearrange, add, or delete steps. For example, the steps described in this disclosure can be executed in parallel, sequentially, or in different orders, as long as the desired result of the technical solution disclosed in this disclosure can be achieved, and this is not limited herein.

[0057] The specific embodiments described above do not constitute a limitation on the scope of protection of this disclosure. Those skilled in the art should understand that various modifications, combinations, sub-combinations, and substitutions can be made according to design requirements and other factors. Any modifications, equivalent substitutions, and improvements made within the principles of this disclosure should be included within the scope of protection of this disclosure.

Claims

1. A method for optimizing a vacuum mixing process, characterized in that, include: The pressure inside the mixing container, jacket temperature, material torque, stirring speed and air flow rate are collected. The pressure change rate, torque fluctuation amplitude, air pumping volume per unit time and pressure fluctuation characteristics are calculated. The vacuum mixing process is divided into the initial exhaust section, bubble release section, viscosity rise section and dense stabilization section. In the initial exhaust section, the vacuum intensity is determined based on the pumping volume per unit time and the pressure change rate, the opening degree of the vacuum valve and the duration of its action are adjusted, and the pressure response delay time after the opening degree change is recorded; in the bubble release section, an alternating control sequence of pumping and stopping is generated based on the pressure fluctuation characteristics, and the actual pumping duration and the actual stopping duration are corrected in combination with the pressure response delay time. During the viscosity ramp-up phase, the target pressure holding range is determined based on the torque fluctuation amplitude, jacket temperature, and pressure change rate. The stirring speed and vacuum valve opening are coordinated and adjusted to maintain the pressure within the target pressure holding range. During the dense and stable phase, a stability criterion is constructed based on the average pressure, pressure fluctuation degree, and continuous range. When the stability conditions are met, the target vacuum level is maintained. When the stability conditions are not met, the vacuum valve opening is adjusted based on the pressure difference direction. When switching between adjacent stages, the target pressure starting point, vacuum valve opening initial value, stirring speed initial value and control cycle of the next stage are corrected based on the end state parameters of the previous stage.

2. The method according to claim 1, characterized in that, The system collects signals of pressure inside the mixing container, jacket temperature, material torque, stirring speed, and air extraction flow rate, and calculates pressure change rate, torque fluctuation amplitude, air extraction volume per unit time, and pressure fluctuation characteristic quantities, including: Discrete sampled values ​​of the current internal cavity pressure and material torque are extracted synchronously at fixed time intervals; A first-order backward difference equation is established to calculate the rate of change of internal cavity pressure; The current sampling time is shifted forward by 3 seconds to form the sliding sampling interval. The maximum and minimum values ​​of the material torque within the sliding sampling interval are extracted, and the difference between the two values ​​is used to obtain the torque fluctuation amplitude. The pumping flow rate signal is numerically integrated within a unit time window to calculate the pumping volume per unit time. Perform a Fast Fourier Transform on the discrete sampled values ​​of the internal pressure within the current control cycle, and extract the maximum amplitude in the 0 to 5 Hz frequency band as the pressure fluctuation characteristic.

3. The method according to claim 2, characterized in that, In the initial exhaust section, the vacuum intensity is determined based on the pumping volume per unit time and the pressure change rate. The opening degree and duration of the vacuum valve are adjusted, and the pressure response delay time after the opening degree change is recorded, including: Based on the pumping volume Q per unit time and the pressure change rate Calculate the vacuum strength The formula is: in, and The set weighting coefficients, This represents the rated maximum pumping capacity per unit time. This is the rated maximum value of the pressure change rate; Target vacuum valve opening calculated based on vacuum pumping intensity. ,in This is the proportionality coefficient. Basic constants; Based on the target opening Current opening degree and valve operating rate Calculate valve action duration : Record the time difference from the moment the valve action command is issued until the absolute value of the rate of change of internal pressure exceeds a set threshold, and use this time difference as the pressure response delay time. .

4. The method according to claim 3, characterized in that, In the bubble release section, an alternating control sequence for pumping and stopping is generated based on the pressure fluctuation characteristics, and the actual pumping duration and actual stopping duration are corrected by combining the pressure response delay time, including: Based on pressure fluctuation characteristics Match the preset control range to obtain the basic pumping time. With basic stop pumping duration ; Utilizing pressure response delay time The base duration is adjusted to calculate the actual pumping time. Actual duration of pumping stop : in, and The compensation coefficient is greater than 0, which is used to generate an alternating control sequence and execute it.

5. The method according to claim 1, characterized in that, During the viscosity ramp-up phase, the target pressure maintenance range is determined based on the torque fluctuation amplitude, jacket temperature, and pressure change rate. The stirring speed and vacuum valve opening are then coordinated to maintain the pressure within this target pressure maintenance range, including: Utilizing jacket temperature Calculate the temperature compensation coefficient Combined with torque fluctuation amplitude and pressure change rate Determine the lower limit of the target pressure holding range. and upper limit : in, This is the ultimate pressure threshold. This is the torque influence coefficient. For bias pressure, For intervals with a magnification ratio greater than 1, The coefficient representing the influence of the rate of change of pressure; When the real-time internal pressure At that time, the vacuum valve opening is increased in fixed steps; When the real-time internal pressure At that time, the vacuum valve opening is reduced by a fixed step size, and the stirring speed is reduced proportionally until the internal pressure rises back to the target pressure holding range.

6. The method according to claim 1, characterized in that, In the dense and stable section, a stability criterion is constructed based on the average pressure, the degree of pressure fluctuation, and the continuous range. When the stability condition is met, the target vacuum level is maintained; when the stability condition is not met, the vacuum valve opening is adjusted based on the pressure difference direction, including: Within the set historical sampling window, extract N pressure sampling points. Calculate the average pressure With pressure standard deviation : Extract the maximum and minimum pressure values ​​within the window and calculate the continuous range. ; when , and When all values ​​are less than the set stable threshold, the stability condition is determined to be met, and the current vacuum valve opening is locked. When the stability condition is not met, based on the current pressure With target pressure Calculate the deviation direction of the vacuum valve opening adjustment amount : in, The current initial opening degree, This is the directional feedback gain coefficient.

7. The method according to claim 1, characterized in that, When switching between adjacent stages, the target pressure starting point, initial value of vacuum valve opening, initial value of stirring speed, and control cycle of the next stage are corrected based on the end state parameters of the previous stage, including: Obtain the end cavity pressure at the end of the previous stage Terminal vacuum valve opening End stirring speed and end-point control cycle ; The initial parameters of the next stage are calculated and assigned values ​​using a preset stage state transition matrix. The target pressure starting point for the next stage is revised to ,in This is the preset differential pressure compensation amount for the corresponding stage; The initial value of the vacuum valve opening in the next stage is corrected to: ,in This is the openness inheritance coefficient; The initial value of the stirring speed in the next stage is corrected to ,in This is the speed inheritance coefficient; The control cycle for the next stage is revised to ,in This is the periodic adjustment coefficient.

8. A vacuum mixing process optimization system, characterized in that, Includes the following modules: The module is used to collect signals of pressure inside the mixing container, jacket temperature, material torque, stirring speed and air flow rate, calculate pressure change rate, torque fluctuation amplitude, air volume per unit time and pressure fluctuation characteristics, and divide the vacuum mixing process into initial exhaust section, bubble release section, viscosity rise section and dense stabilization section. The first correction module is used to determine the vacuum intensity based on the pumping volume per unit time and the pressure change rate in the initial exhaust section, adjust the opening degree and duration of the vacuum valve, and record the pressure response delay time after the opening degree changes; in the bubble release section, it generates an alternating control sequence of pumping and stopping based on the pressure fluctuation characteristics, and corrects the actual pumping duration and actual stopping duration in combination with the pressure response delay time. The determination module is used to determine the target pressure holding range based on the torque fluctuation amplitude, jacket temperature and pressure change rate during the viscosity ramp-up section, and coordinate the stirring speed and vacuum valve opening to maintain the pressure within the target pressure holding range; during the dense and stable section, it constructs a stability criterion based on the average pressure, pressure fluctuation degree and continuous range, and maintains the target vacuum level when the stability conditions are met, and adjusts the vacuum valve opening based on the pressure difference direction when the stability conditions are not met. The second correction module is used to correct the target pressure start point, vacuum valve opening initial value, stirring speed initial value, and control cycle of the next stage based on the end state parameters of the previous stage when switching between adjacent stages.

9. The system according to claim 8, characterized in that, The system collects signals of pressure inside the mixing container, jacket temperature, material torque, stirring speed, and air extraction flow rate, and calculates pressure change rate, torque fluctuation amplitude, air extraction volume per unit time, and pressure fluctuation characteristic quantities, including: Discrete sampled values ​​of the current internal cavity pressure and material torque are extracted synchronously at fixed time intervals; A first-order backward difference equation is established to calculate the rate of change of internal cavity pressure; The current sampling time is shifted forward by 3 seconds to form the sliding sampling interval. The maximum and minimum values ​​of the material torque within the sliding sampling interval are extracted, and the difference between the two values ​​is used to obtain the torque fluctuation amplitude. The pumping flow rate signal is numerically integrated within a unit time window to calculate the pumping volume per unit time. Perform a Fast Fourier Transform on the discrete sampled values ​​of the internal pressure within the current control cycle, and extract the maximum amplitude in the 0 to 5 Hz frequency band as the pressure fluctuation characteristic.

10. The system according to claim 8, characterized in that, In the initial exhaust section, the vacuum intensity is determined based on the pumping volume per unit time and the pressure change rate. The opening degree and duration of the vacuum valve are adjusted, and the pressure response delay time after the opening degree change is recorded, including: Based on the pumping volume Q per unit time and the pressure change rate Calculate the vacuum strength The formula is: in, and The set weighting coefficients, This represents the rated maximum pumping capacity per unit time. This is the rated maximum value of the pressure change rate; Target vacuum valve opening calculated based on vacuum pumping intensity. ,in This is the proportionality coefficient. Basic constants; Based on the target opening Current opening degree and valve operating rate Calculate valve action duration : Record the time difference from the moment the valve action command is issued until the absolute value of the rate of change of internal pressure exceeds a set threshold, and use this time difference as the pressure response delay time. .

Images