A method for forming a fermented soybean cheese by injecting slurry into a mold after rapid fermentation of a large pot of fermented soybean cheese

Abstract

The application provides a quick fermentation post-membrane injection molding method for fermented bean curd large tanks, comprising: collecting pressure signals of each area in the membrane through a multi-point pressure sensor array, calculating a differential sequence to generate a real-time sequence of the pressure in the membrane, simultaneously reading rotor speed and volume flow rate instantaneous readings from the injection pump to construct a current operating parameter combination; processing the real-time sequence of the pressure in the membrane by using a low-pass filtering technology, extracting the position of the pressure slope mutation point and the fluctuation amplitude, combining the flow rate fluctuation interval deviation of the volume flow rate instantaneous readings to generate a dynamic change characterization sequence of the pressure in the membrane for subsequent ratio calculation; according to the updated operating parameter combination, calculating the ratio threshold value duration interval and the pressure flow rate matching error, combining the rotor speed adjustment cumulative frequency and the time window abnormal point density to judge whether the pressure in the membrane and the flow rate reach a stable matching state, providing a basis for subsequent regulation. The application improves the production efficiency, guarantees the product quality and reduces the raw material loss.

Description

Technical Field

[0001] This invention relates to the field of information technology, and in particular to a method for rapid fermentation of fermented bean curd in large tanks followed by intra-membrane grouting. Background Technology

[0002] Fermented bean curd, as a traditional fermented food, plays a vital role in the food processing industry, especially in ensuring product quality and production efficiency. The rapid fermentation process in large tanks followed by in-membrane injection molding is a key technological approach to improving production scale and quality consistency, particularly in large-scale industrial production, directly impacting product taste and market competitiveness. However, current research and application still face numerous challenges, urgently requiring breakthroughs in technological bottlenecks to meet modern production demands.

[0003] Existing methods for fermented bean curd slurry molding often struggle to adapt to the complex changes in the state of the slurry after fermentation, making it difficult to balance production efficiency and product stability. Many processes rely on manual experience or fixed parameters, ignoring the differences in the physical properties of fermented materials at different stages, especially in the slurry injection stage, where they lack responsiveness to dynamic changes in the environment and materials. This static approach not only increases labor costs but also easily leads to quality problems due to improper control, limiting the intelligent upgrading of production lines.

[0004] In this field, the core technical challenges lie in the precise control of the grouting speed and the real-time matching of the fermented material's state. Because the characteristics of the fermented bean curd slurry, such as temperature and viscosity, change continuously with time and the environment, if the grouting speed cannot be adjusted accordingly, the membrane may rupture in the early stages due to excessive speed, or overflow may occur in the later stages due to the undiminished speed. A deeper problem is that the dynamic changes in internal pressure are not effectively sensed and utilized, resulting in a lack of protection against the load-bearing limit during the grouting process, thus affecting the molding effect and equipment lifespan. For example, in actual production, if the initial grouting speed is too fast, the membrane may tear due to the pressure, while in the later stages, if the speed is not reduced in time, overflow will cause material waste and cleaning problems.

[0005] Therefore, how to dynamically adjust the grouting speed according to the real-time state of the fermented slurry and the changes in the membrane pressure during the grouting process, so as to ensure membrane stability, improve production efficiency and avoid material waste, has become a key issue that needs to be addressed in the membrane grouting and molding process after rapid fermentation in large fermentation tanks of fermented bean curd. Summary of the Invention

[0006] This invention provides a method for rapid fermentation of fermented bean curd in large tanks followed by in-membrane grouting and molding, mainly comprising: Pressure signals from various regions within the membrane are acquired using a multi-point pressure sensor array. A differential sequence is calculated to generate a real-time membrane pressure sequence. Simultaneously, instantaneous rotor speed and volumetric flow rate readings are read from the grouting pump to construct the current operating parameter combination for subsequent dynamic analysis. A low-pass filter is used to process the real-time membrane pressure sequence, extracting the location and amplitude of pressure slope abrupt changes. Combined with the flow rate fluctuation range deviation from the instantaneous volumetric flow rate readings, a dynamic change characterization sequence of membrane pressure is generated for subsequent ratio calculations. Based on this dynamic change characterization sequence, the ratio of each pressure slope abrupt change point to the corresponding instantaneous volumetric flow rate reading is calculated, forming a ratio sequence. The number of consecutive frames with abnormal ratios and the time interval between ratio peaks are analyzed. Combined with a preset dynamic ratio threshold, a speed overshoot risk indicator is determined. If the speed overshoot risk indicator is true, the target speed sequence step size and proportional iterative adjustment are calculated based on the pressure slope abrupt change density and speed adjustment trigger conditions. The system generates a speed reduction command and sends it to the grouting pump control unit for real-time adjustment according to command priority. It acquires feedback data from the grouting pump after the speed reduction command is executed, and through closed-loop feedback response delay analysis, determines the actual response time and stable deviation value of the speed. Simultaneously, it re-acquires the real-time sequence of membrane pressure and the instantaneous reading of volumetric flow rate, updating the operating parameter combination for further matching analysis. Based on the updated operating parameter combination, it calculates the ratio exceeding the threshold duration interval and the pressure-flow rate matching error. Combining the cumulative frequency of speed adjustment and the density of abnormal points within the time window, it determines whether the membrane pressure and flow rate have reached a stable matching state, providing a basis for subsequent control. If the membrane pressure and flow rate have not reached a stable matching state, it generates the next cycle's target speed sequence based on the abnormal flow rate fluctuation amplitude and the cumulative mutation amount within the time window. Through dynamically updated target sequences and proportional iterative adjustment amplitudes, it sends the sequence to the grouting pump control unit to form closed-loop control until the matching state is optimized.

[0007] The above-mentioned method for rapid fermentation of fermented bean curd in a large tank followed by membrane grouting involves acquiring pressure signals from various regions within the membrane using a multi-point pressure sensor array, calculating a differential sequence to generate a real-time membrane pressure sequence, and simultaneously reading instantaneous rotor speed and volumetric flow rate readings from the grouting pump to construct a current operating parameter combination for subsequent dynamic analysis. This includes: The pressure signal of the membrane region is continuously acquired by a multi-point pressure sensor array, and the original pressure data of each region is recorded to obtain the initial pressure dataset. Based on the initial pressure dataset, the pressure difference between adjacent time points is calculated, a difference sequence is generated, and the real-time sequence of membrane pressure changes is determined. For real-time sequences, a combination of operating parameters containing multi-dimensional information is constructed by combining instantaneous data of rotor speed and volumetric flow rate obtained from the grouting pump. By performing time-dimensional data alignment on the combination of operating parameters, a unified time-series dataset is obtained, and the synchronization between parameters is determined. If the synchronization is insufficient, the rotor speed and volumetric flow rate data are interpolated to obtain an adjusted set of timing parameters. The support vector machine algorithm was used to classify the adjusted time series parameter set to determine the correlation pattern between membrane pressure changes and operating parameters; For association patterns, generate feature datasets required for dynamic analysis for subsequent system processing.

[0008] The above-mentioned method for rapid fermentation of fermented bean curd in a large tank followed by membrane grouting and molding involves using low-pass filtering technology to process the real-time intra-membrane pressure sequence, extracting the location of pressure slope abrupt change points and the magnitude of fluctuations, and combining this with the deviation of the flow rate fluctuation range from the instantaneous volumetric flow rate reading to generate a dynamic change characterization sequence of intra-membrane pressure for subsequent ratio calculations. The filtered pressure sequence is obtained by processing the real-time intramembrane pressure sequence with a low-pass filter. The pressure slope sequence is obtained by calculating the slope sequence of consecutive adjacent points from the filtered pressure sequence. The pressure slope abrupt change location sequence is obtained by identifying points in the pressure slope sequence where the absolute value of the slope exceeds a preset abrupt change threshold. For each location of a sudden change in pressure slope, the extreme value difference before and after the corresponding location in the filtered pressure sequence is extracted to obtain the fluctuation amplitude sequence. The real-time flow rate sequence is obtained by collecting the instantaneous readings of the volumetric flow rate instrument. The flow rate fluctuation interval sequence is obtained by calculating the difference between the maximum and minimum values ​​in adjacent time periods based on the real-time flow rate sequence. Aligning the flow rate fluctuation interval sequence with the pressure slope abrupt change location sequence by time yields the flow rate fluctuation interval deviation sequence at the corresponding time point. By combining and connecting the pressure slope abrupt change markers, fluctuation amplitudes, and corresponding flow rate fluctuation range deviations at each abrupt change location in chronological order, a sequence representing the dynamic changes in intramembrane pressure is obtained.

[0009] The above-mentioned method for rapid fermentation of fermented bean curd in a large tank followed by intramembrane grouting and molding involves, based on a dynamic change sequence of intramembrane pressure, calculating the ratio of each pressure slope abrupt change point to the corresponding instantaneous volumetric flow rate reading to form a ratio sequence. The method then analyzes the number of consecutive frames with abnormal ratios and the time interval between ratio peaks in the ratio sequence, and, combined with a preset dynamic ratio threshold, determines whether there is a risk indicator of velocity overshoot. This includes: Obtain raw data records of membrane pressure changes over time, calculate the rate of change of pressure value at each time point, form a pressure slope sequence, and determine the set of abrupt change points; Extract the timestamp corresponding to each point from the set of mutation points, obtain the instantaneous reading of the volumetric flow rate at the same time, calculate the ratio of pressure slope to flow rate reading, and obtain the ratio sequence data; For the ratio sequence data, the change range of each ratio is analyzed. If a ratio exceeds the preset dynamic threshold range, it is marked as an abnormal ratio, and the abnormal ratio distribution is obtained. Based on the distribution of abnormal ratios, the number of frames with consecutive abnormal ratios is counted, the cumulative duration of consecutive frames is calculated, and the characteristics of abnormal persistence are determined. By analyzing the abnormal ratio distribution, the time points when the ratio peaks occur are extracted, the time intervals between adjacent peaks are calculated, and the regularity of the peak intervals is determined. By using the characteristics of abnormal persistence and the regularity of peak intervals, and comparing them with preset dynamic threshold standards, if the persistence characteristics or the regularity of intervals exceed the threshold range, it is determined that there is a risk of speed overshoot, and the final risk assessment result is obtained.

[0010] The above-mentioned method for rapid fermentation of fermented bean curd in a large tank followed by in-membrane grouting and molding, wherein if the speed overshoot risk flag is true, the target speed sequence step size and proportional iteration adjustment amplitude are calculated based on the pressure slope mutation density and speed adjustment triggering conditions, a speed reduction command is generated, and the command is sent to the grouting pump control unit for real-time regulation according to the command priority, including: If both speed overshoot and risk flag are detected, pressure slope and mutation density data are retrieved from the system, and the satisfaction of the triggering conditions is determined by comparing and analyzing preset thresholds. If the triggering conditions are met, the target sequence and sequence step size for speed adjustment are calculated based on the changing trends of pressure slope and mutation density using pre-established mapping rules, thus obtaining a preliminary speed adjustment scheme. Based on the calculation results of the target sequence and sequence step size, and combined with the adjustment logic of proportional iteration, the magnitude of speed reduction is determined step by step, and specific speed reduction command data is generated. By sorting the command data according to the priority order, and based on the real-time control requirements, the order of command transmission is obtained, and the final command sending plan is determined. According to the instruction sending plan, the speed reduction instruction is transmitted to the grouting pump control unit in priority order, and the execution process of speed adjustment is controlled in real time to obtain the operating status after adjustment; By collecting monitoring data on the operating status after regulation, and combining the latest changes in pressure slope and mutation density, it is determined whether the target speed sequence needs further adjustment, thus generating a basis for subsequent regulation.

[0011] The above-mentioned method for rapid fermentation of fermented bean curd in a large tank followed by in-membrane grouting and molding includes obtaining feedback data after the execution of the speed reduction command returned by the grouting pump, determining the actual response time and stable deviation value of the speed through closed-loop feedback response delay analysis, and simultaneously re-acquiring the real-time sequence of intramembrane pressure and the instantaneous reading of volumetric flow rate to update the combination of operating parameters for further matching analysis. The data acquisition module obtains the speed sequence and timestamp sequence of the grouting pump after executing the reduction command; The rate of change of rotational speed at each time point is calculated based on the rotational speed sequence and the timestamp sequence to obtain the rotational speed change rate sequence; If the absolute value of five consecutive points in the speed change rate sequence is less than a preset threshold, then the current moment is determined to be a stable moment, and the speed value corresponding to the stable moment is obtained. The stability deviation value is determined by subtracting the target speed value from the stable speed value. The membrane pressure real-time sequence is continuously acquired from the membrane pressure sensor, and the volumetric flow rate instantaneous sequence is continuously acquired from the flow meter. Based on the steady-state time point, 30-second data segments before and after the steady-state time are extracted from the real-time sequence of membrane pressure and the instantaneous sequence of volumetric flow rate to form the steady-state pressure sequence and the steady-state flow rate sequence. The current target speed, stable deviation value, average pressure sequence of stable segment, and average flow rate sequence of stable segment are combined into a new combination of operating parameters.

[0012] The above-mentioned method for rapid fermentation of fermented bean curd in a large tank followed by membrane grouting and molding, wherein the updated combination of operating parameters is used to calculate the interval of continuous exceedance of the ratio threshold and the pressure-flow rate matching error, and combined with the cumulative frequency of rotation speed adjustment and the density of abnormal points within the time window, to determine whether the membrane pressure and flow rate have reached a stable matching state, providing a basis for subsequent regulation, including: Obtain the pressure ratio sequence and flow rate ratio sequence at each time point under the current combination of operating parameters; By judging the pressure ratio sequence by a preset threshold, the continuous time period in which the ratio exceeds the threshold is obtained; For each continuous time period, calculate its duration and record the start and end times. Within the same time range, calculate the error sequence between the actual pressure value and the actual flow rate value to obtain the pressure-flow rate matching error value. Obtain operation records of the most recent speed adjustments, count the total number of adjustments, divide fixed time windows backward from the current time, count the number of outliers within the window and calculate the outlier density; If the total length of the duration of the ratio exceeding the threshold is less than the preset length and the matching error value is less than the preset error upper limit, and the density of outliers is lower than the density threshold, then it is determined that the current matching state is stable. When a stable matching state is determined, a stability flag is output and the current parameter combination is recorded as a stable baseline combination.

[0013] The above-mentioned method for rapid fermentation of fermented bean curd in a large tank followed by intramembrane grouting and molding, wherein if the intramembrane pressure and flow rate do not reach a stable matching state, a target sequence for the rotational speed of the next cycle is generated based on the abnormal fluctuation amplitude of the flow rate and the cumulative mutation amount within the time window. This target sequence and proportional iterative adjustment amplitude are dynamically updated and sent to the grouting pump control unit to form a closed-loop control until the matching state is optimized, including: Obtain the flow rate data sequence within the current time window, and calculate the cumulative mutation amount by summing the absolute values ​​of the differences between adjacent sampling points; If the cumulative mutation amount exceeds the preset threshold and the flow rate fluctuation is greater than the limit, the target generation stage of the next cycle will be entered if the judgment is true; otherwise, the current speed command will be maintained. The values ​​of each point in the target rotational speed sequence are determined based on the magnitude of the cumulative mutation and the amplitude of the flow rate fluctuation. The latest sequence is obtained by dynamically updating the target rotational speed sequence using a linear interpolation method; The latest target speed sequence and the proportional iterative adjustment step size are input together into the closed-loop controller, and the grouting pump is driven to perform speed regulation according to the output command of the closed-loop controller. The cumulative mutation amount within the time window is recalculated from the newly collected membrane pressure and flow rate data after execution to determine whether the current membrane pressure and flow rate have entered the stable matching interval. If the target speed sequence is not matched, the process returns to the dynamic update stage and continues to cycle and adjust.

[0014] The technical solutions provided by the embodiments of the present invention may include the following beneficial effects: 1. This invention collects pressure signals from various regions within the membrane in real time using a multi-point pressure sensor array, and combines this with rotor speed and volumetric flow rate data of the grouting pump to construct a multi-parameter fusion operating state combination. This breaks through the traditional static control mode that relies on fixed parameters or human experience, enabling the grouting process to have real-time perception and adaptive adjustment capabilities for dynamic changes in the physical state of the fermentation slurry.

[0015] 2. This invention employs low-pass filtering and pressure slope mutation detection technology to accurately identify abnormal fluctuations in membrane pressure. Combined with flow rate fluctuation range deviation analysis, it achieves early judgment and warning of speed overshoot risk. Through a closed-loop feedback mechanism, it dynamically generates and executes speed adjustment commands, effectively avoiding problems such as membrane rupture and slurry overflow caused by mismatched grouting speeds, significantly improving the reliability of the molding process and the safety of equipment operation.

[0016] 3. This invention calculates the ratio sequence of pressure slope to instantaneous flow rate and combines it with the analysis of the persistence of abnormal ratios and peak intervals. This allows the system to accurately assess the matching state of pressure and flow rate. Utilizing an iterative adjustment algorithm and closed-loop control strategy, the grouting pump speed is optimized in real time, ensuring that the intra-membrane pressure and grouting flow rate remain within a stable matching range. This guarantees the consistency of fermented bean curd formation and the stability of product quality.

[0017] 4. This invention introduces a response delay analysis and matching error calculation mechanism. The system can promptly evaluate the execution effect after each adjustment and update the combination of operating parameters. If a stable matching state is not reached, the system can automatically generate the target rotational speed sequence for the next cycle based on the flow rate fluctuation amplitude and cumulative mutation characteristics, achieving continuous closed-loop optimization until the optimal process state is reached, significantly improving the system's adaptability and robustness under complex operating conditions.

[0018] 5. This invention organically combines sensor technology, real-time data processing, and closed-loop control strategies to form a complete intelligent control method for grouting molding, which has high industry application value and promotion prospects.

[0019] In summary, this invention effectively solves the core technical problem of speed and pressure mismatch during the slurry casting process after rapid fermentation in large fermentation tanks of fermented bean curd by integrating real-time sensing, intelligent analysis and dynamic control. It has outstanding advantages in improving production efficiency, ensuring product quality and reducing raw material loss. Attached Figure Description

[0020] Figure 1 This is a flowchart of a method for rapid fermentation of fermented bean curd in a large tank followed by in-membrane grouting.

[0021] Figure 2 This is a schematic diagram of an intra-membrane grouting molding method for rapid fermentation of fermented bean curd in a large tank according to the present invention.

[0022] Figure 3 This is another schematic diagram of the method for rapid fermentation of fermented bean curd in a large tank followed by in-membrane grouting and molding according to the present invention. Detailed Implementation

[0023] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be described in detail below with reference to the accompanying drawings and specific embodiments.

[0024] like Figures 1-3 This embodiment of a method for rapid fermentation of fermented bean curd in a large tank followed by in-membrane grouting and molding may specifically include: S101. The pressure signals of each region inside the membrane are collected by a multi-point pressure sensor array, and the differential sequence is calculated to generate a real-time sequence of membrane pressure. At the same time, the instantaneous readings of rotor speed and volumetric flow rate are read from the grouting pump to construct the current combination of operating parameters for subsequent dynamic analysis.

[0025] Continuous pressure signal acquisition within the membrane region is achieved using a multi-point pressure sensor array, recording the raw pressure data for each region to obtain an initial pressure dataset. Based on this initial pressure dataset, the pressure difference between adjacent time points is calculated to generate a differential sequence, determining the real-time sequence of pressure changes within the membrane. For this real-time sequence, instantaneous data on rotor speed and volumetric flow rate obtained from the grouting pump are combined to construct a combination of operating parameters containing multi-dimensional information. By performing time-dimensional data alignment processing on the operating parameter combination, a unified time-series dataset is obtained, and the synchronicity between parameters is assessed. If synchronicity is insufficient, interpolation processing is performed on the rotor speed and volumetric flow rate data to obtain an adjusted time-series parameter set. A support vector machine algorithm is used to classify the adjusted time-series parameter set to determine the correlation pattern between membrane pressure changes and operating parameters. Based on the correlation pattern, a feature dataset required for dynamic analysis is generated for subsequent system processing.

[0026] Pressure signals from various regions within the membrane are collected using a multi-point pressure sensor array. In practice, a 4x4 array consisting of 16 pressure sensors can be used, arranged in key areas within the membrane, collecting data once per second to obtain the pressure value of each sensor. For example, the readings of sensors 1 to 16 are 2.5, 3.1, 2.8, 3.0, 2.7, 2.9, 3.2, 2.6, 2.4, 3.3, 2.5, 2.8, 3.0, 2.9, 3.1, and 2.7 kPa, respectively. These readings are then stored as a time series matrix using a data processing system. Next, a real-time sequence of intra-membrane pressure is generated by calculating the difference sequence. The specific algorithm involves calculating the difference between pressure values ​​at adjacent time points. For example, if sensor 1 reads 2.5 kPa at time t1 and 2.6 kPa at time t2, the difference is 0.1 kPa. This process is repeated for all sensor data, generating 16 difference sequences. The overall intra-membrane pressure trend is then calculated by averaging the results. For instance, if the average of the 16 difference values ​​is 0.05 kPa, it indicates an overall upward pressure trend. Simultaneously, instantaneous readings of rotor speed and volumetric flow rate are read from the grouting pump. Assuming data is acquired once per second via an industrial IoT interface, with a rotor speed of 1200 rpm and a volumetric flow rate of 5.2 L / min, these data are aligned with the pressure data using timestamps to form a unified time-series dataset. Subsequently, a combination of current operating parameters was constructed. Specifically, the average pressure difference (0.05 kPa), rotational speed (1200 rpm), and flow rate (5.2 L / min) were combined into a three-dimensional parameter vector. The parameters were standardized using a data fusion algorithm, such as weighted averaging, to ensure consistent dimensions. For example, the weights were 0.4, 0.3, and 0.3, respectively. The calculated comprehensive operating index was 0.048, used for subsequent dynamic analysis. The above data could also be correlated with the energy consumption data of the grouting pump, such as its power (3.5 kW). Regression analysis was used to explore the relationship between pressure changes and energy consumption, yielding a correlation coefficient of 0.85, indicating that pressure changes significantly affect energy consumption and providing a basis for system optimization. The entire process was implemented through an automated data acquisition and processing system, ensuring real-time performance and accuracy, laying the foundation for subsequent dynamic analysis.

[0027] S102. The real-time sequence of intramembrane pressure is processed using low-pass filtering technology to extract the location of pressure slope abrupt change points and the magnitude of fluctuation. Combined with the flow rate fluctuation range deviation of the instantaneous volumetric flow rate reading, a sequence representing the dynamic change of intramembrane pressure is generated for subsequent ratio calculation.

[0028] The real-time intramembrane pressure sequence is processed by low-pass filtering to obtain a filtered pressure sequence. The slope sequence of consecutive adjacent points is calculated from the filtered pressure sequence to obtain a pressure slope sequence. Points in the pressure slope sequence where the absolute value of the slope exceeds a preset abrupt change threshold are identified to obtain a pressure slope abrupt change location sequence. For each pressure slope abrupt change location, the extreme value difference before and after the corresponding position in the filtered pressure sequence is extracted to obtain a fluctuation amplitude sequence. Instantaneous readings from a volumetric flow rate instrument are collected to obtain a real-time flow rate sequence. The difference between the maximum and minimum values ​​within adjacent time periods is calculated based on the real-time flow rate sequence to obtain a flow rate fluctuation interval sequence. The flow rate fluctuation interval sequence and the pressure slope abrupt change location sequence are aligned in time to obtain a flow rate fluctuation interval deviation sequence for the corresponding time moment. The pressure slope abrupt change marker, fluctuation amplitude, and corresponding flow rate fluctuation interval deviation at each abrupt change location are combined and connected in chronological order to obtain a sequence characterizing the dynamic changes in intramembrane pressure.

[0029] For processing the real-time intramembrane pressure sequence, the pressure data is first smoothed using low-pass filtering to remove high-frequency noise. Assuming the original pressure sequence is data acquired once per second with values ​​of [10.2, 10.5, 11.0, 12.5, 12.0] kPa, a Butterworth low-pass filter with a cutoff frequency of 0.1 Hz is used. The data is then processed using filtering functions from a signal processing library, resulting in a smoothed sequence of [10.3, 10.6, 10.9, 11.5, 11.8] kPa. The filtered data better reflects the pressure trend. Next, the location and amplitude of the pressure slope abrupt change were extracted. By calculating the pressure difference between adjacent points, the slope was determined to be [0.3, 0.3, 0.6, 0.3] kPa / s. The change in slope from 0.3 to 0.6 exceeded the threshold of 0.4, and the third point was identified as the abrupt change point. The fluctuation amplitude was 1.5 kPa, the difference between the maximum pressure of 11.8 and the minimum pressure of 10.3. Subsequently, combined with the instantaneous volumetric flow rate readings, assuming a flow rate sequence of [5.1, 5.3, 5.0, 5.6, 5.2] L / min, the deviation of the flow rate fluctuation range was calculated. The mean was 5.24, the standard deviation was 0.22, and the deviation range was ±0.44 L / min, indicating relatively small flow rate fluctuations. Finally, the location of the pressure abrupt change, the fluctuation amplitude, and the flow rate deviation are integrated to generate a sequence representing the dynamic changes in membrane pressure. The format is [abrupt change time: 3s, pressure amplitude: 1.5 kPa, flow rate deviation: ±0.44 L / min]. This data is automatically stored as structured data by the algorithm for subsequent ratio calculations, ensuring that the correlation between pressure and flow rate changes is quantified. Logically, this forms a complete chain from data smoothing to feature extraction to comprehensive representation.

[0030] S103. Based on the dynamic change characterization sequence of intramembrane pressure, calculate the ratio of each pressure slope abrupt change point to the corresponding instantaneous volumetric flow rate reading to form a ratio sequence. Analyze the number of consecutive frames of abnormal ratios and the time interval of ratio peaks in the ratio sequence. Combined with the preset dynamic ratio threshold, determine whether there is a risk indicator of velocity overshoot.

[0031] Raw data records of membrane pressure changes over time are acquired. The rate of change of pressure is calculated for each time point, forming a pressure slope sequence to identify a set of abrupt change points. The timestamp corresponding to each point is extracted from the abrupt change point set, and the instantaneous volumetric flow rate reading for the same period is obtained. The ratio of the pressure slope to the flow rate reading is calculated, resulting in a ratio sequence data. For the ratio sequence data, the magnitude of change of each ratio is analyzed. If a ratio exceeds a preset dynamic threshold range, it is marked as an abnormal ratio, resulting in an abnormal ratio distribution. Based on the abnormal ratio distribution, the number of frames with consecutive abnormal ratios is counted, and the cumulative duration of consecutive frames is calculated to determine the anomaly persistence characteristic. The time points of ratio peak occurrences are extracted from the abnormal ratio distribution, and the time intervals between adjacent peaks are calculated to determine the peak interval regularity. Using the anomaly persistence characteristic and peak interval regularity, and comparing them with a preset dynamic threshold standard, if the persistence characteristic or interval regularity exceeds the threshold range, a velocity overshoot risk indicator is determined, resulting in the final risk assessment result.

[0032] For the processing and analysis of the dynamic change sequence of intramembrane pressure, we assume we have acquired a pressure data sequence with a time interval of 1 second, and pressure values ​​of 10.5, 11.2, 13.8, 15.1, 16.0, 14.5, and 13.0 (unit: kPa), with corresponding instantaneous volumetric flow rate readings of 2.1, 2.3, 2.8, 3.0, 3.2, 2.9, and 2.6 (unit: L / min). First, by calculating the ratio of the pressure difference between adjacent time points to the time interval, we obtain a pressure slope sequence of 0.7, 2.6, 1.3, 0.9, -1.5, and -1.5 (unit: kPa / s). Using the condition that the slope change exceeds a preset threshold of 1.0 kPa / s, we identify abrupt change points at the 2nd second (slope from 0.7 to 2.6) and the 5th second (slope from 0.9 to -1.5). Next, the instantaneous volumetric flow rate readings at the corresponding moments of the mutation points were extracted, which were 2.3 and 2.9 respectively. These were then compared with the corresponding slope values ​​of 2.6 and -1.5 to calculate the ratio sequence, resulting in ratios of 0.88 and -1.93. Further analysis of the ratio sequence was conducted, setting dynamic thresholds of positive 0.9 and negative -1.8. 0.88 was determined to be below the positive threshold, while -1.93 exceeded the negative threshold, and was therefore marked as an abnormal ratio with a consecutive frame count of 1 frame. Since there were only two mutation points, the peak time interval was 5-2=3 seconds. Based on business logic, if the consecutive frames of abnormal ratios were less than 2 and the time interval was greater than 2 seconds, the risk was initially considered low, but further verification based on historical data trends was required. Therefore, the system automatically retrieved similar ratio sequence data from the past 10 minutes to calculate the frequency of abnormal ratio occurrences. If the frequency was below 0.1 times / minute, no velocity overshoot risk was confirmed; otherwise, an alert was triggered. Through the above algorithm and analysis, the system automatically completed the entire process from data processing to risk assessment, ensuring rigorous logic and traceable results.

[0033] S104. If the speed overshoot risk flag is true, then calculate the target speed sequence step size and proportional iteration adjustment amplitude based on the pressure slope mutation density and speed adjustment trigger conditions, generate a speed reduction command, and send it to the grouting pump control unit for real-time regulation according to the command priority.

[0034] If both speed overshoot and risk indicators are detected simultaneously, pressure slope and mutation density data are retrieved from the system. A comparative analysis of preset thresholds is used to determine if the triggering conditions are met. If the triggering conditions are met, the target speed adjustment sequence and sequence step size are calculated based on the changing trends of pressure slope and mutation density using pre-established mapping rules, resulting in a preliminary speed adjustment plan. Based on the calculated target sequence and sequence step size, and combined with proportional iteration adjustment logic, the speed reduction magnitude is determined step-by-step, generating specific reduction command data. By prioritizing the reduction command data and considering real-time control requirements, the command transmission order is determined, establishing the final command transmission plan. According to the command transmission plan, the speed reduction commands are transmitted to the grouting pump control unit in priority order, real-time controlling the speed adjustment process and obtaining the post-adjustment operating status. By collecting monitoring data of the post-adjustment operating status and combining it with the latest changes in pressure slope and mutation density, it is determined whether further adjustment of the speed target sequence is needed, generating a basis for subsequent control.

[0035] When the speed overshoot risk flag is true, the system first collects pressure data in real time through sensors and calculates the pressure slope mutation density. Assuming the current pressure value sequence is [10.5, 11.2, 13.8, 15.0] MPa, with a time interval of 1 second, the slopes are 0.7, 2.6, and 1.2 respectively. The mutation density is defined as the percentage of times the slope changes by more than 1.0. The calculated density is 50%, which is higher than the threshold of 30%, triggering the adjustment condition. Next, based on the speed adjustment trigger condition, combined with the current speed of 3000 rpm and the target speed of 2500 rpm, the system calculates the target speed sequence step size. Using a linear decreasing algorithm, the step size is (3000-2500) / 5=100 rpm, generating the sequence [3000, 2900, 2800, 2700, 2600, 2500]. Meanwhile, the proportional iterative adjustment range is calculated using the formula Amplitude = current speed × (1 - pressure jump density / 100), resulting in an initial adjustment range of 3000 × (1-0.5) = 1500 rpm. However, limited by a maximum single adjustment of 500 rpm, 500 rpm is actually used. Subsequently, the system generates a speed reduction command with high priority, in the format "RPM_DEC_500_H," and sends it to the grouting pump control unit via the CAN bus protocol. Upon receiving and parsing the command, the control unit adjusts the speed to 2800 rpm in real time and feeds back the adjustment result to the main control system, which records the adjustment log for subsequent analysis. If the pressure slope jump density continues to exceed the threshold, the system will repeat the above process until the speed stabilizes within a safe range. By associating with the grouting pump operating status database, the system can further optimize adjustment parameters, such as adjusting the step size to 80 rpm based on historical data to reduce oscillation risk and ensure the continuity and stability of the control logic.

[0036] S105. Obtain the feedback data after the grouting pump returns the speed reduction command. Through closed-loop feedback response delay analysis, determine the actual speed response time and stable deviation value. At the same time, re-acquire the real-time sequence of membrane pressure and the instantaneous reading of volumetric flow rate, and update the combination of operating parameters for further matching analysis.

[0037] The data acquisition module acquires the rotational speed sequence and timestamp sequence of the grouting pump after executing the reduction command. The rotational speed change rate at each time point is calculated based on the rotational speed sequence and timestamp sequence, resulting in a rotational speed change rate sequence. If the absolute value of the rotational speed change rate sequence is less than a preset threshold for five consecutive points, the current moment is determined to be a stable moment, and the corresponding rotational speed value is obtained. The stable deviation value is determined by subtracting the target rotational speed value from the stable moment value. Real-time sequences of membrane pressure are continuously acquired from the membrane pressure sensor, and instantaneous sequences of volumetric flow rate are continuously acquired from the flow meter. Based on the stable moment time point, 30-second data segments before and after the stable moment are extracted from the real-time membrane pressure sequence and the instantaneous volumetric flow rate sequence to form stable segment pressure and flow rate sequences. The current target rotational speed, stable deviation value, average stable segment pressure sequence value, and average stable segment flow rate sequence value are combined to form a new set of operating parameters.

[0038] First, the speed feedback register value returned from the grouting pump PLC is read via industrial Ethernet. The moment when the speed setting changes from 1200 r / min to 900 r / min after receiving the reduction command is recorded as t0. Then, the actual speed sequence is continuously collected at a sampling period of 5 ms to obtain the time series data speed(t). A first-order inertial element model is used for fitting, i.e., speed(t) = 900 + (1200 - 900) × e^(-t / τ). The time constant τ is solved by the least squares method and is found to be approximately 1.82 s. Then, the actual response time is determined to be the interval from t0 until the actual speed first reaches the target value of 900 r / min ± 5 r / min. The required time was measured to be 2.14 s. Simultaneously, the deviation between the average rotational speed and the set value within 30 s after stabilization was calculated, yielding a stabilization deviation of ±3.7 r / min. Next, the pressure sequence p(t) was synchronously reacquired from the membrane pressure sensor at 10 ms intervals, covering the period from 5 s before to 15 s after the command was issued, collecting approximately 2000 points. Simultaneously, the instantaneous value q(t) from the volumetric flow rate sensor was acquired, in L / min. After preprocessing p(t) and q(t) using a sliding window averaging filter (window length of 50 sampling points), the extracted features included a maximum pressure rise rate dp / dt of 0.42 MPa / s and a maximum flow rate decrease slope of -1.8 L / min. 2The average stable pressure value was 4.86 MPa, and the average stable flow rate was 32.4 L / min. Finally, the above parameters, including response time of 2.14 s, stability deviation of 3.7 r / min, stable pressure value of 4.86 MPa, and stable flow rate value of 32.4 L / min, were combined to form a new operating parameter vector. Euclidean distance was calculated between this vector and the historical best matching library. The three closest combinations were (900 r / min, 4.9 MPa, 33 L / min), (880 r / min, 4.7 MPa, 31 L / min), and (920 r / min, 5.1 MPa, 34 L / min). The combination with the smallest distance was selected as the recommended operating parameter combination for the next stage for subsequent grouting process optimization.

[0039] S106. Based on the updated combination of operating parameters, calculate the interval of the ratio exceeding the threshold and the pressure-flow rate matching error. Combine the cumulative frequency of speed adjustment and the density of abnormal points within the time window to determine whether the membrane pressure and flow rate have reached a stable matching state, providing a basis for subsequent regulation.

[0040] Obtain the pressure ratio sequence and flow rate ratio sequence at each moment under the current operating parameter combination; judge the pressure ratio sequence by a preset threshold to obtain the continuous time period when the ratio exceeds the threshold; calculate the duration of each continuous time period and record the start and end times; calculate the error sequence between the actual pressure value and the actual flow rate value within the same time range to obtain the pressure-flow rate matching error value; obtain the operation records of the most recent speed adjustment, count the total number of adjustments, divide the fixed time window backward from the current moment, count the number of abnormal points within the window and calculate the abnormal point density; if the total length of the duration of the ratio exceeding the threshold is less than the preset length and the matching error value is less than the preset error upper limit, and the abnormal point density is lower than the density threshold, then it is determined that the current matching state is stable; when it is determined to be a stable matching state, output a stability flag and record the current parameter combination as the stable benchmark combination.

[0041] To determine the stable matching state of membrane pressure and flow rate, the system first calculates the interval where the ratio exceeds the threshold using updated operating parameters. Assuming pressure data is collected at one point per minute, the threshold for the pressure-to-flow rate ratio is 1.5 over 60 consecutive minutes. If the ratio exceeds this threshold for more than 10 minutes, it is marked as an over-threshold interval. The system automatically records and calculates that the ratio consistently exceeds the threshold between the 20th and 35th minutes, lasting 15 minutes, exceeding the standard value by 5 minutes, indicating a potential instability. Next, the pressure-flow rate matching error is calculated using the root mean square error algorithm. Assuming a standard pressure value of 10.0 kPa and a standard flow rate value of 5.0 L / min, with measured data of 10.5 kPa and 4.8 L / min respectively, the calculated error is sqrt[((10.5-10.0)^2 + (4.8-5.0)^2) / 2] = 0.36. This error is less than the preset threshold of 0.5, indicating a high degree of matching. Subsequently, combining the cumulative frequency of speed adjustments, the system recorded 8 speed adjustments in the past 24 hours, within the preset threshold of 10, indicating that the adjustment frequency was controllable. Simultaneously, the system analyzed the density of outliers within a 30-minute time window, defining outliers as data points deviating from the mean by two standard deviations. The calculated outlier percentage was 6.7%, below the 10% threshold, indicating that data fluctuations were within acceptable limits. Based on the above analysis, the system calculated a total score of 82 using a weighted scoring algorithm (40% for exceeding the threshold duration, 30% for matching error, 20% for adjustment frequency, and 10% for outlier density), exceeding the stable matching threshold of 75. This indicates that the membrane pressure and flow rate have reached a stable matching state, providing data support for subsequent adjustments. If the score is below the threshold, the system will automatically generate adjustment suggestions and record them in the log, linking them to the equipment maintenance module to ensure a closed-loop operation.

[0042] S107. If the membrane pressure and flow rate do not reach a stable matching state, the target sequence of rotational speed for the next cycle is generated based on the abnormal fluctuation amplitude of the flow rate and the cumulative mutation amount within the time window. The target sequence and the proportional iterative adjustment amplitude are dynamically updated and sent to the grouting pump control unit to form a closed-loop control until the matching state is optimized.

[0043] The flow rate data sequence within the current time window is acquired, and the cumulative mutation amount is obtained by calculating the sum of the absolute values ​​of the differences between adjacent sampling points. If the cumulative mutation amount exceeds a preset threshold and the flow rate fluctuation amplitude is greater than a limit value, the process proceeds to the next cycle target generation stage; otherwise, the current rotational speed command is maintained. The values ​​of each point in the rotational speed target sequence are determined based on the magnitude of the cumulative mutation amount and the amplitude of the flow rate fluctuation. The rotational speed target sequence is dynamically updated using a linear interpolation method to obtain the latest sequence. The latest rotational speed target sequence and the proportional iterative adjustment step size are input into the closed-loop controller, and the grouting pump is driven to perform rotational speed adjustment according to the output command of the closed-loop controller. The cumulative mutation amount within the time window is recalculated from the newly acquired membrane pressure and flow rate data after execution, and it is determined whether the current membrane pressure and flow rate have entered the stable matching interval. If they have not entered the matching interval, the process returns to the rotational speed target sequence dynamic update stage to continue the cyclic adjustment.

[0044] When the membrane pressure and flow rate are not in a stable matching state, the system first collects membrane pressure and flow rate data in real time through sensors. For example, if the current pressure is 2.5 MPa and the flow rate is 10 L / min, and the target matching state is set to a pressure of 2.8 MPa and a flow rate of 12 L / min, the system detects a flow rate fluctuation of ±1.5 L / min, which exceeds the allowable range of ±0.5 L / min. At the same time, the cumulative mutation amount within the past 5-minute time window reaches 3 L / min, which is judged as an abnormal state. Next, based on the fluctuation amplitude and mutation amount, the system uses a linear regression algorithm to predict the target rotational speed sequence for the next cycle. The calculation formula is: Target rotational speed = Current rotational speed + (Target flow rate - Current flow rate) × Coefficient, where the coefficient is 0.8. The current rotational speed is 500 rpm, the target flow rate is 12 L / min, and the current flow rate is 10 L / min, resulting in a target rotational speed of 516 rpm, forming a sequence [516, 518, 520] rpm, corresponding to the next 3 control cycles. Subsequently, the system dynamically updates the target sequence and adjusts the speed using a proportional-iterative adjustment algorithm. The proportional coefficient P is set to 0.5, the integral coefficient I to 0.1, and the derivative coefficient D to 0.05. The system calculates the adjustment increment for each step; for example, if the current deviation is 2 L / min, the adjustment increment is 1 L / min, ensuring gradual approximation to the target value. The system then sends the adjusted target speed value to the grouting pump control unit via a communication interface. Upon receiving the data, the control unit automatically adjusts the motor speed to 516 rpm, forming a closed-loop control. Finally, the system continuously monitors pressure and flow rate data. If the pressure stabilizes at 2.8 ± 0.1 MPa and the flow rate stabilizes at 12 ± 0.2 L / min, the matching state optimization is considered complete. If these conditions are not met, the above process is repeated to ensure system stability. To enhance logic, the system can also correlate with grouting pump temperature data. If the temperature exceeds 50℃, the speed adjustment increment is reduced to 0.5 L / min to prevent overheating and further ensure safe equipment operation.

[0045] Obviously, those skilled in the art can make various modifications and variations to the embodiments of this application without departing from the spirit and scope of the embodiments of this application. Therefore, if these modifications and variations to the embodiments of this application fall within the scope of the claims of this application and their equivalents, this application also intends to include these modifications and variations.

Claims

1. A method for rapid fermentation of fermented bean curd in large tanks followed by intra-membrane grouting and molding, characterized in that, The method includes: Pressure signals from various regions within the membrane are acquired using a multi-point pressure sensor array. A differential sequence is calculated to generate a real-time membrane pressure sequence. Simultaneously, instantaneous rotor speed and volumetric flow rate readings are read from the grouting pump to construct the current operating parameter combination for subsequent dynamic analysis. A low-pass filter is used to process the real-time membrane pressure sequence, extracting the location and amplitude of pressure slope abrupt changes. Combined with the flow rate fluctuation range deviation from the instantaneous volumetric flow rate readings, a dynamic change characterization sequence of membrane pressure is generated for subsequent ratio calculations. Based on this dynamic change characterization sequence, the ratio of each pressure slope abrupt change point to the corresponding instantaneous volumetric flow rate reading is calculated, forming a ratio sequence. The number of consecutive frames with abnormal ratios and the time interval between ratio peaks are analyzed. Combined with a preset dynamic ratio threshold, a speed overshoot risk indicator is determined. If the speed overshoot risk indicator is true, the target speed sequence step size and proportional iterative adjustment are calculated based on the pressure slope abrupt change density and speed adjustment trigger conditions. The system generates a speed reduction command and sends it to the grouting pump control unit for real-time adjustment according to command priority. It acquires feedback data from the grouting pump after the speed reduction command is executed, and through closed-loop feedback response delay analysis, determines the actual response time and stable deviation value of the speed. Simultaneously, it re-acquires the real-time sequence of membrane pressure and the instantaneous reading of volumetric flow rate, updating the operating parameter combination for further matching analysis. Based on the updated operating parameter combination, it calculates the ratio exceeding the threshold duration interval and the pressure-flow rate matching error. Combining the cumulative frequency of speed adjustment and the density of abnormal points within the time window, it determines whether the membrane pressure and flow rate have reached a stable matching state, providing a basis for subsequent control. If the membrane pressure and flow rate have not reached a stable matching state, it generates the next cycle's target speed sequence based on the abnormal flow rate fluctuation amplitude and the cumulative mutation amount within the time window. Through dynamically updated target sequences and proportional iterative adjustment amplitudes, it sends the sequence to the grouting pump control unit to form closed-loop control until the matching state is optimized.

2. The method for rapid fermentation of fermented bean curd in a large tank followed by intramembrane grouting as described in claim 1, characterized in that, The process involves acquiring pressure signals from various regions within the membrane using a multi-point pressure sensor array, calculating differential sequences to generate a real-time membrane pressure sequence, and simultaneously reading instantaneous rotor speed and volumetric flow rate readings from the grouting pump to construct a current operating parameter combination for subsequent dynamic analysis, including: The pressure signal of the membrane region is continuously acquired by a multi-point pressure sensor array, and the original pressure data of each region is recorded to obtain the initial pressure dataset. Based on the initial pressure dataset, the pressure difference between adjacent time points is calculated, a difference sequence is generated, and the real-time sequence of membrane pressure changes is determined. For real-time sequences, a combination of operating parameters containing multi-dimensional information is constructed by combining instantaneous data of rotor speed and volumetric flow rate obtained from the grouting pump. By performing time-dimensional data alignment on the combination of operating parameters, a unified time-series dataset is obtained, and the synchronization between parameters is determined. If the synchronization is insufficient, the rotor speed and volumetric flow rate data are interpolated to obtain an adjusted set of timing parameters. The support vector machine algorithm was used to classify the adjusted time series parameter set to determine the correlation pattern between membrane pressure changes and operating parameters; For association patterns, generate feature datasets required for dynamic analysis for subsequent system processing.

3. The method for rapid fermentation of fermented bean curd in a large tank followed by intramembrane grouting and molding according to claim 1, characterized in that, The process employs low-pass filtering to process the real-time membrane pressure sequence, extracting the location of pressure slope abrupt changes and the magnitude of fluctuations. Combined with the deviation of the flow rate fluctuation range from the instantaneous volumetric flow rate reading, a dynamic change characterization sequence of membrane pressure is generated for subsequent ratio calculations. This includes: The filtered pressure sequence is obtained by processing the real-time intramembrane pressure sequence with a low-pass filter. The pressure slope sequence is obtained by calculating the slope sequence of consecutive adjacent points from the filtered pressure sequence. The pressure slope abrupt change location sequence is obtained by identifying points in the pressure slope sequence where the absolute value of the slope exceeds a preset abrupt change threshold. For each location of a sudden change in pressure slope, the extreme value difference before and after the corresponding location in the filtered pressure sequence is extracted to obtain the fluctuation amplitude sequence. The real-time flow rate sequence is obtained by collecting the instantaneous readings of the volumetric flow rate instrument. The flow rate fluctuation interval sequence is obtained by calculating the difference between the maximum and minimum values ​​in adjacent time periods based on the real-time flow rate sequence. Aligning the flow rate fluctuation interval sequence with the pressure slope abrupt change location sequence by time yields the flow rate fluctuation interval deviation sequence at the corresponding time point. By combining and connecting the pressure slope abrupt change markers, fluctuation amplitudes, and corresponding flow rate fluctuation range deviations at each abrupt change location in chronological order, a sequence representing the dynamic changes in intramembrane pressure is obtained.

4. The method for rapid fermentation of fermented bean curd in a large tank followed by intramembrane grouting and molding according to claim 1, characterized in that, The process involves calculating the ratio of each pressure slope abrupt change point to the corresponding instantaneous volumetric flow rate reading based on the dynamic change sequence of intramembrane pressure, forming a ratio sequence. The number of consecutive frames with abnormal ratios and the time interval between ratio peaks in the ratio sequence are analyzed. Combined with a preset dynamic ratio threshold, the system determines whether there is a risk indicator of velocity overshoot, including: Obtain raw data records of membrane pressure changes over time, calculate the rate of change of pressure value at each time point, form a pressure slope sequence, and determine the set of abrupt change points; Extract the timestamp corresponding to each point from the set of mutation points, obtain the instantaneous reading of the volumetric flow rate at the same time, calculate the ratio of pressure slope to flow rate reading, and obtain the ratio sequence data; For the ratio sequence data, the change range of each ratio is analyzed. If a ratio exceeds the preset dynamic threshold range, it is marked as an abnormal ratio, and the abnormal ratio distribution is obtained. Based on the distribution of abnormal ratios, the number of frames with consecutive abnormal ratios is counted, the cumulative duration of consecutive frames is calculated, and the characteristics of abnormal persistence are determined. By analyzing the abnormal ratio distribution, the time points when the ratio peaks occur are extracted, the time intervals between adjacent peaks are calculated, and the regularity of the peak intervals is determined. By using the characteristics of abnormal persistence and the regularity of peak intervals, and comparing them with preset dynamic threshold standards, if the persistence characteristics or the regularity of intervals exceed the threshold range, it is determined that there is a risk of speed overshoot, and the final risk assessment result is obtained.

5. The method for rapid fermentation of fermented bean curd in a large tank followed by intramembrane grouting and molding according to claim 1, characterized in that, If the speed overshoot risk flag is true, then based on the pressure slope abrupt change density and the speed adjustment trigger condition, the target speed sequence step size and proportional iteration adjustment amplitude are calculated, a speed reduction command is generated, and sent to the grouting pump control unit for real-time regulation according to the command priority, including: If both speed overshoot and risk flag are detected, pressure slope and mutation density data are retrieved from the system, and the satisfaction of the triggering conditions is determined by comparing and analyzing preset thresholds. If the triggering conditions are met, the target sequence and sequence step size for speed adjustment are calculated based on the changing trends of pressure slope and mutation density using pre-established mapping rules, thus obtaining a preliminary speed adjustment scheme. Based on the calculation results of the target sequence and sequence step size, and combined with the adjustment logic of proportional iteration, the magnitude of speed reduction is determined step by step, and specific speed reduction command data is generated. By sorting the command data according to the priority order, and based on the real-time control requirements, the order of command transmission is obtained, and the final command sending plan is determined. According to the instruction sending plan, the speed reduction instruction is transmitted to the grouting pump control unit in priority order, and the execution process of speed adjustment is controlled in real time to obtain the operating status after adjustment; By collecting monitoring data on the operating status after regulation, and combining the latest changes in pressure slope and mutation density, it is determined whether the target speed sequence needs further adjustment, thus generating a basis for subsequent regulation.

6. The method for rapid fermentation of fermented bean curd in a large tank followed by intramembrane grouting and molding according to claim 1, characterized in that, The process involves acquiring feedback data after the grouting pump returns a speed reduction command, analyzing the closed-loop feedback response delay to determine the actual speed response time and stable deviation value, and simultaneously re-acquiring the real-time sequence of membrane pressure and instantaneous volumetric flow rate readings to update the operating parameter combination for further matching analysis, including: The data acquisition module obtains the speed sequence and timestamp sequence of the grouting pump after executing the reduction command; The rate of change of rotational speed at each time point is calculated based on the rotational speed sequence and the timestamp sequence to obtain the rotational speed change rate sequence; If the absolute value of five consecutive points in the speed change rate sequence is less than a preset threshold, then the current moment is determined to be a stable moment, and the speed value corresponding to the stable moment is obtained. The stability deviation value is determined by subtracting the target speed value from the stable speed value. The membrane pressure real-time sequence is continuously acquired from the membrane pressure sensor, and the volumetric flow rate instantaneous sequence is continuously acquired from the flow meter. Based on the steady-state time point, 30-second data segments before and after the steady-state time are extracted from the real-time sequence of membrane pressure and the instantaneous sequence of volumetric flow rate to form the steady-state pressure sequence and the steady-state flow rate sequence. The current target speed, stable deviation value, average pressure sequence of stable segment, and average flow rate sequence of stable segment are combined into a new combination of operating parameters.

7. The method for rapid fermentation of fermented bean curd in a large tank followed by intramembrane grouting and molding according to claim 1, characterized in that, The process involves calculating the sustained interval of the ratio exceeding the threshold and the pressure-flow rate matching error based on the updated combination of operating parameters. Combined with the cumulative frequency of rotational speed adjustments and the density of abnormal points within the time window, it determines whether the membrane pressure and flow rate have reached a stable matching state, providing a basis for subsequent regulation. This includes: Obtain the pressure ratio sequence and flow rate ratio sequence at each time point under the current combination of operating parameters; By judging the pressure ratio sequence by a preset threshold, the continuous time period in which the ratio exceeds the threshold is obtained; For each continuous time period, calculate its duration and record the start and end times. Within the same time range, calculate the error sequence between the actual pressure value and the actual flow rate value to obtain the pressure-flow rate matching error value. Obtain operation records of the most recent speed adjustments, count the total number of adjustments, divide a fixed time window backward from the current time, count the number of outliers within the window and calculate the outlier density; If the total length of the duration of the ratio exceeding the threshold is less than the preset length and the matching error value is less than the preset error upper limit, and the density of outliers is lower than the density threshold, then it is determined that the current matching state is stable. When a stable matching state is determined, a stability flag is output and the current parameter combination is recorded as a stable baseline combination.

8. The method for rapid fermentation of fermented bean curd in a large tank followed by intramembrane grouting and molding according to claim 1, characterized in that, If the membrane pressure and flow rate do not reach a stable matching state, a target sequence for the rotational speed of the next cycle is generated based on the abnormal fluctuation amplitude of the flow rate and the cumulative mutation amount within the time window. This target sequence is dynamically updated, and the proportional iterative adjustment amplitude is sent to the grouting pump control unit to form a closed-loop control until the matching state is optimized, including: Obtain the flow rate data sequence within the current time window, and calculate the cumulative mutation amount by summing the absolute values ​​of the differences between adjacent sampling points; If the cumulative mutation amount exceeds the preset threshold and the flow rate fluctuation is greater than the limit value, the process will proceed to the next cycle target generation stage if the judgment is true; otherwise, the current rotation speed command will be maintained. The values ​​of each point in the target rotational speed sequence are determined based on the magnitude of the cumulative mutation and the amplitude of the flow rate fluctuation. The latest sequence is obtained by dynamically updating the target rotational speed sequence using a linear interpolation method; The latest target speed sequence and the proportional iterative adjustment step size are input together into the closed-loop controller, and the grouting pump is driven to perform speed regulation according to the output command of the closed-loop controller. The cumulative mutation amount within the time window is recalculated from the newly collected membrane pressure and flow rate data after execution to determine whether the current membrane pressure and flow rate have entered the stable matching interval. If the target speed sequence is not matched, the process returns to the dynamic update stage and continues to cycle and adjust.

Images