A predictive conditioning system for environmental parameters in aseptic fruit pulp filling

Abstract

The present application relates to the field of food engineering and aseptic filling, and discloses a predictive adjustment system for environmental parameters of aseptic filling of fruit pulp, which realizes real-time monitoring and preprocessing of environmental, process and equipment state parameters through a data acquisition module; a data correlation analysis module analyzes and generates key control and lag characteristic parameters in multiple dimensions; an adjustment frequency analysis module determines optimal adjustment cycles and control instructions in combination with parameter fluctuation, self-balancing and deviation state; a parameter target value analysis module identifies standard target values of environmental parameters based on characteristics and cycles; a parameter prediction module predicts future state and generates deviation coefficients and adjustment instructions; a self-adaptive adjustment module cooperatively controls basic and fine adjustment units to dynamically correct to parameter stability; and an adjustment frequency optimization module updates cycles in a closed loop to guarantee stable and efficient filling; the system realizes predictive control of the aseptic environment, and ensures that the environment is always in a controlled state of asepsis and stability during food filling.

Description

Technical Field

[0001] This invention relates to the fields of food engineering and aseptic filling technology, and more specifically to a predictive adjustment system for environmental parameters in aseptic filling of fruit pulp. Background Technology

[0002] In the food and beverage industry, the bottling workshop is a crucial link in the product journey from the production line to the consumer. Air disinfection in the workshop is a fundamental aspect of ensuring food safety. Air, as a significant medium for microbial transmission, allows bacteria, mold, viruses, and other microorganisms to proliferate and spread unchecked if not effectively disinfected. Once these microorganisms adhere to the surface of food or penetrate its interior, they can cause spoilage, shorten shelf life, and even trigger food safety incidents, severely damaging consumer health and corporate reputation. Therefore, maintaining constant and sterile environmental parameters in the bottling workshop (such as temperature, humidity, dust particles, and hydrogen peroxide concentration) is a core control point in the fruit pulp production process.

[0003] Although existing technologies are widely used in the food filling industry, they still have the following significant drawbacks when facing complex and ever-changing aseptic filling environments: Firstly, existing technologies mostly adopt traditional single-loop feedback control, that is, the system only starts the adjustment equipment to intervene when the sensor detects that the temperature rises or the humidity exceeds the standard. This lag mode causes the environmental parameters to be in a non-standard state for a long time during the adjustment period. If the pulp filling speed is fast, the instantaneous environmental fluctuation may cause the entire batch of products to be contaminated with microorganisms. Secondly, existing technologies often treat environmental parameters as independent variables for control, ignoring the coupled influence of process parameters and equipment status on the environment. Adjustments are made only based on environmental deviations, which often result in insufficient or excessive adjustments. This fails to accurately eliminate environmental fluctuations caused by process changes, leading to blind spots in the control of the sterile environment. Third, existing control methods usually adopt a single linear control mode, which lacks the ability to identify fluctuations. When faced with overall trend deviations in environmental parameters (such as overall temperature rise due to seasonal changes) and high-frequency random fluctuations (such as instantaneous airflow disturbances caused by personnel movement or door opening), they cannot perform differentiated treatment, making it difficult for the sterile environment to achieve a high-precision dynamic balance. Fourth, existing technologies often rely on manually set fixed adjustment cycles for inspection and control. Fixed cycles cannot adapt to dynamic changes under different operating conditions, and lack closed-loop evaluation and self-correction of adjustment effects. They cannot optimize future adjustment strategies based on historical actual response data, resulting in control accuracy remaining at the initial level forever. Summary of the Invention

[0004] In order to overcome the above-mentioned defects of the prior art, the present invention provides a predictive adjustment system for environmental parameters of aseptic filling of fruit pulp, so as to solve the problems existing in the background art.

[0005] This invention provides the following technical solution: a predictive adjustment system for environmental parameters in aseptic filling of fruit pulp, comprising: Data acquisition module: It monitors and collects various monitoring data during the aseptic filling process of fruit pulp in real time, and preprocesses the collected monitoring data, including environmental parameters, process parameters and equipment status parameters. Data correlation analysis module: By performing multi-dimensional correlation analysis and feature filtering on environmental parameters, process parameters, and equipment status parameters, it generates key control features and hysteresis characteristic parameters; Adjustment frequency analysis module: Based on hysteresis characteristic parameters, environmental parameters, equipment status parameters and key control characteristics, calculate the real-time fluctuation amplitude and fluctuation correlation of monitoring data, and combine the self-balancing ability of the filling process and the deviation status of current environmental parameters to determine whether to intervene in adjustment, and obtain the optimal adjustment cycle and adjustment control command that conforms to the current working conditions. Parameter target value analysis module: Based on key control characteristics, hysteresis parameters and optimal adjustment period, the module performs time-series trend analysis and fluctuation state identification of environmental parameters to obtain the standard target value for environmental parameter adjustment. Parameter prediction module: Based on the current environmental parameters, historical environmental parameters, and standard target values, it predicts the future trend of environmental parameters to obtain the predicted value of the sterile environment state at the next moment. Then, it compares the predicted value of the sterile environment state at the next moment with the standard target value to obtain the prediction deviation coefficient. Combined with the optimal adjustment cycle, it generates predictive adjustment execution instructions. Adaptive adjustment module: Based on the predictive adjustment execution command, it controls the basic adjustment execution unit and the fine-tuning correction execution unit to work together, and monitors the actual response feedback data of environmental parameters in real time, dynamically correcting the adjustment output intensity until the environmental parameters stabilize at the standard target value; Adjustment frequency optimization module: Analyzes the difference between the actual response feedback data and the predicted value of the sterile environment state at the next moment to obtain the frequency matching index, and updates and corrects the optimal adjustment cycle based on the frequency matching index.

[0006] Preferably, the data acquisition module uses a multi-dimensional sensor array distributed in the cleanroom, laminar flow hood, and surrounding area of ​​the filling valve in the aseptic filling room to monitor and collect environmental parameters, process parameters, and equipment status parameters in real time. The environmental parameters include temperature, humidity, pressure difference, dust particle count, and hydrogen peroxide concentration. The process parameters include pulp flow rate, filling speed, and sterilization intensity. The equipment status parameters include fan speed, valve opening, and operating vibration. The preprocessing of the various monitoring data collected includes: The collected environmental parameters, process parameters, and equipment status parameters are sequentially cleaned, transformed, and standardized.

[0007] Preferably, the data correlation analysis module uses the Pearson correlation coefficient to calculate the correlation between each pair of preprocessed environmental parameters, process parameters and equipment status parameters to obtain the correlation coefficient between the parameters. The calculated correlation coefficients are compared with preset correlation thresholds to filter environmental parameters, process parameters, and equipment status parameters. The specific comparison process is as follows: if the correlation coefficient is greater than or equal to the preset correlation threshold, the corresponding parameter is determined to be a strongly correlated factor and marked as a target data factor for retention; if the correlation coefficient is less than the preset correlation threshold, the corresponding parameter is determined to be a weakly interfering factor and removed. Then, time-series alignment analysis is performed on the retained target data factors. Lag characteristic parameters are extracted by calculating the time difference between each target data factor and the environmental parameter changes. Simultaneously, the data fluctuation trends of each target data factor are integrated to construct a multi-dimensional feature vector, thereby obtaining key control features. The data fluctuation trends include the temperature change rate and humidity oscillation amplitude of environmental parameters, the pulp flow rate fluctuation characteristics of process parameters, and the valve opening oscillation frequency of equipment status parameters. The operation of integrating the data fluctuation trends of each target data factor to construct a multi-dimensional feature vector specifically involves: extracting the data fluctuation trends corresponding to environmental parameters, process parameters, and equipment status parameters respectively, and concatenating the extracted data fluctuation trends according to time sequence to generate a multi-dimensional feature vector containing multi-dimensional fluctuation information.

[0008] Preferably, the adjustment frequency analysis module uses the response delay time of each target data factor in the hysteresis characteristic parameter relative to the change of environmental parameter as the prediction lead time for environmental parameter adjustment, and calculates the deviation influence coefficient between the predicted change trend of environmental parameter in the next adjustment cycle and the standard target value by combining the multi-dimensional feature vector in the key control feature. By using real-time collected environmental parameters and equipment status parameters, the real-time fluctuation amplitude of the monitoring data is calculated, and the fluctuation correlation degree is calculated based on the correlation of the fluctuation trends of each data in the multidimensional feature vector. Then, combined with the self-balancing capability of the filling process, the optimal adjustment cycle is obtained by weighted calculation based on the basic adjustment cycle, the normalized real-time fluctuation amplitude, the normalized fluctuation correlation degree, and the normalized deviation influence coefficient. The calculated deviation influence coefficient is compared with the preset intervention threshold. If the deviation influence coefficient is greater than or equal to the preset intervention threshold, it is determined that the current environmental parameters have deviated from the stable range allowed by the filling process. Using the optimal adjustment cycle as the time reference, an adjustment control command containing the parameter adjustment direction and the target set value is generated and sent to the corresponding air conditioning or purification equipment to perform the adjustment operation. If the deviation influence coefficient is less than the preset intervention threshold, it is determined that the current operating condition meets the process self-balancing requirements. No new adjustment command is generated, and the current monitoring status is maintained.

[0009] Preferably, the parameter target value analysis module extracts and analyzes the real-time monitoring values ​​and parameter change rates of environmental parameters within the current optimal adjustment cycle through multi-dimensional feature vectors in key control features, and performs time-delay extrapolation calculation on the parameter change rate by combining the response delay time in the hysteresis characteristic parameters to obtain the time-delay environmental parameter offset value; then, it identifies and analyzes the current fluctuation state of the environmental parameters through the time-delay environmental parameter offset value, determines whether the fluctuation state is a rising edge, falling edge, or steady-state oscillation, and determines the corresponding dynamic correction amount based on the time-delay environmental parameter offset value and the process safety margin, and calculates the standard target value for environmental parameter adjustment by combining the standard operating condition setpoint with the dynamic correction amount; wherein, the standard target value is used to offset the prediction deviation, so that the environmental parameters return to the standard process range after the response time.

[0010] Preferably, the parameter prediction module performs trend fitting calculations on the environmental parameter change data based on the current environmental parameters, historical environmental parameters, and standard target values ​​to obtain the predicted value of the sterile environment state at the next moment. The deviation between the predicted value of the sterile environment state at the next moment and the standard target value is calculated to obtain the prediction deviation coefficient. The initial adjustment amplitude is then calculated based on the prediction deviation coefficient. The initial adjustment amplitude is used to eliminate the prediction environmental parameter deviation corresponding to the prediction deviation coefficient. The initial adjustment amplitude is then calculated by time-series decomposition according to the optimal adjustment period to obtain the predictive adjustment execution instruction used to suppress abnormal fluctuations in environmental parameters.

[0011] Preferably, the adaptive adjustment module extracts the corresponding initial adjustment amplitude and optimal adjustment period according to the predictive adjustment execution instruction, and sends a trend adjustment instruction to the basic adjustment execution unit and a high-frequency fluctuation correction instruction to the fine-tuning correction execution unit, driving the two to work together. The basic adjustment execution unit is used to perform large-amplitude basic adjustment actions to perform differential adjustment on the overall trend of environmental parameters in order to eliminate trend deviations of environmental parameters. The fine-tuning correction execution unit is used to perform small-amplitude correction adjustment actions to perform differential adjustment on the instantaneous fluctuations of environmental parameters in order to suppress random high-frequency jitter of environmental parameters. The actual response feedback data of environmental parameters are sampled in real time and the rate of change is calculated to obtain the current actual regulation response value. The deviation between the current actual regulation response value and the initial regulation amplitude is calculated to obtain the dynamic feedback correction coefficient. The dynamic feedback correction coefficient is then used to perform real-time correction calculation on the current regulation output intensity until the actual monitoring value of the environmental parameters stably converges to the standard target value.

[0012] Preferably, the frequency optimization module aligns the actual response feedback data with the predicted value of the sterile environment state at the next moment, calculates the response deviation between the actual parameter change caused by the adjustment action and the predicted change, and substitutes the response deviation into the frequency matching degree model for mapping calculation to obtain the frequency matching index; it then compares the frequency matching index with a preset frequency threshold to generate a period deviation correction amount, and uses the period deviation correction amount to continuously update and correct the optimal adjustment period.

[0013] The technical effects and advantages of this invention are as follows: (1) By utilizing current environmental parameters, historical data and standard target values, the system predicts the sterile environment status at the next moment through time-series trend analysis and calculates the prediction deviation coefficient in advance. Combined with the lag characteristic parameters obtained from data correlation analysis, the system can predict the inertia of environmental changes and generate predictive adjustment execution instructions. Intervention is initiated before the environmental parameters actually deviate. This predictive control effectively offsets the physical lag time of the adjustment system, ensuring that the environmental parameters are always locked within the standard range during the high-speed fruit pulp filling process, and eliminating the risk of microbial contamination of the entire batch of products due to instantaneous fluctuations.

[0014] (2) By conducting multi-dimensional correlation analysis and feature screening of environmental parameters, process parameters and equipment status parameters, and generating key control features, it is possible to accurately identify whether environmental fluctuations are caused by external factors or internal process changes, thereby eliminating the control blind spot caused by ignoring process changes. This avoids the inability to eliminate interference due to insufficient adjustment, and also prevents energy waste and system oscillation caused by excessive adjustment.

[0015] (3) By decomposing the adjustment action into a basic adjustment execution unit and a fine-tuning correction execution unit, the basic adjustment unit is responsible for performing large-amplitude basic adjustment actions to eliminate the overall trend deviation of environmental parameters; the fine-tuning correction unit is responsible for performing small-amplitude correction adjustment actions to suppress random high-frequency jitter of environmental parameters. This method can quickly correct large trend deviations and smoothly handle subtle instantaneous fluctuations, overcoming the problem of traditional single adjustment mode being unable to address both aspects, and enabling the sterile environment to achieve and maintain a high-precision dynamic balance.

[0016] (4) By analyzing the real-time fluctuation amplitude, correlation degree and process self-balancing capability, the optimal adjustment cycle that meets the current working conditions is calculated. Then, by comparing the actual response feedback data with the predicted state, the frequency matching index is calculated, and the optimal adjustment cycle is updated and corrected accordingly. This allows the adjustment frequency to be automatically adjusted with the dynamic changes in the production conditions. It also has self-learning capabilities and can continuously optimize future adjustment strategies based on historical actual response data. This breaks through the bottleneck of the control accuracy of existing technologies being limited by the initial setting level and realizes high-precision control that is constantly evolving. Attached Figure Description

[0017] Figure 1 This is a system structure block diagram of the present invention.

[0018] Figure 2 This is a diagram illustrating the method steps of the present invention. Detailed Implementation

[0019] The technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings. In addition, the forms of the various structures described in the following embodiments are merely illustrative. The predictive adjustment system for aseptic filling environmental parameters of fruit pulp involved in the present invention is not limited to the structures described in the following embodiments. All other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0020] like Figure 1 The embodiment shown provides a predictive adjustment system for environmental parameters in aseptic filling of fruit pulp, including: Data acquisition module: It monitors and collects various monitoring data during the aseptic filling process of fruit pulp in real time, and preprocesses the collected monitoring data, including environmental parameters, process parameters and equipment status parameters.

[0021] In this embodiment, the data acquisition module uses a multi-dimensional sensor array distributed in the cleanroom, laminar flow hood, and surrounding area of ​​the filling valve in the aseptic filling room to monitor and collect environmental parameters, process parameters, and equipment status parameters in real time. The environmental parameters include temperature, humidity, pressure difference, dust particle count, and hydrogen peroxide concentration. The process parameters include pulp flow rate, filling speed, and sterilization intensity. The equipment status parameters include fan speed, valve opening, and operating vibration. The preprocessing of the various monitoring data collected includes: The collected environmental parameters, process parameters, and equipment status parameters are sequentially cleaned, transformed, and standardized.

[0022] It should be noted that the specific data acquisition and preprocessing execution process of the data acquisition module is as follows: First, a distributed array of multidimensional sensors is deployed in key areas of the aseptic filling room, namely the central area of ​​the cleanroom, below the air outlet of the laminar flow hood, and the surrounding sealed area of ​​the filling valve. This array synchronously and at high frequency collects environmental parameters (such as temperature, humidity, pressure difference, dust particle count, and hydrogen peroxide concentration), process parameters (such as slurry flow rate, filling speed, and sterilization intensity), and equipment status parameters (such as fan speed, valve opening, and operating vibration) to form an initial monitoring dataset containing multidimensional information. Subsequently, the initial monitoring dataset undergoes data cleaning, specifically using... The principle or binning method identifies and eliminates outliers caused by sensor malfunctions or signal transmission interference. For example, when the temperature sensor value around the filling valve suddenly jumps to a range exceeding three standard deviations above or below the historical mean, it is identified as an outlier and eliminated or smoothed. Next, data conversion is performed. For different types of sensor signals, analog electrical signals are converted into uniform digital quantities, and non-numerical state descriptions (such as the "on / off" state of a fan) are mapped to binary values ​​(such as "1 / 0"). Simultaneously, data from different sampling frequencies are aligned to the standard timestamp of the system's main clock using linear interpolation or resampling methods. For example, fan speed data collected every second is interpolated and aligned with dust particle count data collected every minute. Finally, data standardization is performed using the Z-score standardization formula (…). ) / (where x is the current data value, This is the historical average. The monitoring data of different dimensions are normalized (to standard deviation) to eliminate the distortion caused by different units (e.g., standard deviation). and The magnitude difference caused by this will result in the output of cleaned, transformed, and standardized standard time series data.

[0023] Data correlation analysis module: By performing multi-dimensional correlation analysis and feature filtering on environmental parameters, process parameters and equipment status parameters, key control features and hysteresis characteristic parameters are generated.

[0024] In this embodiment, the data correlation analysis module uses the Pearson correlation coefficient to calculate the correlation between preprocessed environmental parameters, process parameters and equipment status parameters, and obtain the influence correlation coefficient between parameters; The calculated correlation coefficients are compared with preset correlation thresholds to filter environmental parameters, process parameters, and equipment status parameters. The specific comparison process is as follows: if the correlation coefficient is greater than or equal to the preset correlation threshold, the corresponding parameter is determined to be a strongly correlated factor and marked as a target data factor for retention; if the correlation coefficient is less than the preset correlation threshold, the corresponding parameter is determined to be a weakly interfering factor and removed. Then, time-series alignment analysis is performed on the retained target data factors. Lag characteristic parameters are extracted by calculating the time difference between each target data factor and the environmental parameter changes. Simultaneously, the data fluctuation trends of each target data factor are integrated to construct a multi-dimensional feature vector, thereby obtaining key control features. The data fluctuation trends include the temperature change rate and humidity oscillation amplitude of environmental parameters, the pulp flow rate fluctuation characteristics of process parameters, and the valve opening oscillation frequency of equipment status parameters. The operation of integrating the data fluctuation trends of each target data factor to construct a multi-dimensional feature vector specifically involves: extracting the data fluctuation trends corresponding to environmental parameters, process parameters, and equipment status parameters respectively, and concatenating the extracted data fluctuation trends according to time sequence to generate a multi-dimensional feature vector containing multi-dimensional fluctuation information.

[0025] It should be noted that the specific execution process and analysis calculation details of the data correlation analysis module are as follows: Based on the preprocessed data, the Pearson correlation coefficient formula is used to calculate the pairwise correlation between each parameter. The specific calculation formula is as follows: ,in and These represent the monitored values ​​of two different parameters (such as cleanroom temperature and fan speed) at the i-th time point. and These represent the average values ​​of the two parameters over a set time period n, where n represents the total number of sampling points. The correlation coefficient r, reflecting the strength of the linear relationship between the parameters, is calculated using this formula. If the absolute value of r is greater than or equal to a preset correlation threshold (e.g., set to 0.7), a strong correlation is determined between the corresponding parameters, and they are marked as target data factors and retained. If the absolute value of r is less than the preset threshold, it is determined as a weak interference factor and removed. Subsequently, time-series alignment analysis is performed on the retained target data factors. The time difference between each target data factor and the change in environmental parameters is calculated using a cross-correlation function to extract lag characteristic parameters. For example, if the calculation shows that when... Ten seconds after the filling valve opening changes, the hydrogen peroxide concentration in the cleanroom begins to show a significant response; this 10-second interval is extracted as a hysteresis characteristic parameter. Finally, to construct key control features, the following parameters are extracted: temperature change rate (the difference between adjacent temperature values), humidity fluctuation amplitude (the difference between the maximum and minimum humidity values ​​within a set time period), pulp flow rate fluctuation characteristics (the standard deviation of flow rate data) from environmental parameters, and valve opening oscillation frequency (the number of cycles of valve opening signal change per unit time) from equipment status parameters. These extracted data fluctuation trend features are then concatenated according to a unified time axis order to generate a multidimensional feature vector containing multidimensional fluctuation information.

[0026] Adjustment frequency analysis module: Based on hysteresis characteristic parameters, environmental parameters, equipment status parameters, and key control characteristics, it calculates the real-time fluctuation amplitude and fluctuation correlation of monitoring data. Combined with the self-balancing capability of the filling process and the deviation status of current environmental parameters, it determines whether to intervene in the adjustment and obtains the optimal adjustment cycle and adjustment control command that conforms to the current working conditions.

[0027] In this embodiment, the adjustment frequency analysis module uses the response delay time of each target data factor in the hysteresis characteristic parameter relative to the change of environmental parameter as the prediction advance amount of environmental parameter adjustment, and calculates the deviation influence coefficient between the predicted change trend of environmental parameter in the next adjustment cycle and the standard target value by combining the multi-dimensional feature vector in the key control feature. By using real-time collected environmental parameters and equipment status parameters, the real-time fluctuation amplitude of the monitoring data is calculated, and the fluctuation correlation degree is calculated based on the correlation of the fluctuation trends of each data in the multidimensional feature vector. Then, combined with the self-balancing capability of the filling process, the optimal adjustment cycle is obtained by weighted calculation based on the basic adjustment cycle, the normalized real-time fluctuation amplitude, the normalized fluctuation correlation degree, and the normalized deviation influence coefficient. The calculated deviation influence coefficient is compared with the preset intervention threshold. If the deviation influence coefficient is greater than or equal to the preset intervention threshold, it is determined that the current environmental parameters have deviated from the stable range allowed by the filling process. Using the optimal adjustment cycle as the time reference, an adjustment control command containing the parameter adjustment direction and the target set value is generated and sent to the corresponding air conditioning or purification equipment to perform the adjustment operation. If the deviation influence coefficient is less than the preset intervention threshold, it is determined that the current operating condition meets the process self-balancing requirements. No new adjustment command is generated, and the current monitoring status is maintained.

[0028] It should be noted that the specific execution process and analysis calculation details of the frequency adjustment analysis module are as follows: First, using the hysteresis characteristic parameters extracted by the data correlation analysis module, the response delay time (e.g., 5 minutes) of the target data factor (such as fan speed) relative to the change of environmental parameters (such as temperature) is directly set as the predictive lead time for environmental parameter adjustment. Then, combining this with the multidimensional eigenvectors in the key control features, the deviation influence coefficient between the predicted change trend of environmental parameters in the next adjustment cycle and the standard target value is calculated. The calculation formula is: ,in To predict the temperature value, the following method is used: Starting from the current moment, using the historical environmental parameter data sequence (such as past temperature values) contained in the multi-dimensional feature vector and process parameters strongly correlated with temperature changes (such as the rate of change of slurry flow rate and the trend of valve opening), a time-series forward extrapolation is performed based on the inertial characteristics of data changes. That is, the data trajectory is extended along the current direction of change to cover the duration of the response delay time, thereby extrapolating the expected value at the future adjustment node time. This expected value is the predicted temperature value. The standard target temperature value, This is used to quantify the severity of future deviations; simultaneously, it utilizes real-time collected environmental parameters and equipment status parameters to calculate real-time fluctuation amplitudes. For example, the sliding window standard deviation method can be used to calculate the fluctuation range of temperature data in the most recent minute, and the fluctuation correlation degree can be calculated based on the correlation between the fluctuation trends of each data point in the multidimensional feature vector (such as the temperature change rate and valve opening frequency). This characterizes the synchronicity of fluctuations between parameters; subsequently, it incorporates the self-balancing capability of the filling process (the system's natural recovery coefficient determined from empirical data). According to the basic adjustment cycle Normalized real-time fluctuation range Normalized fluctuation correlation and the normalized bias influence coefficient Perform weighted calculation of the optimal adjustment period The calculation formula is ,in Using a preset weighting factor, the adjustment frequency is dynamically adjusted according to the current operating conditions using this formula; finally, the calculated deviation influence coefficient is... With the preset intervention threshold (For example, 0.05) Perform numerical comparison, if If the current environmental parameters deviate from the stable range allowed by the filling process, then the optimal adjustment cycle will be used. Using time as a reference, an adjustment control command is generated, containing the adjustment direction of parameters such as "cooling" or "pressure increasing" and the target setpoint, and this command is sent to the corresponding air conditioning or purification equipment to execute the adjustment operation; if If the current operating condition meets the process self-balancing requirements, the system will not generate new adjustment commands and will continue to maintain the current monitoring state, thereby avoiding over-adjustment.

[0029] Parameter target value analysis module: Based on key control characteristics, hysteresis parameters and optimal adjustment period, the module performs time series trend analysis and fluctuation state identification on environmental parameters to obtain the standard target value for environmental parameter adjustment.

[0030] In this embodiment, the parameter target value analysis module extracts and analyzes the real-time monitoring values ​​and parameter change rates of environmental parameters within the current optimal adjustment cycle using multi-dimensional feature vectors in key control features. It then performs time-delay extrapolation calculations on the parameter change rates using the response delay time in the hysteresis characteristic parameters to obtain the time-delay environmental parameter offset value. Furthermore, it identifies and analyzes the current fluctuation state of the environmental parameters using the time-delay environmental parameter offset value, determining whether the fluctuation state is a rising edge, falling edge, or steady-state oscillation. Based on the time-delay environmental parameter offset value and the process safety margin, it determines the corresponding dynamic correction amount. The standard operating condition setpoint and the dynamic correction amount are then calculated to obtain the standard target value for environmental parameter adjustment. This standard target value is used to offset prediction deviations, ensuring that the environmental parameters return to the standard process range after the response time.

[0031] It should be noted that the specific execution process and analysis calculation details of the parameter target value analysis module are as follows: Based on the multidimensional feature vectors in key control features, the real-time monitoring numerical sequence of environmental parameters (such as temperature) within the current optimal regulation cycle is extracted, and the parameter change rate at the current moment is calculated. (i.e., the difference between the current temperature value and the previous temperature value divided by the time interval); subsequently, this is combined with the response delay time in the hysteresis characteristic parameter. (For example, the lag time of cooling capacity delivery in an air conditioning system), the parameter change rate is extrapolated using a time delay, and the calculation formula is as follows: Thus, the offset value of the time-delay environment parameter is obtained. This value represents the expected offset of the environmental parameter after the response delay time, without intervention; the system uses this time-delayed environmental parameter offset value. The sign and magnitude of the value are used to identify the fluctuation state, for example, if If the value is greater than a preset positive threshold, the fluctuation is determined to be a rising edge; if it is less than a preset negative threshold, it is determined to be a falling edge; otherwise, it is determined to be a steady-state oscillation. Simultaneously, the standard operating condition settings for the pulp filling process are retrieved from a preset process safety specification database. (For example, the cleanroom temperature benchmark value of 22 specified by the national food safety standard GB12695) ), and the process safety margin M determined by process risk analysis (e.g., reserves to ensure a sterile environment). (Safety buffer range); Based on the determined fluctuation state, the system determines the corresponding dynamic correction amount based on the time-delay environmental parameter offset value and the process safety margin. The calculation logic is as follows: when it is on the rising edge, When it is at the falling edge, When in steady-state oscillation, Take 0 or a fine-tuning value based on the oscillation amplitude; finally, perform an algebraic superposition calculation of the standard operating condition setpoint and the dynamic correction amount, using the following formula: Thus, the standard target value for environmental parameter adjustment can be calculated. This target value is achieved by pre-applying an adjustment amount that exceeds the current deviation to offset the predicted deviation and cover the safety margin, so that the environmental parameters accurately return to the standard process range after the system's response time.

[0032] Parameter prediction module: Based on the current environmental parameters, historical environmental parameters, and standard target values, it predicts the future trend of environmental parameters to obtain the predicted value of the sterile environment state at the next moment. Then, it compares the predicted value of the sterile environment state at the next moment with the standard target value to obtain the prediction deviation coefficient. Combined with the optimal adjustment cycle, it generates predictive adjustment execution instructions.

[0033] In this embodiment, the parameter prediction module performs trend fitting calculation on the environmental parameter change data based on the current environmental parameters, historical environmental parameters, and standard target values ​​to obtain the predicted value of the sterile environment state at the next moment. The deviation between the predicted value of the sterile environment state at the next moment and the standard target value is calculated to obtain the prediction deviation coefficient. The initial adjustment amplitude is then calculated based on the prediction deviation coefficient. The initial adjustment amplitude is used to eliminate the prediction environmental parameter deviation corresponding to the prediction deviation coefficient. The initial adjustment amplitude is then calculated by time-series decomposition according to the optimal adjustment period to obtain the predictive adjustment execution instruction used to suppress abnormal fluctuations in environmental parameters.

[0034] It should be noted that the specific execution process and analysis calculation details of the parameter prediction module are as follows: First, based on current environmental parameters, historical environmental parameters, and standard target values, a time series dataset is constructed. Then, the least squares method or Kalman filter algorithm is used to perform trend fitting calculations on the environmental parameter change data, establishing a predictive model for environmental parameter changes over time. The model is used to extrapolate the predicted value of the sterile environment state at the next moment. Subsequently, the predicted value of the sterile environment status at the next moment will be... Compared with the standard target value Perform deviation calculations to obtain the prediction deviation coefficient. The specific calculation formula is as follows: Based on the prediction deviation coefficient Calculate the initial adjustment range The calculation logic is as follows: ,in The preset proportional adjustment coefficient is used to calculate the adjustment amount required to completely eliminate the prediction environmental parameter deviation corresponding to the prediction deviation coefficient; then, based on the optimal adjustment period... The initial adjustment amplitude is calculated by time-series decomposition, specifically by... according to The divided adjustment step size is allocated to each time node to generate a predictive adjustment execution instruction containing the sequence of adjustment actions at different times. This instruction suppresses abnormal fluctuations in environmental parameters that may occur at future times by pre-allocating adjustment amounts on the time axis.

[0035] Adaptive adjustment module: Based on the predictive adjustment execution command, it controls the basic adjustment execution unit and the fine-tuning correction execution unit to work together, and monitors the actual response feedback data of environmental parameters in real time, dynamically correcting the adjustment output intensity until the environmental parameters stabilize at the standard target value.

[0036] In this embodiment, the adaptive adjustment module extracts the corresponding initial adjustment amplitude and optimal adjustment period according to the predictive adjustment execution instruction, and sends trend adjustment instructions to the basic adjustment execution unit and high-frequency fluctuation correction instructions to the fine-tuning correction execution unit, driving the two to work together. The basic adjustment execution unit is used to perform large-amplitude basic adjustment actions to perform differential adjustment on the overall trend of environmental parameters in order to eliminate trend deviations of environmental parameters. The fine-tuning correction execution unit is used to perform small-amplitude correction adjustment actions to perform differential adjustment on the instantaneous fluctuations of environmental parameters in order to suppress random high-frequency jitter of environmental parameters. The actual response feedback data of environmental parameters are sampled in real time and the rate of change is calculated to obtain the current actual regulation response value. The deviation between the current actual regulation response value and the initial regulation amplitude is calculated to obtain the dynamic feedback correction coefficient. The dynamic feedback correction coefficient is then used to perform real-time correction calculation on the current regulation output intensity until the actual monitoring value of the environmental parameters stably converges to the standard target value.

[0037] It should be noted that the specific execution process and analysis calculation details of the adaptive adjustment module are as follows: Based on the predictive adjustment execution command, the system extracts the corresponding initial adjustment amplitude and optimal adjustment period. The system then compares the initial adjustment amplitude with a preset trend threshold (e.g., temperature deviation greater than 2). The system allocates adjustment tasks. If the initial adjustment amplitude exceeds the trend threshold, it is determined to be a trend deviation. A trend adjustment command is sent to the basic adjustment execution unit to perform a large-amplitude basic adjustment action (such as significantly adjusting the valve opening of the chiller unit) to change the overall trend of environmental parameters. If the initial adjustment amplitude does not exceed the trend threshold or there are still residual high-frequency fluctuations after the basic adjustment, it is determined to be an instantaneous fluctuation. A high-frequency fluctuation correction command is sent to the fine-tuning correction execution unit to perform a small-amplitude correction adjustment action (such as fine-tuning the spray pulse of the humidifier) ​​to suppress random high-frequency jitter. During the adjustment process, the actual response feedback data of environmental parameters is sampled in real time and the rate of change is calculated to obtain the current actual adjustment response value. (e.g., the actual change in temperature per minute), and the current actual adjustment response value. With initial adjustment range The deviation is calculated based on the expected change in the adjustment target to obtain the dynamic feedback correction coefficient. The specific calculation formula is as follows: Finally, the dynamic feedback correction coefficient is used to perform real-time correction calculations on the current regulation output, that is, to calculate the corrected regulation output. Through this closed-loop feedback, the intensity of the action of the execution unit is continuously fine-tuned until the actual monitored values ​​of the environmental parameters stably converge to the standard target value.

[0038] Adjustment frequency optimization module: Analyzes the difference between the actual response feedback data and the predicted value of the sterile environment state at the next moment to obtain the frequency matching index, and updates and corrects the optimal adjustment cycle based on the frequency matching index.

[0039] In this embodiment, the frequency adjustment optimization module aligns the actual response feedback data with the predicted value of the sterile environment state at the next moment, calculates the response deviation value between the actual parameter change caused by the adjustment action and the predicted change, and substitutes the response deviation value into the frequency matching degree model for mapping calculation to obtain the frequency matching index; it compares the frequency matching index with the preset frequency threshold to generate the period deviation correction amount, and uses the period deviation correction amount to continuously update and correct the optimal adjustment period.

[0040] It should be noted that the specific execution process and analysis calculation details of the frequency adjustment optimization module are as follows: The actual response feedback data (such as the temperature change curve measured by the sensor after the adjustment action) is time-aligned with the predicted value of the sterile environment state at the next moment (such as the expected temperature change trajectory calculated by the parameter prediction module) to ensure that the two are compared on the same time axis. Then, the difference between the actual parameter change caused by the adjustment action (such as the actual temperature drop) and the predicted change (such as the expected temperature drop calculated by the prediction model) within the optimal adjustment period is calculated to obtain the response deviation value. Substitute the response deviation value into the preset frequency matching degree model for mapping calculation. The calculation formula is as follows: ,in A preset sensitivity coefficient is used to quantify the degree of matching between the adjustment frequency and the actual response capability of the system, thus obtaining the frequency matching index. According to the frequency matching index Compare with a preset frequency threshold (e.g., 0.9). If... If it is below the threshold, then use the formula. Generation cycle deviation correction amount ,in To correct the step size factor; finally, the optimal adjustment period is updated and corrected using the period deviation correction amount, for example, the updated optimal adjustment period. This allows for dynamic adjustment of the regulation frequency to adapt to the real-time response characteristics of the system.

[0041] like Figure 2 The present embodiment provides a predictive adjustment method for environmental parameters in aseptic filling of fruit pulp, including the following steps: Step 1: Real-time monitoring and collection of various monitoring data during the aseptic filling process of fruit pulp, and preprocessing of the collected monitoring data, including environmental parameters, process parameters and equipment status parameters; Step 2: By conducting multi-dimensional correlation analysis and feature screening of environmental parameters, process parameters, and equipment status parameters, key control features and hysteresis characteristic parameters are generated; Step 3: Based on the hysteresis characteristic parameters, environmental parameters, equipment status parameters, and key control characteristics, calculate the real-time fluctuation amplitude and fluctuation correlation of the monitoring data. Combined with the self-balancing capability of the filling process and the deviation status of the current environmental parameters, determine whether to intervene in the adjustment and obtain the optimal adjustment cycle and adjustment control command that conforms to the current working conditions. Step 4: Based on key control characteristics, hysteresis parameters, and optimal adjustment period, perform time-series trend analysis and fluctuation state identification on environmental parameters to obtain standard target values ​​for environmental parameter adjustment; Step 5: Based on the current environmental parameters, historical environmental parameters, and standard target values, predict the future trend of environmental parameters to obtain the predicted value of the sterile environment state at the next moment. Then, compare the predicted value of the sterile environment state at the next moment with the standard target value to obtain the prediction deviation coefficient. Combined with the optimal adjustment cycle, generate predictive adjustment execution instructions. Step 6: According to the predictive adjustment execution command, control the basic adjustment execution unit and the fine-tuning correction execution unit to work together, and monitor the actual response feedback data of environmental parameters in real time, dynamically correct the adjustment output force until the environmental parameters stabilize at the standard target value; Step 7: Perform adjustment response difference analysis between the actual response feedback data and the predicted value of the sterile environment state at the next moment to obtain the frequency matching index, and update and correct the optimal adjustment cycle according to the frequency matching index.

[0042] In conclusion, the above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

[0043] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

Claims

1. A predictive adjustment system for environmental parameters in aseptic filling of fruit pulp, characterized in that, include: Data acquisition module: It monitors and collects various monitoring data during the aseptic filling process of fruit pulp in real time, and preprocesses the collected monitoring data, including environmental parameters, process parameters and equipment status parameters. Data correlation analysis module: By performing multi-dimensional correlation analysis and feature filtering on environmental parameters, process parameters, and equipment status parameters, it generates key control features and hysteresis characteristic parameters; Adjustment frequency analysis module: Based on hysteresis characteristic parameters, environmental parameters, equipment status parameters and key control characteristics, calculate the real-time fluctuation amplitude and fluctuation correlation of monitoring data, and combine the self-balancing ability of the filling process and the deviation status of current environmental parameters to determine whether to intervene in adjustment, and obtain the optimal adjustment cycle and adjustment control command that conforms to the current working conditions. Parameter target value analysis module: Based on key control characteristics, hysteresis parameters and optimal adjustment period, the module performs time-series trend analysis and fluctuation state identification of environmental parameters to obtain the standard target value for environmental parameter adjustment. Parameter prediction module: Based on the current environmental parameters, historical environmental parameters, and standard target values, it predicts the future trend of environmental parameters to obtain the predicted value of the sterile environment state at the next moment. Then, it compares the predicted value of the sterile environment state at the next moment with the standard target value to obtain the prediction deviation coefficient. Combined with the optimal adjustment cycle, it generates predictive adjustment execution instructions. Adaptive adjustment module: Based on the predictive adjustment execution command, it controls the basic adjustment execution unit and the fine-tuning correction execution unit to work together, and monitors the actual response feedback data of environmental parameters in real time, dynamically correcting the adjustment output intensity until the environmental parameters stabilize at the standard target value; Adjustment frequency optimization module: Analyzes the difference between the actual response feedback data and the predicted value of the sterile environment state at the next moment to obtain the frequency matching index, and updates and corrects the optimal adjustment cycle based on the frequency matching index.

2. The predictive adjustment system for environmental parameters of aseptic filling of fruit pulp according to claim 1, characterized in that, The data acquisition module uses a distributed array of multi-dimensional sensors in the cleanroom, laminar flow hood, and surrounding area of ​​the filling valve in the aseptic filling room to monitor and collect environmental parameters, process parameters, and equipment status parameters in real time. The environmental parameters include temperature, humidity, pressure difference, dust particle count, and hydrogen peroxide concentration. The process parameters include pulp flow rate, filling speed, and sterilization intensity. The equipment status parameters include fan speed, valve opening, and operating vibration. The preprocessing of the various monitoring data collected includes: The collected environmental parameters, process parameters, and equipment status parameters are sequentially cleaned, transformed, and standardized.

3. The predictive adjustment system for aseptic filling environmental parameters of fruit pulp according to claim 2, characterized in that, The data correlation analysis module uses the Pearson correlation coefficient to calculate the correlation between preprocessed environmental parameters, process parameters and equipment status parameters, and obtains the influence correlation coefficient between parameters. The calculated correlation coefficients are compared with preset correlation thresholds to filter environmental parameters, process parameters, and equipment status parameters. The specific comparison process is as follows: if the correlation coefficient is greater than or equal to the preset correlation threshold, the corresponding parameter is determined to be a strongly correlated factor and marked as a target data factor for retention; if the correlation coefficient is less than the preset correlation threshold, the corresponding parameter is determined to be a weakly interfering factor and removed. Then, time-series alignment analysis is performed on the retained target data factors. Lag characteristic parameters are extracted by calculating the time difference between each target data factor and the environmental parameter changes. Simultaneously, the data fluctuation trends of each target data factor are integrated to construct a multi-dimensional feature vector, thereby obtaining key control features. The data fluctuation trends include the temperature change rate and humidity oscillation amplitude of environmental parameters, the pulp flow rate fluctuation characteristics of process parameters, and the valve opening oscillation frequency of equipment status parameters. The operation of integrating the data fluctuation trends of each target data factor to construct a multi-dimensional feature vector specifically involves: extracting the data fluctuation trends corresponding to environmental parameters, process parameters, and equipment status parameters respectively, and concatenating the extracted data fluctuation trends according to time sequence to generate a multi-dimensional feature vector containing multi-dimensional fluctuation information.

4. The predictive adjustment system for aseptic filling environmental parameters of fruit pulp according to claim 3, characterized in that, The regulation frequency analysis module uses the response delay time of each target data factor in the hysteresis characteristic parameter relative to the change of environmental parameter as the prediction lead time for environmental parameter regulation. It also calculates the deviation influence coefficient between the predicted change trend of environmental parameter in the next regulation cycle and the standard target value by combining the multi-dimensional feature vector in the key control feature. By using real-time collected environmental parameters and equipment status parameters, the real-time fluctuation amplitude of the monitoring data is calculated, and the fluctuation correlation degree is calculated based on the correlation of the fluctuation trends of each data in the multidimensional feature vector. Then, combined with the self-balancing capability of the filling process, the optimal adjustment cycle is obtained by weighted calculation based on the basic adjustment cycle, the normalized real-time fluctuation amplitude, the normalized fluctuation correlation degree, and the normalized deviation influence coefficient. The calculated deviation impact coefficient is compared with the preset intervention threshold. If the deviation influence coefficient is greater than or equal to the preset intervention threshold, it is determined that the current environmental parameters have deviated from the stable range allowed by the filling process. Using the optimal adjustment cycle as the time reference, an adjustment control command containing the parameter adjustment direction and the target set value is generated and sent to the corresponding air conditioning or purification equipment to perform the adjustment operation. If the deviation influence coefficient is less than the preset intervention threshold, the current operating condition is determined to meet the process self-balancing requirements, no new adjustment command is generated, and the current monitoring status is maintained.

5. The predictive adjustment system for aseptic filling environmental parameters of fruit pulp according to claim 4, characterized in that, The parameter target value analysis module extracts and analyzes the real-time monitoring values ​​and parameter change rates of environmental parameters within the current optimal adjustment cycle through the multi-dimensional feature vector in the key control features, and performs time-delay extrapolation calculation on the parameter change rate by combining the response delay time in the hysteresis characteristic parameters to obtain the time-delay environmental parameter offset value. Then, the current fluctuation state of the environmental parameters is identified and analyzed by the time-delay environmental parameter offset value. It is determined whether the fluctuation state is a rising edge, a falling edge, or a steady-state oscillation. Based on the time-delay environmental parameter offset value and the process safety margin, the corresponding dynamic correction amount is determined. The standard operating condition setpoint and the dynamic correction amount are calculated to obtain the standard target value for environmental parameter adjustment. The standard target value is used to offset the prediction deviation so that the environmental parameters return to the standard process range after the response time.

6. The predictive adjustment system for environmental parameters in aseptic filling of fruit pulp according to claim 5, characterized in that, The parameter prediction module performs trend fitting calculations on the environmental parameter change data based on the current environmental parameters, historical environmental parameters, and standard target values ​​to obtain the predicted value of the sterile environment state at the next moment. The deviation between the predicted value of the sterile environment state at the next moment and the standard target value is calculated to obtain the prediction deviation coefficient. The initial adjustment amplitude is then calculated based on the prediction deviation coefficient. The initial adjustment amplitude is used to eliminate the prediction environmental parameter deviation corresponding to the prediction deviation coefficient. The initial adjustment amplitude is then calculated by time-series decomposition according to the optimal adjustment period to obtain the predictive adjustment execution instruction used to suppress abnormal fluctuations in environmental parameters.

7. The predictive adjustment system for environmental parameters of aseptic filling of fruit pulp according to claim 6, characterized in that, The adaptive adjustment module extracts the corresponding initial adjustment amplitude and optimal adjustment period according to the predictive adjustment execution command, and sends trend adjustment command to the basic adjustment execution unit and high-frequency fluctuation correction command to the fine-tuning correction execution unit, driving the two to work together. The basic adjustment execution unit is used to perform large-amplitude basic adjustment actions to perform differential adjustment on the overall trend of environmental parameters in order to eliminate trend deviation of environmental parameters. The fine-tuning correction execution unit is used to perform small-amplitude correction adjustment actions to perform differential adjustment on the instantaneous fluctuation of environmental parameters in order to suppress random high-frequency jitter of environmental parameters. The actual response feedback data of environmental parameters are sampled in real time and the rate of change is calculated to obtain the current actual regulation response value. The deviation between the current actual regulation response value and the initial regulation amplitude is calculated to obtain the dynamic feedback correction coefficient. The dynamic feedback correction coefficient is then used to perform real-time correction calculation on the current regulation output intensity until the actual monitoring value of the environmental parameters stably converges to the standard target value.

8. The predictive adjustment system for aseptic filling environmental parameters of fruit pulp according to claim 7, characterized in that, The frequency adjustment optimization module aligns the actual response feedback data with the predicted value of the sterile environment state at the next moment, calculates the response deviation between the actual parameter change caused by the adjustment action and the predicted change, and substitutes the response deviation into the frequency matching degree model for mapping calculation to obtain the frequency matching index. Based on the comparison between the frequency matching index and the preset frequency threshold, a period deviation correction amount is generated, and the optimal adjustment period is updated and corrected using the period deviation correction amount.

Images